f 1983 ANNALS | Mi THE MISSOURI BOTANICAL GARDEN The ANNALS, published four times a year, contains papers, pri- marily in systematic botany, contributed from the Missouri Botan- ical Garden, St. Louis. Papers originating outside the Garden will also be accepted. Authors should write the Editor for information concerning arrangements for publishing in the ANNALS. EDITORIAL COMMITTEE Nancy Morin, Editor Missouri Botanical Garden CHERYL R. Bauer, Editorial Assistant Missouri Botanical Garden MARSHALL R. CROSBY Missouri Botanical Garden Gerrit DAVIDSE Missouri Botanical Garden JOHN D. Dwyer Missouri Botanical Garden & St. Louis University PETER GOLDBLATT Missouri Botanical Garden Colophon This volume of the ANNALS of the Missouri Botanical Garden has been set in APS Times Roman. The text is set in 9 point type while the figure legends and literature cited sections are set in 8 point The volume has been printed on 70# Centura Gloss, an acid-free paper designed to have a shelf-life of over 100 years. Centura Gloss is manufactured by the Consolidated Paper Company. Photographs used in the ANNALS are reproduced using 300 line screen halftones. The binding used in the production of the ANNALS is a proprietary method known as Permanent Binding The ANNALS is printed and distributed by Allen Press, Inc. of Lawrence, Kansas 66044, U.S.A. © Missouri Botanical Garden 1984 ISSN 0026-6493 ALLEN, ROBERT T. Distribution Patterns among Arthropods of the North Temperate Deciduous Forest Biota AXELROD, DANIEL I. Biogeography of Oaks in the Arcto-Tertiary Prov- ince BARRINGER, KERRY. Polygala dukei (Polygalaceae), a New Species from Panama BOUFFORD, D. E. & S. A. SPONGBERG. Eastern Asian-Eastern North Amer- ican Phytogeographical Relationships— A History from the Time of Linnaeus to the Twentieth Century CASPER, BRENDA B. (See Delbert Wiens, John P. Rourke, Brenda B. Casper, Eric A. Rickart, Timothy R. LaPine, C. Jeanne Peterson & Alan Chan- ning) CHANG, Davip H. S. The Tibetan Plateau in Relation to the Vegetation of China CHANNING, ALAN. (See Delbert Wiens, John P. Rourke, Brenda B. Casper, Eric A. Rickart, Timothy R. LaPine, C. Jeanne Peterson & Alan Chan- ning) CHEN, SING-CHI. A Comparison of Orchid Floras of Temperate North America and Eastern Asia CROAT, THOMAS B. A Revision of the Genus Anthurium (Araceae) of Mex- ico and Central America. Part I: Mexico and Middle America .......... DAVIDSE, GERRIT. Biogeographical Relationships between Temperate East- ern Asia and Temperate Eastern North America: The Twenty-Ninth Annual Systematics Symposium Davis, MARGARET B. Quaternary History of Deciduous Forests of Eastern 5 North America and Europe DELGADILLO, C. (See R. E. Magill, C. Delgadillo & L. R. Stark) 0. DoEBLEY, JOHN F. The Maize and Teosinte Male Inflorescence: A Nu- merical Taxonomic Study FALLEN, MARY E. A Taxonomic Revision of Condylocarpon (Apocyna- ceae) GENTRY, ALWYN H. A new Combination for a Problematic Central Amer- ican Apocynaceae GENTRY, ALWYN H. Alstonia (Apocynaceae): Another Palaeotropical Ge- 2 nus in Central America GOLDBLATT, PETER. (See Ching-I Peng & Peter Goldblatt) HAMILTON, WARREN. Cretaceous and Cenozoic History of the Northern 4 Continents 616 629 203 423 564 HE, SHAN-AN & FRANK S. SANTAMOUR, JR. Isoenzyme Verification of American-Chinese Hybrids of Liquidambar and Liriodendron ... Honc, De-YUAN. The Distribution of Scrophulariaceae in the Holarctic with Special Reference to the Floristic Relationships between Eastern Asia and Eastern North America Hou, HsioH-YU. Vegetation of China with Reference to its Geographical Distribution HsÜ,JEN. Late Cretaceous and Cenozoic Vegetation in China, Emphasizing Their Connections with North America IWATSUKI, KUNIO. (See Masahiro Kato & Kunio Iwatsuki) KATO, MASAHIRO & KUNIO IWATSUKI. Phytogeographic Relationships of Pteridophytes between Temperate North America and Japan ... KIRKBRIDE, H., JR. A New Variety of Declieuxia cacuminis TRER from Bahia KRUCKEBERG, A. R. Temperate Floras: The North Pacific Connection ..... LAPINE, TIMOTHY R. (See Delbert Wiens, John P. Rourke, Brenda B. Cas- per, Eric A. Rickart, Timothy R. LaPine, C. Jeanne Peterson & Alan Channing) LIESNER, RONALD. (See Julian A. Steyermark & Ronald Liesner) „n. LITTLE, ELBERT L., JR. North American Trees with Relationships in Eastern Asia McDOUGAL, KAREN M. (See Clifford R. Parks, Norton G. Miller, Jonathan F. Wendel & Karen M. McDougal) MCKENNA, MALCOLM C. Holarctic Landmass Rearrangement, Cosmic Events, and Cenozoic Terrestrial Organisms MAGILL, R. E., C. DELGADILLO & L. R. STARK. Tortula chisosa sp. nov., a Bistratose-Leaved Species from the United States, Mexico, and South- ern Africa MILLER, NORTON G. (See Clifford R. Parks, Norton G. Miller, Jonathan F. Wendel & Karen M. McDougal) MUHLENBACH, VIKTOR. Supplement to the Contributions to the Syn- anthropic (Adventive) Flora of the Railroads in St. Louis, Missouri, U.S.A. ORNDUFF, ROBERT. Studies on the Reproductive System of Nivenia cor- ymbosa (Iridaceae), an Apparently Androdioecious Species |... La PARKS, CLiFFORD R., NORTON G. MiLLER, JONATHAN F. WENDEL & KAREN M . McDougar. Genetic Divergence within the Genus Liriodendron (Magnoliaceae) PENG, CHING-I & PETER GOLDBLATT. Confirmation of the Chromosome Number in Cephalotaceae and Roridulaceae PETERSON, C. JEANNE. (See Delbert Wiens, John P. Rourke, Brenda B. 748 200 658 170 146 197 Casper, Eric A. Rickart, Timothy R. LaPine, C. Jeanne Peterson & Alan Channing) PHiPPS, J. B. Biogeographic, Taxonomic, and Cladistic Relationships be- tween East Asiatic and North American Cratae RAVEN, PETER H. (See Horoshi Tobe & Peter H. Raven) RICKART, ERIC A. (See Delbert Wiens, John P. Rourke, Brenda B. Casper, Eric A. Rickart, Timothy R. LaPine, C. Jeanne Peterson & Alan Chan- ning) ROURKE, JOHN P. (See Delbert Wiens, John P. Rourke, Brenda B. Casper, Eric A. Rickart, Timothy R. LaPine, C. Jeanne Peterson & Alan Chan- ning) St. JOHN, HAROLD. A New Hesperomannia (Compositae) from Maui Is- land: Hawaiian Plant Studies 116 SANTAMOUR, FRANK S., JR. (See Shan-An He & Frank S. Santamour, f; SODERSTROM, THOMAS R. & STEPHEN M. YOUNG. A Guide to Collecting Bamboos SPONGBERG, S. A. (See D. E. Boufford & S. A. Spongberg) STARK, L. R. (See R. E. Magill, C. Delgadillo & L. R. Stark) STEYERMARK, JULIAN A. The Genus Botryarrhena in Venezuela —... : STEYERMARK, JULIAN A. & RONALD LIESNER. Revision of the Genus Ster- igmapetalum (Rhizophoraceae) STONE, BENJAMIN C. A Guide to Collecting Pandanaceae (Pandanus, Frey- cinetia, and Sararanga) TOBE, HIROSHI & PETER H. RAVEN. An Embryological Analysis of Myrtales: Its Definition and Characteristics WAGNER, WARREN L. New Species and Combinations in the Genus Oe- nothera (Onagraceae) WENDEL, JONATHAN F. (See Clifford R. Parks, Norton G. Miller, Jonathan F. Wendel & Karen M. McDougal) WHITE, PETER S. Eastern Asian-Eastern North American Floristic Rela- tions: The Plant Community Level WIENS, DELBERT, JOHN P. ROURKE, BRENDA B. CASPER, ERIC A. RICKART, TIMOTHY R. LAPINE, C. JEANNE PETERSON & ALAN CHANNING. Non- flying Mammal Pollination of Southern African Proteas: A Non-Co- evolved System Wu, ZHENGYI. On the Significance of Pacific Intercontinental Disconti- nuity WUNDERLIN, RICHARD P. Revision of the Arborescent Bauhinias (Faba- ceae: Caesalpinioideae: Cercideae) Native to Middle America u. 667 71 658 734 YING, TSUN-SHEN. The Floristic Relationships of the Temperate Forest Regions of China and the United States 597 YOUNG, STEPHEN M. (See Thomas R. Soderstrom & Stephen M. Young) 128 ZHONG, CHENG. A Comparative Study of the Vegetation in Hubei Prov- ince, China, and in the Carolinas of the United States 5 Volume 70, No. 3, pp. 421-576 of the Annals of the Missouri Botanical Garden, was pub on 17 July 1984. ANNALS MISSOURI BOTANICAL CARDEN qs 1983 NUMBER 1 ; A in X i $ ! pa Ne, a / Yak 4 AN fo * ] j 4 # ^ »" j P" T | $ p i r i4 | T3 * A M : $ 1 " $ 4 + ; VF i f nr 41 Jp E hd mf E: i y F: $ kra d 1 fest 1 a € TOF te: * Lg ETATIS, CONTENTS Ut : : ARDEN LIBRARY Nonflying Mammal Pollination of Southern CARDEN n Protease: A Non-Coe- volved System Delbert Wiens, John P. Rourke, Brenda B. Casper, Eric A. Rickart, Timothy R. LaPine, C. Jeanne Peterson, & Alan Chan- np — Se l The Maize and Teosinte Male Inflorescence: A Numerical Taxonomic Study John F. Doebley ....—. A SAI E PI DF An Embryological Analysis of Myrtales: Its Definition and Char- acteristics Pine fave & Pea H Kov — — a 71 Revison of the Arborescent Bauhinias (Fabaceae: Caesalpinioideae: Cerci- deae) Native to Middle America Richard P. Wunderlin ........ I a A Guide to Collecting Bamboos Thomas R. Soderstrom & Stephen M. Young ..... DUAE 128 A Guide to Collecting Pandanaceae (Pandanus, Freycinetia, and Sararan- ga) BONN C AINE o i ou ca Meat dean Ea Studies on the Reproductive System of Nivenia corymbosa (Iridaceae), an Apparently Androdioecious Species Robert Ornan — — — — — 146 (Contents continued on back cover) VOLUME 70 198 NUMBER 1 ANNALS OF THE MISSOURI BOTANICAL GARDEN The ANNALS contains papers, primarily in systematic botany, contributed from the Missouri Botanical Garden. Papers originating outside the Garden will also be accepted. Authors should write the Editor for information concerning arrangements for publishing in the ANNALS. EDITORIAL COMMITTEE Nancy Morin, Edito SSe Botanical CHAM MansHALL R. CROSBY Missouri Botanical Garden GERRIT DAVIDSE Missouri Botanical Garden JOHN Dwyer Missouri Botanical Garden & St. Louis University PETER GOLDBLATT Missouri Botanical Garden Published four timesa year by the € Botanical Garden, St. Louis, Missouri 631 ISSN 0026-6493 T. i g Office ofthe A nnals PO. Bos 368, 1041 New Hampshire, Lawrence, Kansas 66044. scription price is $60 per volume U. S., $65 Canada, and Mexi $70 all salen countries. Personal subscriptions are available at $30 and $35, E RS irmail delivery charge, $30 per volume. Four issues per volume. ¢ A zl Lawrence. K 66044 oY F +4 © Missouri Botanical Garden 1983 ANNALS OF THE MISSOURI BOTANICAL GARDEN VOLUME 70 1983 NUMBER 1 NONFLYING MAMMAL POLLINATION OF SOUTHERN AFRICAN PROTEAS: A NON-COEVOLVED SYSTEM! DELBERT WIENS,” JOHN P. ROURKE,* BRENDA B. CASPER,* ERIC A. RICKART,? TiMorHYv R. LAPINE,? C. JEANNE PETERSON,” AND ALAN CHANNING? ABSTRACT Traits parine risa those proteas SA ie for pollination by nonflying mammals include: bowl- shaped heads bearing fleshy bracts, these borne on short, flexible peduncles, often at or near ground level (geoflorous) oo idden beneath Lt overlying foliage (cryptic), and producing copious nectar (ca. 1.8 ml/head, standing crop); individual florets with wiry, yet flexible styles and a nectar-stigma mat ammal visitation to protea flowering heads py the presence of pollen on the rostra (carried in ctar); the transport of fluorescing powders to nequ heads both within and between plants; ey keune tion of small-mammal feces in flowering heads, and the destruction of exclosure bags containing nectar-rich heads. The period of greatest small-mammal activity (1800 hr.) coincides with maximum flower opening. T maze experi- ments showed that small mammals, when given a choice ipia typically bird-pollinated proteas and those having characteristics of flowers pollinated by nonflying mammals, always foraged on Pu latter. That small mammals can effect pollination is indicated by thei foraging behavior on flowering heads while in captivity, the morphological “fit” between individual florets and the rostra of ciel mammals, and by selective exclosure experiments that reduced seed set (50% and 95%) when small y honey bees). The nectar produced idlillildl Viol ta tion The study was supported in part by NSF grant (DEB 78-11624). We thank B. Albee, L. Arnow, H. G. aii E SUM P. Cox, V. Grant, S. D. Hopper, P. H. Raven, and V. Turner for helpful comments ‘on the manuscri too numerous to mention individually assisted the research, in particular the staffs of the Karoo Botanic Gea and Kirstenbosch Botanic Garden and its director Prof. Brian Rycroft. The curator of the Karoo Garden, Mr. Bruce Bayer, provided generous logistical and field assistance and es in stimulating discussions of the problem from its Seng Peter and 1 hospitality "ps Peter's alu Pretoria, assisted in various capacities. The late Fred du Plesis and his wife, pics provided access to the unparalleled study ict on Jonaskop. Dr. Walter Veith, Stellenbosch University, obtained the metabolic rates of Aethomys and the caloric value. of P. humiflora nectar; we are also ned to Stellenbosch University for of. the use of their Prof. H. G. an er, University of California, Berkeley, determined the carbohydrate and amino acid c composi tions of protea one: showed a continuing interest in the problem and offered various helpful — ™ Finally, but Á— nni importantly, we ui indebted to field assistance from Tori Burns, Heidi Fain, and Jenny Turner, who ved the southeaste ? Department of Biology, University of Utah, Salt Lake oe, piam 84112. ? Kirstenbosch Botanic Garden, Claremont, South Africa 7735. * Present address: Department of Biology, University of Pennsylvania, Philadelphia, Pennsylvania 19104. * Department of Biochemistry, University of the Western Cape, Bellville, South Africa 7530. ANN. Missouni Bor. GARD. 70: 1-31. 1983. 2 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 by these proteas meets the energy requirements of the small-mammal community for only several days annually, thus —— is impossible. Proteas adapted for pollination by nonflying mammals al have evo ee decrease in population size [om in response to progressive nt discoveries in the Neotropics of flowers with some similar char- acteristics and also xem inated ^in ps ing mammals support the existence of a worldwide class of flowers adapted for such pollinators The pollination of flowers by nonflying mam- mals was first mentioned by Kerner (1895, v. 2, p. 230) and was discussed nearly 50 years ago by Porsch (1934, 1935, 1936a, 1936b). The subject was not given further attention, however, until Morcombe (1968) suggested pollinatory rela- tionships between various proteaceous flowers and nonflying mammals in the southwestern Australian flora. Rourke and Wiens (1977) re- viewed the problem and noted that various floral features convergent in Australian and South Af- rican Proteaceae suggested adaptations for pol- he following as. Two previous studies of pre- sumably dat pollinated African plants, viz., the baobab (Bombacaceae) (Coe & Isaac, 1965) and Maranthes (Chrysobalanaceae) (Lack, 1977) demonstrated visitation and nectar feeding by bush ies (Galag } / Geoffroy) and genets (Genetta tigrina Schreber), respectively. Recent interest in the subject has resulted in a number of publications: Sussman and Raven (1978) reviewed the problem and reported pol- lination by arboreal Madagascan mouselike le- murs; Sleumer (1955) and Carpenter (19782) re- ported evidence for the pollination of eastern Australian banksias by sugar gliders (Petaurus breviceps Waterhouse) and the indigenous pla- cental bushrat (Rattus fuscipes Waterhouse), re- spectively. Holm (1978) and Ford, Paton, and Forde (1979) commented on the problem and Armstrong (1979) reviewed the subject for Aus- tralia generally. Wiens, Renfree, and Wooller (1979) and Hopper (1980) studied the pollina- tion of Banksia and other flowers by the south- western Australian marsupial honey possum (Tarsipes rostratus = T. spenserae Gray). In the New World, Prance (1980) observed probable pollination by cebus monkeys; and Janson, Ter- borgh, and Emmons (1981) reported extensive visitation and apparent pollination in Bomba- caceae and Combretaceae in the Amazon by monkeys [including the small (100 g) pigmy mar- moset], opossums, and procyonids. Lumer (1980) observed pollination of B/akea (Melastomata- ceae) by rodents in Costa Rica, and Steiner (1981) discovered probable pollination by opossums in Mabea (Euphorbiaceae) in Panama. The sugges- tion of rodent pollination in Hawaiian Freyci- netia (Degener, 1945) is apparently in error (Cox, We report the results of three seasons of field studies on several species of southern African species of Protea (Proteaceae) with characteris- tics indicating adaptations for nonflying mam- Observations cover the winter-early spring (Au- gust-September) field seasons of 1978-79 and the summer (February) field season of 1980 in the Cape region of South Africa. Wiens and Rourke (1978) established that several murid ro- dents and a few additional small mammals vis- ited the flowering heads of Protea amplexicaulis (Salisb.) R. Br. and P. humiflora Andr., as evi- denced by the presence of protea pollen loads on that several rodents foraged readily and non-de- | structively on the nectar from flowering heads of several proteas while in captivity | This study has two broad objectives: first, to test further the proposition that these flowers are pollinated by nonflying mammals and to gain a tem functions as an integrated whole; and sec- ondly, to consider why this unusual animal-flow- ] er relationship may have evolved. Specific areas of study included: (1) the general nature of the plant association in which NMP proteas occur, (2) the composition and activity patterns of the associated small-mammal community, (3) the temporal patterns of anthesis, nectar secretion, and odor production, (4) nectar volume, total. sugar content, carbohydrate and amino acid. composition, (5) stylar and general floral struc- | i 1983] WIENS ET AL.—NONFLYING MAMMAL POLLINATION 3 ABLE 1. The number of small mammals examined for pollen loads and the proteas with which they were associated. Small Mammal d Aog Elephan- Praomys Rhab- Protea spinosus quensis edwardii reauxi pumilio Study Site P. amplexicaulis:4 15 31>Þ 2 3 162 Jonasplaats, see text P. cryophilac r 2 5 Sneeubergnek, Cedarberg Mts., E of Citrusdale P. effusa: l 3 2 4 East bench, Murray Farm, above Gydo Pass N. of Ceres P. humifloras4 13^ 163 DP hag Sh ] 4^ Jonasnek P. recondita: 3 3* 2 4^ Same as P. effusa P. restionifolia 3a Pocskraal, N. shore Stormsvlei res- ervoir, S. of Worcester P. scabra a W. of Villiersdorp P. sulphurea 22b ]* W. of Ouberg Pass, NE of Mon- tague * Captive animals of this species foraged on flowering heads of the respective protea in a manner to assure erase ip i tion. mals with numerous protea pollen grains (7100 and often > 1,000) in fecal samples taken directly ka ce ie * Some heads of these protea species contained fecal pellets. * [n addition to the animals listed, single specimens of Crocidura sp. and Graphiuris ocularis were trapped around P. amplexicaulis and one Dendromus melanotis and a Mus minutoides were captured in stands. ture, (6) genetic compatibility and pollen viabil- ity, (7) the energy resources protea nectar pro- vides to the small-mammal community, and (8) the relative importance of insects as pollinators of NMP proteas. RESEARCH AREAS AND SPECIES STUDIED The NMP proteas utilized in this research are endemic to the southwestern Cape flora (fynbos) of South Africa. This vegetation is unique, with (1) exceptionally rich y (ca. 8,850 Species) and high endemism (73: 1%), (2) adap- tations for periodic burning, (3) a virtual absence of indigenous trees, (4) a low percentage of an- nuals, and (5) restriction to the Table Mountain andstone, which is highly depauperate nutri- tionally especially for P and N. Uplift and ero- sion of this formation have produced a highly dissected landscape comprised of many small mountain ranges with diverse elevations, precip- itation, and soils. Predictably, numerous species exist only as small populations and in isolated i (Goldblatt, 1978; Taylor, 1978; Kruger, eneriec Although the vegetation is unique, the animal P. humiflora community is not. The small-mammal com- munity, for example, is composed largely of species with ranges extending far beyond the dis- tribution of the Cape flora. Likewise, the bee fauna is not especially noteworthy (Michener, ias although bee-pollinated plants such as le- umes constitute an important element of the fori (Goldblatt 1 1978). P protea study sites are indicated in Table à BP proteas used for comparative pur- poses included P. arborea Houtt., P. laurifolia Thunb. (both from Jonasnek), and P. repens (L.) L. (from Jonasplaats, see section on P. amplex- icaulis). These proteas are illustrated in Rourke (1980) PROTEA HUMIFLORA AND P. AMPLEXICAULIS Protea humiflora was studied during late win- ter-early spring (mid-August to mid-September) 1978-79 on J onaskop, a prominent mountain in the R y 50 km south of Worcester and 100 km east of Cape Town. On Jonaskop, P. humiflora forms dense stands along a restricted access road to the sum- mit. Two adjacent study sites (A and B), each 4 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 E2. Density, species composition, and composite home range data from live trapping grids in stands k. TAB of [iris humiflora at Jonasne 1978 1979 Site A Site B Site B Num- Num- Num- Adjusted Home ber Num- ber Num- ber Num- Range Length (m) of ber o ber * of — Species Jnd. Hec Comp. Ind. Hec. Comp. Ind. Hec Comp. N X Range Acomys subspinosus 1 5 2.6 20 95 3.35.1 7 33 5- 339 5s 393 (35.0547 Aethomys namaquensis 1 5 2.6 13 6252228 12° 37061: 0070 6 462 (27.5-58.0) Praomys verreauxi H 52 5289 4 20 7.0 4 35.4 (27.5-45.5) Rhabdomys pumilio 25 419 2658 17 81 29.8 1 5 5.0 17 37.8 (25.0-58.0) Elephantulus edwardii 3 14 5:3 3 346 (27.5-44.2) Totals 38. 181 37: eT 40 95 approximately 2,100 m?, were selected along this road at approximately 700 m (hereafter referred to as Jonasnek). During the 1978 field season the species composition, frequency, and cover were determined for the perennial plants on sites A and B. Both sites were divided into 10 m? quad- rats, each plant was identified and its position and cover plotted on graph paper and the loca- tion of each Sherman trap noted. Various observations were also made on P. amplexicaulis at somewhat higher elevations on Jonaskop (ca. 1,000 m) in an area known locally as Jonasplaats. This locality is relatively flat with many scattered individuals of P. amplexicaulis occurring along the east side of the access road (the west side of the road was burned in 1976). SMALL-MAMMAL TRAPPING SYSTEM METHODS At the P. humiflora study sites at Jonasnek a square grid of 100 Sherman live traps [45 X 45 m (2,025 m?), approximately 5 m trap distances] was used to establish the species composition activity, and movement patterns of the small- mammal community. Each trap was fitted with a switch mechanism and wired to a portable Es- terline-Angus 20 Channel event recorder that in- dicated exact capture times and trap positions within the grid. This system facilitated rapid trap checking while minimizing human disturbances to the grid area. Traps were baited with rolled oats and peanut butter and checked at sunrise, sunset, and at variable intervals of 2 to 8 hours over a 24-hour period. Upon initial capture, an- imals were identified, weighed, sexed, checke for pollen load, marked with numbered ear tags (or toe-clipped), and released at the point of cap- ture. The times and positions of recaptures were merely recorded, but animals were occasionally rechecked for pollen load. Rickart (1981) pro- vides further details. Small mammals associated with the other proteas listed in Table 1 were captured in Sherman live traps (occasionally snap traps) set in irregular transects around, under, or on the branches of flowering proteas. During the 1978 season, the grid system was operated successively at two sites. The first (Jo- nasnek A) consisted of a relatively level area and - the second (Jonasnek B) was on an adjacent, rocky, north- «facing, ru -30° slope (the warm, ary | slope in the 100 m from site A. Vegetation w was qualitatively similar on both sites, each having the same 34 : species dominated by dense stands of P. humi- flora and scattered individuals of Leucodendron salignum R. Br., with generally similar densities of herbaceous ground cover. Soil development was more extensive, however, on site A. The grid | was run continuously in 1978 for 121 hours at- UA 1983] WIENS ET AL. —NONFLYING MAMMAL POLLINATION 40 p eor Er 15 No. of Animal Captures (bars) (Sui) Buiuedo si8wo|J ‘oN UDƏW O Miis O 1200 1800 2400 0600 Clock Time FIGURE 1. Cumulative activity patterns of all five small mammals occurring on the Jonasnek study sites histatins and the opening patterns of P. humiflora flowers (line). site A and 193 hours at site B. In 1979 the grid distance of 5 m) for individuals with five or more Was operated only at site B for a total of 355 captures (Table 2). These values are probably hours. By the end of each trapping period, fewer underestimated because they approach the di- than 596 of the animals captured during a 24- mensions of the grid, which were relatively small hour period were unmarked. for this purpose, particularly with respect to the Cape striped field mouse (Rhabdomys) and Ae- RESULTS thomys. Figure 1 shows the cumulative activity pattern determined from capture times of indi- bl Tue grid-trapping data are summarized in Ta- ossis of all five species. Because some of the € 2. The 1978 data show considerable differ- captures were probably artifacts due to the quick ences in species composition between the two retrapping of previously released animals, recap- id sites, Suggesting that microgeographic vari- tures that occurred less than one hour after the ation affects the species composition of the small- release were eliminated from the histogram. The uaa community within individual stands of diurnal activity shown is almost exclusively due humiflora. The data from 1979 also show a o Rhabdomys. The remaining species are pre- Profound reduction in overall densities on site dominantly nocturnal. Of the 376 total captures, - Only densities of the Namaqua rock moUS€ — 535 (61 706) occurred during the nocturnal-cre- A» € appeared unchanged, while Ver- | cular period from 1800 to 0700 with highest TLS mouse (Praomys) and the elephant shrews activity levels occurring just after dusk. Onion were absent. Praomys was not re- Corded anywhere in the study region the second year, FLORAL MORPHOLOGY OF NMP PROTEAS Mean home csumiated from recapture an by calculating ad- du: between the fart The inflorescence is a many-flowered, bowl- two points of capture plus the average dedica shaped head surrounded by dark, fleshy invo- PROTEA AMPLEXICAULIS— A TYPICAL NMP PROTEA ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 3 IGURES 2-5. Floral and fruiting features of P. amplexicaulis and P. humiflora.—2. Flowering head (P. a.)— 3. Individual floret, arrows indicate pollen presenter and nectar reservoir (P. a.)—4. SEM photo of style tip showing stigmatic slit and grooves of the pollen presenter region (P. a.) (X ca. 2 0).—5. Fruiting head with pistils cut transversely to expose sterile ovaries and those containing endosperm (arrow) (P. h.). lucral bracts (Fig. 2). The peduncle is stout and short (3-4 mm), yet flexible, the heads appearing sessile along the branches. The heads are often borne near ground level (geoflorous) and are typ- ically deeply hidden beneath dense, overlying fo- liage (cryptic). The heads are similar to P. humi- flora in appearance and habit, but are generally more cryptic. 1983] The individual flowers (florets) comprising protea heads are unusual (Fig. 3). The uniovulate pistil is surrounded by four dull-white, non- showy perianth segments, each bearing a single, sessile anther. The unique feature of the floret, however, is the extremely wiry, yet flexible style, which readily withstands rough treatment. It has the dual function of pollen dispersal and recep- tion. Pollen is deposited prior to anthesis onto a specialized, longitudinally grooved, apical region of the style known as the pollen presenter (Fig. 4). Although pollen covers most of the distal portion of the style, it does not initially reach the stigma itself, this being a highly reduced micro- Scopic groove at the apex of the style (Fig. 4). Physical transfer of pollen is thus necessary even for self-pollination (see section on Genetic Com- patibility). Stylar presentation of pollen also oc- curs in Asteraceae and the Campanulales, but the details differ. In proteas the base of the style ruptures the lower perianth segments as it grows laterally from the perianth envelope during late bud develop- ment. Eventually only the base of the style and the pollen presenter remain enclosed by the peri- anth segments and the surrounding anthers. The rest of the style forms a bowlike structure outside the perianth just prior to anthesis (Rourke, 1980). Nectar secretion occurs at this time (see section on Nectar Production). Although anthesis, i.e., the emergence of the pollen presenter from the enfolding antl 1 perianth seg is mildly explosive, the pollen, because it is sticky, is not dislodged from the pollen presenter. In P. am- Plexicaulis and other NMP proteas examined, the three united perianth segments form a nectar reservoir about 10 mm below the stigma (Fig. 3). Because the flowers are in close proximity, however, the nectar often pools, particularly around the bracts. ADAPTATIONS OF PROTEA FLOWERS FOR POLLINATION BY NONFLYING MAMMALS The basic floral structures of proteas inferring pollination by nonflying mammals were summa- "ized by Wiens and Rourke (1978) and discussed and compared to Australian Proteaceae by Rourke and Wiens (1977). These structures as illustrated for p. amplexicaulis include: (1) bowl- Shaped heads borne on short (3-4 mm), stout paies, often with the outside of the bracts ark-colored, (2) copious, sucrose-rich nectar oe with a high (36%) total carbohydrate MPposition, (3) often inflexed, wiry styles ca. WIENS ET AL. -CNONFLYING MAMMAL POLLINATION 7 30-40 mm long, (4) cryptic, geoflorous, axillary positioning of the heads, and (5) a distinctive “yeasty” odor. In contrast, BP proteas produce (1) cylindrical heads with brightly colored bracts, (2) copious, hexose-rich nectar with a low (20— 25%) total carbohydrate composition, (3) straight styles ca. 60-90 mm long, (4) conspicuous, brightly colored, terminally borne heads, and (5) no obvious odor. The following prot hould be added to those previously suggested by Rourke and Wiens (1977) as exhibiting morphologies consistent with non- flying mammal pollination: P. caespitosa Andr., P. convexa Phill., P. cryophila Bolus, P. denticu- lata Rourke, P. effusa E. Mey. ex Meisn., P. pen- dula R. Br., P. piscina Rourke, P. pruinosa Rourke, P. recondita Buek ex Meisn., P. roupel- liae Meisn., P. tenax (Salisb.) R. Br. The follow- a P proteas do not have cryptic heads although they are geoflorous, (e.g., P. cryophila), while one is cryptic but not geoflorous (P. re- condita). Some proteas that are neither cryptic nor geoflorous [e.g., P. nana (Berg.) Thunb., P. pityphylla Phill., P. pudens Rourke, P. witzen- bergiana Phill.] may also be pollinated by nonfly- ing mammals, but verification is needed. Straight, as well as inflexed, styles probably also occur in many presumably NMP proteas; but the rela- tively short styles (ca. 30-50 mm) that maintain an effective stigma-nectar distance of about 10 mm are probably most important for a func- tional rostrum-stigma “fit” (Wiens & Rourke, 78). — Mo Protea cryophila is an impressive exception to the short-style character, yet it appears to prove the rule. Although the styles are ca. 80-90 mm long, the nectar secreted in mature buds is not retained at the point where the style arches out of the perianth tube, as occurs in P. amplexi- caulis or P. humiflora. Instead, the tube is filled with nectar for about 70—80 mm up the tightly stretched perianth tube and approximately 10 mm below the stigma. At this point the tube ends and the perianth flattens out into a strap-shaped structure across which the nectar does not mi- grate, and where it subsequently forms a nectar droplet. A captive spiny mouse (Acomys), when presented with a head of P. cryophila, foraged for nectar among mature buds at this level and contacted previously opened stigmas. Thus the critical stigma-nectar distance of ca. 10 mm is maintained. The structure of the nectar reservoir in P. am- 8 ANNALS OF THE MISSOURI BOTANICAL GARDEN plexicaulis and P. humiflora may also be im- portant in pollination. In these species it forms a shallow, “‘troughlike” structure (Fig. 3), which could facilitate nectar lapping by small mam- mals. In a BP protea (e.g., P. repens) the nectar reservoir, by contrast, resembles a well at the base of the style into which a bird’s beak or tongue could be readily inserted and the stigma-nectar distance ratio approximates the length of the en- tire style (and also the length of the pollinating bird’s beak and extensible tongue). The cryptic positi g of the heads in a num- ber of NMP proteas bonds further description. Different strategies are involved: (1) interior cau- liflory—heads generally borne along older stems densely covered with overlying shoots (P. am- plexicaulis, P. humifloray, (2) interior geoflory— heads borne at or very near ground level and generally covered with overlying shoots, espe- cially in older plants [P. cordata Thunb., P. sub- ulifolia (Salisb. ex Knight) Rourke]; (3) interior enduly—heads borne on pendulous branches generally hidden by overlying shoots, with heads often drooping to near ground level (P. sulphurea Phill., P. witzenbergiana Phill.); (4) exterior ter- minal—heads enclosed by large surrounding bracts (P. recondita, P. foliosa Rourke). Why crypsis evolved is unclear; but a possible explanation is the reduction of occasional nectar robbing by flower birds that presumably cue vi- sually. Dr. E. Granger (pers. comm.), however, suggests that crypsis might reduce predation on small-mammal pollinators by nocturnal raptors (owls), of which there are a dozen species in southern Africa (Oatley, 1971). Small mammals foraging on exposed flower heads should be more vulnerable to aerial predation than animals for- aging inside a foliage cover. These four forms of crypsis also occur in the proteaceous genera Banksia and Dryandra of Australia (George, 1981), which also has a large owl fauna (Mor- combe, 1974). STIGMA MORPHOLOGY METHODS Because of its reduced nature, the stigma mor- phology of the NMP proteas P. humiflora and P. amplexicaulis and the BP proteas P. cyna- roides (L.) L. and P. repens was studied with a Hitachi 450 scanning electron microscope to de- termine whether its structure might provide evi- dence relating to more efficient pollination by particular animals. Proteaceae are generally prot- [Vor. 70 androus and we suspected there might be differ- ing periods of slit opening or receptivity. Seven specimens were collected every 3 hours for a pe- riod of 24 hours. Both individual flowers and entire flowering heads were collected at each sampling, and special care was taken not to dis- turb the flowers prior to fixation. Styles were fixed for 24 hours in one of the following: (1) liquid nigrogen (N;), (2) super- cooled 95% ETOH, (3) a mixture of 50% for- malin and 9596 ETOH, or (4) a solution of mag- nesium phosphate (MgPO,). Styles of individual flowers were prepared by one of the following methods: air dried directly after removal from the fixative, washed in 9590 ETOH prior to drying, duced no detectable changes in stylar structure. Stigmatic morphology was compared among flowers (1) fixed at different times of the day, (2) occurring at varying positions on the flowering head, and (3) representing different age classes. RESULTS Observation of approximately 300 stigmas of P. humiflora and 200 stigmas from the other species (see above), plus P. minor (Phill.) Comp- ton, produced little or no evidence that the stig- matic slit ever opened to any appreciable degree (Fig. 4). Neither did we observe structural changes suggesting differing cycles of receptivity, nor evi- dence indicating differential pollination success - by a certain group of pollinators. Nonetheless, further studies of this subject might prove inter- - esting. [ TEMPORAL PATTERNS OF ANTHESIS AND NECTAR PRODUCTION METHODS To learn whether anthesis and nectar produc- tion are correlated with the activity patterns of | small mammals, we determined the time of flow- | er opening for P. humiflora under field condi- | tions by counting and marking the number of - newly opened flowers at the Jonasnek site A. Flowers on six heads were counted at three-hour intervals over a period of nine days. The erect | styles that identify open flowers were marked with red fingernail polish at each time check. Field observations to determine time of flower opening were also made for P. amplexicaulis at | Jonasplaats and for P. cryophila at the Sneeuberg — site, but at different time intervals. | 1983] TABLE 3. X No. flowers opening/head between prescribed time intervals for nonflying mammal and bird pollinated proteas. WIENS ET AL. —NONFLYING MAMMAL POLLINATION Pollination Species System 0800—0900 1700-1800 2100-2200 P. amplexicaulis NMP 5.4 (N = 24) 2.6 (N = 24) 9.5 (N = 24) P. cryophila NMP 47.4 (N = 50) 4.9 (N = 40) — P. effusa NMP 9.5 (N =31) 4.0 (N = 34) — P. recondita NMP 8.0 (N 4) 3.5 (N = 4) — P. sulphurea NMP 22.6 (N — 46) 6.2 (N — 48) — . arborea BP 19.3 (N = 26) 5.2 (N = 48) 15.7 (N = 26) P. laurifolia BP 9.7 (N = 28) 8.0 (N = 33) 3.7 (N = 27) P. repens BP 0.36 (N = 22) 12.8 (N = 23) 1.1 (N = 22) Additional observations on P. effusa, P. re- condita, and P. sulphurea, were made under lab- oratory conditions, at room temperature, and without special lighting regimes. Flowering heads, which recover easily from wilting and maintain flowering function for a number of days following removal from the plant, were stored in closed plastic bags for 24—48 hours, after which we placed the peduncles in water and cut away the erect styles of open flowers with scissors. Observations were made at 12-hour intervals and the styles of newly opened flowers cut away after each obser- vation. In addition to NMP proteas, three BP proteas (P. arborea, P. laurifolia, P. repens) at the Jonasnek and Jonasplaats sites were studied for comparative purposes. We measured nectar under a dissecting micro- Scope from freshly picked flowering heads using a 15 ul capillary tube. Magnification was nec- essary to ensure that the nectar droplets from the tightly grouped florets had not pooled. RESULTS Time of Anthesis Protea humiflora clearly shows a maximum rate of flower opening between 1800 and 2100, ; y thesis, but the periods cannot be bracketed as Precisely because the observations were not reg- ularly made at short intervals as in P. humiflora (Table 3). Protea amplexicaulis, however, would appear to have a pattern of anthesis similar to P. humiflora, and judging by casual observations of NMP proteas made during the evening hours, it seems likely this is a general pattern. mong BP proteas (Table 3), P. repens has a midday flower-opening pattern, whereas in P. laurifolia Thunb. more florets open in the early morning (0930) and also between 1800 and 2100. The actual time at which anthesis occurred in the newly opened flowers observed at 0930 is unknown because data are lacking for the critical periods (Table 3). Protea arborea appears to have a largely noc- tural anthesis, for which we have no apparent explanation. Presumably the species is pollinated by the Cape sugar bird, as are the other two BP species. No sugar birds were ever observed on P. arborea in the study areas, however, they were common in nearby (ca. 1 km) stands of P. /auri- folia. Nectar Production No nectar production could be detected in the heads of P. humiflora utilized for determining periods of anthesis. Nor were there consistent patterns of nectar production noted in any of the heads on the study areas. A few scattered obser- vations, however, provide some information. Freshly secreted nectar was observed in P. humi- flora flowers on three separate days (Aug. 16, 26, 30) between 1630 and 1800 hours. These days were relatively cold, windy, and g lly stormy, but were without rain during the previous 12 hours. These preliminary observations suggest that nectar secretion is initiated in P. humiflora in late afternoon or early evening during periods of relatively low daily temperature and just prior to the period of greatest small-mammal activity. What appeared to be freshly secreted nectar was also observed in P. cryophila at 2030 hours on Feb. 13, but secretion was confined to rela- tively few flowers on a single head among the 10 under observation. In this instance, however, the [Vor. 70 ANNALS OF THE MISSOURI BOTANICAL GARDEN TABLE 4. Total Sugar tent of (g solute/100 l ) fi fly g l d bird p ll ted proteas. X Total Pollination Sugar Species System Content N A1) Range P. amplexicaulis NMP 37:2 42 6.3 27.9-47.9 P. cryophila NMP 33.3 13 12 25.2-49.4 P. humiflora NMP 37.8 59 6.9 28.2-65.4 P. arborea BP 18.8 5 — 16.8—20.4 P. laurifolia BP 24.4 20 1.0 20.3-26.6 P. magnifica BP 20.6 16 2.9 16.2-24.8 P. repens BP 18.8 45 1.8 14.6-23.6 day was sunny and warm as is common during midsummer. In P. angustata R. Br., P. cryophila, and P. humiflora nectar was first secreted from mature buds, i.e., while the pollen presenter was still enclosed by the perianth segments, although the style had generally already arched laterally away from the — Sepe most of its length. This should n , however, because the Misi styles of open Bowers easily contact the rostra of small mammals foraging for nectar among the outer few whorls of mature buds. Fol- lowing warm days, however, the nectar reser- voirs of some previously opened flowers in P. amplexicaulis, P. cryophila, and P. humiflora ap- peared to develop a moist film by 2000 hours, but no obvious nectar build-up was observed. Laboratory studies such as those by Cowling (1978) may be necessary to obtain an adequate understanding of temporal patterns of nectar se- cretion. NECTAR COMPOSITION METHODS The nectar of NMP and BP proteas was ana- lyzed for (1) total sugar content (g solute/100 g solution), (2) percentage of different sugars com- prising the carbohydrate fraction, and (3) amino acid content. The nectar was extracted from the heads with a capillary tube and the total sugar content was measured with an AO Goldberg re- fractometer (Model 10923) corrected for tem- perature. Because of the high total sugar content of NMP protea nectar, it was often diluted with appropriate parts of distilled water in order to retain the value on the refractometer scale. The values were then corrected by the appropriate dilution factor. Only apparently freshly secreted ponents of the carbohydrate fraction and the paper, quickly dried, and later chromatographed by I. and . Baker following methods they cua described (Baker, Opler & Baker 1978; Baker & Baker, 1979). Considerable care was taken to reduce the possibility of pollen contam- ination in the samples used for determining ami- no acid content. Because freshly secreted nectar was rarely available for analysis of total sugar content, nectar ptt had accumulated in e. h was utilize er flowers is more variable than that freshly se- creted, the sample size was increased. RESULTS Total sugar content of nectar. The three NMP proteas analyzed all have nectar with total sugar content in the mid-thirties (X — 36.196) (Table 4). The four BP proteas have nectar with total sugar content ranging from the high teens to mid- twenties (X = 20.7%), which is typical for most bird-pollinated flowers. There is virtually no overlap in the values between the two BP proteas and the NMP proteas and the differences are statistically highly significant (P < .01). The dif- erence in total sugar content of the nectar in the two groups should be an important distinguish- ing feature between BP and NMP proteas. Carbohydrate composition. The nectar of NMP proteas are generally *sucrose-rich," i.e. the ratio of sucrose to glucose-fructose is >0.5- (Baker & Baker, 1979) (Table 5). Thus all the NMP proteas analyzed for carbohydrate com- position may be characterized as sucrose-rich, with the exception of P. angustata (0.326) which 1983} WIENS ET AL.—NONFLYING MAMMAL POLLINATION 11 TABLE 5. Carbohydrate composition (mean proportions) in nonflying mammal and bird pollinated protea nectar.* Ratio Species and (Sucrose/ Pollination Type Melezitose Maltose Sucrose Glucose Fructose Glu + Fru) P. amplexicaulis (NMP) .036 .029 345 330 .260 .593 P. humiflora (NMP) .019 .021 .394 .364 .203 .728 P. cryophila (NMP) 0335 .015 .276 .343 .333 .410 P. repens (BP) .017 .042 ,123 531 .286 152 P. angustata (NMP?) .049 275 2I .419 .283 .326 a Analyses kindly provided by Prof. H. G. and I. Baker from samples supplied by the authors. is marginal in this respect. Protea repens, a BP (0.152), which is confirmed by other studies (Mostert et al., 1980; Cowling, 1978). Cowling (1978), however, reported that some presumably BP proteas (P. /ongifolia Andr.) also have high Sucrose concentrations. The importance of the differences in the small amounts of melzitose and maltose present in the samples is difficult to as- sess, although P. angustata has considerably higher values than the other species. Amino acid composition. The significance of the amino acid composition of the nectar is dif- ficult to evaluate (Table 6), but the relatively small quantitites present suggest they have no Important nutritional value for small mammals. This should be expected in non-coevolved sys- tems, where the nectar is apparently not an es- sential component of a pollinator's diet (see sec- tion on Reward). The amino acids may, however, impart taste to the nectar (Baker et al., 1978). Nectar odor. The heads of all NMP proteas emit a “yeasty” or fermented odor, as previously mentioned, but two variations are apparent. Pro- tea amplexicaulis, P. cryophila, P. humiflora, P. recondita, and P. sulphurea superimpose a sweet- ish scent to the basic yeasty theme, whereas P. angustata, P. restionifolia (Salisb. ex Knight) Ry- croft, and P. scabriuscula Phill. produce a pun- gent odor reminiscent of rancid butter. Nectar freshly extracted from the heads of P. amplexi- caulis and P. humiflora retains the odor, indi- cating that the volatile fraction occurs in the nec- tar itself, jien heads of P. effusa, P. recondita, and P. Nip maintained in water under laboratory ol wesen emitted perceptively stronger odors iga A similar situation appeared to exist in field ex: amplexicaulis and P. humiflora under nditions. Although these observations are subjective, they complement the data on both flower opening and nectar secretion. That the odor of P. amplexicaulis and P. hu- miflora attracts small mammals first became ap- parent when flowering heads were placed in the cages of captive animals while they were in their sleeping tube. Although the flowering heads were not visible from the tubes, the animals usually emerged within a few minutes, sniffed the air with upraised snouts, and then proceeded di- rectly to the heads and began to forage. FLORAL AND NECTAR PREDATION The loss of reproductive potential in protea through predation was not a major consideration in this study, but several observations warrant mention. The fleshy bracts and styles of flowering heads in the species studied occasionally showed clear evidence of being chewed by small mam- mals, but in P. amplexicaulis this occurred on only about 1-296 of the heads. Thus predation by chewing is probably of little importance, es- pecially since fertilized ovul ight develop if the styles were subsequently destroyed. Cowl- ing (1978) presented evidence that the bracts might act as a carbohydrate sink, thus providing a possible reason for thier occasional exploita- tion by small mammals. A more unusual and significant form of pre- dation in proteas is the removal of entire flow- ering heads. The extent of such predation is in- dicated by the number of heads removed from P. humiflora on study stie B in 1978. Quadrat seven (10 m?) had 307 flowering heads removed from the 45 individual plants occurring on the quadrat; quadrat ten (10 m?) had 164 heads re- moved from 13 P. humiflora plants. Since P. humiflora averages 17.3 seeds per head (Table 9), about 8,148 potential seeds were lost. The num- ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 TABLE 6. Amino acid composition of nonflying mammal and bird pollinated protea nectar, graded on a 1-6.2 scale of increasing concentration from Amino Acid P. amplexicaulis P. cryophila* P. angustata: P. humiflorac P. repens (NMP) (NMP) (NMP?) (NMP) (BP) Alanine 1 2 1 1 1 Arginine — tri ? — p Asparagine 2 1 1-2 1 — Aspartic — — — — oe etc. tr t 2 1 — Glutam 1 2 — 4 1 onc 2 2 1-2 1 2. Glycine 1 3 1 2 2 Histidine — — E ux ns Isoleucine 1 1 — 2 1 Leucine — l l 1(?) — Lysine E 1 — d 1 Methionine 1 2 et 2 1 Phenylalanine == 2 1 1 — Proline 2 1-2 2 1 2-3 rine l 2-3 1 1 2 Threonine — ja ET e" os Tryptophan 1 2 -— oe Tyrosine 2 1 — — 2 Valine 1 1 2) 1 ? Analyses kindly provided by Prof. H. G. and I. Baker from samples obtained by the authors. e 5 = .1212 ug/ul. c 6 = .2424 yg/ul. 4 tr — trace. ber of heads removed by predators in these quad- the immediate study area, ie k overall predation of entire heads was pat he pistons hs heads removed through this form of predation were not obviously chewed, s might be expected from rodent activity, in- stead the heads appeared torn from the plants. Furthermore, the heads were often concentrated in discrete piles ot ind in 15; and —_ showed ge. These observations suggested removal by an organism capable of pulling the heads from the plants. The chacma baboon (Papio ursinus Kerr), which oc- curs in the study area (pers. obs.), is the most likely predator because it forages in this manner (Roberts, 1951). Similar predation was also ob- served on a BP protea, P. repens. Since the heads are either severely disturbed or torn apart (but not eaten) it appears most likely that the pred- ators are seeking either the sweet, copious nectar and/or the large scarab beetles (Anisonyx ursus F.) that occasionally occur in the heads in large numbers. Floral predation by primates is known flora, in the Neotropics (Mori et al., 1978; Janson et al., 1981) Nectar and pollen robbing may be a form of — serious predation in both NMP and BP proteas. When the standing crop of nectar averages sev- - eral milliliters, it offers an unusually rich energy - resource and is heavily exploited by numerous insects. Mostert et al. (1980) found a total of — 2,215 insects in 20 heads of P. repens. The largest | percentages consisted of ants (19%), beetles (67%), — and flies (12%). Whether these insects, which also occur in - NON P. amplexicaulis and P. humi- is un- known. Bees also extract considerablé nectar and | pollen from the heads (Table 15), but whether this should be considered predation is difficult to assess because they probably also contribute to pollination (see sections on Nectar Com- - sumption by Bees and Selective Exclosure Ex- - periments). Predation by ants should be expected in such a rich nectar source, but few ants were noticed - around flowering heads of most proteas observed _ in this study. In P. cryophila, however, the heads - were often heavily infested, especially at night. _ 1983] Mostert et al. (1980), however, reported that nearly 20% of the insects on flowering heads of P. repens were ants. SMALL MAMMAL AND ABIOTIC DISPERSAL OF PROTEA POLLEN METHODS Many ofthe small mammals captured in flow- ering stands of NMP proteas were tested for the presence of pollen on their rostra, and in some inst the feces were al ined for pollen. The presence of pollen on the rostrum was tested by rubbing the area with gelatin blocks (several mm?) containing basic fuchsin stain (Beattie, 1971), but neither the sampling proce- dure nor the size of the gelatin blocks was stan- dardized. The gelatin block was then melted on a microscope slide and spread under a cover glass. Pollen is readily captured and stained by this method and the slides are essentially permanent. Fecal pellets were also analyzed for pollen using the same general technique. To avoid possible imals and partially dissolved in water to soften and spread the fecal material on the gelatin blocks. To eliminate the remote possibility that these pollen loads originated by the chance accumu- lation of abiotically dispersed pollen, the follow- ing materials in dense stands of the NMP protea, P. humiflora (Yonasnek—site A), and a BP pro- tea, P. laurifolia (below Jonasplaats), were ana- lyzed for the presence of (presumably) wind-dis- Persed protea pollen: (1) rocks and leaf litter between protea plants, (2) living leaves not closer than ca. 20 cm from flowering heads, and (3) the bracts of flowering heads. RESULTS Some protea pollen was found on the rostrum of all animals examined, but the amount varied widely, from only a few scattered grains to many thousands (Table 7). Of the 151 animals exam- Re Sin protea pollen, only 15 samples contained hsti àn 100 grains of protea pollen and only 9 «58s than 50. The samples typically also con- hes à few non-protea pollen grains, but they “raged only 3.1 per sample; the highest num- T Was 44. No average for the number of protea Pollen grains was obtained because thousands of Brains were present on many of the slides. The WIENS ET AL.—NONFLYING MAMMAL POLLINATION 13 LE X No. pollen grains in samples (N = 2) collected from various objects in a stand of flowering P. humiflora. Protea Non-protea Pollen Pollen Non-protea exposed leaf litter 6 56 Exposed rocks 2 11 P. humiflora bush? 33 82 P. humiflora heads? 492 75 ? Samples collected from leaves <2 dm from flow- ering heads. ^ Samples taken from bracts of flowering heads; pollen occurred primarily in dense clusters on the tri- chomes of the bracts. number of pollen grains on the rostrum is a func- tion of how recently the animal foraged in rela- tion to the number of groomings and wet-preen- ings. Snap-trapped animals often carried higher pollen loads than live-trapped animals; presum- ably the latter groom and preen while in the traps. During wet-preening, pollen should be ingested and feces did contain protea pollen, often in large numbers. Even in dense stands of protea, non-protea pollen usually predominated on the various ob- jects examined, except on the bracts of flowering protea heads (Table 7). If protea pollen accu- mulated on the rostra of small mammals by chance, then the concentration should be pro- portional to that of non-protea background pol- len which is not the case. Only visitation to flow- ering heads of protea can adequately explain both the size and composition of the small-mammal pollen loads. Small mammals may also visit the flowering heads of BP proteas, e.g., P. laurifolia. In this species evidence of light chewing is occasionally evident on the bracts of the heads, but no indi- cation of intensive rodent activity was observed on any BP protea. We sampled pollen on the rostra of five gerbils (Tatera afra Gray) and a single Rhabdomys trapped in a stand of P. /auri- folia with no NMP proteas within at least 500 m. These animals all possessed low counts of protea and non-protea pollen on their rostra. Three of the animals sampled showed approxi- mately equal numbers of protea and non-protea pollen grains, but had only a low total pollen count (<15 grains/sample). Two of the gerbils, however, showed approximately three to four X Cal/g/hr for Aethomys TABLE 8. Energetic relationships between P. humiflora and small mammals (Jonasnek sites). X Standing (Active & Grooming) No. Protea X No. Heads/ Plant (N = 19) X No. Fls./Head (N = 12) Crop Nectar/Head 20) (ml) (N = 11) X Nectar/ No. Animals Sites A & B X No. Heads/ Quad. Site B Nectar Plants Sites A & B Fl. (ul) (N ANNALS OF THE MISSOURI BOTANICAL GARDEN 4) 11.83 (N = Sites A & B Cal/g = 38 1979 = 58 1978, A A B 56.6 1979, B= 5.2 303 1977, A 3,693 603 = 20 1979, B [VoL. 7053 times more protea than non-protea pollen grains and also relatively larger pollen loads, although these were still small in absolute terms (13:55, 18:68). It is unlikely that the protea pollen in these two samples could have originated from sources other than direct visitation to the heads, since the stickiness of BP protea pollen is similar to that of the NMP proteas. The entire pattern - of small-mammal visitation to various flowers needs further study and can be initially ap- proached by simply measuring pollen loads. POLLINATION ENERGETICS: P. HUMIFLORA AND AETHOMYS METHODS One object of the study was to determine the - potential contribution of NMP protea nectar to — ity. The nec humiflora, Jonasnek site B (for which the small- — mammal composition and density are known), was analyzed in terms of the following param- - eters: (1) the potential volume and caloric con- - tent of P. humiflora nectar, (2) the metabolic rate of Aethomys as a representative small mamm (3) the maximum amount of nectar that Aetho- mys will consume during a given period when - maintained on a strict nectar diet, and (4) the density of the small-mammal community. The caloric content of P. humiflora nectar was the accumulated nectar. This procedure obvious- ly leaves considerable nectar on the head and is thus highly conservative. Individual flowers with | apparently freshly secreted nectar were utilized to determine nectar volume per flower. The annual amount of protea nectar on the- study site was calculated by determining the mean number of flowers per head, and the number of heads occurring in several randomly determined quadrats on grid B as a representative distribu- tion for the entire grid (Table To approximate the amount of nectar Aethol mys might ingest, a single animal was fed a strict diet of P. humiflora nectar (4496 sucrose equiv- alents) from a graduated pipette and the con- | avain 1983] sumption recorded at 12-hour intervals (the quantities of nectar required for the experiment precluded a larger sample of animals). RESULTS The amount of energy in the annual nectar crop of P. humiflora can be roughly approxi- mated from the data in Table 8. Multiply the amount of nectar produced per flower (8.8 ul) by the average number of flowers per head (303). Multiply this result (2.7 ml) by the number of heads per 10 m? quadrat (58) and multiply that sum (156.6 ml) by 25 (the total number of quad- rats). Thus, roughly 3,915 ml of nectar can be expected from the P. humiflora plants on site B. One milliliter of nectar yields 0.31 g of solids, and one gram of solids produces 3.7 Calories. Thus the 3,915 ml of nectar on the study site yields 4,490 C. Aethomys requires approximate- ly 14.1 C. per 24 hours. The 20 animals occurring on grid B in 1979 would require 282 C. per day to satisfy their basic energy requirements. If 50% of the nectar is lost to predation, sufficient nectar would remain to supply the energy requirements of the small-mammal community for 8.0 days from an approximately 45 day flowering period. Because of the many variables involved, how- ever, this calculation provides only a crude es- umate of the actual energy relationships. It does, however, indicate that only a small fraction of the annual energy needs of the small-mamma community can be met by protea nectar. The single Aethomys maintained on P. humi- flora nectar consumed an average of 6.55 ml/ day, which was approximately equal to the amount of nectar taken during the first feeding bout of about 10 minutes at the initiation of the experiment (6.4 ml). Predictably, a strict diet of P. humiflora nectar results in progressive weight ag (16.196 of body wt.) and death (in five ays). SEED SET IN PROTEA METHODS | Seed set was determined by cutting transverse- : through the fruiting heads at mid-ovary level b. Sharp, thin blade. This is possible because e trinte f 1 = s p TA firmly attached to the receptacle of the head for Mos years. Interestingly, the individual fruits ben. attached to the head during this period er or not they develop into viable seed. WIENS ET AL. —NONFLYING MAMMAL POLLINATION 15 Furthermore, viable and sterile achenes do not obviously differ morphologically. When the heads are cut transversely at mid-ovary level, however, viable, endosperm-containing seeds are readily identifiable by their soft, milky-white texture; whereas seeds interpreted as sterile have a dull- white, dry, fibrous content (Fig. 5). Seed set in £A. OlaChenes this study per head containing endosperm expressed as a percentage of the total number of flowers. RESULTS Seed set is generally low in both BP and NMP proteas, and in Proteaceae in general. Excluding species with small sample sizes (<10), average seed set ranges between 6 and 15%, with the exception of P. recondita, which developed 18 and 29% seed set from two populations (Table 9). Although the sample sizes are not large for most of the species (ca. N = 10), we found no significant variation in the 1978 seed crop for P. humiflora when more than 10 heads were sam- pled. While the annual seed set in P. humiflora differed significantly between 1977 and 1979 (P = .01) and between 1978 and 1979 (P = .01), the absolute change itself was not great. No signifi- cant difference was observed between seed set in 1977 and that in 1978 (P = .18), but the annual rainfall differed greatly. Although the percentage of seed set may not vary greatly from year to year, the number of heads per plant may well differ appreciably, but few data are available on this point (Table 8). Little evidence of insect predation was noted in fruiting heads. Larval insects were infrequent- ly observed, and were only occasionally abun- dant in a particular head. Such predation seems unlikely to affect the statistical data significantly. Finally, what factors produce the typically low seed set (1—3096) in these and other proteas (Ta- ble 9)? Lack of pollination and resource avail- ability are commonly accepted explanations for low seed set, but Casper and Wiens (1981) dem- onstrated that in Cryptantha (Boraginaceae) a significantly greater number of embryos are ini- tiated than develop into seeds. More recent stud- ies of Cryptantha (Casper, 1982) showed that embryo reduction cannot be attributed to inad- equate pollination nor to resource availability, thus suggesting genetic control. The consistently low seed set ia both BP and NMP proteas, and the apparent occurrence of similarly low seed set in Australian proteaceous ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 TABLE 9. Seed set in nonflying mammal and bird pollinated proteas. X No. Pollina- X No. Seeds/ % Seed tion Species Flowers/Head Range Head? Range Set Type Study Site P. amplexicaulis 157 (N = 15) 136-176 IL5 0-42 723. NMP Jonasplaats I dn 244 (N = 2) 221-267 9.5 6-13 3.0... BP Jonasnek P. cryophila 1,024 (N = 12) 749-1,217 154 90-227 15.0 NMP Sneeubergnek P. effusa 157 (N = 10) 136-178 17.4 2-30 10.8 NMP Murray Farm P. humiflora 303 (N = 12) 248-373 173 0—95 60 NMP Jonasnek P. laurifolia 283 (N = 10) 230-346 19.4 0-55 TO BP Murray Farm P. magnifica 374 (N = 2) 349-398 5 4-6 E33 pP Jonasplaats P. pendula 157 (N = 10) 136-178 17.4 2-30 11.1 NMP(?) Murray Farm P. punctat 138 (N= 11) 112-155 Ix 2-32 9:0: BP Sneeubergnek P. recondita (1) 258 (N = 10) 118-379 75:9 23-149 293 “NMP Sneeubergnek P. recondita (2 393 (N = 11) 338-436 69.6 18-148 18.1 NMP ay Farm P. scolependrifolia 194 (N = 13) 160-260 16.9 0-51 90 NMP Jonasplaats P. sub lia 105 (N = 21) 88-128 15:5 0-51 14.3 NMP Jonasplaats P. sulphurea 500 (N = 8) 416—588 15,4 0-53 3.0 NMP Ouberg Pass aN same as column | unless otherwise indicated. genera (Rourke & Wiens, 1977; Carpenter & Recher, 1978b) also suggest that genetic factors may play a role in controlling seed set in Pro- teaceae, but experimental verification is needed. GENETIC COMPATIBILITY METHODS Protea humiflora was tested for genetic com- patibility by enclosing individual flower heads in bags of lightweight, small-mesh nylon with a were partially pulled from the heads (presumably by mice attempting to gain access to the highly nectarous heads). Sepang: experiments with these species are difficult to under field con- ditions. Pollen must be physically moved to the stigma to effect pollination (Fig. 4). In 1978 we attempt- TABLE 10. Seed sets of enclosed heads of P. humiflora tested for genetic incompatibility. ed to cross-pollinate flowers by reciprocally ex- changing pollen-laden exclosure bags between heads of different plants. Immediately after transfer the bags were manipulated to distribute — were manipulated inside ihe pollen-laden exclo- - sure bags daily for four days in an attempt to spread pollen over the stigmatic slits of newly opened flowers. The remaining heads were not manipulated. The results were analyzed statis- - tically with the Mann-Whitney U test and x^. RESULTS In 1978 no statistical differences were ob- - served in seed set between selfed and crossed | plants (P = .19), but the sample size was small - (N = 6) (Table 10). Furthermore, in 1978 both — the selfed group and the outcrossed group pro- - duced significantly fewer seeds than the control group (P « .01) H In 1979 the difference in seed set between the - enclosed heads which were manipulated, and | Aoc SERE GUION E NNNM ME Unmanipulated Heads Manipulated Heads Artificially = = Outcrossed nee Year X 96 Control X % Control Heads X Control 1978 4.2 (N = 6) ao IZ 18 (N = 103) 1979 3.4 (N = 12) .16 2.7 (N = 9) .13 21 (N = 35) 1983] WIENS ET AL.—NONFLYING MAMMAL POLLINATION 17 those which were not, was difficult to evaluate. In the non-manipulated group (N = 12) only two heads produced seed, one with 42 seeds and the other with six. In the manipulated group (N = 9), however, six heads produced seed, but none more than 10. The single head with 42 seeds in the non-manipulated group affects the statistics when using the Mann-Whitney U test. Perhaps cross-pollinating insects or some other pertur- bation produced this exceptionally high seed set. To avoid this problem, we compared the per- centages of heads that produced any seed in the manipulated group (60%) and in the non-ma- nipulated group (14.3%) by the McNemar vari- ation of x?. This shows that the percentages are significantly different (P < .01), and suggests that self-pollination (but not necessarily autog- amy) is successful to a limited extent in P. hum- iflora. The data are difficult to evaluate because all heads placed under nylon enclosures (including those artificially crossed in 1978) typically pro- duced significantly fewer seeds than did the con- trol group in 1978 and 1979 (P < .01 in all cases). Thus either the enclosure treatment itself May retard seed set, or manipulation by hand does not effectively transfer pollen to the minute stigmatic slit in P. humiflora. any Proteaceae are protandrous, but this should not have affected the results since in 1979 Manipulations were carried out for four days. In any case, many more of the 300 stigmas in a head should have been receptive than produced seed. Moreover, the pollen remains viable for several days (Table 11), so this should likewise p 2h been a problem. O deters: 1 1 — «E1ne the g Sy in these pro- teas, plants should be grown under greenhouse conditions, Furthermore, a better understanding I5 needed as to how, when, and in what amounts pollen must be deposited on the minute stigmatic Slit to effect pollination. POLLEN VIABILITY METHODS To determine whether residual pollen (i.e., ing) was functional, we tested P. amplexicaulis humiflora pollen for length of viability. Was tested for percent germination on a -enriched sucrose agar medium similar to sed by Taylor (1972), except that the su- concentration was lowered by 75%. All the Pollen ron that y CTOse pollen utilized in this study was collected from owers that opened between 1800 and 2130 on the day the tests were initiated. A single flower of each species was removed from the original group of experimental heads and the pollen plat- ed on agar every 12 to 24 hours for four days, and at 2130 on the sixth day following anthesis. At least two replicates were plated from each flower. Plates were maintained at ambient tem- perature and humidity. The percentage of pollen germinated was determined at least 16 hours af- ter plating by counting the number of germinated and ungerminated grains visible in a single mi- croscope field (X160) in each replicate. RESULTS A substantial proportion of pollen retained the ability to germinate on agar for two full days following anthesis in P. amplexicaulis and at least three days in P. humiflora (Table 11). Thus re- sidual pollen should be capable of fertilization during this period. Some inconsistency both be- tween replicates and for different times of plating is apparent in our results. The extremely low germination of P. amplexicaulis pollen on the morning of the second day, for example, is dif- ficult to explain. Perhaps the viability of pollen from this single flower was anomalously low. Pollen plated in the evening showed a tendency to yield a higher percentage of germination than that plated in the morning, particularly in P. humiflora. Perhaps pollen is physiologically adapted for maximum germination under con- ditions to be expected at the time dispersal nor- mally occurs. SELECTIVE EXCLOSURE EXPERIMENTS METHODS To determine the effects of excluding small- mammal visitation on seed set in P. amplexi- caulis (Jonasplaats) and P. humiflora (Jonasnek), flowering heads were enclosed in a cage of hard- ware cloth (mesh size 13 mm), the base of which was fitted with a nylon skirt that was tightened around the peduncle by a drawstring. The cage was supported over the inflorescence by wiring it securely to adjoining branches, thus preventing movement of the cage and possible damage to the flowers or peduncle. This cage effectively ex- cluded all known mammals in the area (including Mus minutoides A. Smith, the smallest known — ae M V V M AM M M M M M M9 MG NN naaa a TaBLE 11. Pollen longevity in Protea amplexicaulis and P. humiflora. Day No. and Observation Time 3 4 5 P. amplexicaulis 0930 2130 0930 2130 0930 2130 0930 2130 0930 2130 (N = 3) (N=2) (N=3) (N = 2) (N = 2) — (N =2 - (N = 2) X % germination/fl. - .66 .01 .60 .49 € E = P. humiflora 0930 2130 0930 2130 0930 2130 0930 2130 0930 2130 — (N=4) (N=4) (N=4) N=4 N=4 — (N = 4) (N = 4) X % germination/fl. — .68 Al 74 45 57 33 — A TABLE 12. Seed set in protea heads caged to preclude small mammal foraging, but open for visitation by insects. 1978 1979 Control Mammal Excluded Control Mammal Excluded X Range X Range Significance x Range X Range Significance P. amplexicaulis 24.04 1-51 1:33 0-4 P « .001 (N = 27) (N = 9) P. humiflora 18.43 0-95 8.73 0-19 P= .06 2123 0-55 11.56 1-31 P=.01 (N = 51) (N = 11) (N = 35) (N = 16) N3G3VD 'IVOINV.LOS8 INNOSSIN JHL HO STVNNV 0L 10A] 1983] WIENS ET AL.—NONFLYING MAMMAL POLLINATION 19 TABLE 13. Interplant distribution of fluorescing powder by (presumably) small mammals. Approximate Distance from Source Protea Species (m) Points of Deposition P. amplexicaulis? 1.0 Fl. head of another P. a. 0.1 Runways to (but not on) fl. heads of another P. a. 7.0 Runways to (but not on) another P. a. 4.0 Runways to (but not on) another P. a. P. humiflora* 315 Fl. head of another P. A. 8.0 Runways to and on fl. head of another P. h. 0.5 2 fl. heads of another P. h. 25 2 fl. heads of another P. A. P. subulifolias 2.0 Runways to (but not on) another P. s. 0.5 Fl. head of another P. s. 125 Shoots (but not fl. heads) of another P. s. 7.0 Runway and shoots (but not fl. head) of another P. s. 0.5 Shoots and fl. heads of another P. s. 0.5 Runways and shoots (but not fl. heads) of another P. s. 15 Runways and fl. heads of another P. s. * Jonasplaats study area. ^ Jonasnek study area (site B). rodent to occur in the study area), but allowed easy access to insects, especially honey bees. Seed set in the caged heads was compared with that of control heads, i.e., heads on the same or ad- Joining branches left open to natural pollination. The data were analyzed for significance utilizing the Mann-Whitney U test. RESULTS Excluding small mammals from the heads of NMP proteas reduces seed set by approximately 5096 in P. humiflora (Table 12). In 1978 the dif- ferences in P. humiflora between the controls and the rodent-excluded heads approached signifi- cance (P — .06), but in 1979 the differences are considered significant (P — .01). We have no ob- vious explanation for the variable results ob- tained from the experimental groups between 1978 and 1979, except for the reduced sample size in 1978 (N = 11) and general improvement In the technique for placing exclosures over the heads in 1979, Experiments on P. amplexicaulis were con- ducted only during 1979, and seed set in caged eads was reduced more than 95% in comparison to controls (P. < .01). The sample size, however, Was small (N = 9) and the experiments were un- detertaken comparatively late in the flowering Season, FLUORESCING EXPERIMENTS METHODS To determine the extent of intra- and inter- plant movement of pollen and small mammals, the flowering heads of three species of protea (P. amplexicaulis, P. humiflora, P. subulifolia) were dusted with fluorescing powders (Hercules Ra- diant Pigment Type R 103 G) of various colors. Approximately a half teaspoon of powder (or paste produced by adding 50% ETOH) was ap- plied over unopened florets at the center of a flowering head, avoiding the more peripheral open (and presumably nectar-containing) flow- ers. The heads of each plant were treated with a single color of fluorescing powder. Captive Ae- thomys foraged as readily on powdered heads as on non-powdered heads. By using the paste, which left a delicate, easily broken crust, we avoided possible contamination of adjoining plants by wind-blown powder. Heads were normally treated at dusk to avoid possible distribution of the powder by diurnal insects. Observations of other flowering heads, or animals trapped in the area, for traces of flu- orescing powder were made with a long-wave UV lamp late the same night or early the follow- ing morning before the initiation of insect move- ment. Observations on mice, however, were con- ABLE 14. Small mammals captured carrying fluorescing powders applied to P. humiflora flowering heads (Jonasnek, site B).* Approxi- Distance o Source to P Capture imal Carrying Animal Point (m) Fluorescent Powder Acomys 5 Hind feet, tail Acomys 12 ostrum, rear feet Aethomys 9 Rostrum, front fe Rhabdomys 3 Rostrum, chest, front feet Rhabdomys 13 Rostrum, front f Rhabdomys 20 Rostrum, front feet single Aethomys was captured at the P. res- tionifolia study site (Pocskraal) that bore two colors of fluorescing powders « on the rostrum and front feet in- pp y l5 m apart. dicating DALS L tinued on P. humiflora for three days. After the first night, only animals bearing large concentra- tions were recorded because many of the traps became lightly contaminated from recaptured animals carrying fluorescing powder. The high concentrations of fluorescing powder seen on some mice, however, could only have originated directly from the powdered heads. RESULTS Small mammals visit not only different flow- ering heads on the same plant, but also the heads of other protea plants for distances up to 15 m (Tables 13, 14). The large proportion of trapped animals carrying fluorescing powder on their ros- tra (5 or 6) provides further evidence that the small mammals foraged on these heads. In many instances the general movement patterns of the small mammals were clearly evident from the scattered fluorescing particles around the dusted heads and along the rodent runways that often interconnect NMP protea bushes. SMALL-MAMMAL FECES IN PROTEA FLOWERING HEADS METHODS If small mammals regularly frequent the flow- ering heads of proteas, it was reasoned that they might leave behind artifacts as evidence of such visitations. The taking of nectar is difficult to detect, and the removal of pollen from newly ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 TaBLE 15. Number of heads containing any fecal pellets in P. recondita as a function of age.* Fecal Fecal Pellets Pellets Age of Heads Present Absent Flowering heads fully open 6 0 Flowering heads ca. '2 open 2 4 Flowering heads <'2 open 0 6 @ While the data are highly suggestive, the sample sizes are too small to permit a x? analysis for signifi- cance. opened flowers could occur by agents other than small mammals. RESULTS Small-mammal feces in the flowering heads of | P. amplexicaulis, P. cryophila, P. humiflora, P. effusa and P. recondita were the only obvious artifacts discovered. Feces accumulation was found to be age dependent, the older heads con- taining significantly more fecal pellets than youn- ger ones (Table 15). The presence of feces in the heads demands the presence of animals on (or - above) the head at the time of defecation, and fecal accumulation within the heads indicates - frequent and/or relatively long visits. In P. re- condita, however, defecation from above is vir- tually impossible because this protea produces à terminal (but cryptic) head (Rourke & Wiens, vedi f over, protea pollen was present within | | the bs pellets indicating that the animals made - at least two visits to a flowering head: the first when the pollen that later occurred in the feces - was obtained, and the second at the time of dela ecation. NECTAR CONSUMPTION BY BEES METHODS A set of experiments was designed to deter- | mine the extent of nectar consumption in P. humiflora by the African honey bee (Apis mel (accessible to bees) versus control (inaccessible) . heads. Both experiments were conducted on warm, clear, sunny days with the temperature at or above 25°C when bee activity was relatively high. Only flowering heads with at least one row of open, owers and ot hlen VADIC 1 1983] tar were used. Heads were weighed at the begin- ning and end of the experiments. In the first experiment, 20 heads were collected the afternoon of the day before the experiment and kept overnight with their peduncles in water. ext morning all heads were weighed and individually placed in a small can of water with the peduncles submerged. Ten of the containers with their single heads were placed on the ground near a flowering plant of P. humiflora and left uncovered (experimental group). The other ten were similarly positioned but were covered with plastic window screen Mid group) to prevent insect visitation. The experiment je initiated at 1345 and vui i 2 1700. Although the actual number of bee visitations to se experi- mental heads was not recorded, bees only rarely visited the experimental heads clustered in cans. To correct this problem the second experiment was initiated. This experiment was similar to the first except that all heads were collected the morning of the experiment and the experimental heads (after weighing) were wired man onto ore E humi- | P NEUT | on the plant. Sines ‘the experimen heads could not be kept in water, the control heads were sim- ply placed in trays under plastic screens near the bushes with the attached experimental heads. Bees appeared to visit these experimental heads just as readily as the naturally occurring heads. This experiment was initiated at 1030 and ter- minated at 1700. Both experiments were ana- lyzed by the Mann- -Whitney U test. RESULTS ds difference in weight loss between the two s in the first experiment was not significant SRI whereas it was significant in the sec- ond experiment (P < .01). The average weight loss for both experimental and control heads was predictably greater in the second experiment than in the first, since heads in neither the experi- mental nor the control groups were kept in water (Table 16). Decrease in mean weight in the ex- Perimental group is assumed to be the result of nectar loss due to the foraging activities of bees. Interpretation of the second experiment is Complicated because the window screen covering bos Then group possibly reduced evaporative here was little wind on the day of the experiment, however, and the experimental pop- ulation was also mostly shaded by the foliage WIENS ET AL.—NONFLYING MAMMAL POLLINATION 21 TABLE 16. Bee nectar consumption in P. humi- ora Nectar Loss/Head (g) Ex- Resulting peri- from Bee Control Signifi- ment Foraging S.D. (g) cance 1 2.04 0.63 E75 0.86 n.S. 2 3.54 0.36 24) O9! P001 and relatively close to the ground. Thus in the cond experiment we interpret possible reduc- tion in evaporative loss produced by covering the control group with window sorcen as inconse- eight quentia l, loss resulted primarily from the removal of nec- tar by foraging bees. FLOWER PREFERENCE EXPERIMENTS METHODS Choice tests using a T maze were conducted to determine if mice are preferentially attracted to, or preferentially forage from P. humiflora heads (an NMP species) rather than those of P. repens or P. laurifolia (BP species). Several species of rodents from two populations were tested: (1) individuals from the P. humiflora study area mys, Aethomys, Dendromus, Elephantulus) and (2) individuals from an area at least several km away from any known NMP proteas (Aetho- mys, Rhabdomys, Tatera). The first group rep- resented animals that were live-trapped in P. enced" animals), while the second group repre- sented animals that were unlikely to have en- countered P. humi lora or similar flowers (*naive" animals). Animals were maintained in wire cages with cardboard" ‘sleeping tubes” prior to the experimen The sides and s of the T maze were con- structed of masonite and covered above with screen mesh to permit observation. The base and arms were 38 cm long and the individual run- ways 10 cm wide and 15 cm high. Each end of the maze could be opened to facilitate position- ing of test animals and flowering heads. Tests were conducted between approximately 2000 an 2400 hours to coincide generally with the ani- mals’ normal activity periods (Fig. 1). E 22 ANNALS OF THE MISSOURI BOTANICAL GARDEN TABLE 17. Flower preference of small mammals in T maze experiments Bird Nonflying Adapted Mammal Flower- Adapted ing Flowering Signifi- Experimental Group Head Head cance I. “Experienced” Initial arm choice 17 7 n.s. Foraging’? choice l 26 Pe Ol H. “Naive: Initial arm choice 17 17 n.s. Foraging choice 0 21 P< 01 4 Five Aethomys were utilized in the 34 trials. > Trials resulting in no foraging were generally at- tributable to specific anim * Naive animals utilized i in the 34 trials included: five Tatera (21 trials), two Aethomys (11), two Rhab- domys (2 The experimental procedure consisted of plac- ing a freshly picked head of P. humiflora in one arm of the maze and a head of either P. repens or P. laurifolia in the other arm. The position of the heads in the maze arms was determined by a coin toss. A test animal enclosed in its sleeping tube was then transferred into the base of the maze. Initial response time (time before emer- gence from the tube) and subsequent behavior was monitored for up to five minutes. If animals did not emerge within two minutes, the tube was tapped several times. The first direction of move- ment in the arms was recorded as "initial choice" (+ if toward P. humiflora and — if toward the alternative). The first head on which the animals foraged for =15 seconds was recorded as “‘for- aging choice." If animals did not enter the maze arms, or did not forage, their responses were re- were analyzed utilizing the binomial test, with H, that responses are random and thus should be distributed evenly between the two arms of the maze. Analysis of the experienced group ex- cludes the single individuals of Acomys, Ele- phantulus, and Dendromus tested. Analysis of the naive group is broken into Tatera and all others RESULTS Initial choice of maze ends was evenly dis- tributed in both the experienced and naive ani- [Vor. 70 mals (Table 17). In the experienced group, 17 of the 34 animals initially moved toward P. humi- flora and the remaining 17 toward the alternative (P = .13, n.s.). In the naive group, initial choices were also split evenly, 17 to P. humiflora and 17 to the alternative (P — , n.S.). Foraging choice, however, was not random (Table 17). Of 27 foraging responses in the ex- perienced group, 26 were on P. humiflora (P < .01). Of 21 foraging responses in the naive group, all were on P. humiflora (P < .01). These results clearly indicate that the heads of P. humiflora | are the preferred of the two rewards offered for both groups of animals The role of experience in feeding-responses was evaluated by basing expected x? values on the assumption that a// animals should forage and no animals should fail to forage. Of the 36 trials in which experienced animals initially respond- ed, 25 foraged and 11 did not (x? = 3.36, n.s.), and during 38 such trials in the naive group, 21 animals foraged (x? = 7.61, n.s.). DISCUSSION SMALL MAMMAL VISITATION TO PROTEA FLOWERING HEADS Rourke and Wiens (1977) and Wiens and | Rourke (1978) presented preliminary evidence that small mammals regularly visited and pol- | linated the flowers of various South African pro- — teas with which they were associated. That small _ mammals actually visit the flowering heads of | particular proteas is supported by the following new information: (1) the presence of pollen on the rostra and in the gut of small mammals, (2) the nocturnal interfloral and interplant transfer | of fluorescing powder and its occurrence on cap- tured small mammals, and, (3) the accumulation © of rodent feces in the flowering heads of various — proteas Polen loads. Protea pollen was found on the | rostra and in the feces of all small mammals — captured in association with flowering species of | P proteas, although the amounts were highly _ variable (Table 1). Because protea pollen is rath- er sticky, it is not subject to widespread win dispersal as demonstrated by its infrequent oC currence in background samples collected away from the flowering heads (Table 7). Table 7 also _ gives an approximation of the density of back- ground pollen in the environment. The small. percentage of non-protea pollen in the pollen. loads carried by animals tested demonstrates that _ 1983] background pollen does not accumulate in dense concentrations on the rostrum. The non-protea pollen could also originate from animals foraging on these non-protea flowers, but the low con- centration argues against this, or at least suggests a considerable time lapse. If the occurrence of protea pollen on the animals were the result of chance accumulation from background pollen, then the ratios of non-protea and protea pollen on the rostrum should be approximately equal to the background samples. In fact, the ratios differ radically. Protea pollen is present in the samples taken from small mammals in concen- trations averaging at least 100 grains, whereas non-protea pollen averages three grains per sam- ple. This result is greatly underestimated for pro- teas because no more than 500 grains were ever counted per slide, although some slides obvious- ly contained many thousands of protea pollen grains. The pollen in feces is almost exclusively that of protea, and is probably ingested during grooming (little pollen is apt to be ingested during nectar lapping). The data are most consistent with the explanation that the pollen on small mam- mals originated from frequenting the heads of flowering proteas. Fluorescent powder experiments. The results of placing fluorescent powder on the flowering heads of several ground-flowering proteas dem- onstrates that interplant and interhead nocturnal distribution of fluorescing powder occurs up to 15 m from the source (Tables 13, 14). Numerous Particles of fluorescing powder were repeatedly found scattered along rodent runways, strongly Suggesting transport by small mammals. Noc- turnal terrestrial insects would be unlikely to travel equivalent distances in the available time, nor would they be likely to follow rodent run- ways. Furthermore animals with relatively large amounts of flu- Orescing powder on their bodies, especially on the rostrum, supports their role as vectors of most pees fluorescing powder transported noctur- ally, Small-mammal feces in protea flowers. The accumulation of small-mammal feces in protea heads strongly supports the hypothesis that they regularly and frequently visit the flowering heads 9?! NMP proteas. The presence of protea pollen In the feces provides further evidence of multiple n. Visitatio : Exclosure experiments. Experiments de- Ras to test for genetic compatibility inadver- ntly provided additional information relating WIENS ET AL.—NONFLYING MAMMAL POLLINATION 28 to rodent visitation of P. humiflora heads. Twen- ty-four fine-mesh nylon exclosure bags were placed over heads with mature buds, but within six days 14 of the bags (58%) had holes chewed through them or were pulled partially away from the heads, or both. Placing the heads under ex- closures appeared to increase the amount of nec- tar in the heads, thus (presumably) making them more desirable nectar sources. e only animals in the area capable of such behavior might be chacma baboons, but this seems im- probable since they were never observed near the study areas. Althouoh ti D p gularly foraged on NMP protea heads, no wild animals were ever observed visiting the heads. Such observations are technically difficult because the species we studied most extensively (P. amplexicaulis, P. humiflora) have both geoflorous and cryptic heads, and the small mammals, except Rhab- omys, are nocturnal. A single Aethomys was seen in a bush of P. amplexicaulis at approxi- mately 2230 hours, but it did not visit a flowering head during the several minutes of observation. Two rodents, (Aethomys and Rhabdomys) were caught in snap traps high in the branches of P. humiflora near flowering heads, and many ani- mals were captured beneath protea bushes. DOES SMALL MAMMAL VISITATION EFFECT POLLINATION? Several lines of evidence support the conten- tion that small mammals visit the flowering heads of proteas. But what information indicates that these visits also result in pollination? The best evidence originates from observations of: (1) the foraging activities of captive small mammals on the flowering heads of various proteas, (2) floral morphology, and (3) selective exclosure experi- ments. Nectar foraging and putative pollination by small mammals. Some individuals of each species listed in Table 1 (plus Dendromus) for- aged on the heads of NMP proteas when in cap- tivity. The responses of the various species and individuals, however, were not consistent. For example, Aethomys generally foraged readily, whereas Rhabdomys rarely foraged and its activ- ities were sometimes destructive. Elephantulus occasionally licked the surface of the heads with- out actually lapping nectar from the nectar res- ervoirs. Nonetheless, every species in Table 1, at some time, foraged on the heads of NMP pro- 24 ANNALS OF THE MISSOURI BOTANICAL GARDEN teas in a manner that should have effected pol- lination (the process is convincingly recorded cinematographically on 1 film One aspect of rodent visitation io flowering heads mentioned by Rourke and Wiens (1977) needs correction. They speculated that rodent visitation was largely destructive. This is clearly not the case. Except for pollen loss, there is little evidence of rodent visitation following nectar- foraging. The occasional chewing of florets and bracts that prompted the suggestion occurred on only about two percent of the heads in P. am- plexicaulis and is unlikely to have effected pol- lination. The orientation of the styles and nectar res- ervoirs restricts effective foraging only from the center of the head outward along the radii and ensures contact with the stigmatic surface (Figs. 3, 4). The critical stigma-nectar distance of ca. 10 mm makes inevitable the deposition of pollen in the region of the stigmatic slit during nectar lapping and guarantees the maintenance of pol- len loads. Selective exclusion experiments. When small mammals were excluded from flowering heads, seed set was reduced approximately 5096 in P. humiflora and 9596 in P. amplexicaulis. Thus it ever, in the absence of data from the reciprocal experiment involving the exclusion of insects while permitting mammal visitation. Unfortu- nately, experiments along these lines were tech- nically unsuccessful. On the bases of floral mor- phology, physiology, and animal foraging behavior, small mammals should be more effi- cient than insects in pollinating P. amplexicaulis and P. humiflora. More information is needed, however, to establish the relative pollination ef- ficiency of insects and small mammals in this The floral features characterizing NMP pro- teas presumably evolved in response to visita- tion by a variety of small mammals. How some of these features (e.g., easily hes highly concentrated nectar and exposed pollen also promote visitation by numerous insects, es- pecially honey bees, that doubtless effect occa- sional pollination. Based on floral morphology, however, pollination by insects can occur only haphazardly when they forage for pollen or oth- erwise land on the stigma. While foraging for nectar (which they normally do), insects do not [Vor. 70 contact the stigma and pollination is impossible. Bees are possibly the most likely of insects to accomplish cross-pollination since they visit appeared to collect pollen from all species flow- ering during the period in which these studies were conducted (late winter-early spring). As Faegri and van der Pijl (1979) pointed out, “‘so- cial bees < ix — visit any blossom that yields sufficient n " Bees and beetles may, how- ever, eee iia other proteas not closely related to the gre in question, e.g., P. odorata unb. and P. /aetans L. E. Davidson, respec- dde Insects ipei add to pollination suc- cess in some NMP proteas. Traits enhancing their visitation should therefore be selected for, or at least maintained at equilibrium, so long as those characteristics do not retard visitation by small mammals, which are presumably the more effi- cient pollinators. irds are rare and inconsequential visitors to the NMP proteas we studied. During a period 0 four years, involving over 2,000 hours of obser- vation, one malachite sunbird was seen taking - nectar from the rim of a flowering head of P. — humiflora, but the styles were not contacted. Two — Cape buntings (granivores) were captured in — Sherman live traps arround P. humiflora. Al- though they carried protea pollen on their fore- heads, we did not observe them on the inflores- contrast to the visitation pattern in Australia | bine flower birds as well as nonflying mammals — common visitors to proteas and other plants pides 1980, 1982; Turner, pers. comm.). The often cryptic flowering habits of many NMP pro- teas, the absence of visual cues associated with — typical BP proteas, and the high sugar content of | the NMP protea nectar, do not suggest that NMP | i protea flowers are important sources of nectar | for flower birds (Pyke & Waser, 1981). THE ATTRACTING SYSTEM The most obvious attraction is odor, since the — flowers of many NMP proteas are cryptic an | pollination by small mammals is largely noctur- nal. As previously mentioned, the yeastlike odor — 1 with its various modifications attracted captive | | animals. 1l The T maze experiments tested critically both. | olfactory response to protea heads and the ulti- Í = 1983] mate foraging choice. Because the proffered heads could not be seen by the captive animals, olfac- tion is the only stimulus to which they could have responded. The experiment, however, did not test whether the animals were initially respond- ing to the heads of the BP protea (P. repens or P. laurifolia) or the NMP protea (P. humiflora). The results (Table 16) show that initial choice of maze arm was random, i.e., the experimental animals did not initially discriminate between the arm holding the NMP protea and that hold- ing the BP protea, but did so only after receiving additional cues. Since the heads were randomly switched between the two maze arms, the run- ways may have become saturated with the scents of both heads, thus initially precluding selective odor cues. The heads of BP proteas have little scent discernable to humans and have no yeasty odor. While the initial response to the heads was olfactory, the ultimate foraging choice (virtually always the NMP protea) could also have in- volved visual cues. The styles and inner surface of the bracts of P. amplexicaulis, P. humiflora and some other NMP proteas are whitish, whereas the outer sur- face of the bracts is often dark brown or purplish. The contrast produces a “target effect" by em- Phasizing the white center of the head in poor light. This is apparent to the human eye, but we have no experimental evidence regarding its ap- parency to animals, although diff in shades are presumably evident to most mammals. Hawkmoth-pollinated plants are typically white; even generally dark bat flowers often display some white which might act as a nectar guide. Different NMP proteas utilize different strategies to pro- duce the effect and the subject deserves further study. THE REWARD Small mammals visit proteas to obtain nectar (as do myriads of insects and possibly also ba- Mes flowering period of P. humiflora lasts per- i Meu Weeks, and it is reasonable to assume e small-mammal energy budget would be Supplemented during this period by protea nec- he he NMP proteas, however, occur over only infinitesimally small portion of the overall 8°0graphical distributions of the small mammals WIENS ET AL.—NONFLYING MAMMAL POLLINATION 25 in question. In areas no more than several hundred meters from stands of P. humiflora, the small-mammal community is totally without this resource and is presumably not adversely affect- d. The nectar resource from NMP proteas could, however, be more important to the total energy budget than the previous comments indicate. Some local immigration from adjoining areas could increase the density of the small-mammal population on the study areas. That this happens, is suggested by the apparently high density of small mammals on grid B (58 in 1978, 20 in 1979). While we have no comparable population density data for areas without NMP proteas, the small-mammal populations on our study sites appear generally high. We consider the nectar resource of NMP pro- teas as primarily a supplement to the basic small- mammal diet (a sweet treat or junk-food trip?) rather than an important component of their an- nual energy budget. Such a “‘dessert”’ hypothesis is in clear contrast to coevolved systems, where + í 41 nectar ( p ri- ^ P mary or exclusive component of the energy bud- get. Furthermore, the flowering of most NMP pro- teas correlates with the reproductive period of mall l dtł «tra gy resource could be important to females during gestation and lactation, and also in juvenile survival. Blooming at this time should certainly enhance the possibility of flower visitation by small mam- mals, and selection could have shifted flowering by the NMP proteas into this period. Most proteas have peak flowering periods at other times. In unilaterally evolved flowers (see fol- lowing section), the flowering periods appear to coincide with environmental factors that should maximize visitation. For example, pseudoco the dry season when fruits or other flowers are at seasonal lows. The nectar in this latter case could also be an important source of water for relatively large pollinators. As Porsch (1934) pointed out, water is not always so readily avail- able in the tropics as is generally assumed. Mor- combe (1968) made the same argument for southwestern Australia. He commented that the majority of banksias flower during the height of the dry season, although George (1981) indi- 1 1 26 cated flowering in Banksia may not correspond so closely with the dry season as Morcombe sug- gested. If water is an important resource provid- ed by nectar, this should be reflected in a low sugar concentration. In this connection, Schemske (1980) reported that Combretum farinosa Kunth (= C. fruticosum?), which flowers during the pro- nounced dry season in Guanacaste, Costa Rica, produces copious nectar and has among the low- est nectar-sugar concentrations (7.1%) of which tar concentration in unilaterally evolved plants occurring in at least seasonally arid regions (Watt et al., 1974; Pyke & Waser, 1981). THE EVOLUTION OF NMP PROTEAS COEVOLUTION, CO-OCCURRENCE, OR UNILATERAL EVOLUTION? The South African NMP proteas are not co- evolved (Rourke & Wiens, 1977; iens & PE bili bis MERE babe elicits ques- ;nary biologists, suggesting that coevoluGoni 1s perhaps too deeply established as the raison d'étre for the evolution of pollination or dispersal systems as Janzen (1980) has already suggested. Coevolution (Ehr- lich & Raven, 1964) has provided a central para- digm for the study of plant-animal interactions and is an sesto useful concept. Pollination S , however, are also commonly “unilat- erally" evolv pale i.e., plants that have profoundly altered floral function to attract a pollinator that itself has not become specialized to the flower. In other words, the system does not elicit *re- ciprocal selective responses" typical of co- evolved systems, but rather **unilateral selective responses" on the part of the plant. A numbe of biologists have suggested that plants make ini- tial evolutionary adjustments to animals, e.g., Grant and Grant (1965), Baker and Hurd (1968), Baker (1973) and Feinsinger (pers. comm.). Clearly, many generalist animals take nectar (or fruit) during times of seasonal abundance (Snow & Snow, 1971; Heithaus et al., 1974; Carpenter, 1978b). The concept of unilateral evolution is not meant to replace coevolution, but to com- plement it in the sense that it represents one end of a spectrum of biological interactions and co- evolution another. Transitional situations will occur, but this is hardly uncommon in evolu- tionary classifications. Pollination systems in- ie | o ovs \©.5-, ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo 70 Ficus and Yucca). Typically, guilds of pollinators act upon a series of sequentially flowering, co- evolved plant species. The same principles that apply to nectar-foraging animals also apply to fruit- or seed-eating animals. Howe (1980) strongly questioned whether coevolution is al- ways involved in interactions between the plant and the animals that disperse its seeds, preferring to consider the situation an example of **co-oc- currence." Howe does not discuss whether the fruits he studied presumably evolved to entice various animals to eat them, or were at least pre- adapted to the extent that the animals readily | foraged on them. The NMP proteas are clear examples of uni- lateral evolution, since virtually all the basic fea- _ tures of the flowers appear adapted for the at- traction and reward of nonflying mammals that - presumably effect pollination while foraging for | ers and m highly concentrated nectar, however, is the only trait held in common with bee-pollinated flow- — ers. Conversely, good evidence exists that other proteas, e.g., P. repens, have coevolved with the nmr Cape sugar bird (Promerops cafer L.). The latter | commonly takes nectar (and insects) from the — flowering heads and its breeding season is syn- chronized with the peak flowering activity of this protea. Even the young are partially fed on its — nectar (Broekhuysen, 1959; Winterbottom, 1962; Mostert et al., 1980). In our view, all three of these mutualistic situations, i.e., coevolution, sands of plant species, is overdue. While coevo- - lution is a general concept, it was never intended j by its a to s as all-inclusive as the curren ts (Raven, pers. comm.) sugges Of i. pollination systems described by Faegri | and van der Pijl (1979), wind, water, ant, and at — least elements of bee, (social groups), bird, [var- . ious passerine groups, cf. Stiles (1981)], most flies. perhaps beetle systems, and all pollination in- 1 volving mimicry (Wiens, 1978) are probably uni- laterally evolved. Many of the plants known of hought to be pollinated by nonflying mammals are apparently unilaterally evolved (see Intro- duction). Coevolved guild systems probably oc- cur in southwestern Australia, however, where e some Proteaceae may be regularly pollinated bY - eigens di 1983] the marsupial honey possum, Tarsipes (possibly also the southwestern pygmy possum, Cercar- tetus concinnus Gould), which shows clear ad- aptations for nectar and pollen feeding (Wiens et al., 1979; Hopper, 1980). The floral spectrum on which Tarsipes feeds and the degree to which it shares floral resources with birds is currently un- der study (Hopper, pers. comm.; Turner, pers. comm THE ORIGIN OF NONFLYING-MAMMAL POLLINATION SYSTEMS Rourke and Wiens (1977) suggested that at least some of the NMP proteas evolved from bird-pollinated prototypes in both Africa and ustralia, as evidenced by branching patterns of the inflorescences. Many of the NMP proteas pones] y derived xillar y pp by reduction from the large terminal inflores- cences of the BP proteas. Fire was also a possible stimulus in the evolution of NMP proteas through the development of rhizomaty which may have promoted geoflory. Another hypothesis for the origin of NMP pro- teas is suggested by this study. Virtually all NMP proteas are characterized by small, highly local- ized populations often associated with specific soil types. These species are typically low shrubs largely restricted to relatively high elevations in the outlying Cape mountain systems bordering the arid Karoo (Rourke, 1980). As previously Stated, the Cape flora occupies an elevated and greatly dissected landscape closely ass the Table Mountain Sandstone. Consequently, many species in this rich flora occur only as small Populations in scattered, isolated habitats, as do the NMP proteas. In contrast, many of the BP proteas are large shrubs or small trees often oc- curring in dense populations covering many hectares, Flowering stands of these proteas ap- parently attract large numbers of locally migrat- ing flower birds with which they are commonly associated. a Propose that differences in population "d ure may be the single most important factor 'egulating the evolution of pollination by nonfly- LOREMS sth OCIAatCQ Willd e restriction of species having highly he "s ized ecological requirements (especially for "^ may have provided nonflying mam- xe several advantages over birds as pol- vitii. The small mammals involved are ubiq- > Non-hibernating, and non-migratory WIENS ET AL. —NONFLYING MAMMAL POLLINATION 27 residents, and generalist feeders (Roberts, 1951). Such pollinators permit temporal partitioning of flowering times among sympatric NMP proteas, and may be more readily attracted to an ephem- eral and highly restricted resource than more spe- cialized feeders. They ma and beetles) whose activities are often restricted by the long periods of low temperatures char- acteristic of late winter and early spring when most NMP proteas flower. The occasional fluc- tuations in population size of small mammals, however, might be a disadvantage. No other flowers are known to be visited by nonflying mammals in these communities, and sympatric NMP proteas flower sequentially, thus presum- ably no competition exists for mammal polli- nators among NMP species. If the system were, indeed, derived from bird- pollinated species, the prototypes were presum- ably pre-adapted for nonflying-mammal polli- mechanically strengthened tissues (particularly those of the style), which could accommodate a relatively large animal without undue destruc- tion of floral parts. Physiological and structural modifications involved in the change to the pres- ent NMP type include: (1) the shift to nocturnal (p y) nectar production, (2) the production of nectar with high concentra- tions of sucrose and other sugars, (3) reduction of the stigma-nectar distance to about 10 mm, and (4) the production of a volatile, olfactory attractant. The hypothesis explains why pollination by nonflying mammals probably had multiple origins, since species with cryptic, terminal heads (P. recondita), rhizomatous stems (P. angustata); cryptic, geoflorous heads (P. amplexicaulis), and aerial, pendulous heads (P. sulphurea) all rep- resent different lines of evolution in which non- flying-mammal pollination presumably evolved independently. The common denominator among these species is their occurrence in rela- tively small, isolated populations. The shift to the NMP type in protea shows every evidence of being a strategy superimposed on many dif- ferent life forms, each of which responded in differing ways depending on the phyletic con- straints within the system Whether the “restricted population hypothe- sis” is also applicable to other examples of pol- lination by nonflying mammals remains to be anthesis and a 28 determined. Certainly many of the cryptic and/ or geoflorous species of Dryandra and Banksia in Australia occur in small, isolated populations. In Southwestern Australia, Tarsipes (perhaps Cercartetus) is highly adapted for a diet of nectar and pollen and represents a distinct line within its family. This suggests an older and long-es- tablished, flower-animal interrelationship in Australia, as Sussman and Raven (1978) and Ford et al. (1979) suggested, even though many of the cryptic, geoflorous banksias and dryandras are probably specialized within their genera. Sussman and Raven (1978) proposed that nonflying-mammal pollination is an old phe- flower-feeding bats, which are presently un- known in the Cape region of southern Africa or in southern Australia. Less distinctive climatic gradients, however, probably existed in the early Tertiary (Sussman & Raven, 1978). Because of the low frequency of individuals of a species in tropical forests, the "restricted population hy- pothesis" might not be expected to apply unless (1) plants are self-compatible [an uncommon condition according to Bawa and Opler (1975] and produce sufficient flowers to induce regular visitation by nonflying mammals, or (2) the par- ticular plant species do not occur as highly scat- tered individuals in tropical forests. Bats and birds solve the problem of scattered distributions by possessing flight and trap-lining capablilities, which should make them much more competi- tive in tropical forests than nonflying mammals. Since the publication of Sussman and Raven's paper, three examples of pollination by nonflying mammals have been discovered in the New World tropics. Lumer (1980) reported that ro- dents (Oryzomys devius Bangs and Peromyscus mexicanus Saussure) pollinate B/akea (Melas- tomataceae) in a Costa Rican cloudforest. Lumer (pers. comm.) indicates that this B/akea often occurs as clusters of several individuals. But Blakea also grows in a windy, cold environment along the continental divide. Baker (pers. comm.) suggests that the harsh, windy environment may preclude effective bat visitation to such flowers. Little is known about the occurrence of flower- feeding bats in this area, but even if flower-vis- iting bats were present, they might well avoid the environment where these plants occur, thus pro- viding support for Sussman and Raven's argu- ment. Blakea appears to offer some evidence for ANNALS OF THE MISSOURI BOTANICAL GARDEN both the competitive exclusion and restricted population hypotheses. Two additional discoveries of nonflying mam- mal pollination are reported for the New World lowland tropical forests. Janson et al. (1981) in- dicated that in the Peruvian Amazon 13 species Quararibea, and Ceiba (Bombacaceae). Steiner (1981) reported that the red woolly opossum (Caluromys derbianus Waterhouse) visits and [Vor. 70 presumably pollinates Mabea (Euphorbiaceae). | This opossum also visits the typically bat-pol- linated flowers of Ochroma (Bombacaceae) and Trichanthera (Acanthaceae). Steiner also sug- gested that the common opossum (Didelphis vir- giniana Kerr) may likewise be involved in pol- lination. Janson et al. (1981) observed that species of Combretum, Quararibea, and Ceiba are unilat- erally adapted for pollination by various nonfly- ing mammals, as is Blakea (Lumer, 1980) and probably also Mabea (Steiner, 1981). Whether other tropical flowers subject to nectar foraging by nonflying mammals are adapted for pollina- tion by these animals or by bats is uncertain. | Mori, Prance, and Bolten (1978) indicated that - opossums, cebus, and served on Lecythis flowers are feeding opportun- istically, as are (presumably) marsupials on the typically bat-pollinated flowers of Ochroma and Trichanthera (Steiner, 1981). Janson et al. (1981) stated that some primate visitation is clearly Op- portunistic and destructive. With respect to population structure, Steiner squirrel monkeys ob- — i (1981) reported that Mabea may also have à | clumped distribution, and suggested that such distributions may be more common in the trop- ics than was previously thought likely (Hubbell, 1979). Janson et al. (1981) mentioned that Com- bretum and Quararibea are not uncommon if the study area, but that Ceiba has a widely scat- | tered distribution. The reports of pollination of the baobab by _ the bush baby (Coe & Isaac, 1965) and of genets taking nectar from rica need further study to determine the basic pollinator adaptations of the flowers, although Lack indicated that Maranthes is primarily bat- pollinated. Baobabs are bat-pollinated in west ^ Africa (Baker, 1961 and pers. comm.; Jaeger 1954 —cited by Faegri & van der Pijl, 1979). the flowers of Maranthes — (Chrysobalanaceae) (Lack, 1977) in tropical Af — 1983] A number of low or prostrate proteas occur in tropical Africa, e.g., P. enervis Wild, P. heck- manniana Engl., P. paludosa Welw., P. secun- difolia Hauman (Beard, 1963), but the pollina- tors are unknown. A study of these species may well provide additional information on Sussman and Raven's competitive exclusion hypothesis as applied in the Old World tropics. Data currently emerging from the Neotropics indicate that non- flying mammals are not necessarily out-compet- ed by bats. Is THERE A CLASS OF FLOWERS ADAPTED FOR POLLINATION BY NONFLYING MAMMALS? The existence of a class of flowers adapted for pollination by nonflying mammals was suggested by Rourke and Wiens (1977). Their argument was based primarily on the parallel floral evo- lution between South African species of Protea and a number of Australian proteaceous, and et al. (1981), and Steiner (1981), and in Australia (Hopper, 1980, 1982; Turner, in prep.) provide additional information on the subject. Janson et al. (1981) in their extensive studies of Amazonian nonflying-mammal pollinators provided the most important new data. They generally supported the concept ofa class of flow- ers adapted for pollination by nonflying mam- mals, but stressed that sufficient attention has not been given to the possible role of bats in the Pollination of the flowers on which they fre- quently observed nonflying mammals foraging for nectar. In tropical regions, the differential ef- fectiveness of bats and nonflying mammals in Pollinating flowers visited by both groups was discussed by Sussman and Raven (1978), but detailed analyses are needed. In Australia and South temperate Africa a similar problem exists, but it involves birds and nonflying mammals in the former, and insects and nonfl ing mammals in the latter. Janson et al. (198 1) emphasized that Many flowers growing in their study area are id mally similar to those regularly vis- a . Wi s "eem ith floral structure in identifying floral classes. WIENS ET AL.—NONFLYING MAMMAL POLLINATION 29 Janson et al. (1981) suitably amended the list of structural features proposed by Rourke and Wiens (1977) to include characteristics of the Combretaceae and Bombacaceae they studied. Most notably, the stigma-nectar distance is in- creased to accomodate the larger visitors, and the flowers are short-lived. Some flowers lack an odor, but probably possess visual cues that may serve to attract the diurnal primates and pro- cyonids that possibly lack the well-developed ol- factory senses of rodents. This dichotomy of characters among nonflying mammal-pollinated r for mammal pollination, including chiropter- ophily as a subgroup—a suggestion worth recon- sidering. The flowers of Blakea (Melastomata- ceae) (Lumer, 1980) differ in a number of respects from those of Bombacaceae, Combretaceae, Myrtaceae, and Proteaceae, although Mabea (Euphorbiaceae) (Steiner, 1981) appears to have no features departing radically from those men- tioned by Rourke and Wiens (1977) and Janson et al. (1981). The floral features of Blakea need further analysis in this regard, particularly in re- lation to odor and nectar characteristics. As in- formation accumulates, more additions and modifications to the proposed syndrome will probably be necessary. LITERATURE CITED ARMSTRONG, J. A. nisms in the Australian flora—a land J. Bot 17: 467-508. Baker, H. G. 1961. The adaptations of flowering plants to nocturnal and crepuscular pollinators. Quat. Rev. Biol. 36: 64-73. ——. 1973. Evolutionary relationships between ; TU PISA db : a AL 1979. Biotic pollination mecha- review. New Zea- rican tropical forests. Pp. 145-159 in B. J. Meg- gers, E. S. Ayensu & D. W. Duckworth (editors), Tropical Forest Ecosystems in Africa and South ica: a comparative review. Smithsonian In- stitution Press, Washington, D.C. I. BAKER. 1979. Sugar ratios in nectar. Phy- tochemical Bull. 12: 4 : & P. D. Hurd. 1968. Intrafloral ecology. Ann. Rev. Entomology 13: 385-414 A. OPLER & I. BAKER. 1978. Acomparison > . . Cet or me a 30 insect-borne pollen. The Pan-Pacific Entomol. 47: BROEKHUYSEN, G. J. 1959. The biology of the Cape garbird ES ntinah i cafer (L.). Ostrich, Suppl. 3, 180-2 CARPENTER, F. L. 78a. dee for mammal polli- nation? Oecologia a 35: Io Lu 8b. "Plant-pollinator interactions in Ha- (Myrtaceae). Ecology : 57: 1125-11 CARPENTER, F. L. & H. F. RECHER. Hon Pollination, reproduction, and fire. Amer. * dish ees 871- CASPER, B. B. 1982. Ecological sh a of ovule abor- tion and seed dispersal in Cryptantha A sai et ceae). d D. thesis, University ge Uta D. WIENS. Fixed rates oe random ovule abortion in Cryptantha flava (Boraginaceae) and its possible relation to seed dispersal. Ecology 62: 866-869. Cog, M. J. & F. M. Isaac. 1965. Pollination of the baobab (Adansonia digitata L.) by the lesser bush baby (Galago reece evi E. Geoffroy. E. Af- rican Wildlife J. 3: CowriNG, R. M. 1978. pure composition and pro- duction i in Protea L. Honors thesis, Dept. of Bot- Vertebrate pollination and the maintenance of unisexuality i in Freycinetia. Ph.D. thesis, Harvard Universit DEGENER, O. 1945. Plants of Hawai National Park. i Butterflies and plants: a study of coevolution. Evolution 18: 586— FAEGRI k & L. VAN DER Pur. 1979. Principles of Pollination kolory. 3rd ed. Pergamon Press, New Yor Forp, H. A D. C. PATON & N. Forpe. 1979. Birds as pollinators of Australian plants. New Zealand J. Bot. 17: 509-519. GEORGE, A. S. 1981. The genus Banksia. Nuytsia 3: 239-474 a P. . An analysis of the flora of southern Africa: its characteristics, clea eae. and as Ann. Missouri Bot. Gard. 65: 369— cat v. & K. GRANT. 1965. Flower Pollination n the Phlox Family. Columbia U. Press, New Yo rk. HEITHAUS, E. R., P. A. OPLER & H. G. BAKER. 1974. Bat activity ud pollination of Bauhinia pauletia: plant-pollinator ned aa ree 55: 412-419. 8 ad 1982. Feeding behavior of birds and mam- mals on flowers of — gee et Eucalyp- Din Mo nr PaA poe waste of a tropical fruit. Ecology 61: 944. HusBtLL, S. P. 1979. "Tree duper, abundance and ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 | diversity in a tropical dry forest. Science 203: 1299- . 1302. JAEGER, P. la cheiropterogamie. Bu Noire 16: 796-821 JANSON, C. H., J. TERBORGH & L. H. Emmons. 1981. Non-flying mammals as pollinating agents in the Amazonian forest. Biotropica 13 (Supplement): | 1-6. 1954. Les aspects actuels de probleme de ll. Inst. France. Afrique 1980. When is it coevolution? Evolution —612. KERNER, A. 1895. The Natural History of Plants, Volume II. Trans. by F. W. Oliver. Blackie & Son, JANZEN, D. 34: 611 lands of the World. Elsevier Scientific Publ. Co. Amsterdam Lack, A. 1977. Genets feeding on nectar from Ma- ranthes polyandra in northern Ghana E. African Wildlife J. 15: 233-234. LUMER, C. 1980. Rodent pollination of Blakea (Me- | lastomataceae) in a Costa Rican cloud forest. Brit- l tonia 32: 512-517. . 1979. Biogeography of the bees - Ann. Missouri Bot. Gard. 66: 277—347. Morcomse, M. K. 1968. Australia’s Western Wild- flowers. Landfall Press, Pert 1974. sities Press Pty. Ltd., Sydney. Mon, S. A, G.T . PRANCE & A. B. BorrEN. 1978. Additional notes on the floral rcd a neotrop- | ical Lecythidaceae. d 30: MosrERT, D. P RIED rs = "Ng Birds of Australia. Australian Univer — 1980. Protea nectar nn inn fauna i in relation OATEEXG T B 71. Birds of the night. In K man (ait, Birdlife i in Southern Africa. € and Sons SA (Pty) Lt ape Tow PonscH, O. t Sáu unetien und die Frage der E eile: I. Bio 10: 657-685. 294035. die Frage der Sáugetierblume. II. Biol. Gen. 11: 171-188. 6a. Säugetiere als Blumenausbeuter und die Frage der Sáugetierblume. III. Biol. Gen. 12: -21. 2 PRANCE, G.I. e als Blumenausbeuter | 1. Gen. Säugetiere als Blumenausbeuter und — . 1936b. Süugetierblumen. Forsch. & Fortschr. OF, i 1980. A note on the probable poll nat ion of Combretum by Cebus monkeys. Biotro- — Pyke, G. H. pa = M. Waser. 1981. The productio ars by hummingbird and eme meg tees ica 13: 260—270. A. 198 l. Demography and activity p Pro flowers RICKART, E. ince of South Africa. J. Mammalogy 62: vA Rourke, J. P. 1972. Taxonomic studies on Leuco | spermum R. Br. J. S. African Bot. Suppl. 8: 1- 194. : 1980. The Proteas of Southern Africa. Pur- nell, Cape Town 1983] D. Wiens. 1977. Convergent floral evo- lution in South African and Australian Proteaceae and its possible bearing on pollination by nonfly- ing mammals. Ann. Missouri Bot. Gard. 64: 1-17. ROBERTS, A. 1951. Jn R. Bigalke, V. Fitz Simmons D. E. Malan (editors), The Mammals of South Africa. Publ. by the trustees of the mammals of South Africa book fund, Hafner Publ. Co., New York. SCHEMSKE, D. mingbird e ‘of Combre Costa Rica. Biotropica 12: 169— SLEUMER, H. Proteaceae. In C. G. G. J T aO Tion Malesiana. Ser. 1, ien SO Floral — and hum- um farinosum in 181. SNow, B. E & D. W. SNow. 1971. The feeding ecol- ogy of tanagers and honeycreepers in Trinidad. Auk 88: 291-322. STEINER, K. 1981. Nectarivory and potential o: nation by a Neotropical marsupial. Ann. Missouri Bot. Gard. 68: 505-513. STILEs, F. G. —— aspects of bird- flower coevolution, with particular reference to Sots America. Ann. rym Bot. Gard. 68: 3-351. WIENS ET AL.—NONFLYING MAMMAL POLLINATION 31 St ssMAN, R. W. & P. H. RAVEN. 1978. Pollination -73 Phytogograph and ecology of apensis. In M. rger (editor), The me geography and Ecology of Southern Africa. W. Junk, The Hague TAYLOR, R. M. 1972. Germination of cotton (Gos- oa hirsutum L.) pollen on an artificial medi- m. Crop Science 12: 243-244. Warn. \ W. B., P. C. HocH & S. G. Mitis. 1974, Nectar resource use by Colias butterflies. Chemical and visual aspects. Oecologia 14: 353-374 Wiens, D. . Mimicry in plants. Evolutionary Biology 11: 365—403. & J. P. ROURKE. 1978. Rodent era rues in southern African Protea spp. Nature 276: 7 , M. RENFREE & R. O. WOOLLER. 1979. Pus loads of honey possums (Ti arsipes spenserae) and nonflying mammal pollination in southwestern Australia. Ann. Missouri Bot. Gard. 66: 830-838. WINTERBOTTOM, J. M. 62. Breeding season of long- tailed Sugarbird Promerops cafer (L). Ostrich 33: rei THE MAIZE AND TEOSINTE MALE INFLORESCENCE: A NUMERICAL TAXONOMIC STUDY! JOHN F. DoEBLEY? ABSTRACT he genus Zea consists of the peripe ege maize, and its wild and weedy relatives, the kn native to Mexi e Honduras. This study investigates the comparative ogy of the male inflorescence (tassel) from a large sample of teosinte populations and selected maize races. The reason for focusing on the tassel is simply that as an inedible structure it escaped the direct effects of those forces of hu uman selection specifically aimed at its edible female counterpart, ore appropri bioc eran. l, and gene measurements were subjected to canonical variate analysis, : measured on al multivariate statistical ape Consideration of th ic dat Luxurian m ys ssp. mays; (4b) Zea mays ssp. mexicana contains vars. paresis and huehuetenangensis VG fthe following taxa: I section tes— EK. Pa diploperennis (2) Zea perennis, 1 Zea luxurians; II section Zea—(4a) Z ; (4c) Z a mays ssp. parviglumis. The latter dre though morphologically similar, differ sub- nce by some genetic measures. Fach t taxon is described € illustrated with emphasis on the tassel and spikelet. between maize and Zea diploperennis, a fact which casts doubt t on clic recent hypothesis € these annuals represent the hybrid offspring of Zea diploperennis and a "hypothetical wild ma ees thermor is - direct ancestor of maize: either Zea mays ssp. parviglumis or ssp. mexicana gave rise to wise ma The genus Zea consists of a small group of annual and perennial grasses including the eco- nomically important cultigen, maize (Zea mays ssp. mays), and its wild relatives, the teosintes, which are native to Mexico, Guatemala, and Honduras. There are four species included under the popular name teosinte according to the latest taxonomic treatment (Iltis & Doebley, 1980 Thus, Zea includes, in addition to the cultigen listed above, the following taxa: (1a) Zea mays ssp. mexicana from central and northern Mex- ico, including races Chalco, Central Plateau, and obogame of Wilkes (1967); (1b) Z. mays ssp. parviglumis var. parviglumis from southwestern Mexico (= race Balsas of Wilkes); (1c) Z. mays ssp. parviglumis var. huehuetenangensis from western Guatemala (= race Huehuetenango of r many seminal suggestions, critical reading of this manuscript, and introducing me to the — o ial thanks go to Hugh H. Iltis. For guiding me thro ough the inscrutable world of statistics and. diens helpful suggestions on the methodology, I express my gratitude to Robert i uy sed arranging for my teosinte samples to be iscon dison, Wisconsin 53706. Present addres Department of Statistics, North Carolina State University, Raleigh, Non: Carolina 27650. ANN. Missouni Bor. GARD. 70: 32-70. 1983. m indebted to poen: Hi- Br Wilkes), the above three all annuals; (2) Z. lux- | urians, an annual from southeastern Guatemala | .p rennial; and (4) Z. diploperennis, a diploid pe rennial teosinte, the latter two both highly local | species from southern Jalisco, Mexico (Fig. D: - These four species are separated into sect. Zea - containing Z. mays with its three subspecies, and | sect. Luxuriantes containing Z. luxurians, Z | perennis, and Z. diploperennis (Doebley & neis 1980). The genus Zea, along with the closely related genus Tripsacum, which ranges from Sou America to temperate North America, belongs to the tribe Andropogoneae of the Panicoid | subfamily of the family Gramineae. In addition - , to similarities in general aspect, these two genera — Kowal. For their help 1" | mone die to thank Dos ald Duvick and Deal r, and others p helpful commen ed n sin, Ma 1983] SECTION LUXURIANTES O Z. LUXURIANS SECTION ZEA © Z. MAYS SSP, MEXICANA 4 Z. MAYS SSP, PARVIGLUMIS FIGURE 1. Mays ssp. mexicana, the two northern stations represent the NOB DOEBLEY —MAIZE AND TEOSINTE 33 Distribution of native populations of the genus Zea (Ex Doebley & Iltis, 1980, Fig. 29). In Zea OGAME and DURANGO populations respectively, the two southern clusters CENTRAL PLATEAU (western) and CHALCO (eastern). In Zea mays ‘SP. parviglumis, the southern Mexican cluster represents var. parviglumis, and the western Guatemalan one var. huehuetenangensis. shar € in common the highly specialized cupulate fruitcase, but differ from one another in that Tripsacum bears its male and female spikelets in the same inflorescence, while Zea has separate male and female inflorescences. Additionally, the two genera differ cytologically, Zea with a bas chromosome number of ten (x = 10) and 7rip- Sacum with eighteen (x = 18). Some authors segregate Zea and T) ripsacum along with Several monoecious Asiatic genera of the Andropogoneae, in the tribe Maydeae (Hack- el, 1890), However, this appears to be an un- natural group whose members as a result of con- vergent evolution all have male and female Spikelets in Separate inflorescences or separate Parts of the same inflorescence (Celarier, 1957; Clayton, 1973; Smith & Lester, 1980). A system ihe accurately reflects the phylogeny of E 8enera by placing Zea and Tripsacum in ne subtribe, the Tripsacinae, and the Old World o genera in another, the Coicinae, has been pro- posed on the basis of numerical taxonomic evi- dence by Clayton (1973). Although Zea has long held the interest of bot- anists, agronomists, geneticists, cytologists, and ethnobotanists, taxonomists have until quite re- in the 1940s (cf. Iltis & Doebley, 1980), it was not until Wilkes (1967), a geneticist and ethno- tanist, undertook a study of racial diversity among the teosintes that the problem was at all dd dina detailed manner. However, Wilkes, of necessity, largely concerned himself with de- termining the distribution of the teosintes, re- lying heavily on geography for the delimitation of his six races of annual teosinte. Although he employed genetics, physiology, and vegetative morphology to a considerable extent, he paid only scant attention to floral morphology. 34 As a response to the need for further taxonom- ic study of Zea, and especially of its floral mor- phology, Iltis (1971, 1974) initiated a study of Zea systematics and maize evolution at the Uni- versity of Wisconsin in the late 1960s. As a result of this work, two papers (Doebley & Iltis, 1980; Iltis & Doebley, 1980) proposed the taxonomic system for Zea outlined above. The intent of these two papers was to present the new system of classification, discuss the evidence supporting it, and review its implications for Zea phylogeny, but not to discuss in detail the numerical taxo- nomic evidence, nor the methodologies in- volved. The present work concerns these aspects of the project. Building on this base, some new taxonomic conclusions are drawn and sugges- tions for further work propose The work of Iltis and myself spe do on vari- ation: in the tassel e its spikelets. This line of s historical roots in the work of Edgar pe and his associates who in a series of papers (Anderson, 1944a, 1944b, 1951, 1969; Anderson & Cutler, 1942; Anderson & Brown, 1948) called attention to the taxonomic importance of the maize tassel and the need for botanists to re-direct some of their effort away from the maize ear to the tassel. Their reasons for arguing in favor of this approach seem quite reasonable, for the tassel and especially its spike- lets, are HrcA pd easily measured than their highl A i female counterparts (Anderson, - 1944b: T son & Cutler, 1942) In the early 1950s, Anderson's student Reino Alava implemented the first and only taxonomic study of cultivated maize based solely on tassel and spikelet characteristics (Alava, 1952). As a result of his resea arch, Alava identified several Ww those sua maize race classification, partly were rather limited. What prohibited Alava from generating more far- reaching conclusions was that the only specimens available for his analyses were those preserved in the Missouri Botanical Garden herbarium. Many of these were but single specimens from one locality, and thus they were unable to pro- under widely different conditions, giving rise to much environmental variability, which obscures racial boundaries. Despite these restrictions, Alava drew one im- portant conclusion: “Experience with wild grass- ANNALS OF THE MISSOURI BOTANICAL GARDEN on [Vor. 10 l es related to maize points to the male spikelets as one of the most significant features for deter- mining relationships of maize to its possible wild ancestors” (Alava, 1952, p. 90). Two decades later, Iltis, who had formerly taken classes with Edgar Anderson at the Missouri Botanical Gar- den, carried Alava’s line of reasoning a step fur- ther. Iltis (1974) was apparently the first to point out that the inedible tassel was particularly well suited to the task of reconstructing the evolu- tionary bridge between maize and its wild ances- tor, because it was less drastically altered by the effects of human selection than the edible ear. This realization, and sustained emphasis by An- derson himself on the maize tassel, led Iltis (1974) tion could be made through a systematic stud) of the maize and teosinte male inflorescences | Thus, Anderson, who was above all else a pow- erfully influential teacher, some thirty years ago planted in the minds of his students ideas which today are still bearing fruit. Curiously, Anderson, who provided the idea to examine the maize tassel, also provided the stimulus for the development of the statistical | technique used to study this structure in the pres - ent report. This came about when, as a young man, he received a scholarship to study in Ef gland under the direction of the renowned stat istician and evolutionist, Sir Ronald A. Fisher (aen Anderson, 1926), A this time Anderson had tween species and developed a mathematical means to distin guish the two parental types from their hybrids He called this technique the hybrid index, : method of great simplicity as well as practicalit (Anderson, 1936). Exposed to the hybridization problem by Anderson and using Anderson's dat on Iris, Fisher (1936) devised a much more s% phisticated and at the same time far less biase¢ - m 1983] cies, one handed down through several genera- tions of statisticians, the other through botanists. Now joined together, these legacies provide the foundation for the work presented here. MATERIALS AND METHODS SAMPLING AND DATA ORGANIZATION The basic unit of analysis for this research is the local population, here defined as a group of plants growing and presumably breeding togeth- er in an area of 2,000 square meters or less. In the case of hybrid swarms of Zea mays ssp. mays and the various taxa of teosinte, only those in- dividuals whose male and female inflorescences showed no evidence of hybridization with maize were used. Thus, the samples are intended to represent the taxa of teosinte as they maintain themselves in the wild with little or no influence from maize. Non-random samples of eight to twenty in- dividuals from twelve such local populations of teosinte were studied in detail (Table 1, part A). For the most part these plants were collected within 100 meters of major or secondary high- ways. This does not, however, indicate distur- bance in all cases. While some populations grew as weeds in corn fields, others thrived in quite wild places such as dry open savannas or rocky escarpments. To supplement the field-collected local pop- ulations, this study includes herbarium speci- mens collected in the wild by various botanists at different times and in different places. These Specimens were organized into six groups (Table l, part B) each of which consists of individuals from a single political state or from the vicinity of a particular city; each such group was treated as a "local population" in the analysis. Together the field-collected local populations and field- Collected herbarium specimens cover the entire known geographic range of the teosintes (Fig. 2). ll the preceding field-collected plants, having nS of Jalisco and others on dry rocky escarp- i Sin the seasonally arid Balsas river drain- ee Modify the phenotypes to an unknown de- pe This Poses a problem to the interpretation variability among populations which is best DOEBLEY — MAIZE AND TEOSINTE 35 handled by growing seeds from the populations under similar conditions in a garden. For this reason the present research includes twelve sam- ples grown in a common garden at the Pioneer Hi-Bred International Research Station near Homestead, Florida (Table 1, part C). Each of these samples was derived from seeds collected from a single local population. The morphology of these samples indicates how differences in lo- cal environmental conditions may have affected the morphology of the field-collected plants. Another category of population ples i composed of specimens cultivated at various +1 + lal D). Although the environmental component of differences among them is even more complex than for the field-collected population samples, they are included in the analysis because they cover some geographic areas (e.g. Honduras) not represented by the other samples. In order to have a sample of maize varieties with which to compare the teosintes, three so- called “primitive” maize races were grown at the Purdue University Agricultural Alumni Experi- mental Station in southern Florida. These in- clude Nal-Tel, Chapalote, and Palomero Tolu- quefio, the seeds for which were obtained from the International Center for the Improvement of Maize and Wheat (CIMMYT) in Mexico. One other maize population sample collected in the field near Toluca, Mexico, belonging to the or *Conico" race, was ) ically very closely related and therefore members ofthe same taxon in the narrowest sense. In most situations, it is not necessary for the taxonomist to attempt to distinguish between members of one local population, but only to distinguish var- ious local populations from one another. Popu- lations which are very similar may be placed together in the same taxon, while those which are distinct, may, depending on the degree of difference, be placed into separate taxa. CHARACTERS ANALYZED PR 1: 4T. a il sts the 40 populations, 17 quantitative morpholog- ical traits (variates) of the tassel and spikelet were measured. Many of these are based on characters 36 ANNALS OF THE MISSOURI BOTANICAL GARDEN (Vou. 70 . TABLE l. List of populations. A. Populations Collected in the Wild: l. Zea Sabes (DIPLOPERRENNIAL teosinte). Mexico: Jalisco: With tall grass and herbs along small streams in a region of Pinus-Quercus-Carpinus forest, LA VENTANA, Cerro San Miguel, Sierra de Man yit 7 km ENE of El Durazno, 19?31'45"N, 104°13’W, alt. ca. 2,300 m. Sept. 22, 1978. Illis, odiis Doebley & Lasseigne 450. Sample size — 20. . Zea perennis (FERENTIAL, Sees Mexico: Jalisco: In and along a small arroyo with Heteropogon contortus, Chaet ij an area of former pine-oak sacatonal grasslan 9 km WSW of Ciudad Ghi 1 3 km ESE of LOS DEPOSITOS, 19*40'N, 103?35'W, alt. ca. 1,650 m. Sept. 24, 1978. Iltis, Puga, Guzman, Doebley & Lasseigne 550. Sample size = 20. N o2 n orchards on former Pinus-Quercus savanna. Ca. 14 km WSW of Ciudad Guzman, 4 km WSW of Los : iis perennis (PERENNIAL teosinte). Mexico: Jalisco: Along dirt roads, on edges of small maize fields, and - Depusiion 0.2 km due W of PIEDRA sime alt. 2,100-2,200 m. Oct. 1978. R. Guzman s.n. Sample 4A . Zea justis (GUATEMALA teosinte). Guatemala: Jutiapa: Weeds in maize field and hedgerow, 2 km N of senyal rank in EL PROGRESO, 14°22'30’N, 89°51'30’W, alt. 1,025 m. Oct. 22, 1978. K. Lind 419. Sample s . Zea epe (GUATEMALA teosinte). Guatemala: Jutiapa: On sides and top of hot dry lava cliff formed wn an-American Highway road cut. N-side of PAH, 1.6 km E of EL PROGRESO turn-off, 14°21'N, - 89*50'W, alt. 925 m. Oct. 22, 1978. K. Lind 420. Sample size = 8. . Zea luxurians (GUATEMALA teosinte). Guatemala: Jutiapa: In rice and maize fields, and along hedgerows, an by an old road (Camino Viejo), ca. 2 km NW of EL TABLON, 3 km ENE of Jutiapa, 14°18'15’N, - 89°52'45’W, alt. ca. 1,000 m. Oct. 22, 1978. K. Lind 421. Sample size = 11. : . Zea mays ssp. mexicana (CHALCO teosinte). Mexico: Mexico: Weeds in maize field, Valley of Mexico, ca. 5 km SE of CHALCO at km 46.6 on road to Amecameca, 19?12'N, 98?49'W, alt. ca. 2,300 m. Sept 11, 1977. Iltis & Doebley 4. Sample size = 18. - oo . Zea mays ssp. mexicana (CHALCO teosinte). Mexico: Mexico: On unplowed mound in maize field, Valley | of Mexico at km 20.5 on road from LOS REYES to Texcoco. 0.8 km N of Los Reyes, 19?21'N, 98°58'W, i alt. ca. 2,150 m. Sept. 12, 1977. Iltis & Doebley 8. Sample size = 10. o . Zea mays ssp. mexicana (CENTRAL PLATEAU teosinte). Mexico: Guanajuato: Weeds in maize field, at m 57 on road from Morelia to Salamanca, 3 km N of URIANGATO, 20°10'N, 101°10’W, alt. ca. 1,900 | m. Sept. 17, 1977. Iltis & Doebley 96. Sample size = 18 © treeless hillside, lower edge of xerophytic semi-deciduous shrubby savanna, near QUINCEO, 6 km (by air) NW of Morelia, 19°43'N, 101°14’W, alt. ca. 2,000 m. Sept. 18, 1977. Iltis & Doebley 161. Sample size = 10. 100°07'W, alt. ca. 1,050 m. Sept. Ži, 1977. Titis & Doebley R E size — 16. 12. Zea maysssp. parviglumis var. p . Zea mays ssp. mexicana (CENTRAL PLATEAU teosinte). Mexico: Michoacan: In maize fields, on mostly — . Zea mays ssp. parviglumis var. parviglumis (BALSAS teosinte). Mexico: Guerrero: Weeds in maize field in 7 region of semi-deciduous savanna of Leguminosae, Bursera, Pseudomodingium, and tree Ipomoea, at km 103 on road from Iguala to Arcelia, 11 km (by road) W of ACAPETLAHUAYA turn-off, 18°23'N, o: Michoacan: Very steep and ungrazed i i south facing rocky slope with thor shrubs, small Pinar trees and grasses, ca. 1 km S of TZITZIO on road to Huetamo, 19°34'N, 100°55’W, alt. ca. 1,500 m. Sept. 15, 1977. Iltis & Doebley 87. Sample | 0. size = 2 B. Populations of Field Collected Herbarium Specimens: 13. Zea perennis (PERENNIAL teosinte). Mexico: Jalisco: Among sunflowers, Bidens and grasses, 1 mile S of 4 railway station of CIUDAD GUZMAN, 19?41'30"N, 103?28'40"W, alt. 1,520 m. Oct. 28, 1921. G. N. E Collins, s.n. (US). Sample size = 3. 1 14. nn luxurians (GUATEMALA teosinte). Guatemala: Jutiapa: Near El Progreso, 7/tis G-5 (WIS), Wilkes 1 3118 (2 sheets MO, US), 43122 (F); near El Tablon, Standley 75842, 75876 (F); near Jutiapa, Standley — nid 76109, 76052 (F); Chiquimula: near Ipala, Steyermark 30287 (2 sheets F). Sample size — 11. 15 8 (US). Sample size = 1 16. bu mays ssp. mexicana (CENTRAL PLATEAU teosinte). Mexico: Durango: Ne DURANGO. E. Palme! ; ; i : em luxurians (GUATEMALA teosinte). Mexico: Oaxaca: Near SAN AUGUSTIN, Oct. 1840. Liebman 743 (4 sheets US, MO, WIS), G. Collins 15 (5 sheets US, WIS). Sample size . Zea mays ssp. mexicana (NOBOGAME teosinte). Mexico: Chihuahua: Near NOBOGAME, Wilkes s.n. e sheets F, US), H. S. Gentry 17973 (US). Sample size — 3. L . Zea mays ssp. parviglumis var. huehuetenangensis (HUEHUETENAGO teosinte). Guatemala: Huehuete- 1983] DOEBLEY —MAIZE AND TEOSINTE 37 TABLE l. (Continued). nango: Near SAN ANTONIO HUISTA, /ltis & Lind G-120 PA, sheets WIS), Wilkes s.n. (2 sheets WIS), 43603 (US), McBryde 81862 (2 sheets NA, F). Sample size — C. Grown in a Common Garden at Pioneer Hi-Bred International Research Station near Homestead, Florida from October 1978 to February 1979: 19. Zea luxurians (GUATEMALA teosinte). Seeds from Guatemala: Jutiapa: Along hedgerow, small stream nd in maize field ca. 5 km W of AGUA BLANCA, 14°29'N, 89*42"W, alt. ca. 920 m. Jan. 1, 1976. Titis G-38. Cult. Doebley 376. Sample size = 8. 20. Zea luxurians (GUATEMALA teosinte). Seeds from Guatemala: Jutiapa: Along hedgerow, small stream and in maize fields 1.2 km N of EL PROGRESO on road to Jalapa, alt. 1,040-1,060 m. Dec. 1975. Iltis G-5. Cult. Doebley 377. Sample size — 21. Zea luxurians (GUATEMALA teosinte). Seeds from Guatemala: Jutiapa: Weeds in maize fields and hedge- rows on slopes above laguna, 1 km E of south entrance pass, LAGUNA RETANA, 5~7-9 km N of E Progreso, alt. 1,150 m. Dec. 29, 1975. Iltis G-20. Cult. Doebley 378. Sample size = 2. Zea ies (GUATEMALA teosinte). Seeds from Guatemala: Chiquimula: Maize ficlds and hedgerows, 2 km N of IPALA on road to Chiquimula, 14°38'N, 89°38’W, alt. ca. 800 m. Jan. 2, 1976. Iltis G-42. Cult. Panis, 379. Sample size = 23. Zea mays ssp. mexicana (CHALCO teosinte). Seeds from Mexico: Mexico: Edges of maize fields 0.5 km SE of LOS REYES on road to Amecameca, 19?20'N, 98°57’'W, alt. ca. 2,225 m. Dec. 1, 1971. Iltis & Cochrane 175. Cult. Doebley 374. Sample size — 24. Zea mays ssp. mexicana (CENTRAL PLATEAU teosinte). Seeds from Mexico: Michoacan: Weeds in maize field, near QUINCEO, a pueblito on the slope of Pico de Quinceo, 6 km (by air) NW of Morelia, 19?43'N, 101*14'W, alt. ca. 2,000 m. Dec. 4, 1971. Iltis & Cochrane 276. Cult. Doebley 375. Sample size = 15. 25. Zea mays ssp. mexicana (NOBOGAME teosinte). Seeds from Mexico: Chihuahua: Near NOBOGAME, obtained from G. Beadle s.n. Cult. Doebley 370. Sample size — 8 26. Zea mays ssp. parviglumis var. parviglumis (BALSAS teosinte). Seeds from Mexico: Guerrero: 4.5 km E of Mazatlan on road to EL SALADO. Nov. 1972. Wilkes s.n. (USDA Plant Inventory No. 181, p. 220, Accession No. 384061). Cult. Doebley 372. Sample size — 27. Zea mays spp. parviglumis var. parviglumis (BALSAS teosinte). Seeds from Mexico: Michoacan: Km 127 on road from HUETAMO to Morelia, International Center for the Improvement of Maize and Wheat Accession No. 8761. Cult. Doebley 373. Sample size — 28. Zea mays ssp. parviglumis var. parviglumis (BALSAS teosinte), Seeds from Mexico: Michoacan: Roadside cliffs, | km S of TZITZIO toward Huetamo, 19°34'N, 100°55'W, alt. ca. 1,500 m. Dec. 6, 1971, Iltis & Cochrane 308. Cult. Doebley 380. Sample size = 8. 29. Zea mays ssp. parviglumis var. / is(HUEHUETENANGO Guatemala: Huehuetenango: Along trail at TZISBAJ, alt. l, 510 m. Feb. 1964. Wilkes s. s "(USDA Plant Gene No. 177, p. 132, Accession No. 343233). Cult. DoR 371. Sample size = - Zea mays ssp. parviglumis var } (HUEHUE Seeds from Guatemala: Huehuetenango: Maize fields, ca. 1.5-2.5 km ENE of aN ANTONIO HUISTA on road to Jacaltenango, 15*40'N, 91*45"W, alt. ca. 1,300-1,400 m. Jan. 1976. Iltis & Lind G-120. Cult. Doebley 381. Sample Size = 8 N D. Populations of Herbarium Specimens Cultivated under Dissimilar Conditions: 31. Zea perennis (PERENNIAL teosinte). Seeds or rhizomes from Mexico: Jalisco: 1 mile S of railway station of CIUDAD GUZMAN, Oct. 28, 1921. Collins s.n. Cult. Iltis & Cochrane 26376 (3 sheets WIS), 26372 (2 sheets WIS); Jitis s.n. (WIS); S. Calderon s.n. (F); Peebles & Harrison 3527 (US); Silveus 798 (US). e size = 32. Zea luxurians (GUATEMALA teosinte). Origin of seeds uncertain (see Wilkes, 1967: 11): Cultivated at different localities throughout the world. Brazil: Luederwaldt 22277 (US), Reits 4744 (US), Oliva 83 (US). Cuba: J. G. Jack 6340 (US). Hawaii: Hitchcock 14889 (US). Phillipines: E. D. Merrill 11222 (US). USA: $i Florida: (US sheet 727070). Sample size — 7. Zea luxurians (GUATEMALA teosinte). Honduras: Choluteca: Near SAN ANTONIO DE PADUA (Stand- a dies Standley 27317 (2 sheets US), Freytag s.n. (2 sheets MO), Molina 5881, 5882 (US). Sample 34. Ys Moye ssp. mexicana (NOBOGAME teosinte). Seeds from Mexico: Chihuahua: Near NOBOGAME, Obtained from G. Beadle s.n. grown at € E v. Agric. Alumni Exp. Stn., Florida City, Fla. Feb.-May 1978. Cult. Iltis & Doebley s.n. Sample 35. Zea 4 Mays ssp. parviglumis var. pesos scd teosinte). Seeds from Mexico: Jalisco: LA HUER- “ee PII ANNALS OF THE MISSOURI BOTANICAL GARDEN TABLE 1. (Continued). TITA, Cerro La Petaca, et. trail to El Durazno, 8 km (by air) ESE of r"— Castillo, 19*33'30"N, 104?22'30"W, Dec. 14 7. Guzman s.n. Cult. paella: 178. Sample si 36. Zea mays ssp. pe eet aed HUETENANGO ecl e: atemala: [Vor.70 Gu Huehuetenango: Near SAN ANTONIO HUISTA. piscis s.n. (MO sheet 1168094), Weatherwax s.n. (MO sheet 1213458). Sample size — E. Maize Populations: 37. Zea mays ssp. mays (Race PALOMERO TOLUQUENO). Mexico: Mexico: International Center for the Improvement of Maize and Wheat No. BA 70-539# (Mex 5). Grown at Purdue Univ. Agric. Alumni Exp. Stn., Florida City, Fla. Feb.-May 1978. Sample size — 16. 38. Zea mays ssp. mays (Race NAL-TEL). Mexico: Campeche: International Center for the Improvement of Maize and Wheat No. tep 62A 906# (Campeche 42c). Grown at Purdue Univ. Agric. Alumni Exp. Stn., Florida City, Fla. Feb.-May 1978. Sample size — 19. 39. Zea mays ssp. mays (Race CHAPALOTE). Mexico: Sinaloa: International Center for the Improvement of e and Wheat No. TL 7B 5# (Sin 2). Grown at Purdue Univ. Agric. Alumni Exp. Stn., Florida City, a. Feb.-May 1978. Sample size = 19. 40. eis mays ssp. mays (Race Conico). Mexico: Mexico: Depauperate plants 1-1.5 m tall. 2.5 km W of VILLA VICTORIA, 48.5 km W of Toluca on road to Morelia, 19°25’N, 100°02’W, alt. ca. 2,600 m. Sept. 28, 1978. Iltis, Doebley & Lasseigne 760. Sample size — 7. studied by Alava (1952) and by Wellhausen and his associates (1952). However, since differences exist in how they were applied here and else- where, a discussion of each of the 17 traits is necessary. The traits measured for this study may be di- vided into three categories: (1) tassel, (2) spikelet, and (3) outer glume. In order to reduce error variability when measuring these characters only pedicellate spikelets from the middle portion of the central spike (see Alava, 1952) and only tas- sels terminating the main culm (see Wellhausen et al., 1952) were used. However, on some her- barium specimens, one cannot determine if the tassel came from the main culm or from a lateral one, so a few tassels from lateral branches may have been inadvertently included. A. TASSEL CHARACTERS: [With the exception of branch number these characters were all mea- sured in millimeters (mm) and without magni- fication.] 1. Branch Number. Total number of branches excluding the central spike. Thus, for the tassel in Figure 3 the branch number is fourteen. No attempt was made here to distinguish between primary, secondary, and tertiary branches. Such distinctions, however, are not unimportant tax- onomically. In fact, Wilkes (1967, p. 110) has noted differences in degree of tassel branching between his races of teosinte. But, as Wilkes also observed, the distinctions are often obscured by environmental variability. he terms primary, secondary, and tertiary are used differently by Anderson (1944a, 1951) and Alava (1952) on the one hand, and Wellhausen and his associates (1952) and the successive au- thors of the Races of Maize series on the other. - What the former authors refer to as the primary — branch or central spike, the latter authors term ! simply the central spike (axis). Then, single i branches departing from the central axis are call secondary branches by Anderson and Alava, and | primary branches by Wellhausen. In general, An- derson and Alava always designate branches onè order higher than Wellhausen. The Wellhausen — system is adopted here (Fig. 3) because most Ie searchers (Wellhausen et al., 1952; Goodman. - ; Wilkes, 1967; oodman & Pateni 1969) worki ng with maize an it, and because this system has the ud P" vantage of considering the central spike not a$2 branch, but as an axis (in the same way that one would not ordinarily refer to the main stem % - any plant as a branch). A 2. Central Spike Length. Distance from the tp — of the uppermost spikelet of the central spike 10 i the point of departure of the uppermost primary — tassel branch (length A-D on Fig. 3). A few aù- thors (Wilkes, 1967; Mangelsdorf, 1974; Bir - 1978) reported that the southeastern Guatema- lan teosinte (Z. luxurians) lacks a central spike | However, all teosintes possess a central spike, although it is generally less prominent in tassels of sect. Luxuriantes. Photographs of southeast | ern Guatemalan teosinte tassels often show the central spike, which usually projects beyond the | other branches (Wilkes, 1967: plate XVII; Fig | 16, this report). The confusion surrounding this LH — bel b 1983] FIELD SPECIMENS © POPULATIONS FROM ONE LOCALITY ^ POPULATIONS OF HERBARIUM SPECIMENS @ MAIZE POPULATIONS GARDEN SPECIMENS © POPULATIONS GROWN UNDER UNIFORM CONDITIONS ^ POPULATIONS GROWN UNDER NON-UNIFORM CONDITIONS O MAIZE POPULATIONS DOEBLEY — MAIZE AND TEOSINTE FIGURE 2. Distribution of the populations of Zea sampled for this research. matter has been discussed fully elsewhere (Doe- bley & Iltis, 1980). 3. Branching Axis Length [or Length of Branching Space (Wellhausen et al., 1952)]. Dis- tance along the central axis of the inflorescence Tom the point of departure of the uppermos tassel branch to that of the lowermost (distance E on Fig. 3). Srouped together, such as the Condensation In- dex applied by Anderson (1944a) to the lowest Prt branch. The teosintes, however, nearly in iis have only one spikelet pair per node (as ranch N of Fig. 3) so Anderson's Conden- on Fig. 3) divided by five. Because the tassel branches, as opposed to the central spike, do not show th ffects of d tication, this trait, in contrast to Central Spike Internode Length, should be useful for comparing maize and teo- sinte. B. SPIKELET CHARACTERS: [Measured under a binocular dissecting microscope (15X).] 6. Spikelet Width. Maximum width of the fully developed spikelet just prior to anthesis, (dis- tance A-B in Fig. 4). Since many herbarium spec- imens were past anthesis with shriveled spike- lets, there exist iderabl variability for this trait. 7. Pedicel Length. Pedicel length from the ra- chis to the base of the spikelet (distance D-E on e ig. 4). 5 ye Length. Length ofthe longest mature anther in the chosen spikelet. When the central spike of a herbarium specimen had long passed anthesis, an anther from a spikelet of a major lower primary branch was used. C. OUTER GLUME CHARACTERS: 9. Glume (Spikelet) Length. Length from tip to base of spikelet (distance C-D on Fig. 4). Mea- sured under a dissecting microscope (15X). 10. Glume Width. Maximum width of glume 40 ANNALS OF THE MISSOURI BOTANICAL GARDEN Bi ee T9 URE 3. Diagramatic drawing of the Zea male iniae F, G, and N are primary branches; H is ch; I and J are tertiary branches; and ch; A-D = len ngt th of es peso branchi -] es of five internodes on a major lower primary branch. (distance A-E on Fig. 5). The glume was re- moved from the spikelet, flattened, and mounted on a microscope slide in Permount, a histological mounting medium. 11. Wing Width. Width of wing measured where widest, usually 1 or 2 mm below spikelet tip. Wings are projections of the major nd nerves of the outer glume and are most prom nent along the upper one-third of the ipikelet Measured under a dissecting microscope (40X) Only spikelets of the perennial teosintes pos- sess well-developed, easily measured wings (Hitchcock, 1922; Iltis et al., 1979; Doebley & Iltis, 1980). Zea luxurians has much less prom- inent wings, while all three subspecies of Z. mays basically lack this structure or show only traces of it. 12. Width of the Shoulders. Distance between the two major lateral nerves (distance B-D on Fig. 5). Measured under a compound microscope (50X) on the same glume that was used for char- acter 10. Although the terminology of Alava (1952) was followed here, the “shoulders” might [Vo.. 70 f Fh ape he Set hee ee g Tra Saree 7 T, i Mh FiGURE4. Drawing of a male pedicellate spikelet and its punit qe rachis internode with the sessile Ke . A-B = spikelet width; C-D = glumè (spikelet) lenti: and D-E = pedicel length. also be referred to as the “back” of the glume Using the term “back” would be especially ap- ad ol for the teosintes of sect. Luxuriantes ose glumes are flattened on the “back,” and Mes lack shoulders as found in sect. Zea 13. Shoulder Vein Number. Number oe veins inefatts ing the mid-vein, shoulder veins, majo ateral veins, and margin veins (e.g., total veit number is ten in Fig. 5). Counted under a com- - pound microscope (50X) on glumes mounted on slides 15. Lateral Vein Width. Width of one of the | two major lateral veins. Measured with a com - erem microscope (50X) on glumes mounted ki es. TI contains C Comune error variability because (1) the mar gins of the veins are often irregular; (2) many - older t veins; and (3) eV ven i at D magnification the units of measurement - were too large to record this trait accurately: ' Necerta as Alava (1952) attributed som c con- 1983] importance to this character, and as there are clear differences between the taxa of teosinte in this regard, lateral vein width was measured. 16. Shoulder Vein Width. Width of an average vein on the shoulders (Fig. 5). Measured in the same way as lateral vein width (character 15). 17. Mid-Vein Width. Width of the mid-vein (Fig. 5). Measured in the same way as lateral vein width (character 15). Some teosintes, and Zea luxurians in particular, lack a distinct mid-vein. In these cases, the vein closest to the center of the glume was measured. DATA ANALYSIS The quantitative morphological data were Studied using canonical variate analysis, a multi- variate technique, which computes synthetic give the largest possible ratio of the among-pop- ulation to the within-population variance, an F-ratio. In this sense, the first canonical variate best separates all populations. Additional ca- nonical variates are computed that, although also maximizing the F-ratio, are not correlated with Previous ones, and thus contain different infor- mation (Seal, 1964; Kowal et al., 1976). One can obtain a picture of how the local populations relate to one another by graphing the first ca- nonical variate against the second for all indi- viduals, Graphs incorporating the third or even urther canonical variates may also be useful. Under some circumstances one wants to look Only at a graph of a particular subgroup of local Populations included in the overall analysis. To 9 $0 one can perform additional canonical vari- ate analyses or close-ups including only the local Populations in the subgroup of particular inter- est. In this way one can get a better picture of bn relationships between the local populations Ithin the subgroup. Such close-up analyses can all Shay in two ways. (1) One may include ‘ginal populations in the close-up analysis d ate among and graph only those in diene in question. Here, the pairwise dis- mensio tween the populations in the full di- Nep: space will be the same as for the Vitintes analysis. However a new set of canonical sal be computed that maximally sepa- is best n the gr aphed populations. This option Populati en the variances and covariances of the Ons in each subgroup are similar, be- DOEBLEY— MAIZE AND TEOSINTE 41 — IGURE 5. Semi-diagramatic drawing of a flat- tened outer glume of the Zea male spikelet. A and E are the edges of the glume; B and D are the lateral veins; C is the mid-vein; area A-B = left margin; area E = right margin; area B-C = left shoulder; area C-D = right shoulder [terminology following Alava (1952)]. cause it uses the best estimate of the error vari- ation. (2) However, if the variances and covari- ances are not similar, then including all populations would give a biased estimate of the error variation in the smaller group of popula- tions that are being separated by the close-up analysis. Under these conditions one should ex- clude populations outside the subgroup in ques- tion from the close-up analysis. This will result in some changes in the between population dis- tances. The new analysis should more accurately discriminate among the populations as error due to dissimilarity of the variances and covariances has been reduced. For the present data set, dis- similarities in the variances and covariances ex- ist between the pop lati in diff t bgr ps. so the second option was used. The canonical variate analysis program used (CANCOV, Kowal, unpublished) standardizes the canonical variates so that the pooled within- population variance of each canonical variate equals one. When this is done each distance be- tween a pair of populations is measured in units of the pooled within-group standard deviation (Kowal et al., 1976). This distance is referred to as Mahalanobis's distance. This is a very useful measure of distance, because if the canonical scores are approximately normally distributed, one can use properties of the normal distribution 42 ANNALS OF THE MISSOURI BOTANICAL GARDEN to deduce roughly how well separated two pop- ulations are. For example, for normally distrib- uted populations 99.7% of all individuals fall within three standard deviations (equivalent to a Mahalanobis’s distance of three) of the mean. Local populations separated by twice this amount (a Mahalanobis’s distance of six) will be clearly distinct from one another. R th easily interpreted when individuals within the populations are multinormally distributed with equal variance-covariance matrices, all charac- ters were log transformed except Shoulder Vein Number and Wing Width, which were square- root transformed. It should be realized that mul- tinormality and equality of the variance-covari- ance matrices, while generally desirable, are necessary only if one wishes to use the technique to make statements. about the probability of cor- rectly Ifcanonical variate orl is used : solely as a method of projecting points (observations) in a multidimensional space onto a plane (a graph) as done in this paper, then normality and equality of the variance-covari- ance matrices are irrelevant. Canonical analysis has three features of im- portance for taxonomists. First, since taxono- mists tend to look for characters which show little variability within taxa and much variability between taxa, they in a sense perform implicitly in their minds what canonical analysis does ex- plicitly, i.e., maximize the ratio of among- to within-population variance. Principal compo- nents analysis, which discriminates on the basis of variability among individuals, does not gen- erally work as well with taxonomic problems in which the major goal is seeing how well the taxa are separated (Kowal, pers. comm.; Steudel, 1978). Secondly, canonical analysis does not re- quire the systematist to assign local populations to any taxon, and thus allows all local popula- tions to be judged independently of the research- er’s bias as to where they belong in the taxonomic hierarchy. Finally, since canonical analysis pro- vides a graph on which each specimen is repre- sented by a separate point, it allows the taxon- omist to keep track of individual plants. For these reasons, canonical analysis is well adapted to sys- tematic endeavors. Indeed, it has been fairly ex- tensively applied by both plant and animal tax- onomists (Oxnard, 1969; Robinson & Steudel, 1973; Kallunki, 1976; Riggins et al., 1977; Kel- ler, 1979; Kowal & Ruffin, 1979; Price, 1980; [VoL. 70 | Doebley & Iltis, 1980; Bell & Lester, 1980; Bow- - man, 1980; Reynolds & Crawford, 1980). RESULTS FIELD DATA A graph of the first and second variates from an initial canonical variate analysis of the 18 field and four maize populations (Fig. 6) reveals two clusters: one corresponding to sect. Luxuriantes, which contains Zea diploperennis, Z. perennis, d mays. The coefficients of the eigenvectors mea- sure the contribution of each trait to each ca- nonical variate, and each eigenvalue, expressed as a percentage of the sum of all eigenvalues, measures the relative amount of the total among- pulation Sud A explained by a canonical edle (Ta . Each eigenvector gives the weights (in EM form) used to calculate — the weighted sum of the original traits, that i5 the canonical score, for an individual. Frg m the eigenvectors, one can see that ind on Fig ure 6 with long lateral branch internodes, narrow © glumal wings, wide mid-veins, short glumes, and a small total vein number receive the largest val- ues for canonical variate one. These traits best separate he. two sections of the genus, Dochig | and Iltis (19 ful in PUE Rec s the two sections. The second canonical variate on Figure 6 sep- i arates the three subspecies of Zea mays and shows Z. mays ssp. mont to nus closest morphologically i to Z. mays ssp. xicana. For this canonical | variate, the Meri ru (Table 2) indicates that indiv aa , like those of the maize races i volved, wath short central branch internodes, long glumes, few tassel branches, and a long central spike receive the largest scores. Zea mays SSP- _ parviglumis, on the other hand, with small glumes, — ong internodes on the central spike, many tasse — branches, and a short central spike, receives tht _ smallest values. The three species that compose the lower clus _ cm ' ter of individuals (sect. Luxuriantes) do not se — arate from one another along either axis, al- PV M E41 AE 1 LEONE 1: dci th species are large (Table 3). In fact, Zea luxuria | is very well separated from the perennial ta - (D = 6.1). The explanation for this distortion is simply that because Figure 6 is in reality a 17 dimensional figure compressed into two dimen- 1983] DOEBLEY~— MAIZE AND TEOSINTE FIRST CANONICAL VARIATE 8 FIRST CANONICAL VARIATE Figu 7 ZER-SECT. LUXURIANTES: FIELD DATA ZEA: FIELD DATA e BGO. A (ET D.la2:- A4 D. Bo Id p uM UI Ne TCU. EE y. O 15.0 12 412.5 10 Sect. J10.0 c z Zea z i OD a 7.5 * ES X S. B ^ di i inue UE F SEC ae 2. e am a a tie - XX 312.5 a® a^a x * ^ a" - .0 -2. wire x 4-2.5 -5. 3 -5.0 * S XX 4-7.5 x ze -10 Luxuriantes ; -10-0 4-12.85 ee E e: Wir mE T d-15.0 SECOND CANGNICAL VARIATE emcees. ee eee eee oe ee wou NO; e ing ANCHA "p; parviglumis (^), and ssp. mexicana a (t), an tions of the wild taxa Nobogam SECOND CRNONICRL. VARIATE ZEA SECT. ZEA: FIELD DATA 2. 4. 10 abs Uu j r 4 x 3 X Xx x -8- x 4 x x a Xx X 1 X "o n 4. " + ge A 4 + + i 2 * -o. +H + ] peel 5 n + m + + Eis. a ttt ar 1 4-4 a E as 1 ^ Jis s y J 1 1 ao Ee We g^ 2. 4. 4m OR 5 (*), ssp. parviglumis (O) an Z. diploperennis (+); Z. perennis non CIU . luxurians ( 3 field popula of Zea mays including Z. mays me (Y); and Z. mays ssp. parviglumis vars. parvi FIRST CANONICAL VARIATE ERBE r o iT te eet Pe a e S EEE oe TEE E Ue UECDEÍÁ Eees 42. 10 Deis $9 J. 1 PUn pet e a ea BA A J RENN a T TEE. * P ds ^ A 44 Ps -4* a T A^ a o A a^ R 43 A^ 73 - A a 2. 4^ * ^ 4 4 si. ^ a) z 10 . d-1. x x = x xx x + d: + à ++ x + +o x E x RCM z ns + x + eas x x 4 x EX 4-s. x m x x 4-6: is x 4-7. fi "ES --8. [ e Hu > i w U N -L o Lx EA A WE telo sue lua Jj. Xe. 3. SECOND CRNONICRL VRRIRTE FIRST CRNONICRL VRRIRTE nay 2. Graphs of the first and second variates from canonical an aize populations inciadine Zea diploperennis Me p. mexicana (A).— Qo. DADG tions of Zea sect. Zea including Z. mays ssp. mays > bogame WILD ZEA MAYS: FIELD =DRIA 358,58. eee eh .0.. 1| eee. 4. De id ' * i LU ' L LU L LI L ] -6. x ssp. mex B ^ ut x Mw 4X ] ~ Be 42. x X x x M. nee. 4 x x+ 4 + x n ] +x x O19 x s [u] * 4 S E à ^ n J H-2. a * 4, gg 4 a "has O á a a di r A 1:59 -4 A yx ^ ^ ^ ssp parvi -B. -B. 1 1 1 L UA FTE FE . -4. d = 0 be 2 $5 «4 B5. SECOND CRNONICRL VRRIRTE Z. peren alyses.—6. 18 field populations of nis (X); Z. joris (Y); and Z. sect. Luxuriantes EPOS SITOS SO), and PIEDRA 9 field populations of Zi LOS DE of Z. mays ssp. mexicana. —9. Central reis xicana races Chalco (+), iglumis (A) and huehuetenangensis (D). Taste 2. Eigenvectors (standardized such that the largest element in absolute value equals one) and percent variance for the can Figures 6 to 13. onical variates from Figure 10 11 Canonical Variate l 2 l 2 l 2 1 2 1 2 1 2 Characters l. BRAN-# .015 —.500 .068 115 —.411 —.756 —.176 —.751 —.437 .799 —.707 —.439 2. CNSP-L .514 .459 .184 .802 315 —1.000 —.212 —.378 —.736 .487 —.820 7.357 3. BRAX-L .246 .120 .266 .134 —.020 —, 101 .145 .371 —.350 .528 —.133 72:320 4. CSIN-L —.089 —1.000 .080 .027 — 1.000 stat —.272 .054 —.514 .468 —.479 —.809 5. LBIN-L 1.000 .222 .404 — 1.000 —.145 —.286 —312 —.218 —.046 —.135 .056 —.281 6. SPIK-W —.262 4433 —.178 .134 .399 = 116 12 —.064 .259 —.251 :261 —.045 7. PEDI-L —.062 —.335 SER —.038 7.210 .342 .105 .168 .174 .030 .188 —.775 8. ANTH-L .065 .209 .027 .126 —.115 341 == 119 .013 .068 —.109 —.094 —.418 9. GLUM-L —.766 .673 —.400 Det» .952 .597 1.000 .011 1.000 — 1.000 1.000 1.000 10. GLUM-W .094 .332 7.365 .241 .285 .197 ,132 .894 222 —.265 .189 —.614 11. WING-W —.964 —.346 —1.000 .282 = 185 .259 —.065 —.210 .540 .092 251 .016 12. SHLD-W —.341 — 215 .058 —.827 33.193 .396 .144 —.956 Un 7051 .361 .024 13. SHD-V# —-.238 —.172 .865 .079 .041 —219 .074 .665 .736 .510 .050 —.264 14. TOT-V# —.705 —,021 .160 .962 .095 .091 .200 — 1.000 .054 .600 —.024 —.878 15. LATV-W —.422 —.098 ~,110 31 .109 .004 .190 .404 .236 .454 .099 —.007 16. SHVN-W .097 .025 .099 ~.587 .039 —.086 —.078 — 292 —.331 E215 —.121 —.115 17. MIDV-W .801 .066 153 —.452 =.259 .038 —.240 .820 —.252 —.277 —.211 —.127 recent variance 46.26 33.10 66.81 12.65 72.93 9.21 52.08 16.34 31.97 10.70 60.72 17.46 vt N3GIVO TVOINV.LOS P3IOOSSIN JHL AO STVNNV 0L 10A] 1983] DOEBLEY — MAIZE AND TEOSINTE AAanhal 152 25. os at. 45 ee Figure 6. Data given below are the means TABLE 3 g popul grap (and ranges) for pairwise interpopulation distances, grouped by taxa. Zea mays Zea mays ss S Zea Zea mays p. sp. Species diploperennis | perennis luxurians ssp. mays | mexicana parviglumis Number of populations l 3 5 E 6 3 Zea diploperennis ~- Zea perennis 3.9 4.2 (3.5—4.1) Zea luxurians 5.9 6.3 3.9 (4.6-7.9) (5.1-8.6) Zea mays ssp. mays 11.9 12.0 : 5.6 (11.5~12.2) (11.1-13.0) (9.4—13.9) Zea mays ssp. mexicana 9.9 9.4 (8.7-10.7) — (6.3-10.2) Zea mays ssp. parviglumis 12.7 12.0 (11.6-13.6) — (9.4-14.4) 9.0 9.3 4.1 (7.0-12.3) (7.0-11.3) 11.3 12.4 6.1 4.1 (9.4-13.9) (9.6-14.4) (4.573) sions, some pairwise distances between the pop- ulations and especially between those popula- tions within the same cluster are not accurately represented. One may obtain a better view of ow the taxa within the two sections relate to one another by performing close-up analyses of each cluster as discussed in the Materials and Methods. Figure 7 is a plot of the first two variates from 9 close-up analysis of the subgroup correspond- ing to sect. Luxuriantes. On this graph the five populations of Zea luxurians, all represented by triangles, receive positive values for the first ca- E : x he Second canonical variate discriminates ks ids à four perennial populations, but fails niam Es e perennis from Z. iplope- "WS : ^ hi$ 1s not particularly surprising as these "MER are very closely related, Z. perennis az x ly being the autotetraploid derivative diis Ploperennis. The characters that best dis- Suish these two taxa are their rhizomes (the like short shoots) and general robustness Plants (the diploid having wider leaves and Stems) (Iltis et al., 1979). In inflorescence Of the taller morphology, Z. diploperennis has on the average a greater number of tassel branches and wider spikelets with wider shoulders. However, these differences are not large enough to allow its sep- aration from Z. perennis on Figure 7. The basic identical, one might treat them as two subspecies of Z. perennis. However, the great dissimilarities in their dg laspect justify main- taining them as separate species (Iltis et al., 1979). A canonical graph ofthe taxa of sect. Zea shows the maize populations separating well from the two wild subspecies of Zea mays, but these two wild taxa overlap somewhat with one another (Fig. 8). Of the two wild subspecies, Z. mays ssp. mexicana is closest to the cultigen maize, and closest to the Nal-Tel race of maize, in particular (Table 4). The Nobogame specimens, which have slightly smaller spikelets than the Chalco and Central Plateau races, still appear closest to Z. mays ssp. mexicana rather than Z. mays ssp. parviglumis (Fig. 8). Part of the reason for this is that, like Z. mays ssp. mexicana and unlike Z. mays ssp. parviglumis, race Nobogame char- acteristically has few tassel branches. Maize separates from the teosintes along the first canonical variate primarily on the basis of its shorter central spike internodes and longer glumes (Table 2; Fig. 8). These two traits, which are correlated with the number of kernels per ear and kernel size, respectively, probably reflect the indirect effects of domestication. As the pre-Co- 46 ANNALS OF THE MISSOURI BOTANICAL GARDEN 1 e i. 4 [VoL 70 - . Fi TABLE4. M g populations grap (and ranges) for pairwise interpopulation distances, grouped by taxa. Figure 8. Data given below are the means — Zea mays Zea mays | Zea mays Race ssp. F Taxa ssp. mays! Nal-Tel mexicana parviglumis Number of populations 3 1 6 3 Zea mays ssp. mays! 3.88 Race Nal-Tel 6.7 — (6.0-7.3) Zea mays ssp. mexicana 9.3 E3 3.8 (8.6-10.9) (6.4—8.5) Zea mays ssp. parviglumis 12.8 9.3 5.1 4.0 (11.8-13.9) (8.9-9.8) (4.4-7.0) ! Excluding race Nal-Tel. lumbian Mexican Indians selected for larger size and a reduction in the length of internodes on the central spike (Iltis, 1971, 1981; Galinat, 1971; Beadle, 1972; Doebley & Iltis, 1980). Thus, maize differs from the annual teo- sintes of Mexico and west-central Guatemala mainly in characters that were transformed by human selective pressures, supporting the hy- pothesis that maize is domesticated teosinte. Pinpointing the particular taxon of teosinte which was transformed into maize is more dif- ficult. Ni approach thig promem, one may use d with Sard halanobis *s dist ance )t lut vergence. This proceda is reasonable though phologically quite distant from maize (Fig. 6; Table 3) cannot be considered likely candidates for the direct ancestor of maize (Doebley & Iltis, 1980). Of the two remaining taxa, Z. mays ssp. mexicana and ssp. parviglumis, the former is morphologically closest to the Mexican maize races studied here (Fig. 6; Table 3). This close- ness reflects the facts that Z. mays ssp. mexi- cana, like these Mexican races of maize, has larg- er spikelets, fewer tassel branches, and a longer central spike than Z. mays ssp. parviglumis. In Figure 8, race Nal-Tel of maize separates out from the other three maize races (Conico, Palomero Toluquefio, and Chapalote). This is of interest because races Nal- Tel and Chapalote have generally been regarded as closely related (Well- hausen et al., 1952, p. 58; cf. Goodman 1972, p. 179). Yet, the great dissimilarities between these mandates that we consider the possibility that their similarity in ear traits results, at least in part, from convergent evolution. The fact that b the two races are so widely separated geograph- ically — Nal- Tel from the Yucatan and Chapaloté from northwest Mexico— supports this hypoth- esis. Results from the study of this small sample of maize races points to the potential of tassel and r spikelet traits in helping to understand the tax- onomy and evolution of maise races. Such an approach in somè | ways more otsjoctive means of measuring the dé mendous differences between the female inflo- : rescences of these two taxa (Anderson, 19444 _ 1951). 3 Figure 9 shows in greater detail the relation ship between Zea mays ssp. mexicana and SSP: parviglumis, with each of Wilkes’s (1967) races given a separate symbol. Again, the two subsp® cies separate fairly well with a small area of ove lap. Similar results are obtained by graphing thes | populations using spikelet length versus spikelet width as done by Iltis and Doebley (1980). Figure 9 also reveals that within Z. mays SSP: mexicana, races Chalco and Central Plateau do not separate at all, demonstrating that there 15 no male inflorescence morphological basis for considering them separate taxa. Within Z. m4 | ssp. parviglumis, vars. parviglumis and huehue tenangensis do diverge to some degree along © nonical variate two, but with so few specimens 1983] of the latter the case for granting the two races, Balsas and Huehuetenango, even varietal status is weak. GARDEN DATA To supplement the field data and help eluci- date the degree to which environmental differ- infl phology of the taxa, plants grown in a common garden were studied. Un- fortunately, the perennials grew very poorly in the garden and often failed to produce tassels. For this reason only the annual teosintes are in- cluded here. An initial canonical variate analysis of all 12 garden populations (Fig. 10) shows the same ba- sic pattern as the field data. Namely, Zea lux- urians is very distinct from the other annuals, which in turn separate into Z. mays subspecies mexicana and parviglumis. The major characters involved in separating Z. luxurians from Z. mays along both the first and second canonical variates are glume length, shoulder vein number, total vein number, wing width, central spike length, branching axis length, lateral vein width, branch number, and length of internodes on the central spike (Table 2). This corresponds well with the field data, which showed Z. /uxurians to be dis- tinct on the basis of its higher vein number, wider wings, shorter branching axis, and broader lateral nerves. One difference between the garden and field Populations is that the garden populations of the two wild subspecies of Zea mays separate much More effectively than the field populations, and Mies any Overlap. The distance between these bi Subspecies is greater for the garden data (Ta- es 3 and 5). This sharpening of the differences oy bia garden data, so the sharpening of dif- ces probably does not reflect dissimilar sampling, oo more closely at the two wild subspe- biis Ll mays (Fig. 11), one observes the same except “ese of variation as seen in the field data Subspec; more effective separation of the two the feta d other difference, as compared to rather di a IS that race Nobogame appears Mays ss Stinct from the other two races of Z. ation iei mexicana and shows a closer associ- : Mays ssp. parviglumis (Table 5). The DOEBLEY —MAIZE AND TEOSINTE 47 eigenvectors involved in Figure 11 (Table 2) in- dicate that while races Chalco and Central Pla- teau differ from Zea mays ssp. parviglumis by their longer spikelets, shorter branching axes, and fewer branches, race Nobogame is distinguished from Z. mays ssp. parviglumis by a combination of these characters, as well as fewer nerves on its outer glume and shorter internodes on the central spike. The distinctiveness of race Nobogame for the garden data does not agree well with the field data, which showed complete overlap of this race with races Chalco and Central Plateau. The prob- able reason for this is that race Nobogame, being the northernmost teosinte and adapted to flower with longer days, is forced to flower prematurely in Florida before it attains much vegetative de- velopment. Thus, the plants are much more de- pauperate and distinct from specimens of races Chalco and Central Plateau, which flower much later in the garden and obtain a normal amount of vegetative development. For this reason, and because herbarium specimens of races Chalco, Central Plateau, and Nobogame cannot for the most part be told apart, it seems best to include these three races within a single subspecies. MISCELLANEOUS POPULATIONS Six populations of plants cultivated under non- uniform (uncontrolled) conditions are of interest as some of them are from areas not otherwise sampled, such as Honduras. Figure 12 includes three such populations, as well as all other pop- ulations of sect. Luxuriantes. On this graph the Honduras population clusters with the other populations of Zea luxurians as does the strictly artificial population composed of Guatemala teosinte plants (diamonds of Fig. 12) cultivated at different localities around the world. Clearly, while the Honduras population does belong in Zea luxurians (Guatemala teosinte), where it was placed by Wilkes (1967), more and better ma- terial is necessary to determine if it deserves dis- tinction at the subspecific level. Another pattern of interest revealed in Figure 12 is that the garden populations group together, as do the field populations and the uncontrolled cultivation populations. This demonstrates the extreme effects of the environment on the mor- phology of the plants, emphasizing the impor- tance of comparing only plants grown under the same conditions, especially when looking at dif- ferences at the subspecific level. 3 1 E. 48 ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 - 10 ZER: GRRDEN DRTR | 11 WILD ZEA MAYS: GARDEN DATA | in: oe 5. 4. 73. ees -1. Oe d. 2. 3. 4... ce a rene 0. 2. Sree T T T T T T T T TOTI i PENA 1 ! l ' ' A | E J L ; | * ] 8. + es 8.L- -78 | F + 5 * d e L * ] n * * + E. 46 Br os = tuo ad 6 x X x zx s | - E Q4 = x $ d s "E F i ar = zi dd x E 4, i gon | x 368 PIEN l | > i i > We * 4 ¥ a ete * J za Lp n Hho J f o 2 ] a E m^ H = E x Š% x 4 2 iP Y 1 © HL n x x zig. oO o.L [ul T - od. z x = ng ] c F x” m 1 T m Y Y A e H Xa F E e F SAC ^ J - [ m o 98 — ^ p EN m -2 [7] Bx ae ex ^ ^ nim Qo ses & ^m = [ eure Na D a uu "T 1 | u [ og? WL noa Sie n J -4.} “ong n Eu aes d T* F D m o m L 4 | F on8 - - 3 b m ful a [ul L By -B.[ a Te ce 4 8 | 1 1 1 L Us as BS M apy oss HUC LAVA 8 gu i ere + * "y Me ir E39 Comic SEIS PTS 0. 2 4. 6. SECOND CRNONICRL VRRIRTE SECOND CRNONICRL VRRIRTE 12 zeA SECT. LUXURIANTES: ALL DATA 13 WILD ZEA MAYS: ALL DATA EI =A. “2. 0. 2. di sb.xdiie3. 22. bs 0... 1902548. As L 1 I ' T Lu Li j I $73. 3443-7 8 1 I LU 1 1 Lu FUE LU B e.- 3 E 1 i D T 6.} : ii i AE om oh ! $ x a g "jJ Ssp. mex X x x E ë * m m m2 O 4 Lu xX d =x a} [1 [e x e Srt € XXE X* x 42 d L yes Eabb xXx x - = [4 Cn m E + x di *à o^ i = 2. a ++ Mew xt +m tH E 5 x + -o gd 0. E x 40 uud Tat * X x 4 [s - eg. + » E: F x * oJ o DEEP «s 3 1 ] m La m " $ d Y --2. S , Lo "e 88 S "m | I L 5 2 = H v^ om "JË og, © B 4 + ar XX J s [ a a 5g LA [s ra S Yy y x d * x x du be w es a P 4 wo Y * ue Rs wn as m^ S J ma E y T XX Set 4 ac E at wa” 4 u x* " 1 = a, e Aa ] 6.F x E. » ag EN ^ ^ d = x 1 ^ 4a A 4 [ x XX ^a 4 £ A a J " x | ~T x Ax X PES. we Ae ssp. parvi. Bi ENWWNNWWRM a a IST xix Pid iu . ms a: ae a — D T "T T. gc. [(. 72. 5. 4. E 2 SECOND CRNONICRL VRRIRTE SECOND CANONICAL VARIATE —— FiGURES 10-13. Graphs of the first and second v. ariates from canonical analyses.— 10. 12 garden lations y obe including Z. luxurians (+); Nobogame (A) of 2 Z. mays ssp. mexicana (X), and ssp. hipeehunis (C), and rat Z. mays ssp desire na.— 11. 8 garden ponniations of the wild taxa of Zea mays includ Z. Mays ssp. mexicana races Ch o (+), Central Plateau (X (X ight vars. parviglumis (A), and rerit ensis (D). — 12. e populations of Zea a sect. _Luxuriantes including Z. diploperennis (+); Z. perennis, 2 field populations (X). contr m Collins’s CIUDAD GUZMAN collection (Y) and Collins's CIUD GUZMAN field population (A); and Z. prapa field populations (*), Honduras population (*), cont 1983] DOEBLEY — MAIZE AND TEOSINTE 49 Je Mahalanobis’s distances among populations graphed on IE 10. Data given below are the means (and ranges) for pairwise interpopulation distances, grouped by ta Zea mays Zea mays Zea SSp. Race ssp. Taxa luxurians mexicana Nobogame parviglumis Number of populations 4 2 l 5 Zea luxurians 31 Zea mays ssp. mexicana! T2 4.1 (6.5—7.9) Race Nobogame 10.0 6.1 7 (9.7—10.4) (5.4—6.9) Zea mays ssp. parviglumis 10.5 6.9 227 3.4 (8.9-12.0) (5.5-8.1) (5.3-6.3) ' Excluding race Nobogame. On the lower portion of Figure 12, the Zea perennis population composed of plants culti- vated from seeds and rhizomes collected by Col- lins in 1921 shows no closer relationship to Col- lins's field specimens from the same locality than to the perennial teosinte populations from other Stations. In fact, this field collection of Zea pe- rennis from Ciudad Guzman is separated by a greater distance from its corresponding cultivat- ed population (D = 6.0) than from the La Ven- tana population of Z. diploperennis (D = 5.2). Here then, two samples from the same local pop- ulation fail to cluster closest to one another due 2 differences in the environments in which they Figure 13, which includes the two wild sub- Ken of Zea mays, shows a pattern of variation ee to Figure 12, n the cultivated qun ations ofeach subspecie subspecif . Again, this suggests that, at ds Nn iic level of evolutionary differentiation by a genetic differences can be easily obscured eee sabe modifications of the pheno- bei € characters which seem to vary most een field and garden populations of both Z. luxurians and Z. mays are vein number and spikelet width, the ratio of vein number to spike- let width being larger for garden grown plants. Finally, Figure 13 shows that the annual teo- sinte population from southwestern Jalisco re- cently discovered by Guzman (1978) at La Huer- tita near Casimiro Castillo clusters with Z. mays u arviglu n supported by their similar p aive hologies and the fact that the La Huertita population, like other ssp. parviglumis popula- tions, ws in seasonally moist habitats with rainfall in excess of 1,200 mm annually (Doe- bley, 1984) SUMMARY OF RESULTS The preceding analysis of tassel and spikelet morphology shows that the genus can be divided into sections Luxuriantes and Zea. The former section is characterized by few tassel branches, a short branching axis, short internodes on the tassel branches, and highly nerved outer glumes on which the prominent lateral nerves are de- veloped into wings near the apex of the spikelet. Section Zea, on the other hand, is characterized y a larger number of tassel branches, a longer branching axis, longer internodes on the tassel — ees Populations (O) and uncontrolled garden population (Q).— 13. 9 field, o controlled garden and 3 uncon- (+ () an garden populations of the wild taxa of Zea oni —— Z. sae . mexicana, 5 field populations wes 2 garden Po (X); Z. mays ssp. mexic ce No 5 beid co pulation (*) and 2 garden ulations a mays -— Anis ipe: 4 field pésténede: (A), 5 saa re (CJ) and the LA 50 ANNALS OF THE MISSOURI BOTANICAL GARDEN branches, and fewer-nerved outer glumes that lack wings. Within sect. Luxuriantes, Zea lux- urians is completely distinguished from the pe- rennials on the basis of its much more numer- ously nerved outer glumes with smaller ie could not be distinguished on the basis of tassel morphology. Within sect. Zea, the wild popu- lations (teosintes) can be divided into two groups corresponding to Z. mays ssp. mexicana and ssp. parviglumis. On the oe heces cmd h aphs, these two sub mall a of overlap for field data but no DUE pe wie grown in a uniform garden. For both the field and garden data, the two subspecies were distin- guished in that Z. mays ssp. mexicana has fewer branches, a shorter branching axis, and larger spikelets and anthers than Z. mays ssp. parviglu- mis. Within Z. mays ssp. mexicana, races Cen- tral Plateau, Chalco, and Nobogame were com- pletely indistinguishable for field data; however, for the garden data, race Nobogame separates from the other two. Garden specimens of race Nobogame tend to have fewer tassel branches and smaller spikelets than typical Z. mays ssp. mexicana. This seems to be an artifact of its adaptation to the long days of northern Mexico, which forces it to flower prematurely in Florida. Within Z. mays ssp. parviglumis, no consistent differences were found between vars. parviglumis and huehuetenangensis. Finally, Z. mays ssp. mays was shown to differ from its two conspecific wild subspecies only on the basis of characters that could have been altered during the domes- tication process. These characters include the length of internodes on the central spike, which are shorter in maize, and the length of the spike- lets, which are longer in maize. Internode length would have decreased as man selected for a con- densed ear with a large number of grains. Male spikelet size would have increased as man se- lected for larger female spikelets (kernels). In both cases, the parallel variation in the tassel and ear results because these structures are homologous (Anderson, 1944a). ILLUSTRATION OF TAXA Iltis and Doebley (1980) provided a taxonomic synopsis of Zea, while Doebley and Iltis (1980) illustrated the male spikelets of Zea to some ex- tent, providing a dichotomous key to the group. The goal here is to describe and illustrate further the taxa, with special reference to male floral [VoL. morphology. Those aspects of vegetative and f male inflorescence morphology that are essenti to the circumscription of the taxa are also d d scribed. However, the female inflorescence, 4 especially its immensely complex branchi pa terns in the teosintes are slighted here, and à in need of a separate detailed study. Similarl hybrids of the various teosintes and maize, though common in the field, are not describe below Zea Linnaeus, Species Plantarum 971. 175 Genera Plantarum, ed. 5, 419. 1754. I. Section Luxuriantes Doebley and Iltis, Am | J. Bot. 67: 982. 1980 | 1. Zea diploperennis Iltis, Doebley, and man, Science 203: 186. 1979. : Male inflorescences with (0—)3-13 + diverge to bcm ens branches, these 6-15 cm long, 12- wide, the central spike barely exceeding t midi branches in length (Fig. 14); branching ad a 1-4 cm long; spikelets 8.5—11.5 mm long, in S sile-pedicellate pairs, these distichously arrang on the branches; branch internodes scaberul 0 with prominent abscission layers between the! broad (1 mm wide), in cross-section triangu. with ciliate edges, and short (2-6 mm), the sp lets therefore crowded and overlapping (€. go spikelet pairs in 4 cm), with the tip of the se ; spikelet of each pair reaching the base of ther higher pair above on the same side (Figs. 23); pedicels scabrous, enlarged below the s pik let, 1.5-3.5 mm long; glumes of the spikelet £ brous (rarely scaberulous), sublustrous, of purple tinged, stiffand firm and somewhat brit when old; outer glume flat on back and tigh enclosing the inner, strongly green-nerved, ! nerves often clustered marginally near the cally ciliate prominent lateral wings (Fig. inner glume keeled along the mid-nerve, sca? patie ae it contacts the rachis. e inflorescences consisting of slenc pass spikes, each with 5-10 cupulate cases, these trapezoidal-cylindric, 6-9 mm on long side, 2.5-4.5 mm on the short side, 4-5m Loosely clump-forming perennial, with bO cord-like and tuber-like rhizomes, both of th” with short (2-6 mm) internodes. Rare, endemit to Sierra de Manantlan, Jalisco, Mexico, 1,400-2,400 m. DOEBLEY — MAIZE AND TEOSINTE S 14-17. Pressed male inflorescences.—14. Zea a ehh Iltis, Guzman, ier 3» sad — (plant GG).— 15. Zea perennis, Guzman s.n. (plant I).— 16. eg ine n n E ; MAYS ssp. mays (Mexican Pyramidal), I/tis and Doebley 405 ‘saan B). (Scale in c Fig ioe 45 4 s 52 ANNALS OF THE MISSOURI BOTANICAL GARDEN 2. Zea perennis (Hitchcock) Reeves and Man- gelsdorf, Amer. J. Bot. 29: 817. 1942. Euchlaena perennis Hitchcock. Male inflorescences similar to those of Zea diploperennis, except with fewer, (0—)3-8, erect or rarely somewhat nodding branches (Fig. 15), these 6-12 cm long, 10-20 mm wide; branching axis 1—2.5 cm long (see also Figs. 24, 25, 38). Female inflorescences similar to those of Zea diploperennis, except with the fruitcases com- monly gray or sepia, speckled with dark brown; weight of 100 mature fruitcases 8.3 g. Densely sod-forming perennial, with cord-like rhizomes, these with long (1—6 cm) internodes. Rare, endemic to the vicinity of Cuidad Guz- man, Jalisco, Mexico, alt. 1,520—2,200 m. 3. Zea luxurians (Durieu and Ascherson) Bird, Taxon 27: 363. 1978 Euchlaena luxurians Durieu and Ascherson. Male inflorescences with (0—)4-25 erect branches, these 7-20 cm long, 8-15 mm wide, the central spike often exceeding the branches in length (Fig. 16); branching axis 1-9 cm long; spikelets 8.5—12.5 mm long, in sessile-pedicellate pairs, these distichously arranged on the branch- es (Figs. 26-27); branch internodes densely sca- berulous or strigose with prominent abscission layers between them, broad (1 mm wide), tri- angular in cross-section with ciliate edges, + short (3-6 mm), spikelets therefore + crowded and overlapping (e.g. 10 pairs in 4 cm) with the tip of the sessile spikelet of each pair reaching the base of the next higher pair above on the same side (Figs. 26-27); pedicels scaberulous or stri- gose, enlarged below spikelet, 1.5-3.8 mm long; glumes of the spikelet scaberulous, rough, stiff and somewhat brittle when old; outer glume flat on back, tightly enclosing inner glume, and hav- ing numerous (10-28) indistinct nerves between its two prominent ciliate lateral nerves, the latter produced into narrow wings apically (Fig. 39); inner glume with 5-12 nerves, keeled along mid- rve. Female inflorescences consisting of slender distichous spikes, each with 5-9 light brownish gray t l cupulate fruitcases, these 7-1 1.5 mm on tons side, 3.7—6.5 mm on short side and 3-5 mm in diameter; weight of 100 mature fruit- cases 7.6-9.9 g. Robust annual, main culm usually highly branched and without any or only few tillers in [Voi m e wild. Honduras, southeastern Guatemala (Provinces of Jutiapa, Chiquimula, and Jalapa and southeastern Mexico (Oaxaca, San Augustin _ only), alt. 410-1,100 m. H. Section Zea. 4. Zea mays Linnaeus, Species Plantarum 97 1753 Male inflorescences with numerous branché : (in well grown plants many more than 12, exce] in certain races of cultivated Z. mays), these lak (stiff in some races of maize), the central spike - occasionally somewhat stiffer, stronger, and mort - densel i i : . mays) (Figs. 17-21); branching axis usua (1-)3-18 cm long; spikelets 4.6-13.0 mm lo usually in sessile-pedicellate pairs; branch int nodes variously scabrous-pubescent, slender (I s than 1 mm wide), rounded on the back, not strongly flattened, elongated (3. 5-8. 2 mm © T -1 B 3 o 3 qa 13 3 a o = S 2 oO 3 m oO n g 5 variously scabrous-pubescent, rounded on b the outer glume only loosely embracing the in one, both flexible and + papery in texture; outer the keels hair-like and soft, and not scaberul and tooth-like (Figs. 40-45); inner glume nerved. Female inflorescences either (in maize) a si branch, or (in teosinte) consisting of slender tichous spikes, these composed of 5-12 or mof + triangular cupulate fruitcases. 4a. Zea mays L. ssp. mays. Male inflorescences highly variable, with (02. | 3-30 or more branches, these 5-30 cm long. s central spike almost always strongly disti" | guished from the tassel branches (Fig. 17); spike | lets 6-13 mm or more long, in sessile-pedicella® _ 1 DOEBLEY — MAIZE AND TEOSINTE C ^ ^ dabis ` 4 S "s NE NN, r i Fa iape Vestae —18. Zea mays ssp. mexicana nee Chalco, J/tis and Doe Mey nays s exica ce Central Plateau, Palmer 743.— 20. Zea mays ss S iglumis "en at. parviglumis (race Baloo. This dbaker 362 (plant A).—21. Zea mays ssp. Suena var. huehuetenan- g Sis $ (r race Huehuetenango), Doebley 371 (cultivated in Homestead, Fla.). (Scale in cm. 54 pairs, these distichously arranged on the lateral branches (Figs. 28-29) and polystichously ar- ranged on the central spike (here single spikelets or groups of three spikelets are not uncommon the sean one sometimes Sd. sessile), the spikelets on the latter extremely crowded and pee ru (up to 26 or more spikelet pairs in 4 cm); glumes of the spikelet glabrous, scabrous, lanate, sericeous or villous (Fig. 40). Female inj ngle ter- minal spike (“ear”), this aisea (8-20 or massive central axis (“cob”), laterally displaced from their reduced, collapsed, empty, and hidden cupules, the entire spike tightly enclosed in a series of 8-12 or more leaf sheaths (“‘husks’’), all borne on a thick peduncle (“shank”). Monopodial or sparsely tillered annuals. World-wide cultigen. 4b. Zea mays L. ssp. mexicana (Schrader) Iltis in Annual Rev. Genet. 4: 450. 1971; Phyto- logia 23: 249. 1972; emended circumscrip- tion Amer. J. Bot. 67: 1001. Male inflorescences with 0-20(-35) nodding branches, these 5-25 cm long, the central spike identical with to slightly larger than the tassel branches (Figs. 18-19); spikelets (6.6—)7.5—10.5 mm long, in sessile-pedicellate pairs, these dis- tichously arranged on all branches (Figs. 30-31); glumes sparsely to densely scabrous (Figs. 41- 43). Female inflorescences consisting of slender distichous spikes; each spike consisting of 9-12 or more triangular cupulate fruitcases, the former enclosed in a single sheath and borne on a short to elongate slender peduncle; fruitcases 6-10 mm long, 4-6 mm wide, often pointed or “pinched” on the axial side, variable in color, gray, tan, or dark brown, and mottled or speckled with dark brown; weight of 100 mature fruitcases 5.6— 10.4 4b-I. Race Chalco of Wilkes. Plants vegetatively robust, maize-like, gener- ally with no or only few (1-6) tillers; leaf sheaths densely pilose and often dark red in color; leaves sparsely pilose; Valley of Mexico and its slopes, Mexico (cf. Wilkes, 1967), alt. 2,150—2,500 m. 4b-II. Race Central Plateau of Wilkes. Essentially identical to race Chalco except leaf ANNALS OF THE MISSOURI BOTANICAL GARDEN A [Vor. 70 | sheaths less densely pilose and not as deep red in color. Some populations are vegetatively somewhat depauperate and resemble race No- ogame. Michoacan, Guanajuato, and eastem Jalisco (cf. Wilkes, 1967), alt. 1,750-2,100 m. — 4b-III. Race Nobogame of Wilkes. Similar to race Chalco and especially race cdi tral Plateau except the plants vegetatively less robust with fewer tassel branches, narrower shorter leaves, and slightly smaller male and fe- male spikelets, the male spikelets having a fewer- - nerved outer glume. Nobogame, Chihuahua (d. Wilkes, 1967), alt. ca. 1,900 m. 4c. Zea mays L. ssp. parviglumis Iltis and Doc Í bley, Amer. J. Bot. 67: 1001. 1980. Male inflorescences similar to those of t more densely "branched (up to 100 or more branches) with tertiary branching much morè frequent (Figs. 20-21); spikelets markedly small- i er than those of ssp. mexicana 4.6—7.2(-7.9) mm - long (Figs. 32-35, 44—45), 1.6-2.8 mm wide. Female inflorescences similar to those of ssp. mexicana except with fewer (5—10) triangular cu- pulate fruitcases per spike; fruitcases smaller (5. | 8.0 mm long, 3.0-5.0 mm wide) and blunt 0 — the axial side; weight of 100 fruitcases 3. 1-5.6 + .9) B. 4c-I. Zea mays L. ssp. parviglumis Iltis and Dot bley var. parviglumis. Race Balsas of Wilkes: | Leaves commonly at least sparsely pilose 10 velvety pubescent, plants up to 3.5 m tall. Plans | of thorn scrub and open summer-green tropical ' deciduous forest, as well as maize fields and theif , edges, on well-drained slopes of mountains and hills in the Rio Balsas valley of Guerrero, Mi- | choacan, and Mexico at 600-1,600 m alt. (-1. 950 | m ? cf. Wilkes, 1967, p. 119), these in huge dens - and clearly wild stands, and rarely in south- western Jalisco at 400—1,200 m, this a distinc population with a moister ecology and later flow ering date (Guzman, 1978, 1982); flowering fro September through October, with ripe fruit b December. 4c-Il. Zea mays L. ssp. parviglumis Iltis and Doebley var. huehuetenangensis Iltis and Doebley, Amer. J. Bot. 67: 1002. 1980. Ra® Huehuetenango of Wilkes Leaves essentially glabrous, plants up to 5 " l tall, the tallest ofall teosintes. Plants of old HH ppm pn rp nn É ia. DOEBLEY —MAIZE AND TEOSINTE “my m á FIGURES 22-23. Lateral branch segment of a male inflorescence of Zea diploperennis, Iltis, Guzman, m. Doebley and Lasseigne 450 (plant GG).—22. Abaxial view.—23. Adaxial view. (Scale in m fields, edges of fields and oak forest, Province of uehuetenanago, Guatemala, from 900-1,650 m alt. (fide Wilkes, 1967, p. 56); flowering from late November to January and fruiting from Jan- uary through February, depending on the onset of the rainy season, which here, as in the habitat a the typical variety, is tremendously variable from year to year. Lm the time of its original description, it has Ps iini that Zea mays ssp. parviglumis var. tyes! : nangensis can be distinguished from the fae i. CEN by its essentially glabrous leaves though th i4 It has also been learned that, al- nae i abitat of Zea mays ssp. parviglumis e A river valley is highly seasonal and gion recei T geli. the winter months, this re- making ám 120-160 cm of rainfall annually bley, ee mally, at least, quite moist (Doe- Pániilen. ) m this sense then, Zea mays ssp. cupies 5 var. parviglumis (Balsas teosinte) oc- a habitat (seasonally mesic) similar to the mesic : th not abitat of var. huehuetenangensis. This is hok "prising, considering their similar mor- phologies, THE PHYLOGENY OF ZEA Littl E i * attention has been devoted to discussion * Beny ofthe teosintes, ostensibly because SPoradi i regarded them as anything more than © hybrids of maize and either Tripsacum ) or some hypothetical, primeval teosinte. The first of these ideas, the theory of Mangelsdorf and Reeves (1939), that teosinte is a maize-Tripsa- cum hybrid, has now been dismissed as unten- able even by its author (Mangelsdorf, 1974). The various other proposals that racial diversity in teosinte represents nothing more than the incor- poration of varying amounts of maize germ- plasm into some primitive teosinte still have their supporters. In this section I will examine the morphological and other evidence in relation to one of these theories, namely Wilkes (1979) and Mangelsdorfs (Wilkes & Mangelsdorf, 1979; Mangelsdorf et al., 1981) most recent hypothesis that all the races of annual teosinte are the prod- ucts of the hybridization of Zea diploperennis and maize. Further, I will articulate my own view that all the teosinte taxa represent the products of adaptive radiation and geographic speciation with only minor if any maize introgression. If one believes with Wilkes and Mangelsdorf that the annual teosintes were conceived through the miscegenation of maize and diploperennial teosinte, then they should exhibit intermediacy between these two taxa for at least some traits. Intermediacy may result from processes other than hybridization, and therefore cannot on its own establish the occurrence of hybridization. Nevertheless, intermediacy for at least some traits is a common outcome of hybridization. Thus, 5 ud Ll ae a mmm. mmt RR Rene AR am > imeem Rer C4 agente Urs P SERE bm Ed Cem. vien amengan nama Sie IE d Ld RT d asna e _ SRO TOREEN IN oo À—— (mta ceca Lord Ld — — Li E mer emen Ed pis E mE CRT de An mST US imn ——u€ Ed Ed End Lr nm X e — nase o ma nnÀ norms —— ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 $5. s 24-27. Lateral branch segment of a male gettin — 24-25. Zea pies Iltis, Puga, O, man, Doebley and Lasseigne 550 Agy t C).—24. Abaxial v —25. Adaxial view.— 26-27. Zea luxurians, K nd 419.—26. Abaxial view. — Adaxial view. (Scale in poe DOEBLEY — MAIZE AND TEOSINTE N NO co Hn | - £ o R —— RR as. < —— ——]ááÁ— m—— Se, —n — — ntl —ááá eee, n Sna es — — -m m. ee —— a —À —— ee —— — in um (ee — Fic T Tel (c RES 28-31. Lateral branch segment of a male inflorescence. —28- 29. Zea mays ssp. mays race Nal- race (Cultivated in Homestead, Fla.).—28. Abaxial view.—29. Adaxial view.— 30-31. Zea mays ssp. mexicana entral Plateau, I/tis and Doebley 96.—30. Abaxial view.— 31. Adaxial view. (Scale in mm.) ANNALS OF THE MISSOURI BOTANICAL GARDEN LLLI | e —ÜQ ee mee ma Car ——— o me ee —À c ah ee rd Lond MÀ e Man a ——- — C —— ce-—— a d 8) 1 | (dd 2-35. Lateral branch segment of a male inflorescence. Wee Zea mays ssp. par viglu - huehuetenangensis (race Huehueten nango), Doebley 417 (cultivated in Hom d, Fla.). — 32. Abaxial v t . Zea mays ssp. pisi var. parviglumis (race Balsas), Iltis and Doebley 362 p Me Abaxial view. 4M. Adaxial view. (Scale m.) one can check Wilkes and Mangelsdorfs hy pothesis by inspecting the mean values for some important taxonomic traits of the taxa presum- ably involved in this event (Table 6). In order to give this hypothesis the best possible chan of validation, I have listed on this table two r4" er extreme forms of so-called “‘primitive’ ' maie Mangelsdorf, himself, has used one of these mái 1983] races (Palomero Toluquefio) to test this hypoth- esis (Camara H. & Mangelsdorf, 1981). In Table 6, there is scarcely any evidence to bolster the theory of Wilkes and Mangelsdorf. For glume length, spikelet width, and vein num- diploperennis and maize. For pedicel length and length of internodes on both the central spike and the lateral branches, all of the Mexican an- nuals have values larger than either maize or diploperennial teosinte. For caryopsis weight and central spike length, Z. mays ssp. parviglumis has values smaller than either of the two hypo- thetical “parental” species. For pedicel length, vein number, and length of internodes on the central branch, Z. luxurians has values outside of the range between diploperennial teosinte and maize. Clearly then these data in no way verify the hypothesis that the annual teosintes sprang forth from the hybridization of maize and Z. diploperennis. On the other hand, if one cares to hypothesize that Zea luxurians and not Z. diploperennis ful- filled the role of the primeval pure teosinte in- volved in the hybridization with maize, the same difficulties arise. As can be seen from Table 6, the Mexican annual teosintes are no more inter- mediate between Z. luxurians and maize than they are between maize and Z. diploperennis. This conclusion is further supported by Figure 6 Which demonstrates that the Mexican annual teosintes are in no way morphologically inter- mediate between maize and the teosintes of sect. Luxuriantes, Other independent evidence also reflects dis- Paragingly on Wilkes’s hypothesis. First, analysis of isoenzymatic variation in Zea shows a large number of alleles restricted to sect. Luxuriantes and absent from all taxa of sect. Zea, as well as à close association between maize and Mexican eee teosinte, and thus, no evidence for the *rmediacy of the annual teosintes (Doebley et al., 1984), Similarly, the seed proteins ofthe taxa of sect. 7 is et al., 1979. Timothy et al., 1982). The tre- we diversity of cytoplasm DNAs is Pecially damaging to Wilkes and Mangelsdorf's DOEBLEY — MAIZE AND TEOSINTE 59 36 FicunEs 36-39. Outer glumes of Tripsacum and teosinte.— 36. Tripsacum australe from Peru (culti- vated at Fairchild Tropical Garden Redlands Nursery, Row 44 Space 2, FG67-257).—37. Zea diploperennis, Iltis, Guzman, Doebley and Lasseigne 450 (plant O).— 8. Zea perennis, Iltis, Puga, Guzman, Doebley and Lasseigne 550 (plant V).—39. Zea luxurians, Doebley 376 (cultivated in Homestead, Fla.). (Bar = 2 mm.) hypothesis, because these DNAs have strict ma- ternal inheritance, so their diversity can be ex- plained only by gradual evolution. Hybridizati could not have been involved. Much more sense can be made of the available information on variability in Zea if we simply treat the various taxa of this genus as the prod- ucts of allopatric, geographic speciation and adaptive radiation. First, morphological evi- J 60 ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL 70 | shed p e yl Homestead, Fla. . Zea mays ssp. mexicana race Central Plateau, Doebley 375 (cultivated in me en mexicana race Nobogame, Doebley 370 (cultivated in Homestead, Fla. var. parviglumis, Doebley 372 (cultivated in Ho gensis, Doebley 371 (cultivated in Homestead, Fla.). (Bar = 2 mm. yan glumes of Zea mays sensu lato.— 40. Zea mays ssp. mays race C (cultivaté -ea Mays ssp. mexicana race Chalco, Doebley 374 — in seg stea ad pa ma ad, Fl Zen See s ssp. varié 4 mestead, Fla.).—45. Zea mays ssp. E var. huehueten? — 1983] DOEBLEY — MAIZE AND TEOSINTE 61 TABLE 6. A comparison of maize and teosinte for some important taxonomic traits (field specimens only). terial 1 Methods All measurements are in mm unless otherwise noted. For a key to character RA Taxa Zea mays Zea mays are Pes Zea diplo- Ze ssp. ssp. Palomero Characters perennis luxurians parviglumis mexicana Nal-tel Toluqueño 9 GLUM-L 9.3 9.8 6.1 9.0 9.0 11.9 6 SPIK-W 24 2.5 1.8 hd 2.6 3:5 7 PEDI-L 22 2.7 4.3 4.2 1.9 1.8 13 SHD-V# 7.2 12.7 2.5 3.4 3.2 3.3 14 TOT-V# 16.3 Zhe 8.1 10.6 M 10.9 Caryopsis wt. (mg) 27.0 36.0 23.0 42.0 70.0 145.0 | BRAN-# 6.5 13.3 223] 19.7 42.2 2.8 4 CSIN-L 3.2 4.4 51 5.2 1.9 1.6 5 LBIN-L 3.1 4.5 5.8 6.0 5.2 4.8 2 CNSP-L 114.6 136.4 99.5 136.3 221.6 261.7 3 BRAX-L 23.8 45.3 73.5 86.5 166.5 15.2 dence Presented in this paper demonstrates that Zea divides into two clearly defined groups, sec- tions Luxuriantes and Zea (Fig. 46). Data from ‘isozymes (Doebley et al., 1984), cytoplasm ge- the genus. Of these sections, sect. Luxur which rather closely resembles the related genus Tripsacum, is undoubtedly the more primitive. The many morphological features held in com- mon by Tripsacum and sect. Luxuriantes, but lacking in sect. Zea, bespeak the phyletic affin- ities between these two taxa. These features in- clude typically many-nerved, flattened outer Miis ofthe male spikelets with two prominent nerves developed into wings (Figs. 36- Lene habit. Further, Tripsacum and sect. Mi RAE characteristically have many ter- ma eterochromatic regions on their chro- Galin es (Doebley & Iltis, 1980; Pasupuleti & at, 1982). Paesi, initial and probably quite ancient "El Ee of sections Luxuriantes and Zea, di- x n continued within each of these two probati In sect. Luxuriantes, Zea luxurians niala. p amed quite early from the peren- us tha Pass ithe p Sarin exist o - to the dry and highly seasonal €ntin southeastern Guatemala. The di- vergence between Z. diploperennis and Z. peren- nis apparently came somewhat later by means of autopolyploidy (Shaver, 1962; Galinat, 1971). ithin section Zea, Z. mays ssp. parviglumis var. huehuetenangensis is probably the most primitive taxon (Fig. 46). Evidence for this con- clusion comes in part from cytology. Kato (1976; cf. Longley, 1941b) has shown that this teosinte has many terminal heterochromatic regions (knobs) like the teosintes of sect. Luxuriantes. Further, isoenzymatic data show this teosinte to have substantially diverged from the other taxa of section Zea (Doebley et al., 1984). On the other hand, tassel morphology (this paper), cy- toplasm genomes studies (Timothy et al., 1979), and seed protein work (Smith & Lester, 1980; Mastenbroek et al., 1981), all show the clear re- lationship of this teosinte to the other teosintes of sect. Zea and to maize. Thus, it seems that Z. mays ssp. parviglumis var. huehuetenangensis is to some degree intermediate between the two sections though clearly belonging to section Zea (Fig. 46). Given the extent of the genetic dis- tinctiveness of this variety, it might best be el- evated to a subspecies. Zea mays ssp. parviglumis var. huehuetenan- gensis shows its closest relationship to Z. mays ssp. parviglumis var. parviglumis, although these two teosintes are distinct both in terms of their genetics and their vegetative morphology. They are in part distinguished by the many interna heterochromatic regions (knobs) found on the chromosomes of the typical variety but lacking T 4 | 62 ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL 70 - SECT. LUXURIANTES SECT. ZEA ZEA MAYS DIPLOPERENNIS PARVIGLUMIS | OAM PERENNIS LUXURIANS HUEHUE. PARVI. MAYS MEXICANA URE 46. A phylogeny for Zea that is supported by much of the available data. Dashed lines indicat’ | s. Whi e figure seems to indicate Z. mays ssp FIG possible alternative frui TASSELS FEM STIFF BRANCHES SPIKELETS FLAT ON BACK & WINGED GLUMES STIFF & BRITTLE FLEXIBLE & + PAPERY FRUITCASE TRAPEZOIDAL HABIT MUCH TILLERED | MONOPODIAL ANNUALS ) PERENNIALS KNOBS TERMINAL [ KNOBS TERMINAL & INTERNAL mays has acutely triangular i ays. et : : : fruitcase, the fruitcase is essentially absent from maize, being represented only by the cupules which are hidden . within the ear. on those of Huehuetenango teosinte (Fig. 46). They are similar in tassel morphology (this pa- per), seed proteins (Mastenbroek et al., 1981), and cytoplasm genomes (Timothy et al., 1979). These two varieties also share an adaption to a somewhat similar environment and this may partially account for their similar inflorescences. Both grow in low altitude (400—-1,600 m), warm, seasonally moist sites with a long growing season. As discussed by Doebley (1984), the small seeds, glabrous, green or dilute red sheaths, small male spikelets, and many tassel branches characteris- tic of these two varieties may be adaptions to warm, moist environments. The typical variety is most common in the Balsas river valley; how- ever, subpopulations of it inhabit similar envi- ronments in the lower, tropical deciduous forest of southwestern Jalisco, Mexico. Zea mays ssp. parviglumis var. huehuetenangensis is disjuncted from the Mexican population, occurring in the moist, warm montane oak forest of western Gua- temala —the wettest of all teosinte stations. ? i The next group to diverge in Fig. 46 is Z. ma ssp. mexicana. This subspecies shows à much i different adaptation than Z. mays ssp. parvigli- | mis. It has colonized the higher, colder, drier sites with a shorter growing season, and is characte" ized by large seeds, red hairy sheaths, large m?" _ spikelets, and fewer tassel branches as an adap- l tation to this environment (Doebley, 1984). This | sinte. Both have internal chromosome kn? | (Kato, 1976) and they are similar though distint | 1983] isoenzymatically (Doebley et al., 1984). Figure 46 indicates this affinity in that ssp. mexicana and ssp. parviglumis var. parviglumis are placed near one another. Although the evolutionary scheme just out- lined best accommodates the morphological and other information, other possibilities cannot be entirely dismissed. For example, Zea /uxurians, while being related to the perennials, may also be the ancestor of the other annual teosintes. As indicated by the dashed line on Figure 46, Zea luxurians may have emerged from the mainstock of Zea after the perennials and then the other annuals may have evolved from it. Although this scheme requires the annual habit to be derived only once, one would expect a closer relationship between Z. luxurians and Z. mays sensu lato were this system correct. In fact, its annual habit notwithstanding, Z. luxurians shows no closer relationship to Z. mays than does Z. diploperen- nis. Indeed, the many-nerved outer glumes, high- ly elongate trapezoidal fruitcases, and large ter- minal chromosome knobs of Z. /uxurians show 1t to have diverged further from sect. Zea than the perennials. WHENCE CAME MAIZE? teosi : : Osinte is the ancestor of maize (Beadle, 1939, Bess 1956; Miranda Colin, 1966; Iltis, wae 2, 1983; Galinat, 1971, 1975; DeWet e 1972; Harlan et al., 1973; Kato, 1976; Mic - 1976; Doebley & Iltis, 1980). Al- iss a mer years the tide of opinion has dE ^al of the latter theory, the oppo- "eis di € teosinte theory have not regarded a with equanimity and the subject re- controversial. In the following discussion €w the implications of the present study debate. As i .. Mentioned above, any explanation of the Origin of ^ maiz : ; - dersta : € must of necessity provide an un if the DOEBLEY — MAIZE AND TEOSINTE 63 has never been documented, and more impor- tantly, the teosintes are quite clearly not hybrids of maize and Tripsacum. Similarly, if we hy- pothesize as Wilkes (1979, p. 12) has that prior to man’s meddling there were only two forms within the genus Zea, (1) Z. diploperennis, the primitive Tripsacum-like perennial, and (2) Z. mays, the polystichous wild maize, and that these lystichy, a trait of obvious utility to man, arose in the wild among this group of grasses, the An- dropogoneae, in which distichy is the universal rule, while the distichous wild relatives of maize, the annual teosintes, arose under domestication when maize crossed with Zea diploperennis. In this sense Wilkes’s theory turns the most prob- able explanation inside-out by having the do- mesticated species emerge in the wild and the highly successful wild taxa emerge under do- mestication. A much more parsimonious interpretation of the facts is to view the distichous taxa, the teo- sintes, as the products of natural selection within the purely distichous Andropogoneae, and the lous polystichy of maize that is found only in the cultigen as the utilitarian artifact of do- mestication (Doebley & Iltis, 1980). The present study supports this view by revealing teosinte (Zea) to be a genus with a relatively complex internal structure, as foreshadowed by Wilkes (1967), and one in which each population is ge- d hologicall sculptured to meet netically P 5 J the demands of its particular environment. In addition to implicating teosinte as the direct ancestor of maize, the biosystematic data aid in pinpointing the exact locality in which the cul- tivation of maize began. If one considers the cultigen. The genetic evidence discussed in the previous section also enables us to rule out Z. mays ssp. parviglumis var. huehuetenangensis as a possible ancestor of maize. Of the remaining taxa, Z. mays ssp. mexicana and ssp. parviglu- mis var. parviglumis, the former shows the clos- est morphological relationship to maize (see Re- sults). Certain populations of race Central Plateau in particular display the smallest Mahalanobis's distances from race Nal-Tel of Z. mays ssp. mays. 64 ANNALS OF THE MISSOURI BOTANICAL GARDEN Although this suggests that maize could be do- mesticated Z. mays ssp. mexicana race Central ateau, the nature of the resemblance cautions us to look more closely. Zea mays ssp. mexicana displays greater affinity to maize than does Z. mays ssp. parviglumis primarily because of its larger spikelets and fewer tassel branches. As these two traits have themselves undergone consid- erable transformation during the domestication process, they can provide only flaccid evidence on which to pinpoint either Z. mays ssp. mex- icana or ssp. parviglumis as the ancestor of maize. Other available data augment the morpholog- ical evidence. First, the isoenzymatic work of Senadhira (1976, Fig. 12) showed that some pop- ulations of Balsas teosinte (Zea mays ssp. par- viglumis) cluster closest to the maize races Cha- l (Doe (1976) cytological studies revealed that both B chromosomes and abnormal chromosome 10 (type I), which are ai common in Mexican maize races, co-occur only in teosinte from the Balsas region and are unknown in the Chalco area. This again points to the Balsas river drain- age as the cradle of — €— Thus, most g t with the theory that maize arose from Bulsas teosinte, while morphology would suggest that Z. mays ssp. mexicana is closest to maize. Prudence would advise that further information be gathered be- fore rendering a final verdict. While it has long been recognized that the bio- systematic evidence points to teosinte as the ancestor of maize, the archaelogical evidence available apparently does not. The earliest ar- chaeological maize specimens recovered from Tehuacan possess long soft glumes, a narrow flexible rachis, and shallow non-indurate cupules (Mangelsdorf, 1974), traits which would be dif- ficult to derive from the female teosinte spike. It is for this reason, among others, that Mangels- dorf and his supporters have remained adamant in their belief that maize evolved not from teo- sinte, but from a “wild maize." Recently, Iltis (1981, 1983) has proposed a new theory, his Cat- astrophic Sexual Transmutation Theory, which attempts to explain both the biosystematic and archaeological evidence. Quite simply, he sug- gests that the teosinte female spike is not the ancestor of the maize ear, as has long been thought, but rather that the central spike of the LOCORIL [VoL. 70 | teosinte tassel, which normally terminates a lat- eral branch, gave rise to the familiar maize ear. This new theory draws in part from older the- ories on the origin of the maize ear (Kellerman, 1895; Montgomery, 1906; Iltis, 1911; Weath- erwax, 1918). l This theory has an initial appeal to the mor- f phologist because the maize ear is terminal on | lateral branch, a position which in teosinte is usually occupied by a tassel. Thus, the early do- mestication of maize would have involved a se change of the central spike of the terminal tassel on the lateral branches from male to female. Iltis | believes this change and the subsequent (or con current) condensation of the central spike inl the maize ear happened rapidly; thus we have , the Catastrophic Sexual Transmutation Theory. The aspect of this theory that most stirred my imagination was its implications for the archae ological specimens from Tehuacan. As described above, they do not fall neatly into a sequent - between the teosinte female spike and the mau ear. However, under Iltis’s theory, they are ex ^ actly the intermediates one would predict. The | central spike of the teosinte tassel has a flexible rachis, shallow non-indurate cupules, and spike l lets with long soft glumes [much like Mangels- l dorf's (1974) ancestral pod-corn!], all characte - of the Tehuacan specimens. Further, the centfil } tassel spike of teosinte has two functional spike lets per node (cupule) as does the maize ear ar bu unlike the teosinte ear, so that by applying Trist | theory the reactivation of a suppressed spik is not necessary (cf. Beadle, 1972). While the Catastrophic Sexual -— — 3 Theory has some attractive features, it needs &* perimental verification. The developments morphology of the male and female inflore® cences of Zea should be studied with this theo | in mind. Such work might not only resolve the | origin of the maize ear, but may also provide? | yardstick for the assessment of primitive versis advanced traits of the ear. Such a yardstick b been sadly lacking in all studies of racial V tion in maize to date. LITERATURE CITED ALAVA, R. O. 1952. Spikelet variation in Zea ma L. Ann. Missouri Bot. Gard. 39: 65-96. the ANDERSON, EDGAR. 1936. Hybridization in American Tradescantias. Ann. Missouri Bot. Gart 23: 511-525. ip . 1944a. Homologies of the ear and tassel Zea mays. Ann. Missouri Bot. Gard. 31: 325- ———-—. 1944b. Two collections of prehistoric cot 1983] tassels from southern Utah. Ann. Missouri Bot. Gard. a 345-354. The Sacred Plume: A Description of the Sete Tassel with Some Indications of Its Im- portance. Pioneer Hi-Bred Corn Company, Des Moines, Iowa. 1956. Why botanists visit math departments. Missouri Bot. Gard. Bull. 44: 148-151. — ——. 1969. What I found out about the corn plant. Missouri Bot. Gard. Bull. 57: 6-9. — A W, L. Brown: 1948. A morphological analysis of row number in maize. Ann. Missouri Bot. Gard. 35: 323-336. — — & H.C. Cutter. 1942. Races of Zea mays I. Their recognition and classification. Ann. Mis- souri Bot. Gard. 29: 69-88. BrApLE, G. W. 1939. Teosinte and the origin of maize. E sco sa 245-247. 972. The mystery of maize. Field Mus. Nat. __ His. Bull. yen 2-11 1980. Hg ancestry of corn. Sci. Amer. 242: 2 . LEsrER. 1980. Morphological and allozymic evidence for Sabatia formosa (Gentia- M in section Campestria. Amer. J. Bot. 67: BIRD, R. McK. 19 A name change for Central merican teosinte. Taxon 27: 361-363. 1980. Maize evolution from 500 B.C. to the present. “ob 12: 30-41. Bowman, R. 1 m sect. Zausch- EL Bot. 67: 671-685. ANGELSDORF. 1981. Perennial j^ and annual teosinte phenotypes in crosses of Zea diploperennis and maize. Bussey Institution, Harvard Univ. Publ. No. 10, pp. 3- CELARIER, EP! Cytotaxonomy of the Andro- Pogoneae II. ae Ischaeminae, Rottboellinae cee Maydeae. Soe 22: 160-183. few W. D. 1973. The awnless species of An- DA TOpogoneae. der Bull. 28: 49- RLINGTON, C. D. 1956. Chromosome Botany. Al- mai eJ FEA 1972. Origin of 279. the tripartite hypothesis. Euphytica 21: 271- Doeme, J. F J. F 4. Maize introgression into teo- a reappraisal. Ann. Missouri Bot. Gard. (in — Su & H. H. I 1980. Taxonomy of Zea. I. J. “orig Classification with key to taxa. Amer. ER 982-993. ma M. M. Goop C. W. STUBER. Isoenzy- tic variation in Zea (Gramineae). Syst. Bot. (in FISHER, R. A. 19 The mati in 179-188. Gaunar, w. c. 1971, E - Genet. 5 : ES Thee evolutionary emergence of maize. E. m ot. Club 102: 313-324. AN, M 1967. The races of maize: the use (nobis generalized distances to measure 36. The use of multiple measure- taxonomic problems. Ann. Eugenics 7: The origin of maize. Annual 47-478. DOEBLEY —MAIZE AND TEOSINTE 65 Fitotecn. Latinoamer. 4: J jua ———. 1972. Distance analysis in biology. Syst. Zool. 21: 174-186. . PATERNIANI. 1969. The races of maize. 978. Una nueva localidad para el teosinte Zea perennis y primer reporte de Zea mexicana para Jalisco. Bol. Inform. Inst. Bot., Univ. Guadalajara 6: 9~10. 982. El teosinte en Jalisco: su distribucion türlichen Pflanzenfamilien by Lamson-Scribner and Southworth]. Henry Holt and Co., New York. HARLAN, J. R., J. M. J. De WET . PRI RICE. 1973. Comparative evolution of cereals. Evolution 27: 311-325. HITCHCOCK, A. S. 1922. A perennial species of teo- sinte. J. T Acad. Sci. 12: 205-208. ILTIS, HUGO. . Uber einige bei Zea mays L. beo- e Verursa über die ee der Indukt. Abstam AME 5: 38-57. Iris, H. H A The e ma i mystique —a reapprais- al of the origin of corn abstract of a lecture Ms at the Corn Conference, Un iv. of Illino ois iM in 1969, and Iowa State 970. Botany Dept., Univ. of Wis. “Madiso on, "Wi is. —. 1972. The taxonomy of Zea (Gramineae). rtu 23: 248-249. Morphologic-systematic studies of the ius Si scans of teosinte. National Science Foundation Grant Proposal. Unpublished, Univ. of Wis. Herbarium, Madison. Offset. 1981. Thecatastro ophic sexual transmutation theory (CSTT): the epigenesis of the teosinte tassel spike to the ear of corn Oc. Amer. Misc ries Publ. 160: 70. . 1983. From teosinte to maize: the strophic sexual transmutation. Science (in cias jJ 80. Tax axonomy of Zea (Gramineae). I. Subspecific categories in the Zea mays complex and a generic synopsis. Amer. J. Bot. 67: 994-1004 — — , ——,, R. Guzman M. & B. P 1979. Zea diploperennis Piin a new i acu from exico. Science 203: 186-188. KALLUNKI, J. A. 1976. Aegea studies in Good- yera (Orchidaceae) with emphasis on the hybrid origin of G. ao y 28: 53-75. Cytological studies of maize. 35. KELLER, SUE. 1979. A revision of the genus Wisli- zenia (Capparidaceae) based on S cmc stud- ies. Brittonia 31: 333-351 KELLERMAN, W. A. 1895. Primitive corn. Meehans’ Monthly 5: 44, 53. Kowar, R. R. UFFIN. 1979. Generic circum- scription in the Xanthocephalum complex risen eae, 66 ANNALS OF THE MISSOURI BOTANICAL GARDEN of paper presented at Bot. Soc. of Amer. meeting, Stillwater, Ok. Bot. Soc. Amer. Misc. Series Publ 157; , M. L. LeEcHowicz & M. S. ApAMs. 1976. The use of canonical analysis to compare response curves in physiological ecology. Flora 165: 29—46. LANGHAM, D. G. 1940. The inheritance of interge- neric differences in Zea-Euchlaena hybrids. Ge- 1941a. Knob gig mc on teosinte chromosomes. J. Agric. Res. 62: 401-413. Chromosome morphology in maize and its wild relatives. Bot. Rev. (Lancaster) 7: 262- 9, MANGELSDORF, P. C. 1974. Corn: Its Origin, Evo- lution and Improvement. Belknap Press of Har- vard Univ. Press, Cambridge, Mass. R. G. Reeves. 1939. The origin of Indian corn spe its relatives. Texas Agric. Exp. Sta. Bull. 574: 1-315. m M. ROBERTS & J. S. ROGERS. 1981. The probable origin of annual teosintes. Bussey Insti- tution, Harvard Univ. Publ. No. 10, pp. 39-69. MASTENBROEK, INEKE, CE CoHEN & J. M. J. DE WET. 1981. Seed pr in and of s and its closest relatives. Biochem. Sys delitos Eco 9: 179-183 MIRANDA COLIN, S. 1966. Discusion sobre el origen y la evolucion del maiz. Memorias del Temco Mont errey, I p. 233 -251. d Nee ie Colegio ja Postgraduados, Chapingo, MoNTGO E. G. 19i : What | is an ear of corn? Po opa sig Monthly 68: 55-62. OXNARD, C. E. 69. Mathematics, shape and func- tion: a study in primate anatomy. Amer. Sci. 57: 5-9 PASUPULETI, C. V. & W. C. GALINAT. 1982. Zea dip- loperennis. I. Its chromosomes and comparative cytology. J. Heredity 73: 168-170. Price, R. A. 80. Draba streptobrachia (Brassica- ceae), a new species from Colorado. Brittonia 32: LPH, L. R. 1976. Contributions of wild rela- tives of maize e to the padi eos history of do- ti a synthesis of divergent hy- a 30: 321 -34 REYNOLDS, J. F. & D. J. CRAWFORD. 1980. A quan- titative study of variation in Chenopodium atro- virens-dessiccatum-pratericola complex. Amer. J. Bot. 67: 1380-1390. i [Vor. 70 I RIGGINS, R., R. A. PIMENTAL & D. R. WALTERS. 1977. Morphometrics of Lupinus nanus (Leguminosae). I. Variation in natural populations. Syst. Bot. 2: 317-326. RoBiNSON, J. T. & K. STEUDEL. 1973. Multivariate discriminant analysis of dental data bearing on early hominid affinities. J. Human Evol. 2: 509- 527. Sear, H. L. 1964. Multivariate Statistical Analysis i London Genetic variation i n corn and its relatives. Unpublished Ph.D. thesis, University of California, Davis. SHAVER, D. L. 1962. A study of meiosis in perennial teosinte their tetraploid hybrid. rbi 15: 43-5 SMITH, J. S. C. & R. N. LESTER. eem Biochem systematics and evolution of Zea, Tripsacum related genera. Econ. Bot. 34: 201-218. STANDLEY, P. C. 1950. Teosinte in Honduras. Cite } 1: 58-61. STEUDEL, K. A multivariate analysis of the | pelvis of early hominids. J. Human Evol. 7: 587 | | 595. a D.H; COC S EEVINGS; D. R. PRI d DE & J. 4 KERMICLE. 1979. Organell T DW Ps teosinte. Proc. Natl. Acad. U.S.A. 76: on j 4224. ———, — —, A. K. WEISSINGER, R. R. SEDEROFF, M u. 1982. Compa isons of mitochondrial DNAs of Zea mays L. and its relatives. Bot. Soc. Amer. Misc. Series Publ 8. WEATHERWAX, P. 1918. The evolution of maize. Bull. ) Torrey Bot. Club 45: 309-342. WELLHAUSEN, E. J., L. M. ROBERTS & E. HERNA ANDEL | (in collaboration with P. C. Mangelsdo | 1952. Races of Maize in Mexico. Bussey Inst i S. WiLKEs, H. G. 1967. Teosinte: The Closest Relativ? | of Maize. Busey Institution, Harvard Uni* | Cambridge, Mas: . 1979. nie and Central America as 3 for the origin of agriculture and the aio maize. Crop Improv. 6: 1-18. | & P. C. MANGELSDORF. 1979. Zea dip ennis: the “missing link” in corn’s geneal stract of paper presented at the 12th annual ing of the Soc. for Econ. Bot., Raleigh, N.C. | ^ am — 1983] DOEBLEY — MAIZE AND TEOSINTE 67 APPENDIX A. Means (and standard deviations) for all characters and all populations. For key to population and character numbers see Materials and Methods. All measurements are in millimeters except characters 15, 16 and 17 which are in microns. Population Character 1 2 3 4 5 6 T 8 l 6.5 3.8 4.5 11.8 13.6 14.5 10.6 15.9 (2.7) (1.2) (1.1) (6.1) (7.2) (5.2) (6.7) (8.1) 2 114.6 98.8 84.7 113.3 138.5 137.5 120.0 112.3 (15.5) (8.1) (9.4) (25.0) (29.5) (19.2) (42.0) (29.0) 3 23.8 122 13.6 45.6 43.6 46.8 47.5 71.8 (11.7) (5.6) (5.7) (25.1) (13.4) (16.3) (25.0) (26.6) 4 7 4.0 4.5 4 (.55) (.43) (51) (.67) (.80) (.49) (1.22) (1.83) 5 2.9 4.3 4.5 (.44) (.32) (.61) (.83) (.77) (.60) (.92) (1.22) 6 2 335 2.4 2.5 (.18) (.17) (.23) (.17) (.09) (.27) (.31) (.27) 7 22 2.4 24 2.9 2.6 l (.57) (.56) (.60) (.65) (.45) (51) (1.27) (.63) 8 52 5.2 5.2 (.28) (.21) (.28) (.42) (.60) (.29) (.65) (.49) 9 9.3 10.2 9.5 (.90) (.48) (.68) (1.00) (1.14) (.69) (.83) (.75) 10 3.6 37 3.4 4 (31) (.16) (.29) (.23) (.22) 622) (.30) (.27) 11 1 390 1 .200 196 0 (.038) (.056) (.046) (.039) (.052) (.028) (.028) (.034) 12 1.6 4 2.0 2.0 (.34) (.26) (.31) (.22) (.28) (.30) (.22) (.32) H 5.8 5.6 7 11.9 13.5 23 (1.01) (1.15) (1.42) (1.56) (2.30) (1.43) (.84) (.92) n 1 15.0 13.5 19.9 21.9 21.8 9.5 (1.33) (1.07) (1.42) (1.59) (1.89) (1.94) (1.58) (1.13) " l 28 109. 114. 101 1 109. 108 (14.6) (16.6) (12.2) (15.3) (9.2) (16.2) (13.3) (15.2) 16 55 61. 71. 85 (11.6) (8.4) (12.9) (8.2) (6.9) (12.7) (18.0) (27.7) 17 57. 50. 63. 69. 63. 100. 96. (15.9) (12.3) (11.1) (7.9) (4.9) (12.3) (17.6) (14.0) Population Character 9 10 T 12 13 14 15 16 l 28.4 24.0 35.0 10.2 3.3 22.0 10.7 (12.7) (6.7) (21.5) (5.5) (1.5) (3.7) — (7.7) 2 141.1 172.5 118.4 100.1 105.7 109.5 97.0 119.1 (27.8) (28.2) (29.2) (23.1) (20.0) (26.5) - (48.6) 3 106.4 120.0 105.5 45.5 3 49.0 48.9 (42.8) (21.6) (36.4) (26.7) (8.0) (23.0) - (37.0) 1 4.8 53 4.8 4.9 4.4 (.99) (1.26) (1.18) (1.22) (.44) (.76) -— (.63) 5.4 8 6.0 4.4 4.6 4.8 (.87) (.97) (.74) (81) (.12) (1.05) _ (.71) 68 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo.. 70 APPENDIX A. (Continued). Population Character 9 10 1 12 13 14 15 16 6 24 2.5 1.8 1.8 2.5 2.6 2.4 23 (.34) (.21) (.21) (.12) (.07) (.25) - (22) 7 4.5 4.2 44 4 0 9 2.8 2.9 (.99) (1.08) (.87) (.77) (.88) (.74) be (.86) 8 3 5.5 3.7 3.9 5.1 4.7 4.8 4.1 (72) (.60) (.44) (.29) (.55) (.45) = (.98) 9 9.1 9.6 6.0 5.8 5 9.5 9.8 8.1 (.75) (1.29) (.55) (.53) (.29) (.62) = (1.00) 10 3.7 3.3 9 2.6 4 6 3.4 3.4 (.34) (.33) (.34) (.34) (.32) (.27) = (.27) 1 .024 .040 .020 .026 413 .156 12 049 (.028) (.033) (.020) (.027) (.101) (.033) $e (017) 12 1.7 8 1.5 3 2.1 2.0 L7 1.7 (.21) (.26) (21) (.12) (.25) (.28) s (27) 13 2.9 4.8 2.1 24 6.7 11.3 14.0 27 (1.02) (1.32) (1.00) (1.01) (2.08) (2.83) = (1.27) 14 9.0 3.9 8.3 73 14.7 19.1 22.0 10.6 (1.64) (2.08) (1.89) (1.77) (1.53) (2.88) a (1.13) 15 112. 83. 87. 90. 121. 27: 110. 88. (14.9) (9.2) (16.6) (8.7) (9.3) (15.0) = (12.4) 16 91. 73. 81.1 80. 68. 61. 40. 72. (20.6) (9.6) (18.2) (10.6) (10.7) (7.6) wt (10.2) 17 112. 79. 78. 91. 59. 66. 40. 84. (17.1) (4.9) (13.9) (11.1) (19.3) (14.0) — (15.4) Population Character 17 18 19 20 21 22 23. 2 1 5.7 16.6 16.6 12.0 16.0 15.5 13.5 8.3 (1.5) (7.5) (5.3) (4.3) (3.6) (6.4) (9.0) (2.5) 2 92.0 79.8 105.6 88.0 106.4 94.3 139.3 173.9 (22.5) (22.0) (20.2) (15.7) (13.6) (13.8) (75.8) (41.5) 3 36.0 69.4 62.6 56.0 59.5 66.5 61.7 57.5 (7.0) (31.3) (14.3) (15.2) (10.8) (23.4) (27.4) (22.6) 4 47 4.9 5.6 5.3 4.9 5.6 5.7 7.9 (.93) (.82) (.82) (.45) (.82) (.39) (1.48) (1.51) 5 5.3 4.8 6.2 7 5 6.4 6.7 7.8 (.42) (1.00) (.85) (.99) (.30) (71) (1.08) (1.56) 6 24 1.9 2.9 T 9 8 0 3.1 (.22) (.29) (.30) (.16) (18) (.24) (.32) (39) 7 32 4.1 3.9 3 2.9 4.4 4.1 4.9 (1.09) (.96) (.76) (.82) (.85) (78) 4 (234 (84 8 4 3.9 5.1 5.1 5.2 5.3 9 5.8 (.59) (.57) (.45) (.49) (17) (.29) (.89) (7D 9 7.5 6.5 10.0 10.9 10.2 10.7 10.6 11.1 (.57) (.83) (.50) (.54) (58 © (10) a4») Uf 10 3.0 2.8 4.3 4.1 4.1 34 3.9 41 @ (.26) (.33) (.38) (.51) (.29) (.28) (.72) (.48) T 013 .020 475 T 160 160 024 064 (.023) (.020) (.020) (.056) (.043) (.042) (.027) (042. —'—— oM 7 = T 1983] DOEBLEY —MAIZE AND TEOSINTE 69 APPENDIX A. (Continued). Population Character 17 18 19 20 Zi 22 23 24 12 1.5 1.4 2.7 2.4 T8 2 2 (.23) (.25) (.32) (.53) (.34) (.20) (.52) (.40) 13 2.3 15.6 1 15.8 14.2 (1.15) (.98) (2.07) (3.56) (2.31) (1.40) (1.25) (.80) 14 1 24 26. 11.0 1 (1.15) (1.42) (2.53) (3.56) (1.69) (2.07) (3.09) (1.45) 15 l 0 114. 10 113. l 0 (9.3) (19.0) (11.9) (9.9) (9.2) (13.3) (9.0) (15.5) 16 74 72 79. 80. 74 (19.4) (12.5) (9.6) (11.0) (8.5) (8.8) (13.5) (15.7) 17 ? 78. 8.3 7.1 9.8 (10.7) (22.9) (13.8) (8.5) (11.6) (7.7) (15.7) (14.8) Population Character 25 26 27 28 29 30 31 32 l 9.8 31.3 28.9 16.3 23.8 37.4 5.3 11.4 (4.0) (12.2) (9.9) (7.5) (7.2) (12.1) (3.9) (5.5) 2 83.8 81.2 83.6 118.9 80.2 85.9 97.4 116.6 (19.0) (18.5) (19.1) (44.6) (18.8) (40.6) (17.4) (37.5) 3 39.3 3 0 64.4 94.6 132.3 4 3 (11.7) (22.8) (22.7) (25.2) (15.4) (16.4) (24.6) (15.1) 4 4 6 5.7 4.9 (.80) (.49) (.77) (1.29) (.43) (.87) (.68) (.41) 5 6.2 6.8 7.0 0 5.7 4 (.63) (.92) (1.19) (1.36) (.67) (1.05) (.43) (.55) 5 24 23 2.0 2.2 21 2 (.20) (.25) (.22) (13) (.22) (.33) (.34) (.12) 7 6 5. 4.0 2 (.61) (.82) (.92) (1.17) (1.14) (1.11) (.60) (.31) 8 4.1 0 4.3 4. 4.0 .0 (.65) (.37) (.43) (.49) (.42) (.66) (.60) (.40) 9 7.4 9 2 5.9 6.8 8.4 (.59) (.60) (.34) (.92) (.74) (.94) (.48) (.68) S 2] 2.7 29 3.0 3.2 2.6 2.8 (.42) (.33) (.28) (.53) (.29) (.22) (.30) (.26) H 025 016 02 020 030 340 1 (.030) (.020) (.024) (.021) (.021) (.028) (.021) (.015) > 4 1.3 1.5 5 1.4 (.30) (.24) (.18) (.32) (.15) (.19) (27) (.20) 13 2f 3.4 3.0 6 1 a "Gy QA Op 020 (93) (O6) (1.98) b 10.5 11.3 Hi I2 9.0 15.4 18.9 (.83) (1.88) (2.34) (1.64) (1.70) (2.20) (1.23) (2.67) be 86. 85. J. 33. ^ (6.9) (12.4) (9.0) (11.1) (10.1) (12.9) a A 8) 73. 79. 74. 84. 80. 77. : i (10.5) Ts (7.1) (13.1) (9.5) (12.9) (18.0) ( "p 8.4 86. 82. 93. 85. 89. (15.3) (10.2) (11.3) (7.0) (11.7) : s : 70 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 APPENDIX A. (Continued). Population Character 33 34 35 36 37 38 39 40 l 10.3 11.0 35.2 15.5 2.8 42.2 7.4 4.9 (4.3) (9.8) (11.1) (3.5) (.86) (11.9) (4.0) (3.0) 2 121.7 97.4 67.7 85.5 261.7 221.6 311.1 2584 (26.8) (8.3) (17.7) (10.6) (23.1) (37.9) (50.8) (29.4) 3 48.3 37.6 98.8 91.5 15.2 166.5 52.6 264 (24.9) (23.9) (13.9) (23.3) (13.6) (19.5) (26.0) (19.7) 4 5.6 4.2 4.7 5.0 1.6 1.9 2.6 2.1 (.81) (.71) (.47) (.61) (.35) (.51) (.82) (25) 5 5.1 4.6 6.3 53 4.8 5.2 5.0 22 (.56) (.86) (.61) (.28) (.82) (.68) (.49) (77) 6 6 25 2.1 1.9 3.5 2.6 3.2 0 (.42) (.33) (.15) (0.0) (.27) (.23) (.36) (17) 7 3.5 2.1 3.3 6.0 is 1.9 2.8 2.9 (1.09) (1.21) (.31) (.18) (.87) (1.04) (1.3) (1.7) 8 5.3 4.9 4.1 3.4 6.8 5.5 7.1 7.6 (.62) (.77) (.51) (.18) (.48) (.51) (.49) (.48) 9 10.9 8.4 75 6.8 11.9 9.0 12.9 12.8 (.97) (1.16) (.39) (.35) (.72) (.71) (1.30) (1.19) 10 3.0 3.3 2.8 27 3.9 38 4.5 4.0 (.53) (.37) (.19) (.09) (.40) (.31) (.59) (61) 11 173 .008 .033 .020 .028 1.07 .023 .023 (.033) (.017) (.030) (.028) (.024) (.020) (.020) (.021) 12 1.8 1.6 1.5 1.4 2.0 1.6 1.8 2.1 (.27) (.32) (.15) (.20) (.32) (.31) (.32) (35) 13 13.0 28 3.8 .5 33 15 4.1 3.7 (2.83) (1.30) (1.60) (.70) (1.13) (1.11) (1.55) (1.60) 14 22.5 9.0 11.3 8.5 10.9 11.1 13.5 10.6 (3.56) (2.12) (1.75) (2.12) (1.88) (2.06) (2.25) (1.62) 15 102. 106. 5 74. 92. 97. 103. l. (11.7) (24.3) (7.8) (13.1) (8.6) (11.3) (16.0) (12.8) 16 42. 93. 84. 70. 84. gi. 91. 3. (15.2) (16.1) (19.5) (6.6) (10.2) (12.0) (9.6) (11.3) 17 50. 106. 85. 65. 90. 95. 94. 82. (12.7) (22.4) (18.1) (0.0) (12.7) (13.6) (14.0) (13.6) pi em rim ici AN EMBRYOLOGICAL ANALYSIS OF MYRTALES: ITS DEFINITION AND CHARACTERISTICS! HIROSHI ToBE? AND PETER H. RAVEN? ABSTRACT A combination of embryological characteristics clearly defines Myrtales as comprising Combre- taceae, Lythraceae (including Puni dS i ), Melast yrtaceae, On ceae, Oliniacea d Trapaceae, a circumscription that agrees with that of the **core Myrtales given by Dahlgren and Thorne (1983). The ordinal characteristics are: 1) anther tapetum glandular, 2) ovule crassinucellate, 3) inner integument 2-layered (except in Syzygium), 4) micropyle formed by both integuments (except in Syzygium and Trapa), 5) antipodal cells ephemeral or absent, 6) endosperm formation Nuclear type and 7) seed exalbuminous. Haloragaceae, Lecythidaceae, and Thymelaeaceae definitely should be excluded from Myrtales on the basis of differences in three or more of these primary defining characteristics. On the other hand, embryological evidence does not contradict the possibility of a relatively close relationship between Elatinaceae and Myrtales, even though an overall consideration of their features seems to make such a relationship seem less likely. Embryological evidence indicates a considerable degree of heterogeneity in Rhizophoraceae, a family a » or group of families that is clearly not assignable to Myrtales. INTRODUCTION Although embryology has been an important ea) » Maheshwari (1950), and Davis inier a: angiosperms, and their books mue of use as guides to the literature. A (1964) 4 aher authors, including Maheshwari Philipso Tee (1966), Brewbaker (1967), and of indiv E (1974), have evaluated the significance ado ual characters or assemblages of char- tł lati 343 CAiffronut 0 Es : Soups of angiosperms. In 1967, a symposium emb; : Tyological characteristics have a great deal to offer aS sources of evi i ; idence for the systematics Of higher-level ore, pip y In thi & Ps among g plants ing we have analyzed the available ation concerning the embryology of Myr- Ese a tales and a few other families that have been thought to be closely related to this order. Our primary purpose has been to characterize Myr- tales embryologically and to chart the main out- lines of relationship within the order from an embryological perspective. As defined by Dahl- gren and Thorne (1983), the families of core Mvrtales are: Combretaceae, Crypteroniaceae, Lythraceae (including Punicaceae and Sonnera- tiaceae), Melastomataceae, Myrtaceae, Onagra- addition, Chrysobalanaceae, Coridaceae, Elaeag- naceae, ‘Haloragaceae, Lecythida- ceae, Rhizophoraceae, Thymelaeaceae, which are excluded from the order by Dahlgren and Thorne (1983) but are “in various respects conspicuously similar to Myrtales," are analyzed from an em- bryological point of view in the light of the avail- able information. We offer the present critical review to bring together all available literature and to serve as a guide to the most appropriate directions for future studies of the embryology of Myrtales. METHODS In the course of this review we have analyzed nearly all of the references cited by Davis (1966), as well as the subsequent publications that have ndation. We are grateful to and to Barbara Palser and t. a University, 1-33 Yayoi-cho, Chiba 260, Japan. LSA. 3 : : Missouri Botanical Garden, Post Office Box 299, St. Louis, Missouri 63166, U.S ANN. Missouri Bor. GARD. 70: 71-94. 1983. 72 been available to us. Not all of these publications included information that was valuable for eval- uating the embryology of Myrtales, and many had incomplete or inadequate information. For example, since the work of Geerts (1908) it has been understood that Onagraceae have the dis- tinctive Oenothera-type embryo sac develop- ment. Earlier references such as those of Hof- meister (1847, 1858), Vesque (1879a, 1879b), Ward (1880), and Guignard (1882) were not able to explain this type of embryo sac development fully owing to the less precise techniques used in the nineteenth century (Maheshwari, 1948). Similarly, most other nineteenth-century embry- ological studies have relatively little to offer for current evaluations of et relationships of the families of angiosperm In general, only unde genera from in- dividual families have been examined. The pow- er of our comparisons between families is de- rived from the depth and scope of earlier investigations. We have attempted to take this factor into account in utilizing and evaluating the published information about the embryology of Myrtales and allied groups as follows: Total number of embryological studies anthers, ovules, and seeds of genera Level of : : of a given family k ledge — x neg be 3 X (number of ims genera of the family) We have divided the embryological data re- viewed into three parts: namely, that concerning the anthers, ovules, and seed development. Most of the references we consulted refer to only one or two of these three major components of em- bryology. Comprehensive studies of all three classes of data provide the only sound basis for evaluating the relationships of genera and fam- ilies. Even if some information is available, it may be strictly limited, and may therefore be of relatively little use. It is very rare, for example, for an earlier study to describe accurately the thickness of the integuments. A second problem in utilizing published data concerning embryology concerns nomenclature Names have sometimes changed so often in the past that it is difficult to be certain how many taxa are involved among those studied earlier. Except for Onagraceae and Penaeaceae, we have in general accepted the nomenclature used in the ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot 70 — 3 f f l articles involved as a basic guide to the number of taxa studied earlier. For the number of genera in a given family, we have used Cronquist (1981) as a reference. Despite these difficulties, we hope that our evaluation of the amount of knowledge available concerning the embryology of particular families — will be useful in arriving at a sound understand- __ ing of the thoroughness of the earlier studies of the group and thus help the usefulness of the conclusions drawn for an evaluation of relation- ships. EMBRYOLOGICAL CHARACTERS USED FOR ANALYSIS . Í The characters that are treated as most mM- portant in making these evaluations are basically those that Maheshwari (1964), Davis (1966), and Palser (1975) have considered *embryologici characters of taxonomic significance." We have elected to use the whole set of characteristics $0 as to make what we believe to be the most ef fective comparison between families. Cons quently, we have added several characters, suc! as thickness of integuments and presence or ab sence of fatty globules in megaspores of embry? sacs, to those treated as fundamental by the at- thors just mentioned. Specifically, we have deal! with the following 35 characters: Anthers: — . more. N yledonous, Md Redu Epidermis: persistent or . Endothecium develops ind thicken- ings or not. . Middle layers: persistent or not. . Tapetum: glandular or amoeboid. . Number of nuclei in a tapetal cell: om two, or more. . Cytokinesis in a pr sea oe mother cell: simultaneous or succes Shape of microspore dic tetrahedril decussate, isobilateral, or otherwise. . Num ead of cells in a mature pollen: ont or tw (o 0 9 s e Ovules: 11. Degree of ovule curvature: anatropoU* campylotropous, or otherwise. 12. Tenuinucellate or crassinucellate. 13. Number of integuments: one or two. . Type of wall development: Basic, Ee l Number of sporangia per anther: four! ! 1983] > — cn — ON it; 18. 19, N © aii N N N Ww N AB N Cn N on N N N oo TOBE & RAVEN — EMBRYOLOGY OF MYRTALES . Thickness of integuments: two- or multi- layered. . Presence or absence of vascular tissue in integuments. . Micropyle: formed by inner, outer, or both nts. integument Nucellar beak formed or not. Chalaza with hypostase or not. Endothelium formed or not. . Archesporium one- or multi-celled. Cytokinesis in a megaspore mother cell: occurs or not. . Shape of a tetrads: linear, T-shaped, or otherwi . Position of "reni i ORDAINEN micro- l pylar or chalaza . Type of megagametophyte development: Polygonum, Oenothera, Penaea, or other. . Fatty globules in megaspores and embryo Sacs present or absent : Characteristics of synergids: hooked, pyr- iform, or otherwise - Characteristics of antipodal cells: persis- tent or ephemeral. Definitions of the term "Persistent" and “ephemeral” were some- times vague, so in this work the antipodal cells that degenerate and disappear before fertilization are referred to as *ephemer- al,” whereas those that persist up to fer- tilization and postfertilization are referred to as “persistent.” - Number of constituent nuclei or cells in a mature embryo sac: eight (as is usual in the Polygonum-type embryo sac), five (due to early disintegration of three antipodal Cells in the Polygonum-type embryo sac), four (as in the Oenothera- -type sac), 16 (as 2 the p teme embryo sac), or oth- rwise : Path of pollen tube: porogamous, chala- zogamous, or mesogamous - Type ul endosperm formation: nuclear or cellul . D or absence of endosperm in ma- seed. ture - Type of embryogeny: Onagrad, Solanad, Asterad, or otherwise. . Characteristics of suspensor: short, mas- Sive, haustorial, or otherwise. mbryo with te equally developed cot- yledons or n ' r common or not. RESULTS AND DISCUSSION Mee reales have been tabulated in order to g families (Tables 1- 4). Of the families of interest, Crypteronia Psiloxylon and Heteropyxis (Myrtaceae) are also unknown embryologically. The investigation of these taxa in relation to the characteristics pre- sented in Tables 1—4 obviously is a matter of high priority. Our embryological analyses, even though they were in most cases based on inadequate infor- the definition of Myrtales as including a certain group of families linked together by their com- mon possession of a set of shared embryological characteristics. This set of families agrees with stomataceae, Myrtaceae, Onagraceae, Olini- aceae, Penaeaceae, and Trapaceae. The bryological cl teristi mon to this set of families are: (1) Tapetum pnan (Table 2; Penaeaceae and Punicaceae are unknown in this respect). (2) Ovule sleet (Table (3) Inner integument two- levered: (Table 3). The only known exception among the core fam- ilies of Myrtales is Syzygium (Myrtaceae). Both in its initiation and subsequent early growth, the inner integument is consistently two-layered in all members of Myrtales except Syzygium. It forms a marked contrast with the outer integu- ment, the thickness of which not only varies from genus to genus but which also tends to become thicker in the course of development. The anal- yses of Davis (1966, p. 15) indicated to her that the number of integuments present should be treated either as a generic or as a specific char- acteristic, but for Myrtales it appears to be of more fundamental significance. It seems clear that the unitegmic condition of the ovule in Sy- zygium must have originated secondarily within Myrtaceae; the distribution of this feature in the family should be studied further (4) Micropyle formed by both integuments 74 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot 70 . | | | (93€ TABLE l. General information and references of the embryology of Myrtalean and non-Myrtalean families analyzed. Number of Genera Studied Level of | Families So Far Knowledge Selected References i Myrtales | 1. Combretaceae (20/400) 10 50% Brewbaker (1967); Fagerlind (1941); Karsten (1891); Nagaraj (1954a, 1954b, 1954c, 1955), Pal (1951); Rao (1963); Venkateswarlu (1952b); Venkateswarlu and Rao (1972). 2. Lythraceae (23/500) 13 15% Brewbaker (1967); Joshi and Venkateswarlu (1935a, 1935b); Smith and Herr (1971); Souéges (1925); Tischler (1917); Venkateswar- lu (19372); Warming (1878); Mauritzon (1995 1939). | 3. Melastomataceae (200/4,000) 19 3% Brewbaker (1967); Crété (1956, 1957, 1960a, | 1960b); Iconomides (1958); Ruys (1925); Sub ramanyam (1942, 1944, 1946, 1948, 1951); Ziegler (1925). Myrtaceae (140/3,000) 29 9% Brewbaker (1967); Davis (1968, 1969); Greco l (1930); Mauritzon (1939); Narayanaswaml | and Roy (1960a, 1960b); van der Pijl (1934) | Polunina (1957a, 1957b, 1957c, 1958a, ! 1958b, 1959, 1964); Prakash (1969a, 1969, 1969c, 1969d, 1973); Roy (1953, 1955, 1960, 1961, 1962a, 1962b); Roy and Sahai (1962) Souéges (1940a); Tiwary and Rao (1934). 5. Oliniaceae (1/8) l 66% Mauritzon (1939). 6. Onagraceae (17/675) 12 37% Beer (1905); Bonnet (1912); Brewbaker (1967) Gates (1911); Geerts (1908, 1909); Haberlandt (1927); Håkansson (1925); Hulbary and Ra? (1959); Ishikawa (1918); Johansen (19282, 1928b, 1929, 1930a, 1930b, 1931a, 1931, 1931c, 1933, 1934); Kahn (1942); Lebégue (1948a, 1948b); Maheshwari and Gupta (1934); Modilewski (1909); O'Neal (1923) — Pagni (1958); Renner (1914, 1921); Seshav4 taram (1967, 1970); Souéges (1920, 1935. 1946); Subramanyam and Govindu (1948): Tackholm (1914, 1915). 7. Penaeaceae (7/20) 4 3896 Stephens (1909). 8. Punicaceae (1/2) l 33% Brewbaker (1967); King (1947); Mauritzon 9. Sonneratiaceae (2/8) 2 83% Joshi (1939); Karsten (1891); Mauritzon (1939) — Venkateswarlu (1936a, 1936b, 1937b). i 1983] TOBE & RAVEN—EMBRYOLOGY OF MYRTALES TABLE l. (Continued). ber of Genera Studied Level of Families So Far Knowledge Selected References 10. Trapaceae 1 100% Brewbaker (1967); Ghosh (1954); Gibelli and Ferrero (1891); Ishikawa (1918); Ram (1956); Trela-Sawicka (1978). Non-Myrtales ll. Elaeagnaceae (3/50) 1 33% Rau and Sharma (1970); Sharma (1966). 12. Elatinaceae (2/40) 2 83% Dathan and Singh (1971); Frisendahl (1927); Kajale (1939); Lemesle (1929); Raghaven and Srinivasan (1940). 13. Haloragaceae 8/100) 3 38% Bala-Bawa (1969a, 1969b, 1970); Brewbaker (1967); Kapil (1962); Kapil and Bala-Bawa (1968); Nagaraj and Nijalingappa (1967a, 1967b, 1974); Nijalingappa (1967); Souéges (1940b); Stolt (1928). 14. Lecythidaceae 0/400) 4 1396 Brewbaker (1967); Mauritzon (1939); Treub (1884); Venkateswarlu (19522). I5: Rhizophoraceae 4/100) 6 1296 Brewbaker (1967); Carey (1934); Cook (1907); Juncosa (1982); Karsten (1891); Mauritzon 16. Thymelaeaceae (50/500) 18 1796 Brewbaker (1967); Fagerlind (1940); Fuchs heera except for the unitegmic Syzygium). failing cf to Davis (1966, p. 16), in 88 of 189 by the im angiosperms the micropyle is formed both in inner integument alone; in 74 families fia are involved; and in only four integ ne 1s the micropyle formed by the outer Rite ent. In the other families, the integu- stati E iam of the micropyle are con- fM. dS m individual genera but vary from ge- suggest E within the family. These results Bee t variation in the participation of the the micr Outer integuments in the formation of dije hs.s à an important embryological i y With- (C yrtales, the cases of Guiera senegalensis Da mbretacea 4" Winia fascicularis and D. micropetala (Myr- (1938); Guérin (1913, 1915); Kausik (1940); Mauritzon (1939); Osawa (1913); Souéges (1942); Strasburger (1884); Venkateswarlu (1945, 1946, 1947a, 1947b); Vesque (1879a, 1879b); Winkler (1904). taceae; Prakash, 1969c), Stenosiphon linifolius (Onagraceae; Johansen, 1930b), in which the mi- cropyle is apparently formed by the inner integ- ument alone, should be regarded as isolated ex- ceptions. In some of these instances, the formation of the micropyle is probably a sec- ondary characteristic caused by the spatial con- dition of the ovarian locule (see footnotes 15 and 18 in Table 1). In Trapa (Trapaceae) the micro- pyle is not formed by integuments owing to the production of an elongated nucellar beak. The nucellar beak is evidently a secondary charac- does not hinder the inclusion of Trapaceae in yrtales. (5) Antipodal cells absent or, if present, ephemeral (Table 3). Of the families of Myrtales TABLE 2. Embryological data of anthers. Num md Anther Wall Anther Middle Tapetal Cytokinesis Microspore Mature Pollen Families Sporangia Development Epidermis Endothecium Layers Tapetum Cell in Meiosis Tetrad Grain Myrtales 1. Combretaceae (20/400) 4 Basic type ex- persistent (?) fibrous ephemeral glandular 2-nucleate simultaneous tetrahedral or 2-celled! cept in isobilateral Guiera sene- galensis 2. Lythraceae (23/500) 4 Dicotyledonous — fibrous? ephemeral? glandular? 2-6-nucleate? simultaneous? tetrahedral or 2-celled type? isobilateral? 3. Melastomataceae ; (200/4,000) 4 =- persistent fibrous except ephemeral ex- glandular l-nucleate simultaneous tetrahedral 3-celled in Melasto- cept in ma and Melastoma Oxyspora malabath- ricum? 4. Myrtaceae (140/3,000) 4 Basic type persistent or fibrous ephemeral glandular 2-nucleate, simultaneous tetrahedral or 2-celled ephemeral but 1-nu- te cleate in Eucalyptus melliodora 5. Oliniaceae (1/8) 4 - — fibrous (?) ephemeral glandular 2-nucleate (?) — — — Onagraceae (17/675) 4or _ persistent fibrous ephemeral glandular 2-nucleate, simultaneous tetrahedral, 2-celled many* but multi- isobilateral nucleate in or decussate some species of Oenothera and Fuch- sia 7. Penaeaceae (7/20) ~ ~ - E zu E d = S HE 8. Punicaceae € — — 2-celled VM — 2-celled N3GP3IVOD 'TVOINV.LOS INNOSSIN JHL AO STVNNV 0L 10A] Tase 2. (Continued). Number of ther Wall Anther Middle Tapetal Cytokinesis Microspore Mature Pollen Families Sporangia Development Epidermis Endothecium Layers Tapetum ell in Meiosis Tetrad Grain 10. Trapaceae (1/15) 4 — persistent fibrous ephemeral glandular multi-nu- simultaneous tetrahedral or 2-celled cleate ecussate Non-Myrtales 11. Elaeagnaceae (3/50) 4 = — fibrous ephemeral glandular 2—4-nucleate -— tetrahedral or 3-celled decussate 12. Elatinaceae (2/40) 4 Basic type persistent fibrous ephemeral glandular 2-nucleate in simultaneous tetrahedral or 2-celled in Bergia but isobilateral Bergia, but —4-nu- 3-celled in cleate in Elatine Elatine 13. Haloragaceae (8/100) 4 Monocoty- persistent fibrous ephemeral glandular 1-5-nucleate simultaneous tetrahedral or 3-celled ledonous ecussate type 14. Lecythidaceae (20/400 — — — fibrous ephemeral amoeboid 2-nucleate — — 2-celled or 3-celled 15. Rhizophoraceae (14/100) — _ - — — — — — — 2-celled 16. Thymelaeaceae 4 Monocoty- persistent fibrous ephemeral glandular Gopr simultaneous tetrahedral or 3-celled ledonous but 2-6 isobilateral or Basic type cleate in Wikstroe- 7 Pal (1951) reported the three-celled condition in Terminalia catappa, the o (Nagaraj, 1954c; Venkateswarlu & Rao, 1972) and on other species of aminah Ne 1954a; Venkateswarlu "s Rao, "1972: Brewbaker, 1967) reported only the two-celled conditions. only on Ammannia baccifera (Joshi & Venkateswarlu, 1936 Combretaceae. Other papers on Terminalia catappa Ba ). 3 adine to Subramanyam (1948), five to seven middle layers are formed in Melastoma melabathricum, the upper two or three persisting and the remainder ultimately ag C alylophus, Clarkia, Gaura, Hauya, Heterogaura, and two species of Ludwigia have polysporangiate anthers in which microsporogenous tissue is divided by sterile septa into many distinct packets (Raven, 1969; Eyde, 1978). [e861 SITYLYAN JO ADOOTOAHSNS — NHAV?3I 7? JJOL 3, part A. Embryological data of ovules. Abbreviations: i.i., inner integument; o.i., outer integument. See Table 3, part B, beginning on page 80, TABLE A for additional characters; see page 82 for footnotes. oo Number of Thickness = Vasculature of In- Micropyle Nucellar Families Curvature! Nature of Nucellus Integuments Integumen teguments rmation Beak Hypostase Myrtales 1. Combretaceae (20/400) anatropous crassinucellate z i.i. 2-layered; o.i. absent i.i. and o. i c not formed present in -layered in nly i.i some gen- ost genera Padi sene- but 3-layered galensis in Terminalia and Bucida z 2. Lythraceae z (23/500) anatropous crassinucellate* 2 i.i. 2-layered; o.i. absent i.i. and o.i. not formed absent m -layered in most except in o genera but 5-lay- Ammannia A ered in Cuphea T 3. Melastomataceae z Q anatropous, or crassinucellate 2 i.i. 2-layered; o.i. absent i.i. and o.i. not formed absent Iz campylo- 2-3-layered ó js m (Memecylon) = 4. Myrtaceae S (140/3,000) anatropous crassinucellate 2, but i.i. 2-layered; o.i. present in the i.i. and o.i. with not formed present in a only 1 2-layered in most single integu- few exceptions!’ some gen- z in genera but 2—4 ment in Syzy- o Syzygium? layered in breed gium C genera Q 5. Oliniaceae ; z (1/8) campylotropous __ crassinucellate 2 i.i. 2-layered; o.i. present in o.i. i.i. and o.i. not formed absent m 4-layered » 6. Onagraceae (17/675) anatropous crassinucellate 2 i.i. 2-layered; o.i. absent i.i. and o.i., but not formed present i 2- or multi-lay- Li most genera ered Stenosiphon"? 7. Penaeaceae (7/20) anatropous crassinucellate 2 i.i. 2-layered; o.i. absent i.i. and o.i. not formed — 2-layered 8. Punicaceae ix (V2) anatropous crassinucellate 2 i.i. 2-layered; o.i. absent i.i. and o.i. not formed absent $ 4-layered P : E à Mug ast RM e ttti qt mento tt MR s icum ires s re ai pt ti uà ERES vt aad iiid — à iiis : i LE 3, part A, continued from page 78. Embryological data of ovules. Abbreviations: i.i., inner integument; o.i., outer integument. See Table 3, part B, beginning on page 80, for additional characters; see page 82 for footnotes. Number of Thickness xin Vasculature of In- Micropyle Nucellar Families Curvature! Nature of Nucellus Integuments Integuments? teguments Formation Beak Hypostase 9, Sonneratiaceae (2/8) anatropous crassinucellate i.i. 2-layered; 0.1. absent i.i. and 0.1. not formed absent 2-layered 10. Trapaceae (1/15) anatropous crassinucellate i.i. 2-layered; o.i. absent not formed formed present 5-14-layered Non-Myrtales 11. Elaeagnaceae (3/50) anatropous crassinucellate — — — not formed present 12. Elatinaceae (2/4 anatropous crassinucellate i.i. 2-layered; o.i. absent i.i. and o.i not formed absent 2-layered 13. Haloragaceae (8/100) anatropous crassinucellate i.i. 2-layered; o.i. absent i.i. and o.i. not formed present -layered 14. Lecythidaceae (20/400) anatropous tenuinucellate l.i. multi-layered; present in o.i. LL not formed absent i. multi-layered 15. Rhizophoraceae (14/100) anatropous crassinucellate i.i. massive; o.i present in o.j. in i.i. and o.i. in formed in Aniso- — massiv some genera Bruguiera and phyllea Rhizophora; only i.i. in Gyno- troches; not formed in Aniso- phyllea (?) 16. Thymelaeaceae (50/500) anatropous crassinucellate i.i. 3—4-layered; o.i. absent Li. not formed present in 3—4- layered man species [£861 STIV.LHANW JO ADOTOAUAWA— NHAV3 7? IJOL part B. Embryological data of ovules. Abbreviations: i.i., inner integument; o.i., outer integument. See Table 3, part A, beginning on page 78, E 3, oc for additional characters; see page 82 for footnotes. o Pattern of Fatty Globules Embryo in Megaspores Number of Nuclei Cytokinesis Megaspore Functional Sac and Embry in Mature Families Endothelium Archesporium? in Meiosis — Tetrad^* Megaspore? Formation Sacs Synergids Antipodal Cells Embryo Sac Myrtales Combretaceae (20/400) not formed l-celled$ occurs linear chalazal cell Polygonum absent hooked and ephemeral, but 5, but 8 or more type” i persistent in in Guiera se- Guiera sene- negalensis galensis » 2. Lythraceae : z (23/500) not formed multi-celled; ^ occurs linear chalazal cell Polygonum absent hooked ephemeral 5 Z only one type O functions T 3. Melastomataceae T (2 not formed l-celled occurs linear chalazal cell Polygonum absent hooked ephemeral 5 m type = Myrtaceae O (140/3,000) not -celled occurs linear chalazal cell Polygonum absent, but diverse in form ephemeral 5 = formed!! type present in E Psidium E guajava Ad 5. Oliniaceae zZ (1/8) not formed l-celled occurs linear chalazal cell Polygonum absent — ephemeral 5 Q ty = 6. Onagraceae g) (17/675) not formed l-celled occurs linear micropylar Oenothera absent filiform absent 4 m cell z 7. Penaeaceae s: (7/20) not formed l-celled does not decussate'? all 4 nuclei Penaea absent — absent 16 Occur type 8. Punicaceae (1/2) not formed l-celled occurs linear chalazal cell Polygonum absent elongated ephemeral 5 type 9. Sonneratiaceae (2/8) not formed multi-celled; ^ occurs linear chalazal cell Polygonum present in hooked ephemeral 5 only one type Sonneratia — functions but absent $ in Duaban- Jm ga a MM qii ————À 5007 ——— rÀ csi —— ——— — — €—— —— cate eina nn atl aM Goth don capras go ids b t — EE part B, continued from page 80. LE 3, Embryological data of ovules. Abbreviations: A, V tomi on page 79, for additional characters; see page 82 for footnotes. T; inner integument; O.i., outer integument. See Table 3, part Pattern of Fatty Globules mbryo in Megaspores Number of Nuclei Cytokinesis Megaspore Functional Sac and Embryo in Matur Families Endothelium Archesporium? in Meiosis — Tetrad* Megaspore? Formation Sacs Synergids Antipodal Cells Embryo Sac 10. Trapaceae (1/15) not formed 1-2-celled occurs linear chalazal cell Polygonum absent pyriform ephemeral 5 type Non-Myrtales 11. Elaeagnaceae (3/50) — 1—3-celled occurs linear chalazal cell Polygonum absent — ephemeral 5 type 12. Elatinaceae (2/40) not formed multi-celledi ^ occurs linear or chalazal cell Polygonum absent hooked and ephemeral 8 only one T-shaped type pyriform in functions Bergia am- manioides 13. Haloragaceae (8/100) not formed l-celled occurs linear chalazal cell Polygonum absent hooked and persistent 8 type pyriform 14. Lecythidaceae (20/400) formed in multi-celled; ^ occurs linear chalazal cell Polygonum absent pyriform ephemeral 5 many only one type species!* functions 15. Rhizophoraceae (14/100) formed in 1-celled occurs linear chalazal cell Polygonum absent pyriform ephemeral in Sorg Carallia, Ceriops but Gyno- rsistent in troches, Gynotroches Bruguiera and Cas- sipourea 16. Thymelaeaceae (50/500) not formed l-celled occurs linear chalazal cell Polygonum absent hooked, rarely persistent; cells 8 or more type filiform usually a i (Daphne can- fy (up to nabina) many as 30 in [£861 STIV.LHAN JO ADOTOAYAWA—NAAVA Y 3HOL TABLE 3, parts A & B, concluded. Footnotes. ! Predominant or usual conditi ? Based on the original thickness ofi integuments seen at the initiation and early growing stage. diti “commonly several archesporial cells differentiate” in Combretaceae. In contrast, ore recent pape (1972), ihe wie of Moe dealt with 18 species in 9 genera pine Terminalia d qd blot made it clear that the one-celled archesporium is common and ne multi-celled one (i.e., consisting of two or more cells) is auritzon (1939) i ied the 16-nucleate Penaea-type em tts o sac in two species of Combretum, C. paniculatum and C. pincianum. These results to "Eh reconfirmed because other authors have reported ecd Polygonum-type embryo sacs in Combretum. Variation in female gametophyte formation is iens in some genera, however (see Hjelmquist, 1964), for re 8 Smith and Herr (1971, p. 198), contrary to TE other authors, stated that “the nucellus of Ammannia coccinea was tenuinucellate.” But tenuinucellate ovules have never been reported in other species of Ammania nor in other Lythraceae. The drawings of sections of ovules (Smith & Herr, 1 . 167, Fi 4 and 5) which they intended to document the existence met tenuinucellate condition qe this species indicate rather that their material was clearly crassinucellate; the pore iet ao that - beers with the archesporial cells are evidently megaspore mother cells cut off from the actual archesporia s. sid pecies of Syzygium, bacio paniculata — 1939), E. Jambos (Pijl, 1934), E. malaccensis (Pijl, 1934; Roy, 1960), E. fruticosa (Roy, 1961) ue x eee Roy, 1962b) are reported to have a unitegmic ovule. But all of these Eugenia are assigned to Syzygium (sensu Schmid, 1972): ac ino to Schmid (pers. comm.), Eugenia denk is assigned to Syzygium paniculatum Gaertner, Eugenia jambos to Syzygium jambos (L.) Alston, Eugenia malaccensis to Syzygium malaccense (L.) Merr. & Perry, Eugenia fruticosa to Syzygium fruticosum DC., and Eugenia myrtifolia to Syzygium us Boh ris (Roxb.) DC. On the other hand, Eugenia bracteata, which should remain in Eugenia sens. str. sensu Schmid, has a bitegmic € (Roy, | 55). or i integument." ' In fact, Darwinia JUS ccs and D. micropetala have dne micropyle formed by the inner a alone Paak, 1969c), whereas Darwini uis i Rises in the formation of the mic e The e secon growth of the ram integument combined with the suppression of growth of the — integumen of the limited space available within the ovarian locule may result in the formation of the micropyle by the inner integument only. In species of Darwinia that have the Demi ui formed by the inner omen only, the inner integument develops into "collar-like lips" (Prakash, 1969c), a pekenn that clearly seems to on !! Endo hei has not been duong in €—— but in Leptospermum, Kuntzea, Agonis, Callistemon, and Melaleuca, the — part of the nucellus tends sd be destroyed by the growth of the mbryo sac, a feature that y accompany ium. In Melaleuca particularly, the poen re embryo sa € borders directly on hei inner integument (Mauritzon, ie “beak-like process" of the inner integument (Johansen, 1930b) a characteristic that seems to be soak or in the family. 13 Cell walls are absent, but four megaspore nuclei are arranged in such a way that one lies at the top, one lies at the bottom, and the other two lie at the sides. The decussate arrangement is acquired by oblique divisions of the micropylar nucleus and of the chalazal nucleus that were formed by meiosis I, and not by a combination of vertical and transverse divisions. 14 According to Mauritzon (1939), an endothelium (called a “mantle layer" by him) is formed in Couroupita guianensis, X arborea, and Barringtonia bot and in speciosa but not in other species of Barringtonia nor in Gustavia angusta. Venkateswarlu (1952) reported its occurrence apoleona imperialis Barringtonia acutangula. Yhe absence of an endothelium in some species of fennen and Gusiavia should be reco 2 ed, because the presence of end: Eibrlium.i js otherwise a family « characteristic of Lecythida proc (Davis, 1966, p. 16). Pro nen ritzon's observations had not extended to old enough stages of ovule development to observe the and it is likely that it is "eB tah Birman ceae. N3QVO TVOINV.LO8 IYNOSSIN JHL AO STVNNV 0L 10A] 1983] mentioned above, all except Onagraceae and Penaeaceae consistently have ephemeral anti- podal cells that degenerate and disappear before fertilization. In these families, therefore, the ma- ture embryo sac comprises only five nuclei or cells: an egg; two synergids; and two polar nuclei, which may fuse into a secondary nucleus before fertilization. This relationship suggests that for yrtales the antipodal cells are unnecessary in the organized mature embryo sac. Exceptionally, the antipodal cells of Guiera senegalensis (Com- bretaceae) persist into the postfertilization phase, a feature that is regarded as unique for Combre- laceae (Venkateswarlu & Rao, 1972). In Ona- graceae and Penaeaceae, antipodal cells are ab- sent throughout megagametophyte development. In the 16-nucleate Penaea-type embryo sac, one might regard the three chalazal cells as corre- sponding to the antipodal cells. Stephens (1909) mentioned, however, that these three chalazal cells resembled an egg apparatus more or less Closely. In fact, the three chalazal cells are pro- duced in the same way as the three cells ofa true micropylar egg apparatus, but from a different mégaspore nucleus. E. Nuclear-type endosperm formation (Table (7) Exalbuminous seed (Table 4). Core Myrtales E together, these features characterize the Ma. yrtales group defined by Dahlgren and à € (1983). Within this assemblage, Lythra- ae are further characterized (except for Punica) Y having a multi-celled archesporium in the M P a EN that is not common elsewhere Nod. er. Although this characteristic has in Feceived less attention than other em- i an features, it does appear to have some k €ast in particular cases. Thus, Vijayara- ind dea adopted it as one of the bases for clusion of Paeonia from Ranunculaceae. A for a close relationship among these 19365 EM, 1939; Venkateswarlu, 1936a, Vationg 7b). In agreement with these obser- ies tad analysis of the embryological liter- them Ec no basis for distinguishing between j Supports their inclusion in a single TOBE & RAVEN—EMBRYOLOGY OF MYRTALES 83 family (Dahlgren & Thorne, 1983). Sonneratia and Duabanga differ in that the former consis- tently has fatty globules in its megaspores and embryo sacs whereas these are lacking in Dua- banga (Karsten, 1891; Venkateswarlu, 1937b; Mauritzon, 1939). This accords with certain oth- er lines of evidence in suggesting that these gen- era may not be directly related to one another. Punica differs from other Lythraceae in having a uni-celled archesporium, and a thick multi- layered outer integument. Nonetheless, it is in- cluded in Lythraceae by Dahlgren and Thorne (1983). Further embryological studies on Punica are clearly needed. Although many of the embryological charac- teristics of Penaeaceae are unknown, those that The only distinctive feature among those that have been reported is the characteristic 16-nu- cleate Penaea-type embryo sac. This type of em- bry levelop is characterized by the fact that meiosis in the megaspore mother cell is not accompanied by cytokinesis. Following meiosis, all four of the decussately arranged megaspore nuclei function, each dividing twice to produce four nuclei or cells. The quartet associated with the true micropylar egg apparatus is always de- rived from the micropylar megaspore nucleus (Stephens, 1909). In both the frequency of di- vision of the megaspore nucleus and the origin of the egg apparatus, the Penaea-type embryo sac development resembles the Oenothera-type of embryo sac development that is characteristic of Onagraceae, discussed below (see also Ma- heshwari, 1948). Melastomataceae exhibit a distinctive embry- ological feature that is unknown elsewhere in the order, uni-nucleate anther tapetum cells in place of the bi- or multi-nucleate ones characteristic of other families of Myrtales. In addition, Me- lastomataceae are the only family of Myrtales in which the endothecium sometimes develops fi- brous thickenings. Another distinctive feature that has been claimed for Melastomataceae is three-celled mature pollen, in place of the two- celled mature pollen characteristic of all other families of Myrtales. A careful review of the few published reports, however, has indicated that Melastomataceae, like all other core Myrtales, have two-celled mature pollen (Tobe & Raven, in prep.). In this connection, it seems to be worth- while to note the fact that the pollen cell con- TABLE 4. Embryological data of seeds. Endosperm Endosperm in Path of Families Pollen Tube Formation Mature Seed Embryogeny Suspensor Embryo Polyembryony Myrtales 1. Combretaceae 20/400 porogamous Nuclear type absent Asterad type short and small 2 cotyledons with rare same size 2. Lythraceae (23/500) porogamous Nuclear type absent! Onagrad type! short and small! 2 cotyledons with absent 3. Melastomataceae (200/4,000 porogamous Nuclear type absent Onagrad type short and mas- 2 cotyledons with occasional sive 4. Myrtaceae (140/3,000) porogamous Nuclear type absent Onagrad type short and small, 2 cotyledons with usual in Sy- or absent same size zygium 5. Oliniaceae (1/8) — Nuclear type — — — — — 6. Onagraceae porogamous Nuclear type absent Onagrad type short and small 2 cotyledons with rare same size 7. Penaeaceae (7/20) porogamous Nuclear type absent Asterad type absent 2 cotyledons with = same size 8. Punicaceae 2/8 — Nuclear type — — — — = 9. Sonneratiaceae (2/8) porogamous Nuclear type absent Onagrad type short and small 2 cotyledons with absent same size 10. Trapaceae (1/15) porogamous Nuclear type absent Solanad type long, coiled, 2 cotyledons with absent haustorial; the upper part forming collar extremely differ- v8 N3QGUVD TVOINV.LOS IXNOSSIN JHL HO STVNNV AL 10A] [£861 STYIV.LHAIN AO ADOO'IOASIHIW3 — NJAV?H X 3HOL TABLE 4. (Continued). Path of ndosperm Endosperm in Families Pollen Tube Formation Mature Seed Embryogeny Suspensor Embryo Polyembryony Non-Myrtales 11. Elaeagnaceae (3/50) — Nuclear type absent Asterad type short and mas- 2 cotyledons with absent sive same size 12. Elatinaceae 2/40 -— Nuclear type present Solanad type short and small 2 cotyledons with absent same size 13. Haloragaceae (8/100) porogamous Nuclear type in present Caryophyllad 2-celled, en- 2 cotyledons with absent aurembergia type larged, hausto- same size but Cellular rial type in Halora- gis and Myriophyllum? 14. Lecythidaceae (20/400) — Nuclear type — — short and mas- — — sive in Bar- ringtonia vrie- sei 15. Rhizophoraceae (14/100) porogamous Nuclear type present — short and mas- 2 cotyledons with rare sive, or long same size 16. Thymelaeaceae (50/500) porogamous Nuclear type present? Asterad type short and small, 2 cotyledons with rare or absent same size ! Based only on Duabanga idi (Venl , 1937b). ? Even in the Nuclear type, c ly as the 8 leate stage (Bala-Bawa, 1969a; Nagaraj & Nijalingappa, 1974), a phenomenon that demonstrates a strong te endency towards the Cellular ? According to Guérin (1915), in Thymelaeaceae ae that completely lack endosperm are rather exceptional. Even in the species of Phaleria which nearly lack endosperm, two- to five-layered endosperm ussue was present on the surface of the cotyledons in the mature seed. 86 dition is highly consistent in the order (cf. Brew- baker, 1967). The pollen cell condition of Melastomataceae should be studied further. Although little is known about the embryology of Oliniaceae, what is known clearly supports their inclusion within Myrtales, as do their other features (Rao & Dahlgren, 1969). The only known distinctive feature is the thick, vascularized outer integument. Closer comparisons with Melasto- mataceae ought to be made when more infor- mation is available, following the suggestions of Rao and Dahlgren (1969); the comparisons with Thymelaeaceae and Rubiaceae suggested by these authors cannot be taken as indications of genuine relationship in view of the embryological and other features of Olinia. Combretaceae have been studied as exten- sively from an embryological point of view as any family of Myrtales. Guiera differs in several respects from the other members of the family and indeed stands out within Myrtales, but in general, Combretaceae agree closely in their em- bryological features with other Myrtales. Guiera differs from all other Myrtales in having persis- tent antipodal cells and a micropyle formed by the inner integument only (Venkateswarlu & Rao, 1972). Based on these features, Venkateswarlu and Rao (1972), taking into account also evi- dence from floral morphology and anatomy, pro- posed the establishment of a unigeneric tribe, Guiereae, for Guiera. In its overall features, how- ever, Guiera clearly fits into the subtribe Com- bretinae together with Combretum and other genera (Stace, 1965; Exell & Stace, 1966, 1972), so that its unusual embryological attributes must almost certainly be secondary. The persistence of antipodal cells, which multiply after fertiliza- tion, seems clearly to be secondary, and it ap- pears logical to consider the lack of participation of the outer integument in the formation of the micropyle secondary also. From an embryological point of view, Tra- paceae are the most distinctive family included in Myrtales. A long nucellar beak is formed in Trapa as a result of an extension of the apex of the nucellus. The formation of this beak evi- dently precludes direct participation of the in- teguments in the formation of the micropyle. A second very distinctive feature of Trapa is the long, coiled haustorial suspensor, the upper part of which forms a collar supporting the embryo proper. A third distinctive feature is the asym- metrical embryo, which has two very unequal cotyledons. Some or all of these features might ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor 70 i be related to the aquatic habitat of the family but this suggestion needs further evaluation. Other embryological characteristics of this monogeneric family include the thick, nonvascu- larized outer integument and the multinucleate anther tapetal cells. The embryological charac teristics of Trapaceae do not support a dired relationship either with Onagraceae (a family in which Trapaceae were formerly included) or with Lythraceae (contrary to the views of Miki, 1959). Myrtaceae, like Combretaceae, have few dis- tinctive features that discriminate them from other Myrtales. Syzygium is unique in the order in its unitegmic ovules. Doubtless this represents l a derived feature as it does in angiosperms gêi- — erally (Bouman, 1977), and it probably will be found in other genera when the family is better known embryologically. Mauritzon (1939, P. 102-103) stated that in “Eugenia paniculata” (=Syzygium paniculatum) “two two-layered m teguments have fused together to form a single | integument of four layers.” Unitegmy could also | be derived by the loss of one of the two integl” ments. A third way in which unitegmy could derived has recently been suggested by Bouman (1977) and Bouman and Schier (1979). The authors have concluded that both in Ranuncl laceae and in Gentianaceae unitegmy has omer nated following a complicated process involving the fusion of the primordia shifting of the inner integument and an arre that has led to a | sted ! development of it subsequently. If unitegmy has | originated in this way in Syzygium, a major i of its single integument would be composed 0 the outer integument. This integument might then represent an ancestral form of outer integument pears virtually certain that unitegmy in Syzye um has been derived from bitegmy within Eo taceae because the genus is not otherw! $ remarkable within the family. The distributio? of unitegmy in Myrtaceae should be investiga further. tive 4-nucleate Oenothera-type embryo $aC © known elsewhere in angiosperms. This has been considered to have a close relat » to Lythraceae, and several embryologists e: Tischler (1917), Mauritzon (1934), and Josh! j Venkateswarlu (1936) considered that the H thraceous embryo sac with ephemeral ant! cells forms phylogenetically an intermediate Onagraceae are characterized by their € fami | jonshiP — 1983] between the 4-nucleate embryo sac of Onagra- ceae and the normal 8-nucleate embryo sac. In fact, ephemeral antipodal cells are common to most members of the order Myrtales, and the condition in Lythraceae cannot, therefore, be taken as indicative of a direct relationship be- tween this family and Onagraceae. The Onagrad- type embryogeny common to both families oc- curs in Myrtaceae and Melastomataceae as well. Thus none of the embryological attributes of On- agraceae suggests a particular relationship either to Lythraceae or to any other family. Embryo- logically, the relationship between the 4-nucleate Oenothera-type of embryo sac in Onagraceae and the 16-nucleate Penaea-type embryo sac in Pen- aeaceae appears to be of more interest, and per- haps more suggestive of a direct relationship. These types of embryo sac development resem- ble one another in that antipodal cells are not formed at all duet . ome e inthat the mi ropylar megasp : baih families divides only twice to form an egg ap- paratus (see also Maheshwari, 1948). Although these similarities may have originated as a result of convergent evolution, they do at least indicate that the loss of the antipodal cells by omission e om nuclear division in megagametogene- occurred twice in Myrtales, since Pen- Nee and Onagraceae are manifestly not re- mes _ to one another. At present, we Meo at ephemeral antipodal cells constitute . Ave condition in Myrtales that was pres- ent in the common ancestor. Evaluation of Other Families Nes differ fundamentally from Myrtales, (2) a. r, in having (1) persistent antipodal cells; "er Cellular- or Nuclear-type endosperm ormation (Nu : i clear-t tion in Halora ype endosperm forma th b Tom à core Myrtales in exhibiting the Monocot- type of anther wall formation; Cary- ype embryogeny; and a 2-celled, en- austorial s This type of hangs: uspensor. 1s typ torial suspensor may be related to the ad- TOBE & RAVEN—EMBRYOLOGY OF MYRTALES 87 aptation of Haloragaceae to a more or less aquat- ic habitat. In this connection it may be noted that Trapaceae, which are also strictly aquatic, also have a very distinctive type of haustorial suspensor. Taken together, these embryological features virtually exclude the possibility of any direct relationship of Haloragaceae with Myr- le Insufficient information concerning Rhizo- phoraceae (including Anisophylleaceae) is avail- able to characterize the family or its constituent parts. Notwithstanding this, in their massive in- ner integument and albuminous seeds, Rhizo- phoraceae differ sharply from Myrtales. The available embryological information indicates a mr which might not be related directly to one another. In Bruguiera and Rhizophora (Rhizophoreae) the micropyle is formed by both integuments (Cook, 1907; Carey, 1934; Mauritzon, 1939), and in Gynotroches (Gynotrocheae) it is formed by the inner integument alone (Mauritzon, 1939). In Anisophyllea (Anisophylleeae) the extension of the nucellar apical tissue (? = nucellar beak) pre- vents the formation of the micropyle by integ- uments (Karsten, 1891). Furthermore, an en- dothelium is formed in Carallia and Gynotroches (Gynotrocheae; Karsten, 1891; Mauritzon, 1939), Cassipourea (Macarisieae; Juncosa, 1982) and in Bruguiera (Rhizophoreae; Mauritzon, 1939) but not in the other genera. The antipodal cells are ephemeral in Ceriops and Bruguiera (Rhizopho- reae) but persistent in Gynotroches (Gynotro- cheae; Karsten, 1891; Mauritzon, 1939). The outer integument is vascularized in Rhizophora, Ceriops, and Bruguiera (Rhizophoreae; Carey, 1934; Mauritzon, 1939), but probably not in the other tribes. The endothelium and the nature of the integuments that form the micropyle appear to be characteristic of Rhizophoraceae at the family level (Davis, 1966). Those members of Rhizophoraceae with persistent antipodal cells differ from Myrtales in this respect. Further de- tailed study on the embryology of the various genera of Rhizophoraceae should be valuable in evaluating the apparent heterogeneity of this Sy mily. a citendns differ from Myrtales in pos- sessing (1) a thick, 3- to 4-layered inner integu- ment; (2) micropyle formed by the inner integ- ument alone; (3) persistent antipodal cells that often multiply; and (4) albuminous seeds. Em- 88 ANNALS OF THE MISSOURI BOTANICAL GARDEN bryologically, these differences are decisive in uling out any direct relationship between Thymelaeaceae and core Myrtales. Lecythidaceae have many distinctive features embryologically. They are: (1) an amoeboid an- ther tapetum; (2) tenuinucellate ovule; (3) a thick, multi-layered inner integument; and (4) micro- pyle formed by the inner integument alone. In addition, the family is characterized by having an endothelium (Mauritzon, 1939). This com- bination of features seems decisive in precluding any direct relationship between Lecythidaceae and Myrtales. Ithough its two constituent genera differ in several respects, Elatinaceae agree with Myrtales in many embryological features. Among them are the glandular anther tapetum; crassinucellate ovule; a 2-layered inner integument; a micropyle formed by both integuments; ephemeral anti- podal cells; and Nuclear-type endosperm for- mation. The only known embryological differ- ence between Elatinaceae and Myrtales concerns the albuminous seeds of Elatinaceae compared with the exalbuminous seeds of Myrtales. Fri- sendahl (1927) observed a one-layered endo- sperm in ripe seeds of Elatine spp. Dathan and Singh (1971) observed a three- to five-layered endosperm in mature seeds of both Bergia odor- ata and B. ammanioides and a one- to two-lay- ered one in those of B. aestivosa; in contrast, Raghaven and Srinivasan (1940) did not observe endosperm in the mature seeds of Bergia capen- sis. Despite the difference in persistence of endo- sperm, on the basis of embryological features alone we would suggest that Elatinaceae might be placed adjacent to core Myrtales. The clas- sifications of Takhtajan (1980) and Cronquist (1981) regard Elatinaceae as having a close re- lationship with Clusiaceae-Hypericoideae. Em- bryologically, Hypericoideae differ markedly from Elatinaceae in having tenuinucellate ovules and an endothelium, and we suggest that this relationship suggested by Takhtajan and Cron- quist is probably incorrect. Elaeagnaceae have been studied very inade- quately from an embryological point of view. gh infi ioni ilable, particularly concerning the integuments, to evaluate properly the possibility of a relationship between the Elaeagnaceae and Myrtales on that basis. Avail- able information indicates that Elaeagnaceae share the following embryological features with Myrtales, however: glandular anther tapetum; >| | [VoL. 70 | | crassinucellate ovule; ephemeral antipodal cells, | Nuclear-type endosperm formation; and exa. buminous seeds. The embryology of Elaeagna- ceae should be studied in more detail and the | possibility of a relationship of this family with Myrtales should be evaluated further in the light | of this information. | In summary, Myrtales are clearly circum | scribed by embryological evidence to include | Combretaceae, Lythraceae (including Punica- l ceae and Sonneratiaceae), Melastomatacea®, | Myrtaceae, Onagraceae, Oliniaceae, Penaea and Trapaceae. We further suggest that the pu sibility of a direct relationship between Elatini- | ceae and Myrtales is supported by embryologi? data and should be investigated further from ? l number of other points of view. E | Within Myrtales, the large families Myrtaceae and Combretaceae seem not to be distinctiveané | to agree with the generalized characteristics | the order in their embryological features. D^ | d Sonnerati- | | aceae) seem to form another relatively general - ized group within the order, standing 4p somewhat from Myrtaceae and Combretaceae i" their multi-celled archesporium. This m ti-celled archesporium might be taken either as a second- ary or as a primary characteristic, and its u | should be investigated further in the context phylogenetic studies of the group. Melastomaté ceae deviate from all other families of Myrta in the characteristics of their anthers. Penaeact# and Onagraceae, although the latter are m more generalized than the former in their em bryological characteristics, are both relativell specialized in their loss of antipodal cells. j small families of Myrtales that otherwise Se" not to be directly related to one another— Olim ceae, Punicaceae (included by Dahlgren b eiii pr ERE o ie pq ti t assumed that Oliniaceae, Punicaceae, and a paceae may retain the primitive integumen” condition for Myrtales. In that case, Punicace®® although they might still legitimately be p as belonging in Lythraceae, might be viewed | a distinctive, archaic offshoot within that x 1983] Trapaceae are more distinctive in their embryol- ogy than any other family of Myrtales and differ more from the core Myrtales in their embryo- logical features than do Elatinaceae. FUTURE DIRECTIONS FOR RESEARCH More embryological information about Myr- tales should prove useful in elucidating further the relationships of the families of the order in the future, and in turn in evaluating the rela- tionships of this order to other groups of angio- sperms. Even though the available embryologi- cal evidence has allowed a clear circumscription of Myrtales on embryological grounds alone, our comparisons of many taxa are based on incom- plete information. It is particularly critical that no information at all has been available on a number of very interesting genera of Myrtales such as: Alzatea, Axinandra, Crypteronia, Dac- tylocladus, and Rhynchocalyx, one, several, or all of which may constitute Crypteroniaceae (Van Beusekom-Osinga & Van Beusekom, 1975; va 3 's available on Heteropyxis and Psiloxylon, gen- era that might be rega ra bau eee families or be included within Myrtaceae de- ere upon one’s point of view (Schmid, 1980). wi Rt all of these genera clearly belong * order Myrtales, a study of their em- to xU has the potential of contributing much .,.. Overall understanding of the elements Within this order on, Oliniaceae, Penaeaceae, and Puni- thori inis poorly known embryologically, and uation rmation is necessary for a proper eval- of their status and relationships. Even Nee dad and Myrtaceae, for which the . . «nowledge from an embryological point of iiiv ^n imated as only 396 and 996 re- Since eY, Should be studied more extensively order ga are the two largest families of the tributes ation of their embryological at- ne Critical, i "he the groups that have been included lai e Said to be related to Myrtales, no de- mbryological information is available 'n8 Chrysobalanaceae and Coridaceae. Chrysobala Naceae are sometimes considered to TOBE & RAVEN—EMBRYOLOGY OF MYRTALES 89 be closely allied to the ancestor of Myrtales be- cause of their possession of many common fea- tures (Dahlgren & Thorne, 1983). Dahlgren and Thorne (1983) indicated a few gross embryolog- ical features of these families including the fact that in Chrysobalanaceae the ripe seeds are ex- albuminous, and in Coridaceae, the ovules are bitegmic and tenuinucellate. As regards Rhizo- phoraceae, embryo l cords with the conclusions of Dahlgren and Thorne (1983) and others about the basic het- erogeneity of the family. Detailed information about the embryology of the constituent parts of Rhizophoraceae sens. lat. would no doubt con- tribute in an important manner to the resolution of the problem of relationships of Rhizophora- ceae and the proper constituency of this family. A few comments on the references we con- sulted during the course of preparation of this paper are now appropriate. Most of these refer- ences were incomplete, even though the respec- tive authors might have drawn a great deal more tophyte than to other features of great systematic interest such as the development of the nucellus and that of the integuments. As mentioned above, very few studies refer to the thickness of the in- teguments except in descriptions of the structure of seed coats. Despite this, a proper understand- ing of the thickness of the integuments seems clearly to be of some value as an indicator of affinity and phylogenetic advancement of re- spective groups. The thickness of secondarily amplified integ t to be of less interest as a systematic characteristic than is the original thickness as the integuments are initiated and starting to grow. The thickness of mature integ- uments is directly related to the structure of the mature seed coats, which may be highly diverse even within a single genus or a group of related genera; e.g., “Eucalyptus” (Gauba & Pryor, 1958, 1959, 1961). More information about the thick- ness of the integuments and their structure before they begin to thicken during the course of seed coat formation will probably prove to have con- sons. In this connection, Bo ] stated that “seemingly identical multi-layered in- teguments (such as those of J uglandaceae and Liliiflorae) originate by means of different on- 90 ANNALS OF THE MISSOURI BOTANICAL GARDEN togenetic processes” and that “characters de- rived from the structure of mature seed coats are of doubtful taxonomic significance at and above the family level, aa they are amplified by en togenesis should be incorporated routinely into embryological studies, particularly those intend- ed to have systematic Wopbcatiett: ted to ab- normal or exceptional cases, and there is less attention generally given to those characteristic features for particular groups that may have much greater taxonomic importance. In many refer- ences, it is difficult or impossible to determine which was the characteristic state for the taxon being investigated. Thus, the archesporium may be described as one- to three-celled or as one- celled but occasionally multi-celled, but without any explanation as to which was the character- istic state for that taxon. Exact statements about Many evaluation of embryological reports and for pro- moting their utility in the future. Similar obser- clude that embryological studies should be more extensive in respect of the features that they con- sider and report, and more precise in terms of the way in which they handle anomalous or ex- ceptional situations encountered. In conclusion, embryological features afford a sound basis for the delineation of Myrtales. There re, however, many gaps to be filled in our knowledge of the taxa that clearly belong within this order, and other taxa that may or may not be directly related to it. A careful choice of the characters to be examined and critical descrip- tions of these characteristics will assist greatly in our ongoing evaluation of relationships within Myrtales and allied groups. LITERATURE CITED BALA-BAWA, S. 1969a. Embryological studies on the Ha iloragidaceae. Ir. Laurembergia brevipes ues dl. Proc. Natl. Acad. Sci. India 35: 273-290. 969b. Embryological studies on the Hal- oragidaceae. III. MATO tesi aut DC. Beitr. Biol. Pflanzen 45: 447—464. —— 1970. Hoi gi cie In Proceedings of the [Vor. 70 Symposium on Comparative Embryology of An- giosperms. Bull. Indian Natl. Sci. Acad. 41: 226- 229 Beer, R. 1905. On the development of the pollen grain and anther of some Onagraceae. Beih. Bot. Centralbl. 19: 286-313. BEUSEKOM-OSINGA, R. J. VAN & C. F. BEUSEKOM. 1975. Delimitation and subdivision of the Crypteroni- aceae (Myrtales). Blumea 22: 255-266. | BOESEWINKEL, F. D. 981. Development of ova and dissertation, Univ. Amsterdam. : | BoNNET, J. 1912. Recherches sur l'é évolution | des cellules-nourriciéres du pollen, chez les Angio- | spermes. Arch. Zellf. 7: 604-722. Bouman, F. 1971. The application of tegumentary studies to taxonomic and yet Probi | Ber. Deutsch. Bot. Ges. 84: 169-17 977. Integumentary dene third way | to unitegmy. Ber. Deutsch. Bot. Ges. 90: 15-28. & S. ScHiER. 1979. Ovule ontogeny : and ssi on the evolutionary origin of the single integ | BREWBAKER, J. L. 1967. The distribution and ph* | logenetic significance of binucleate and a | pollen grains in the angiosperms. Amer. J. Bo. | 54: 1069-1083. Carey, G. 1934. Further invention on the ei — bryology of viviparous se Linn. Soc. New | South Wales 59: 392-410. Cook, M. T. 1907. The embryology of Rhizophor . mangle. Bull. Torrey Bot. Club 34: 271-277. | CrÉTÉ, P. 1956. A propos de l'embryogénie du Se nerila wallichii. Bull. Soc. Bot. France 103: dl 603. i 957. Embryogénie des Melastomacées. 1 i veloppement de l'embryon chez le ne "e | Compt. Rend. Hebd. Séances Acad. Sci 3 f 376. 960a. Embryogénie du Calvoa rien 1 Compt. Rend. Hebd. Séances Acad. Sci. 250: 37 | T 1960b. Embryogénie des Melastomacés Développement de l'embryon chez le Bertolo | | maculata. Compt. Rend. Hebd. Séances s Acad. 251: 1907-1909. Cus CRONQUIST, A. 1981. An Integrated System ofc a sification of Flowering Plants. Columbia Univer 4 sity Press, New Yor DAHLGREN, R. 1975. The dinpiburion aff characitf | within an angiosperm syste embi” oed characters. Bot. Not. 128: 81197. R. F. Torne. 1983. The order! M circumscriptio: ,and Missouri Bot. Gard. 70. (in sean : 1971. eme 1 Systematic wee "Td t. 16: 19- Floral viórplicli ad ‘the develo? | 1983] t y te p ] +4. SEE 2 8 Austral, jJ. Bot. 17: 177- 190. ExELL, A. W. & C. A. SrAcE. 1966. Revision of the Combretaceae. ue Soc. Brot., Ser. 2, 40: 5-26. & Patterns of distribution in the CB EG Pp. 307-323 in D. H.Valentine (editor), Taxonomy, Phytogeography and Evolu- tion. Academic Press, London EYDE, R. H. 1978. Reproductive structures and evo- lution in Ludwigia (Onagraceae). I. Androecium, placentation, merism. Ann. Missouri Bot. Gard —655. : 644. FAGERLIND, F. 1940. Zytologie und hari cidit bi ildung i in der Gattung Wikstroemia. Heredita 26: 23-50. 1941. Der Bau der Samenanlage und des Macrogametophyten bei Quisqualis indica. Bot. Not. 1941: 217-222. FRISENDAHL, A. 1927. Über die Entwicklung chas- mogamer ur P Gattung Elatine. Acta Horti Góthob. 3: 99-142. FucHs, A. 1938. Beitráge zur Embryologie der Thymelaeaceae. Oesterr. Bot. Z. 87: 1-41. Gates, R. R 11. Pollen formation in Oenothera gigas. Ann. Bot. (London) 25: 909-940. GAUBA, E. & L. D Pryor. 1958. Seed coat anatomy and taxonomy in Eucalyptus. I. Proc. Linn. Soc New South Omes 83: 20-32. Be 1959. Seed coat anatomy and tax- n Eucalyptus. II. Proc. Linn. Soc. New South Wales 84: 278-291. Ure Seed coat anatomy and tax- nomy in Eucalyptus. TW Proc. Linn. Soc. New So outh Wales 86: 96-1 GEERTS, J. M. 1908. Sei zur Kenntnis der cy- tologischen Entwicklung von Hence lamarck- lana. Ber. gee Bot. Ges. 26: 6 eil Trav . Bot 1 5: 93-206 4. The embry ptg Bae systematic pion of of Trapa E Current Sci. 23: 24- hs — G. & F. F ERRERO. 1891. Intorno allo svi- ppo dell’ovule e del seme della Trapa natans. Malpighia 5: 156-218. 19 mbriologia del Myrtus communis. du Be Giom. Bot. Ital. 37: 609-630. ied 913. Le tégument séminal et les trachées He aires des Th ymelaeacées. Compt. Rend. - Séances Acad. Sci. 156: 398-4 iss 1915, Reliquiae Legio 1, Recherches a Structure anatomique de l'ovule et de | ri des Thymela oe Ann. Jard. Bot. Bui- Tg 29: 1-35. —— 27. Zur Zytologie und Paiol oa Gametophyten von due i . Preuss. Akad. 1927: 33-4 , Zur Zytologie der CAPE uuu dug Hereditas 6: 257-274 a velo ST, H. 4. Variations in embryo sac de- pment. Phytomorphology 14: 186-196. TOBE & RAVEN—EMBRYOLOGY OF MYRTALES 91 HorMEISTER, W. 1847. Untersuchungen des Vor- gangs bei den Befruchtung der Oenothereen. Bot. Zeitung (Berlin) 5: 785—792. Neuere Beobachtungen über Embryo- bildung der Phanerogamen. Jahrb. Wiss. Bot. 1: 2-188. HULBARY, R. L. & A. N. Rao. 1959. Flower devel- iru and gametogenesis in Oenothera laciniata Hill. Proc. Iowa Acad. Sci. 66: 91-97. Fusce te J 19 La formation de l'ovule et la éminogenèse dans le Clidemia hirta. Bull. Soc. Bot. France 105: 230-234 ISHIKAWA, 1918. Studies « on the embryo sac and fertilization i in Oenothera. Ann. Bot. (London) 32: 2 joa D. A. 1928a. The hypostase: its presence and function in the ovule of the Onagraceae. Proc. Natl. Acad. U.S.A. 14: 710-71 19 28b. The hypostase and seed sterility in the iib aeons Madrofio 1: 165-167. 1929. Studies on the morphology of Onagra- ceae. I. The megagametophyte of Hartmannia tet- raptera. Bull. Torrey Bot. Club 56: 285-298. 1930a. Studies on the morphology of the sem agraceae. II. Embryonal manifestations of fasci tion in Clarkia elegans. Bot. Gaz. i ri eibuteville 1930b. Studies on the morphology of the Onagraceae. IV. Stenosiphon linifolium. Bull. Tor- rey Bot. Club 57: 315-326. ———. 1931a. Studies on the morphology of the On- agraceae. III. Taraxia Ae (Nutt.) Small. Ann. Bot. (London) 45: 111-1 1 . Studies on n the rphology of the Onagraceae. M Zauschneria latifolia, typical ud a egular embryology. New York Acad. Sci. 33: 1-26. ——. 1931c. Studies on the morphology of the On- agraceae. VI. Anogra pallida. Amer. J. Bot. 18: 854—864. . 1933. Studies on the morphology of the On- agraceae. VII. Gayophytum ramosissimum. Bull. Torrey Bot. Club 60: 1-8. . 1934. Studies on the morphology of the On- agraceae. VIII. Circaea pacifica. Amer. J. Bot. 21: 508-510. JOSHI, re C. 1939. Embryological evidence for the rela sg of the Lythraceae and related family. Cu : 112-113. a 1935a. Embryological ras n the Lythraceae. 1. Lawsonia inermis. cad. Sci. India 2: 3. c——— Ap eee. 1935b. d studies in the Lythraceae. 2. Lagerstroemia. Proc. Natl. Acad. Sci. India 2: 523- . 1936. Embryological studies in the Lythraceae. Proc. Natl. Acad. Sci. India 3: 377- Juncosa, A. 1982. Embryo and seedling develop- in the Rhizophoraceae. Ph.D. Thesis, Duke a ity. Kann, R. of Jussiaea repens. J. Indian A contribution to the embryology Bot. Soc. 21: 267- KAJALE, L.B. 1939. A contribution to the life history 92 ANNALS OF THE MISSOURI BOTANICAL GARDEN of Bergia ammanioides. J. Indian Bot. Soc. 18: 157-167. KaPiL, R.N. 1962. M Ea NN. Aue of ic ei in relation to taxonomy. Bull. Bot Surv. India 4: 5 . BALA- BAWA. 1968. Embryological stud- ies on the Haloragidaceae. I. Halor ragis colensoi -28. ber die 3534 tiim -Vegetation im malayische- Archipel. Eine morphologische- biologische Studie. Biblioth. Bot. (Stuttgart) 22: 11-18, 31-41. KAUSIK, S. B. the ovule and embryo sac of d erioceph- alus. Proc. Natl. Inst. Sci. India 6: 117-132. Dicen d A. 194 ee A Oenothéracées. 6vclop Dn de l'embryon chez l'Epilobium montanum L. Bull. Soc. Bot. France 95: 173-178. Paot des Oenothéracées. Dé- veloppement de l'embryon chez l’ Epilobium spi- catum Lam. Compt. Rend. Hebd. Séances Acad. Sci. 226: 429-430. LEMESLE, R. . Embryologie des Elatinacées. Dé- veloppement de l'embryon chez l'E/atine alsinas- trum. Compt. Rend. Hebd. Séances Acad. Sci. 189: 1569. sre P. 1948. The m embryo sac. v. (Lancaster) 14: 1-56. : ovi An Introduction to the Embryology of Angiosperms. McGraw-Hill, New York. Vistas Bot. 4: 55-97. he fe Jussiaea repens. Current Sci. 3: 107-108. 1934. Zur Embryologie einiger Ly- raceen. Acta Horti Góthob. 9: 1-21. 1939. Contributions to the embryology of the orders Rosales and Myrtales. Acta Univ. Lu nd. 35: 1-121. Miki, S. 1959. Evolution of Trapa from ancestral Lythrum pre. oo Proc. Imp. Acad. Japan 35: 289-2 Moun aware ]. 19 e Zur Embryobildung einiger Onagraceen. Ber. Deutsch. Bot. Ges. 27: 287-292. MULLER, J. 1975. NAGARAJ, M. ontribution to the flo ral ı mor- phology of Terninaho ehebula Retz. Current Sci. 23: 299-300. 1954b. Megas sporogenesis and m tophyte development in Combretum ovalifolium Roxb. Current Sci. 23: 370. c. Contribution to the floral morphology of Terminalia dos appa L. Current Sci. 1935. ral morphology of Terminalia bie- lerica. ratte Sci. 24: 89-90. & B.H IJALINGAPPA. 1967a. Em logical studies i in Myriophyllum intermedium Proc. Indian bs Sci. 65: 210-220. l -A note on the rahe me mbryo- DC. anri tha A ra ME (W. & A.) Schindler. Cien. Sci. 36: 494—495, 1974. Embryological studies in 1940. Structure and development of 1964. Embryology in relation to taxonomy. | [Vo.. 70 | | drin hirsuta. Bot. Gaz. (Crawfordsville 135: 19-28. Nativ .& S. K. Roy. 1960a. Embryol- y of the genus Psidium. J. Indian Bot. Soc. X. | 35.45. — ——-. 1960b. Embryo sac developmen and polyembryony i in Syzygium cumini. Bot. Ni. | 1960: 273-284. Wero ea B.H. M. 1967. Mopar : : priophyllum intermedium DC. C : 104- O'NEAL, e E "1923, A study of the embryo sac de | < A nether eben Bull. Torrey} Bot. Club 50: 13 46. dasa T-9137 On a yr mds of the pall grain and embryo sac of Daphne, with special erence to the tertile ef Daphne odora. J. Col. | Agric. Imp. Univ. Tokyo 4: 237-264 | PacGNi, P. 1958. Richerche cito- embriologii Oenothera acaulis Cav. Caryologia 11: 181- On the panela t of T e | 1 9. pa. PALSER, B. Fe 1975. e bases of angios rm i | logeny: embryology. Ann. Missouri Bot. p 621-646. PHILIPSON, W. R. 1974. Ovular morphology an er major classification of the dicotyledons. Bot. Linn. Soc. 68: 89-108. ü PIJL, L. VAN DER. 1934. Uber die Polyen aa genia. Recueil Trav. Bot. Néerl. 31 be ral PoruNiNA, N. N. 1957a. A comparative swor embryology and morphology of flowering in Bot species of Eucalyptus. Proc. Acad. Sci. USSR, . Concerning the biolog and embryology of Callistemon os a Proc. Acad. Sci. USSR, Bot. Sci. Sect. 170. (In Russian.) of 57c. On the flowering and emb yo Í Feijoa. Izv. Glavn. Bot. Sada SSSR 29: 167-! 5 (In Russian dem 19582. On the flowering biology wo bryology of Eugenia myrtifolia Sims. Izv. Bot. — SSSR 30: 59-65. (In Russian.) el On the flowering biology and | Olay of E get lanceolatus MA w Žurn. (Moscow & Leningrad) 43: 1169-1178. | tus "x 116: 167- . On the flowering biology and e] I ogy of uci Trudy Glavn. Bot. Sada 210. (In Russian with FAgish di mmary.) im - 1964. A comparative study on the em i ogy, the biology of blooming, and the fruit in Feijoa sellowiana under rtp conditi growth. Pp. 153-158 in V. V. Popov et Problemy sovremennoi embriologii. Izda mr H Moskovskogo Universiteta, Moscow. (In sian. | PRAKASH, N. a. The quiim of i» capitata Reichb. Austral. J. Bot. 17: 97-1 19 me aspects of e pe history morgane ta citrinus (Curt.) Skeels. Austral. J. 17 7-1 Reproductive heus usd udge (M dd mi | J. Bot. 17: 215-227. dl 1969d. A contribution to the life piston ^ 1 1983] TOBE & RAVEN— poo NES xe Sweet (Myrtaceae). ustral 1973. A MH «4 is id Ad of Mela- leu euca. Malayan J. Sci. 2: 81-85. RAGHAVEN, T. S. & V. K. SRINIVASAN. 1940. A con- tribution to the life history of Bergia capensis. J. Indian Bot. Soc. 19: 283-291. Ram, M. 1956. Floral morphology and embryology of Trapa bispinosa, with a discussion on the e Sys- dario of the genus. Phytomorphology Rao, P. S. P. 1963. So ome embryological observations 1 of Guiera senegalensis. Current Sci. 32: 30-31. ^0, V. S. &R. DAHLGREN. 1969. The floral anatomy rs Duns of Oliniaceae. Bot. Not. 122: — A. & V. K. SHARMA. 1970. Elaeagnaceae. s UU siad Symposium on Comparative ngiosperms. Bull. i Sci. Acad. 41: 185-187. eaa es pis -H. 1969. A revision of the genus Camis- v la ise ramnecse). Contr. U.S. Natl. Herb. 37: R me O. 1914, Befruchtung und Embryobildung enothera la amarckiana und einiger verwand- en va Flora 107: 115-150. "tg He uer He eterogamie im wwethlichon: se Nes ius Z. Bot. 13: 609-621. 1953. Embryology of E __ Caren E iri gy ugenia jambos. Embryolo Siar d el 22e 2 of Eugenia bracteata. 60. roe of Eugenia malaccensis. qe daa » — fE E 2 & Cult 2x " MM myrtifolia bryo E aly 1962. emb and em- Pone "E hyllifoliuens A ee Bot. 925. Cont ib i ribution à Phistoire du dé- Bu Oppement des slélastomatacées, Ann. Jard. Bot. 19 » resolution of the Eugenia-Sy- 2. muc controversy (Myrtaceae). Amer. J. Bot. 59: 423-436 Be. 1980, Co By of Psil subfamilia] a Sone pn 29: MARP, K. es, omparative ases and morphol- Xylon and Heteropyxis, and the E Ew choc. of Myrtaceae. Vergleichende Embryol d ios ryologie der Nswavatanan, V. Te cbrüder Porniresger Bene. A contribut Pe hell Ludwi igia octovalvis Sube sesion Gabaa) Pe - Proc. 54th Indian Sci. Con 70. III. Abstr., pp. 328-329. e O Onagraceae. In Proceedings of the «B n Comparative Embryology of An- 225. ull. Indian Natl. Sci. Acad. 41: 220- SHARMA, V fra TO 66. Brae ap i of Elaeagnus con- . HERR. 1971. : ive develop- EMBRYOLOGY OF MYRTALES 93 ment, megagametogenesis and early embryogen in Ammannia coccinea Rothb. J. Elisha Mitchell : 192-1 20. Embryogénie a t éveloppement de l'embryon z l'Oenothera lepras i Compt. Rend. Hebd. Brea Acad. Sci. 170: 946-948. 1925. Embryogénie w Lythracées. Dével- oppement de embryon chez le Lythrum salicaria. Compt. Rend. Hebd. Séances Fines Sci. 180: 1417- 1418 1935. Embryogénie des Oenothéracées. Les principaux termes du développement de l'em- bryon chez le Ludwigia palustris Elliott. Compt. Rend. Hebd. Séances Acad. Sci. 200: 1626-1628. 1940a. Embryogénie des Myrtacées: dével- L Compt. Rend. Hebd. Séances Acad. Sci. 210: 550. Ob. Embryogénie des Haloragacées. veloppement de l'embryon chez le Mop end. Hebd 5- Embryogénie ‘des Thymélaeacées. Dé- veloppement de l'embryon chez le Daphne me- pt. Rend. Hebd. Séances Acad. Sci. 214: 569-571. . 1946. Embryogénie des Oenothéracées. Dé- veloppement de l'embryon chez le Circaea E tiana L. Compt. Rend. Hebd. Séances Acad. 223: 700-702. Stace, C. A. 1965. The significance of ME an id dermis in the taxonomy of the Combre A general view z tribal, generic and a ze acters. J. Linn. t. 59: 229-252. Bossen B. L. 1909. cnim embryo sac and emb of certain Penaeaceae. Ann. Bot. (London) 23: 363- 8. SToLT, K. A. H. 1928. Die Embryologie von Myr- iophyllum alterniflorum DC. Svensk Bot. Tidskr. 22: 305-319 STRASBURGER, E. 1884. Die ETEY bei Daphne. Ber. Deutsch. Bot. Ges. 2: 112-1 aphn Sum wav AM. K. 1942. Gametogenesis aay em- CW Terie I lodian Bot. Soc. 21: 69-85. A contribution to the life — of 115- Sonerila wallichii. Proc. Indian Acad. Sci. 1 : 1946. Further contribution to the embryol- ogy of the genus Osbeckia. Half-Yearly J. Mysore Mn V 1-11. An embryological study of Melastoma 27: 11- melabahricum J. Indian Bot. mbryology of Oxyspora paniculata. 2 e | 205- & H. C. Govınpu. 1948. The development of the female ong ge in Oenothera odorata an Bot. Jacq. J. Indian Soc. 27: 25-30. TAcKHOLM, G. 1914. LE e vis der Embryosack- aie von Lopezia coronata Andr. hak Bot. Tidskr. 8: 223-234. a ae über die Samenent- n. Svensk Bot. Tidskr. wicklung einiger Onagrace 9: 294-361 TAKHTAJAN, A. L. 1980. Outline of the argon, of flowering plants Mew iophyta). Bot. (Lancaster) 46: 225-35 94 TiscHLER, G. 1917. Uber die Entwicklung und phy- logenetische Bedeutung des Embryosacks von Lythrum salicaria. Ber. Deutsch. Bot. Ges. 35: 233- 246. Tiwary, N. K. & V. W. Rao. 1934. A preliminary note on the embryo sac development of Calliste- mon linearis. Current Sci. 2: 3 TRELA-SAWICKA, Z. 1978. Embryological studies in Trapa natans L. Acta Biol. Cracov., Ser. Bot. 21: 01-108 TREUB, M. 1884. Notes on l'embryon, le sac em- bryonnaire et l'ovule. 5. L'embryon du Barring- tonia vriesei. Ann. Jard. Bot. Buitenzorg 4: 101- 106. VENKATESWARLU, J. 1936a. A preliminary note on the embryology of Duabanga sonneratioides. Cur- rent Sci. 4: 742-743. 1936b. Some observations on the ovule and embryo sac of Sonneratia apetala. Current Sci. 5: 20 1937a. Structure and development ofthe em- bryo sac of Pemphis acidula. J. Indian Bot. Soc 16: 259- 937b. A contribution to the embryology of the Sonneratiaceae. Proc. Indian Acad. Sci. 5: 206- 262. 1945. Embryological studies in the Thyme- laeaceae. 1. Thymelaea arvensis. J. Indian Bot. Soc: 24: 6. 1946. A case of er ayes og in Daphne cannabina. Current Sci. 15: 169. 947a. Apian a sR in the Thyme laeaceae. 2. Daphne cannabina d Wikstroemia canescens. J. Indian Bot. Soc. 26: 13-3 47b. A contribution to the pie of Pisonia aculeata. J. Indian Bot. Soc. 26: 182-194. ANNALS OF THE MISSOURI BOTANICAL GARDEN phology 2: 231-240 1952a. Embryological vit in Lecythida- ceae. I. J. Indian Bot. Soc 6. 19 Combretaceae. 1. Poivrea coccinea. Phytomor- P. S. P. Rao. in some Combretaceae. Bot. Not. 125: 1972. Embryological studies 161- 179, 509. Nouvelles recherches sur le dével- oppement du sac embryonnaire des Phanero- games angiospermes. Ann. Sci. Nat. Bot., Sér. 6. 8: 261-390. VIJAYARAGHAVAN, M. R. 1970. Ron In roceedings of the Symposium on Comparative Embryology of Fonda pm Bull. Indian Natl. Sci. dem 41: 45-52. VLIET, G. VAN. 1975. Wood anatomy of E insti ai sensu lato. J. Microscopy 104 65-82 & P: B 75. Comparative anatomy of the Crypteroniaceae sensu lato. Blumea 22: 17 195. Warp, H.M. 1880. A contribution to ven pee of the embryo sac in Angiosperms. J. Linn. Soc. Bot. 17: 519-54 WARMING, E. 1878. De l'ovule. Ann. Sci. Nat. Bot. Sér. 6, 5: 177—066. WINKLER, H. 1904. Die Parthenogenesis bei Wiks- troemia indica. Ber. Deutsch. Bot. Ges. 22: 513 580. ZIEGLER, A. 1925. _ Beiträge zur Kenntnis des i lastomataceen. Bot. Arch. 9: 398-467. 52b. Contributions i bis embryology of — H [VoL. 70 I | — ieee REVISION OF THE ARBORESCENT BAUHINIAS (FABACEAE: CAESALPINIOIDEAE: CERCIDEAE) NATIVE TO MIDDLE AMERICA! RICHARD P. WUNDERLIN? ABSTRACT Twenty-seven arborescent species of Bauhinia native to Middle America, of which 20 species are t I t 1 3 «1 1 112.9 MW. | 4 i th 14 endemic, are recognized in a systematic treatmen of ee of a few ancestral species from South America, followed by rapid speciation in the lii nov demic to S i new habitat. Ba Luis Potosí, Mexico is described and illustrated s pe chapulhuacania nom. nov. (= B. dipetala var. macrophylla Wunderlin) is proposed. ys, descriptions, distributional data, and taxonomic index are provided. _ The caesalpinioid tribe Cercideae is divided Into two subtribes, the Cercidinae and Bauhi- ninae (Wunderlin et al., 1981). The Cercidinae consist of three small distinctive genera; Cercis a the large, diverse, pantropical pl n inia and the monotypic Madagascan " terea. The approximately 300 species "issu d os middle elevations throughout bees t € Species are about evenly divided ies. ova hemispheres; however, they are route Ae a pe Duis teri a sc Pon Aane Und. calyx spathaceous; 3 tile samene 3 or 19 em A in SHECMHON Am ican h blunt, spinelike infratectal processes. ipid Suballiance— Fertile stamens 1 B. andrieuxii Hemsl. dub e their length (rarely 1e lobed less n half), ub between lobes usually obtuse to rounded at B. deserti 4b. Corolla de of 5, subequ pota. 9a. Petals ree pink, vni, EE i ves bifoliolate, 1-2 cm per leaflets 2-(3)-nerve d 23. B. ramosissimé etala > m > g. 5 = wn g 5 10b. esa bilobate or entire (rarely bifoliolate in some forms of B. macranthera), 4 cm long or longer, 7—-9-nerved. lla. Inflorescence less than 10-flowered; petals pink, blade subequalling the pilose claw; lower ie of young leaves glabrous or pubescent, but not wit yel- lowish-brown hai 15. B. macrantheta 11b. Aip aria 20- His red or more; petals brick red, blade about twice as "d as glabrate claw; dr surface of young leaves with ye d hairs rubeleruziant Í. P I B. divaricata but HER branches and undersurface of young leaves never with yellowish- brown hairs), or white with à í. n eo —— 12a. Leaves bifoliolate 14. B. lunarioidë 12b. Leaves bilobate or entire. E lig eaves ONS a 1o H jennings! | 13b. Leaves bilobat 14a. Petals white with conspicuous red median stripe .. pri B. pring 14b. Petals gein B. divaricata), ier a m ~ median stripe. 15a. Petals 16a. rs buds alw 16b. Eaves ovate to anne. bilobate for 3⁄4 or more deir pint flowers buds reflexed when young, erect at anthesis - "19. B. pes cu l wn T Petals glabrous or glabrate. 17a. Calyx bright red, free tips obsolete 17b. ao gere or rarely with dull red tinge, free c m lon : ea f dra 3b. Fertile stamens a or 10. 18a. Fertile stamens 3. J Petals fs (rarely white) sessile; leaves miris Ane —6) cm long ..... 6. B. 19b. Petals white, short-clawed; leaves ovate, 12-22 c = A É pansamala™ e RA ee wenn for about half € e: olia flowe 6.5.5 e 10. B. erythroci | 1983] 18b. Fertile stamens 0 WUNDERLIN—ARBORESCENT BAUHINIAS 101 Leaves entire Somes bilobate on same plant). 2la. Sho mens connate for about half their TN leaves with inner pair of e S tant or closer to outer aciei nerves than to midri uidis 20b. qr, always bilobate. 22a. Infl .. 20. B. petiolata orescence, lower surface of leaves, and fruits brown tomentose .. 25. dern 22b. Inflorescence, lower surface of leaves, and fruits other than brown tomen Ja. Leaves less than 2.5 cm long, apex of lobes rounded .. m or longer, apex of lobes eens to caudate. 24a. Petals light sulfur-yellow; fruits 5-7 cm 23b. Leaves 5 cm long o B uni usi cookii 24b. Petals white or pinkish with roseate center, fruits over 10 cm n iong: 25a. Cal DE. linear to filiform 26a. Flower buds 4-5 cm long; fruits ca. 1 cm wide 27. B. “ungulata 26b. Flower buds 7-10 cm long; fruits ca. 2 cm wi 6. B. multinervia . Calyx spathaceous; petals with Tn ae 27a. Staminal sheath distinctly cy 27b. Staminal sheath not ligulat l. Bauhinia aculeata L. , Sp. Pl. 374. 1753; non ell., 1825. TYPE: Celoftdi ia or Venezuela: : Eu UM Paeys, BM; photograph of olotype Bauhinia emarginata Mill., Gard. Dict. ed. 8. 1768: non Jack, 1822: nec Wall., 1831; nec Roxb. ex G. ace. B YPE: Colombia. Bolivar: Cartagena, lotype n s.n. (holotype, BM; photograph of ho- B hia rotundata Mill., Gard. Dict ed. 8. 1768. rre: l Bolívar: Cart ena, Houstoun s.n. (ho- bilis BM; photograph of holotype, U “i ngula Jacq., A Bot. 22. 1801; non Willd IR ud., 1840. TYPE: Venezuela: Jacquin s.n. Olotype, W, not s pus ^ inia glan glandulosa H.B.K., Nov. Gen. Sp. Pl. 6: 314. >. 344 rM slandilosa (H.B.K.) DC., Prodr. Humboldt E a : Venezuela. Sucre: Cumaná, MO: Dije ^n iti otype, P, not seen; microfiche, , NY, U8). seen; photograph of iso- WRG Vog., Linnaea 13: 302. 1839. : Venezuela. Fe Federal District: Near Caracas, er s.n. (ho D fi nt of otype not seen; fragme Pe. F Aou ii of holotype, F, MO, Bauhinia : ma ungula Willd. ex Steud., Nom. Bot. ed. 2. 1: Pi an & Hammon 1024 (holotype, GH, not seen: NY, US). Ba “in pends Pittier, Trab. Mus. Com. Venez. 1: ches, on Pe: Venezuela. Miranda: Los Mari- (holoty Ev road to Santa Lucía, Pittier 11783 phot EN, not seen; isotypes, NY, P, US; Bauhinia Raph of NY isoty ype, NY). E belio Pittier, Trab. Mus. Com. Venez. San Rama, 4. cr: Venezuela. Cojedes: Between ael de Onoto and Agua Blanca, Pittier 24. B. Piste tale 11726 (holotype, VEN, not seen; isotypes, NY, pew sohultzei Harms, Repert. Spec. Nov. Regni Mods 24: 210. 1928. TYPE: Colombia. Magdalena: evada de Santa Marta, upper Río Frío, Schutz 428 (holotype, B, not seen; fragment of otype, F; photograph of holotype, F, MO, NY, TYPE: El Sal r. Sonsonate: Vicinity of aoe Sandey p (holotype, US; iso type, NY). Shrubs or small trees up to 8(-10) m tall; branches tomentose when young, soon glabrate, with 1-2 intrastipular spines up to | cm Bauhinia albiflora Britt. Mie: N. Amer. FL 23: 203. 1930. armed 4 long. Leaves c eous ly ipu emarginate or bilobate for '/^, eel to !^ their length, lobes parallel or slightly vergent, 3-6 cm long, 2.5-5.0 cm wide, base cor- date to rounded, apex of lobes rounded to acute, margins smooth, tomentose, strigose or glabrate on both surfaces, 7—9-nerved; petiole 0.5-3.0 cm long, slightly canaliculate, tomentose, strigose, or glabrate; stipules linear, ca. m long, ca- ducous; adpetiolar intrastipular excrescence forming a frequently reflexed spine, others mi- nute. Inflorescences short racemose, less com- monly geminate or 1-flowered, subterminal, ax- illary, racemes 3-10-flowered, rachis tomentose, strigose, or glabrate, buds linear-lanceolate 3-4 cm long, tomentose, strigose, or glabrate, free tips up to 1.5 mm long; bract linear, 2-5 mm long; bracteoles similar to bract; pedicels 0.5-1. 0 cm long, tomentose or strigose; hypanthium tubular, 0.3-1.0 cm long; calyx spathaceous at anthesis; petals 5, subequal, white, 3-8(-10) cm long, blade 102 narrowly elliptic to obovate, 2-10 times as long as claw, (0.5—)1.0-2.5 cm wide, glabrous, claw sparsely pilose to glabrous; fertile stamens 10, er, filaments slender, arcuate, sparsely pilose at m linear, 4-7 mm long; pollen spheroid, inaper- turate, sexine reticulate with clavate supratectal processes; gynoecium nearly equalling stamens, arcuate, ovary and gynophore densely pilose, style glabrate, gynophore nearly equalling to '2 as long as style, stigma capitate. Fruit linear to narrowly elliptic, apiculate with persistent style, 7-15 cm long, 1.5-2.5 cm wide, light brown, pubescent at maturity, sometimes becoming glabrate in age, gynophore 1-2 cm long, puberulent; seeds ob- long to suborbicular, 6-11 mm long, 5-8 mm wide surface dull, dark brown, funicular aril-lobe scars equal, 2.0~2.5 mm long. Bauhinia aculeata consists of two subspecies, the typical form, to which the Middle American material is referred, and subsp. 2 liflora (Juss.) Wunderlin, which occurs in Bolivia, Ecuador, and Peru. The latter is distinguished by larger leaves, fruits, and by flowers with a conspicu- ously elongated hypanthium. In Middle America B. aculeata is native to El Salvador and Panama in Central America and Barbados and Grenada in the Lesser Antilles. It is also known from Anti- gua as an escape from cultivation. Outside Mid- dle America, this species is native to Colombia and Venezuela where it is especially common. The species is found at elevations ranging from near sea level up to 1,700 m in open tropical deciduous forest or in disturbed sites. It flowers throughout the year but most abundantly from March to June. It is sometimes cultivated within and outside its natural range as an ornamental. The taxonomy of the B. aculeata complex is still only partly understood; there are a number of names applied to entities in South America that may or may not be distinct. Bauhinia acu- leata is closely related to about 30 additional South American and two Old World species. Britton and Rose described Bauhinia albiflora as having five fertile stamens, but an examination of the type material revealed it to have ten. Bauhinia aculeata is highly polymorphic as is evidenced by the extensive synonymy. However, it is easily distinguished from all other Middle American species because it is the only spiny species with large, white, showy flowers and ten ANNALS OF THE MISSOURI BOTANICAL GARDEN E [Vor. 70 — ——— fertile stamens. The only other spiny species in Middle America, B. pauletia, has inconspicuous green flowers with filiform petals and only five fertile stamens. | Bauhinia aculeata is believed to be bat-polli- ) nated and possibly also moth-pollinated, as ap- | pear to be many of its allies in South America. ! In Venezuela this species is known as “Pata de Vaca," “Dibrito,” *Quimbrapoto," “Guara- pa," or ““Urapa.” | Representative specimens: EL SALVADOR. SAN ) MIGUEL: Vicinity of San Antonio N of Ciudad Barrios, | Allen & Armour 6867 (F). SONSONATE: Vicinity of Aca jutla, Standley 21983 (NY, US). PANAMA. COcLE: la | Pintada, León 12 (MO). PANAMA: E of Pacora, Wood- | son et al. 735 (MO, NY, US). ANTIGUA. Upper Cedar | Hill, Box 1193 (F, UC, US). BARBADos. St. Thomas, Hopewell Gully, Bovell 208 (UCWI). GRENADA. Sl George's, Broadway s.n. (F, NY, US). 2. Bauhinia andreiuxii Hemsl., Diag. Pl. Nov- | 48. 1880. T photograph of G isotype, F, MO, US). ail Shrubs or small trees; branches tuberculate, glabrate. Leave gJ. o c & [md e = ai > o S — eo 2 = —— e 3 ye 2 bas g un v & Sat e cordate to rounded, apex of lobes rounded, upper surface glabrous, lower surface tomentellous, r 9-nerved; petiole 5-15 mm long, slightly cana ulate, tomentellous; stipules ovate, ca. 1 mm lon | caducous; adpetiolar intrastipular excrescente | fused and forming a subulate projection C? mm long, others minute. Inflorescences race | mose, congested, terminal, 3-8-flowered. rachis strigose or tomentellous, buds linear-lanceolat® 12-18 mm long, strigose or tomentellous, x tips ca. 1.5 mm long; bract lanceolate, Cà- l. mm long; bracteoles similar but slightly smalleí than bract; pedicels ca. 5 mm long, strigose - tomentellous; hypanthium cyathiform, 2-3 mn long; calyx spathaceous at anthesis; petals 5 sub equal, white, broadly elliptic, ca. 1 cm long. b -6 mm wide, glabrous, claw ca. | mm » j glabrous; fertile stamens 10, 5 nearly equallif* | petals, alternate 5 ca. 1 the length of petals: yi equally short-connate at base for 1-3 mm. ments slender, glabrous, anthers linear oblong ca. 2 mm long on longer stamens, ca. !- m ‘ long on shorter stamens; pollen spheroid, inap?" turate, sexine reticulate with blunt, spinelike. ™ — ee — — — gjin 1983] fratectal processes; gynoecium nearly equalling petals, tomentose, gynophore nearly equalling style in length, stigma capitate. Fruit linear, apic- ulate with persistent style, 8-10 cm long, ca. 1.5 cm wide, dark brown, tomentose; seeds not seen. Bauhinia andrieuxii is a wide-ranging but in- frequently collected species occurring in western Mexico from Sinaloa to Oaxaca. It occurs in de- ciduous thorn scrub at elevations from near sea level up to around 1,800 m. This species is distinguished from other species of Bauhinia with ten fertile stamens in the area by its small leaves, which are suborbicular or broader than long. It is most closely related to B. seleriana. The pollinators of this species are probably butterflies or possibly bees. Lee fšamined: MEXICO. MORELOs: Cañón de ; on Cuernavaca-Yautepec road, Fryxell U, MO); Mitla, Smith 107 WS iig Cerro Llano Redondo, W of Caima- drieux WIG). 8 (NY). Without precise locality: An- 3 H ` * . ire beguinotii Cuf., Arch. Bot. Sist. 9: - 1933. TYPE: Costa Rica. Limón: Be- ines Waldeck and Río Barbilla, C ufondon- f 64 (holotype, W, not seen; photograph of holotype, F, MO, US). Sh d oic or small trees up to 6 m tall; branches omentose to tomentellous when young, long, slightly canaliculate, glabrate; stipules sistent; ien i ihe 1.0-1.5 mm long, per- arged and E ar intrastipular excrescence en- Tming a subulate projection up to Others minute. Inflorescences race- nal, or subterminal and axillary, ra- tomentellons t" "^ rachis brown tomentose to long. br a A uds elliptic-lanceolate, 1.5-2.0 cm ovate, ca. | Ordi free tips minute; bract pedicels "d m long; bracteoles similar to bract; Panthiun, E long, brown tomentose; hy- thaceous fia panulate, 5~6 mm long; calyx spa- 2.5.3.9 ^n E petals 5, subequal, white, Wide, he Ong, blade oblanceolate, 8-13 mm TOUS, claw 3-5 mm long, glabrous; fer- cemes 5.10 WUNDERLIN — ARBORESCENT BAUHINIAS 103 tile stamens 10, the 5 longer ?4 the length to nearly equalling the petals, alternate 5 stamens shorter, short-connate for ca. !^ the length of shorter stamens, fil ts sparsely pil 1 base, anthers triangular at anthesis, ca. 3 m long on longer stamens, ca. 2 mm long on shorter stamens, sparsely pilose to glabrate; pollen sphe- roid, inaperturate, sexine reticulate with blunt, spinelike processes; gynoecium nearly equalling Ww AUOY U aiu ear, apiculate with persistent style, 12-15 cm long, 2 cm wide, brown, minutely strigose to glabrate, gynophore 1.0-1.5 cm long, glabrate; seeds not seen. Bauhinia beguinotii is a rare species known in Middle America only from Costa Rica and Pan- ama, where it occurs in tropical evergreen forest at elevations from near sea level up to about 400 m. Bauhinia beguinotii is a member of the Petio- lata alliance, which is essentially a South Amer- ican group and includes B. petiolata and B. cook- ii. It is less closely related to B. pansamalana, also of this alliance, and differs in having ten rather than three fertile stamens. Bauhinia be- guinotii is distinguished from B. petiolata, its most closely related species in Central America, by shorter stamens connate for about half their length rather than being free nearly to the base and by leaves with the inner pair of nerves closer to the midrib than to the outer adjacent nerves rather than equidistant or closer to the outer nerves. Bauhinia beguinotii is known outside Middle America only from Gorgona Island, Narifio Province, Colombia. The Colombian material is recognized as var. gorgonae (Killip ex Cowan) Wunderlin and is distinguished by leaves deeply bilobate or bifoliolate rather than entire or only slightly bilobate. Since the degree of lobing in leaves of Bauhinia is often quite variable, further collections may prove var. gorgonae not to be distinct from the typical form found in Panama and Costa Rica. pecimens examined: Costa RICA. LIMON: Drainage of the Rio Estrilla Valley, Shank & Molina 4466 (F, US). PANAMA. PANAMA: El Llano-Cart-Tupile road, 16 km N of Pan-American I (MO); 14.8 km N of the Pan-American 5 Highway, Nee et al. 8881 (MO); 10-1 American Highway, Mori & Kallunki 2847 (MO). 104 4. Bauhinia chapulhuacania Wunderlin, nom. nov. Bauhinia dipetala var. macrophylla paso a st. Nat. 13: 104. 1968 TYPE: Mex idalgo: Near Chapulhuacan, Moore pa fiacheyeiee GH). Trees up to 6 m tall; branches strigose when young, soon glabrate. Leaves subcoriaceous, broadly ovate, bilobate about 4—'4 their length, lobes slightly divergent, (6—)10-15(—30) cm long, ovate- lance olate, 2. 0-2. 5 mm long, Seen adpetiolar int forming a subulate projection ca. 1.5 mm m long, others minute. Inflorescence racemose, terminal, or subterminal and axillary, 20—30-flowered, rachis strigose, buds linear, 2-3 cm long, strigose, free tips up to 2 mm long; bract lanceolate, ca. 1.5 mm long; bracteoles similar to bract but smaller; pedicels 1.0—1.5 cm long, strigose; hy- panthium elongate-campanulate, 6-10 mm long; calyx spathaceous at anthesis; petals 2, subequal, linear, tapering to base, 1.0-2.0 cm long, 0.5-1.5 mm wide, brown pilose on inner surface; fertile stamen 1, 2.5-4.0 cm long, filament slender, ar- cuate, glabrate, short-connate at base with stami- nodes, anther linear-oblong, green, 8-12 mm long, glabrate; staminodes 9, subequal, 4.5-13.0 mm long, connate for ca. 3⁄4 their length, brown pilose on inner surface, abortive anthers absent; pollen spheroid, 3-colporoidate, sexine striato-reticu- late with elongate supratectal processes; gynoecium ca. equalling fertile stamen, arcuate, cm wide, dark brown, sparsely brown strigose, gynophore 2.5-4.0 cm long; seeds broadly oblong, 10-12 mm long, 8- 10 mm wide, surface dull, dark brown, funicular aril-lobe scars subequal, ca. 3 mm long Bauhinia chapulhuacania is a rare species en- demic to southeastern San Luis Potosí and north- ernmost Hidalgo, Mexico where it occurs in open, mixed, mesic forest at 700 to 900 m elevation. This species is most closely related to B. di- petala and B. deserti, both of which are charac- terized by two inconspicuous green petals. Bau- ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL 70 » hinia chapulhuacania occurs over 100 miles to the north of the range of these other two related species and differs in larger leaves and fruits. Because of the close resemblance of its flower to that of B. PM B. chapulhuacania was orig- inally a variety of that species. How- ever, M enci of additional collections in- dicates that it is best treated as a distinct bul | closely related species and it is here elevated t0 - specific rank. Since the name B. macrophyllais preoccupied (B. macrophylla Poir., Encyc. Supp. 1: 600. 1810 [= B. guianensis Aubl.]), a new name must be chosen. Bauhinia chapulhuacania is undoubtedly bat pollinated as are its close relatives B. deserti and dipetala. — ee ~~ — Ina pecimens examined: MEXICO. HIDALGO: SofTa zunchale, Barkley & Webster 7264 (F, MEXU, TEX) near Hidalgo-San Luis Potosi border, Km 343- 344 on highway below Pao teatro Moore & Wood 3652 | (MEXUL UC: US) 7 of Chapulhuacan along | get 85, Wunderlin 5050 (MO, USF). SAN LUIS POT : 31 mi. S of Ciudad Valles along pu EC MGR et al. 857 (US); Puerto Verde, 64 i Ci sc Valles on highway to Río Verde, Fryxell fl Anderson 3605 (MO, USF); near the waterfall a6 | Salto, pss 3920 (F, NY, TEX, UC, US). \ — 5. Bauhinia cookii Rose in Britt. & Rose, N mer. Fl. 23: 205. 1930. TYPE: Costa Ri Alajuela: Cuesta de la Vieja, road 1o San Carlos, Cook & Doyle 41 (holotype, US fragment of holotype, NY; isotype, US). Casparia calderonii Rose in Britt. & R se, N. e | “aon ‘ 0. Bauhinia calderonii (Roe | Lundell, Phytologia 1: 214. 1937. E >= = &E- Sze = dell. Type: El Salvador. Sonso fat Calderón 2226 (holotype, US; medios US; ment of isotype, NY). regis um Lundell, Phytologia 1: 213. 1937. p Belize. El Cayo: Valentin, Lundell 6298 to type, ‘MICH: isotypes, NY, US). Trees up to 15(-40) m tall; branches strig?“ -— ves or puberulent when young, soon glabra dl submembranaceous or chartaceous, broat" | divaricate, 5-10(-15) cm long, 7-10(-12) ^ wide, base cordate to truncate, apex of lobes acu to obtuse, margins smooth, upper surface gr brous, lower surface strigose, at least on nerve il 7-9-nerved; petioles 2-3(-5) cm long. sligh 15 canaliculate, strigose; stipules lanceolate, © : mm long, caducous; adpetiolar WE e í crescence enlarged and forming a subulate l — 1983] jection up to 1.5 mm long, others minute. Inflo- rescences paniculate, congested into fasciculate clusters, sometimes produced before the leaves and appearing cauliflorus or ramiflorus, 15—60- flowered, rachis and rachilla strigose, buds lin- car-elliptic, 8-10 mm long, strigose, free tips minute; bract ovate-lanceolate, ca. 1 mm long; bracteoles similar but slightly smaller than bract; pedicels 5-7 mm long, strigose; hypanthium cy- athiform, 2-3 mm long; calyx spathaceous at an- thesis; petals 5, subequal, light sulfur-yellow, el- liptic to oblanceolate, sessile or short clawed, 10- 15 mm long, 5-7 mm wide, glabrate; fertile sta- a 5 nearly equalling petals, alternate ones z à sa length, glabrate, short-connate at base Mi is lacinate rim projecting above point ae 10n, filaments slender, anthers linear-ob- ad ose of longer stamens ca. 2 mm long, of ES oo ca. 1.5 mm long; pollen sphe- IL e. sexine reticulate with blunt, del. cur epal processes; gynoecium ca. 1⁄2 nen gth of petals and ca. equalling shorter sta- an Emoe ca. equalling style, stigma small, asl E us linear, apiculate with persistent adi m long, ca. 1.5 cm wide, light brown, om monere 5-10 mm long; seeds ob- deck Rt Tong. 5-8 mm wide, surface dull, ; , funicular aril-lobe scars equal, form- Ing an E a e arcuate white line for ca. % the length of M. one " widespread species from Si- Collections sali Mexico southeast to Costa Rica. ut it is to Cui not been seen from Nicaragua, deciduous or E there. It occurs in tropical from near sea E foresta atetevation commonly fr €vel up to about 2,000 m but more flowers Sit. 300 to 1,000 m. It apparently from March to id ut the year but most commonly duced before th uly. Flowers are sometimes pro- veloped, e leaves appear or are much de- Thi E, oes E is part of the Petiolata alliance but the Other tup a be closely related to any of ica. It is E of this alliance in Central Amer- use it ud recognized in Middle America Sufur.yeos Ihe only tree Bauhinia with light h Owers in fasciculate clusters along the brane es. In Guat m emala and El i 5 "Pata qe ES. Salvador the local name *prese R ; tat; lY of La osi specimens: MEXICO. CHIAPAS: Vicin- H4 mi, WNW c. Miranda 6485 (US). coLima: Ca. Santiago on road to Cihuatlán, Jalisco, WUNDERLIN —ARBORESCENT BAUHINIAS 105 McVaugh 20776 (MICH, US). GUERRERO: Cañón de Benito Juárez, Sebástopol, Textepec, Dioscorea Bri- gade 2741 (MEXU). siNALOA: Mazatlán, Reko 4520 MEX i TALHULEU: Río Talculán, 5 mi. W of Retalhuleu, Standley 87377 (F). SUCHITEPEQUEZ: Along Rio Madre Vieja above Patulul, Standley 62207 (F). HONDURAS. COMAYAGUA: Pito Solo, Lake Yojoa, Edwards P-404 sta RICA. ALAJUELA: Plains of the Rio San Tarcoles, Póveda 1097 (MO). 6. Bauhinia coulteri Macbr., Contr. Gray Herb. 22. 1919. Based on Bauhinia platypetala Benth. ex Hemsl. Shrubs to 0.8 m or small tree up to 6 m tall; branches tomentose when young, soon glabrate. Leaves subcoriaceous, suborbicular, bilobate up to !^ their length, (2-)3-5(-6) cm long, (2-)3-5 (-6) cm wide, base cordate, apex of lobes round- ed, margins smooth to strongly crisped, upper surface glabrous, lower surface tomentose to gla- brate, 5-9-nerved; petiole 1-2 cm long, slightly canaliculate, tomentose; stipules lanceolate, ca. 1 mm long, caducous; intrastipular excrescences minute. Inflorescences racemose, terminal, or al and axillary, 2-10-flowered, rachis bracteoles cels slender, 1-2 cm long, tomentose; hypan- thium cyathiform, 1-2 mm long; calyx spatha- ceous at anthesis; petals 5, subequal, pink or rarely white with darker venation, narrowly elliptic, sessile, (1-)2-3(-4) cm long, 0.5-1.5 cm wide, or strigulose near base on inner surface; rt-connate with stami- . V» as long as petals, glabrous, lateral stamens 5-7 mm long, stout, glabrous, anthers as i staminodes 7, subequal, 4-6 mm long, connate for ca. 2 their length, glabrous, abortive anthers present; pollen spheroid, 3-colporoidate, sexine reticulate; gynoecium ca. equalling petals, ar- cuate, strigose, gynophore subequalling style, stigma not much differentiated from style. Fruit linear, apiculate with persistent style, (5-)6-8(-15) 106 cm long, 1.0-1.5 cm wide, light brown, tomen- tellous to glabrate, gynophore ca. 1-2 cm long, tomentellous; seeds elliptic, 7-8 mm long, 4-5 mm wide, dark brown, funicular aril-lobe scars subequal, 2-3 mm long. Two varieties are recognized. KEY TO VARIETIES OF BAUHINIA COULTERI la. Shrubs up to 0.8 m tall; leaves subglabrous Or sparsely p l bel i I e c 6a. var. coulteri 1b. Small trees 5-6 m tall; leaves densely to- 6a. Bauhinia coulteri Macbr. var. coulteri. Bauhinia platypetala Benth. ex Hemsl., Diag. Pl. Nov. 49. 1880; non Burch. ex Benth., 1870. TYPE: Mex- ico. Hidalgo: Zimapán, Coulter 531 (holotype, K; Casparia coulteri (Macbr.) Britt. & Rose, N. Amer. Fl. : , 1930. Bauhinia coulteri forma albiflora Wunderlin, Rhodora 70: 286. 1968. TYPE: Mexico. Tamaulipas: 19 km SE of Miquihuana on road to Palmillas, Stanford et al. 851 (holotype, MO; isotypes, GH, NY, UC). Bauhinia coulteri is a central Mexican species of open oak-pine, juniper, or deciduous forests and occurs at elevations of 600 to 2,400 m. It flowers from April to September. This species is easily recognized by its subor- bicular leaves and pink (rarely white), sessile pet- als, and three fertile stamens. No other species of Bauhinia in Mexico has three fertile anthers. It is readily distinguished from var. arborescens by the characters given in the above key. Representative specimens: MEXICO. HIDALGO: 16 km S of Jacala, Frye & Frye 2251 (GH, MO, NY, UC, US). QUERÉTARO: 12 mi. E of Landa de Matamoros, Crutch- field & Johnston 6136a (TEX). sAN Luis POTOSI: Minas de San Rafael, Purpus 5187 (F, GH, MO, NY, P, UC, US). TAMAULIPAs: 18 mi. by road SE of Bustamante toward La Presita and Tula, Johnston et al. 11154 (TEX). 6b. Bauhinia coulteri var. arborescens Wunder- lin, Phytologia 15: 120. 1967. TYPE: Mexico. Querétaro: Ca. 80km NE of! Querétaro, above Pilón on road to Pinal de Arnoles, Mc Vaugh 10351 (holotype, MO; isotypes, DUKE, LL, MICH, TEX, US) Bauhinia coulteri var. arborescens is known only from the type collection. It occurs in the lower limits of the pinyon pine-juniper belt at about 2,700 meters elevation. It was collected in flower in April. ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo.. 70 This variety is easily distinguished from the typical by the characters in the above key. The local name is “Manito de Cabra.” 7. Bauhinia deserti (Britt. & Rose) Lundell, Publ. Carnegie Inst. Wash. no. 478. 211. 1937. J Casparia deserti Britt. & Rose, N. Amer. FL 23: 216. 1930. Bauhinia dipetala var. deseri (Britt. & Rose) Wunderlin, Phytologia 15: 53. 1967. TYPE: Mexico. Puebla: Near Te huacan, Rose & Rose 11375 (holotype, US; | ) isotype, NY). Shrubs to 5 m tall; branches strigose or pu- berulent when young, soon glabrate. Leaves sub- coriaceous, broadly ovate, bilobate 12 to ^5 their length, lobes parallel or slightly divergent, (3-)5- 8(-12) cm long, (3-)5-8(-12) cm wide, base cor date, apex of lobes rounded or obtuse, margins slightly crisped, upper surface glabrous, lowe! intrastipular excrescence enlarged and forming 4 subulate projection up to 1.5 mm long, others minute. Inflorescences racemose, terminal, s subterminal and axillary, 20—30-flowered, rachis strigose, buds linear, 2—3 cm long, strigose, ye tips up to 2 mm long; bract lanceolate, Ca l. mm long; bracteoles similar to bract, but smaller pedicels 0.6-15 mm long, strigose; hypanthium elongate-campanulate, 6-8 mm long; calyx spa thaceous at anthesis; petals 2, subequal, lineal, tapering to base, 1-2 cm long, brown pilose 0? inner surface; fertile samen 1, 2.5-4.0 cm - filament slender, arcuate, glabrous, shore nate with staminodes at base, anther linear long, green, 8-12 mm long, glabrous; staminod® 9, subequal, 8-12 mm long, connate for e their length, brown pilose externally, wee anthers absent; pollen spheroid, 3-colporoidi ii processes; gynoecium ca. equalling stamen, f. cuate, ovary tomentose, style and gynophore mentellous, gynoph equalling tyle ee subclavate, hardly differentiated from style. F di linear, apiculate with persistent style, 10-18 € long, ca. 1 cm wide, dark brown, sparsely k^ gose, gynophore 2.5-3.5 cm long; seeds oblong —10 cm long, 5-8 cm wide, surface dull, 3 brown, funicular aril-lobe scars subequal, € mm long. of Bauhinia deserti is restricted to the states Oaxaca and Puebla in Mexico in dry dec! 900 M- forest at elevations of about 1,000 to 1,993 ! — M — ——MÀÀ —Ó M — ag, i | 1983] It apparently flowers throughout much of the year but most abundantly from August to No- vember Bauhinia deserti is most closely related to B. chapulhuacania and B. dipetala; all three are characterized by green flowers with two petals. Itis distinguished from B. chapulhuacania by its smaller leaves and fruits and from B. dipetala by Its coriaceous leaves with the apex of the lobes usually rounded to obtuse. Bauhinia deserti is geographically separated from B. chapulhuaca- nia and occurs in a dryer habitat than B. dipetala. Bauhinia deserta was reduced to a variety of B. dipetala by Wunderlin (1967) but examination of additional collections leads to the conclusion that it should be retained at the specific level. This species is believed to be bat-pollinated, khi the related B. chapulhuacania and B. di- a. Pe ep specimens: MEXICO. OAXACA: Monte NY, P. UC B^ Pringle 6065 (A, F, MO, MS EBLA: Teh (F, GH, MO, NY, UO. ehuacan, Purpus 11 90 8. Bauhinia dipetala Hemsl., Diag. Pl. Nov. ind Mura dp dipetala (Hemsl.) Britt. & Rad poe Fl.23: 216. 1930. TIE - Veracruz: Valle de Córdoba, Bour- DA 1713 (holotype, K; isotypes, F, GH, P, n o unilateralis Britten & Baker, p Bot. 35:233. & RN M mn Baker) Britt eye: cd Fl. 23: 214 er TYPE: Mex- BM). Z: Veracruz, Houstoun s. n. (holotype, b e chartaceous, broadly ovate, bilo- allel to ud rarely to V^ their length, lobes par- (3-)5 HEC divergent, (3-)5-8(-12) cm long, apex of pem wide, base rounded to cordate, or slight] S acute to obtuse, margins smooth Y crisped, upper surface glabrous, lower Bose or cama m long, slightly canaliculate, stri- mm long. per ent; stipules lanceolate, ca. L5 "rescence dé sistent; adpetiolar intrastipular ex- jection arged and forming a subulate pro- "ud zl a mm long, others minute. Inflo- ose, terminal, or subterminal and axil] "ea i oo 30-fowered, rachis strigose, buds ca. 1 mm long, strigose, free tips up to br one bract lanceolate, ca. 1.5 mm long; Similar to bract but smaller; pedicels WUNDERLIN—ARBORESCENT BAUHINIAS 107 ca. 5 mm long, Plath hypanthium elongate- campanulate, 4-6 mm long; calyx spathaceous at anthesis; petals 2, pic linear, 6-8 mm long, ca. 2 mm wide, brown pilose on inner sur- face; fertile stamen 1, 2.0-2.5 cm long, filament slender, arcuate, glabrate, short-connate at base with staminodes, anther linear-oblong, green, 8— 10 mm long, sparsely brown pilose, staminodes 9, subequal, 3—4 mm long, connate for ?4 or more their length, brown pilose on inner surface, abor- tive anthers absent; pollen spheroid, 3-colporoi- date, sexine striato-reticulate with elongate su- pratectal processes; gynoecium ca. equalling stamen, arcuate, ovary tomentose, style and gy- nophore tomentellous to glabrate, gynophore ca. equalling style, stigma subclavate, hardly differ- entiated from style. Fruit linear, apiculate with persistent style, (8—)10—15(-18) cm long, 1.0-1. 5 m wide, dark brown, puberulent to glabrate, he esee 2.0-2.5 cm long; seeds oblong, 7-8 mm long, 6—7 mm wide, surface dull, dark brown, funicular aril-lobe scars subequal, ca. 3 mm long. Bauhinia dipetala occurs in southern Mexico and Belize along the edges of moist deciduous forest at elevations of 200 to 1,100 m. It has not been collected in Petén, Guatemala, but is to be expected there. It is sometimes cultivated as an ornamental and has become naturalized in Cuba. It flowers from December to May. Bauhinia dipetala is most closely related to B. deserti and B. chapulhuacania, which also have green flowers with two petals. The species has green flowers with a musky odor, which are bat-pollinated. Representative specimens: MEXICO. CHIAPAS: Along a small river in the barrio of Tih Ha', paraje of Mah- bene hauk, psig ae a iege vni Ton 2050 (F, MSC, US). MORELOS: Oax randa 1209 (MEXU). OAXACA: Near Oa i i 2112 (P). PUEBLA: Te- (P). VERACRUZ: Ca. 10 mi. Eo 6136 (F, Havana, León 6241 (NY). 9. Bauhinia divaricata L., Sp. PI. 374. 1753; non Lam., 1785; nec Hort. ex Steud., 1840. weed divaricatus (L.) Raf., Sylva Tel- 122. 1838. Casparia divaricata i H.B.K. ex Jackson in Index Kew 1895, nom. inval. d divaricata A H.B.K. ex Britt. & R : r F2 215. 1930. TYPE: ici n Clifford (holotype, BM). 108 Bauhinia enn [var.] beta L., Sp. Pl. ed. 2. 535. 1762. Tv maica: Herb. Miller (holotype, BM). Bauhinia pan Kos Ho Kew. 2: 48. 1789. Casparia ta (Ait.) Griseb., Fl. Brit. W. Ind. 213. 1860. Based on Bauhinia divaricata irt, beta L. Bauhinia divaricata Lam., : 389. 1785; non 1 Joo; nee Hort. ex d. 1840. TYPE: ?Ja- aica: He MD (holotype, FI, not seen). —— porrecta Sw., Prodr. 66. 1788. Casparia por- a (Sw.) Kunth ex Griseb., Fl. Brit. W. Ind. 213 1860. Cane adh: (Sw HBK. ex doe] in Index Kew. 1: 449. 1895, nom. inval. TYPE: ?Jamaica: pairs n. t oiotype, ?S, not seen). Bauhinia latifolia Cav., Icon. 5: 4. 1799. Casparia la- JAB. K. ex Jackson i in Index Kew. 1: TYPE: Mexico: Nee s.n. (holot pe, MA, not seen Bauhinia sack Poir., Encyc. Suppl. 1: 599. 1811; non Ham. ex Roxb., 1832. type: Herb. Jussieu (ho- otype, P-JU, not seen) Bauhinia racemifera Desv., J. Bot. 3: 74. 1814. TvPE: America. deli Desvaux Moretus Bauhinia a americana Laun., Herb. Amat. et 1821. TYPE ropical America. -— Launay vibus P-JU Ae Leno DC., Prodr. 2: 512. 1825. Based Bauhinia divaricata Lam. Bauhinia Repti DG, Prodr. 2: 512. 1825: Cas d (DC. ) Britt. & Rose, N. Amer. Fl 1930. TYPE: Based on a Sessé and Mocifio plate ar G, not seen; photograph of copy at G, F, MO, US; blueprint of tracing of copy at G, O MO, : Bauhinia furcata Desv., Ann. Sci. Nat. (Paris) 9: 429. Casparia furcata (Desv.) Jackson in Index Kew. 1: 449. 1895, nom. inval. TYPE: Tropical America. Herb. Des deg r aoni e. P). Poni adansoniana Gui Perr. Fl. Seneg. Tent. 1/269 3 i RAE but the pati i is of B. divaricata (fide Pellegrin, 1927 Bauhinia versicolor Bertol., Hort. Bonon. Pl. Nov. 1: . Casparia versicolor (Bertol.) Britt. & Rose, N. Amer. Fl. 23: 215. . TYPE: Mexico. Ta- maulip jm sone a vate domi "n collected by Bos- toloni T n. (hol OLO, not seen). Bauen mexicana Vog., pesi 15: 299. 1839. cof paneer et (Vog.) Britt. & Rose, N. Am . Veracruz: Nes ear Papantla, aeda n. (holotype, B, not seen, pre- Dauhints tanta var beta ccm & Arn., Bot. Cap. Beech YPE: Mexico. Nayarit: San kas nd "Ten. Lay 4 Collie s.n. (holotype, ?GL, not seen Bauhinia schlechtendaliana Mart. & Gal l. Brux. 1843. Casparia schlechtendaliana (Mart. & Gal.) Britt. & Rose, N. Amer. FI. 23: 215. 1930. TYPE: pres xico. Oaxaca: Rio ré las Vueltas, Gal- otti 9 (holotype, BR, not seen; isotype, P). Bauhinia pact srt Harms in Loess., Bull. Herb. Boissier 7: 548. 1899. pesos! amblyophylla (Harms) Britt. & “pes : > a Nene Between Te- quisistlán and J alapa, Seler 1689 (NY). LECTOTYPE ANNALS OF THE MISSOURI BOTANICAL GARDEN i [Vo.. 70 ) fadi dE tvnes: Seler 1689 | and 1 890, B, ése in World War Il. Sean sp se Rose, Contr. U.S. Nat. Herb. 10:97. 1906. C. dg confusa (Rose) Britt. & Rose, N. Amer. Fl. 23: 215 0. TYPE: Mexico. San Luis Potosí: Tomasopo Canyon, Pringle 3104 (holo- type, US; isotypes, CM, F, GH, MO, M SC, NY, VE). — MÀ — U i Bauhinia goldmanii Rose, Contr. U.S. Nat. Herb. 10: | 97. 1906. rvPr: Mexico. Chiapas: Tuxtla Gutiér- rez, Goldman 742 (holotype, US. fragment of ho- | otype, NY). it: Bauhinia me Jennings, Ann. Carnegie Mus. 127. . TYPE: Cuba. Havana: N of Caleta i Gra apii 630 (holotype, CM). | — epii Brandeg., Univ. "Calif. Publ. Bot. | : 183. 1922. Casparia peninsularis (Brandeg) | a8 & Rose, N. Amer. Fl. 23: 1930, TYRE | Mexico. Baja California Sur: W E of Cape rè | on, Brandegee s.n. (holotype, UC). l Bauhinia divaricata var. angustiloba Ekman ex Urban, ) Ark. Bot. 24A: 8 1. TYPE: Haiti | de la € western group Cap-à-faux, Ekman 10535 (holotype, NY, l stroyed in World War II; isotypes, F, LL, | US). Shrubs or small trees up to 8 m tall; branches strigose or tomentellous when young, S00? gla- brate. Leaves chartaceous to subcoriaceous, orbicular to narrowly ovate to lanceolate, ema ginate to bilobate about 3⁄4, rarely nearly l entire length, lobes parallel or divaricate, 2-12 cm long, 2-10 cm wide, base rounded to cores | apex of lobes acuminate to rounded, margi smooth to slightly crisped, upper surface tomet = wn = T. ——— — bi idi or tomentellous, 5—9-nerved, petiole 0 l 4.0 cm long, slightly canaliculate, strigose, 1 | mentellous, or glabrate; stipules ovate to linea | 1-3 mm long, deciduous; adpetiolar inta excrescence enlarged and forming a subula jection up to 2 mm long, others minute. mo 1 rescence racemose, terminal, or subterm rminal P | axillary, (10-)20-50-flowered, rachis strigoS | tomentellous, buds linear, 12-20 mm long, stri | gose or tomentellous, free tips up to 2 mm wa bract 1 late to linear, 1.5-4.0 m m lon de | teoles similar to bract but smaller; pedicels z | 2.0 cm long, strigose or tomentellous; hyp l thium cyathiform, 2-3 mm long; calyx SP? | ceous at anthesis; petals 5, subequal, white. e ing pink in age, 1.5-2.5 cm long, blade ovate | narrowly elliptic, 1-3 times the length of cla» 1.5-6.0 rap wide, glabrous, claw glabrous; fe of stamen 1, subequalling to twice the len ! petals, filament slender, slightly arcuate. ^ rous, connate at base with staminodes, 407 linear-oblong, ca. 2 mm long, black. glabro | : 7 e 2 oo É Ue o E C © EE © 5 e o - qo — £e S © & z o = 3 Lon} 2 eu E 1983] staminodes 9, subequal, 1.0~1.5 cm long, con- nate for 4 or more their length, white, petaloid, abortive anthers present or absent; pollen sub- spheroid, 3-colporoidate, sexine striate; gynoe- cum ca. equalling or slightly exceeding stamens, arcuate, short pilose to tomentose, Ovary more pubescent than gynophore and style, gynophore ca. equalling style, stigma capitate, green. Fruit linear, apiculate with persistent style, 2.0-2.5 cm long, brown, puberulent to glabrate, gynophore -0-2.5 cm long; seeds oblong to suborbicular, 8-10 mm long, 6-8 mm wide, surface dull, dark brown, funicular aril-lobe scars subequal, ca. 3 mm long. e Greater Antilles, except for Puerto Rico. It D in St. Christopher and St. Thomas E * Lesser Antilles; its absence from Puerto e psc that it is introduced there. It is the Belize M species of the genus in Mexico, tilles i ius Guatemala, and the Greater An- habilis u inia divaricata occurs in a variety of S Tanging PW eka icata is most closely related to temala do pam of southwestern Mexico, Gua- guished b nee from which it may be distin- brown ds ite flowers, lack of dense yellowish- divaricata "ence, and smaller leaves. Bauhinia exico Spa bly originated in the southern Spread north ; uatemala area and from there Central Am in Mexico, south to Costa Rica in as St, Pr and east in the Antilles as far Bauhinig E of the Lesser Antilles. Aus kafsize shana cta is highly polymorphic in ; "ape, and vesture. The flowers are also a races intergrade considerably and no . ly method of distinguishing them tax- lime to crea been found. It seems best at this In Mexico E. as belonging to a single taxon. "Pata de Cat, the local names for this species are Ta, "Pie de Cabra," “Pata de Vaca," WUNDERLIN — ARBORESCENT BAUHINIAS 109 "Pie de Vaca," “Ts’ulubtok’,” and *Sakts'ulu- btok’,” in Belize “Pata de Vaca," “Pie de Vaca," “Bull Hoof,” “Ts’ulubtok’,” and in the Cayman Islands “Bull Hoof,” in Jamaica “Bull Hoof” and *Mocco John," and in the Dominican Re- public “Pata de Chivo” and “Huella de Chivo.” Representative specimens: MEXICO. BAJA CALIFOR- SUR: Buena Vi É . SE of Chi de Corzo on Mexican Highway 190, Raven & Breedlove il mi 10959 (F, GH, MICH, NY, P, UC, US). HIDALGO: Santa Ana, on Pan-American Highway, Frye & Frye 2652 (GH, MO, NY, UC). jALisco: 18-20 mi. SW of o QUINTANA ROO: Cozumel! Island, vicinity of San Miguel de Cozumel, L 56 (DUKE, MO); j 2 mi. W of Xilitla on road to Jalpan, King 14308 (F, MICH, NY, TEX, UC, US). sinaLoa: El Monte, near Los Labrodos, Mexia 939 (A, GH, MICH, MO, NY, UC, US). rABAsco: Tenosique, Matuda 3397 (A, F, LL, MICH, MO, NY). TAMAULIPAS: Vicinity of Ciudad Victoria, Palmer 4 (CM, F, GH, MO, NY, UC). vERA- Rio Sebo] between Sebol and rrizal, N of Sebol, Steyermark 45765 (F, LL). cni- RESO: Between Tululmajillo and Finca Montanita, Steyermark 43343 (A, F). HUEHUETENANGO: Paso del Boqueron, along Río Trapichillo, below La Libertad, Steyermark 51130 (F). PETÉN: Vaxactun, Bartlett 12511 TALHULEU: Between Nueva UC (MICH, NY, UC). ORANGE WALK: Honey Camp, dell 640 (A, F, GH, MO, NY, UC, US). STANN CREEK: Stann Creek Railway, Schipp 221 (A, F, , MICH, MO, UC). TOLEDO: Forest Home, Schipp 8502 (F). H 8801 (F, GH, MICH, MO, NY, UC, preghi n ruins, Molina 6624 (F, MO y n: Quebrada Contarranas, NE of Contarranas, Molina 7808 (F, US). SANTA BÁRBARA: San Pedro Sula, 110 Thieme 5183 (GH, NY, US). OLANCHO: Vicinity of Catacamas, Standley 18410 (F). yoro: Agua Blanca, 10 mi. a Record & Kuylen H.35 (GH, NY, NSONATE: Near Yuayua, Pittier Along road Stevens 3822 (MO, USF). CosrA os GUANACASTE: 11 of ruz, 0.5 mi. W of main road, Liesner 4851 (MO “USB. "CUBA. HAVANA: Isla de Pi- nos, Caleta Crocodrilos, Britton et al. 15316 (NY). ATANZAS: Penínsola de Zapa C Shafer 11 168 E. MO, NY, US). CAYM Little Cayman, near Snipe ay Proctor 28040 (J, River, Wedder- Antonio, Hitchcock s.n. (MO, US). s- ST. ANDREW: North- ern end of Lo Long Mountain, just E of Kingston, Crosby et al. 106 (DUKE, F, LL, MICH, MO, MS UG; US, USF). sr. ANN: 1 mi. S of eins Proctor 6624 (IJ, LL, US). sr. CATHERINE: Rodney's Lookout near Port Henderson, Wunderlin 5073 (MO USF). sr. ELIZABETH: ie ipsins Maxon & Killip 1501 (F, ILL; US). ST AMO e River, near Montego Bay, Maxon & Killip 1668 (F. “ILL. US). st. THoMAs: Along road between Holland Bay and Morant Point, Proctor 27843 (IJ, LL, MICH, US). TRELAWNY: Road N Spring Garden toward Burnt Hill, Anderson & Stern- berg 1239 (DUKE, LL, MO, US). HAITI. GONAVE ISLAND: Vicinity of Etroite, Leonard 3283 (NY, son s.n. (photograph of specimen at BM, IJ). Sr. Tagua. Reidle s.n. (P). 10. Bauhinia erythrocalyx cies aa Sida 5: 353. 1974. TYPE: Mexico. Cam e; "Cen- tral Buenfil" (?Puerto Buen Fiel), Lundell 1152 (holotype, F; isotypes, F, MICH) Bauhinia Baien Lundell, Wiee d 344. 1977. TYPE: a. Petén: , 1 km N of vil- e Lu ndell & Contreras 20475 (holotype, LL, not seen; isotype, LL). Shrubs or small trees up to 15 m tall; branches strigose when young, soon glabrate. Leaves char- taceous, ovate, bilobate for '^ their length, lobes divergent, 3.5-7.0 cm long, 4-6 cm wide, base cordate, apex of lobes obtuse, margins slightly crisped, upper surface glabrous, lower surface strigose, 7-nerved; petiole 1.5-3.0 cm long, ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 slightly canaliculate, strigose; stipules lanceolate, ca. 1.5 mm long, caducous; adpetiolar intrastipu- lar excrescence enlarged and forming a subulate projection ca. 1.5 mm long, others minute. In- florescences racemose, terminal, or subterminal and axillary, 10—20-flowered, rachis strigose, buds linear, 1.5-2.0 cm long, strigose, red, free tips minute; bract ovate-lanceolate, ca. 1 mm long bracteoles similar to and slightly smaller than bract; pedicels 1-2 cm long, strigose, reddish; hypanthium cyathiform, 2-3 mm long; cal spathaceous at anthesis; petals 5, subequal, white or cream, ca. 1.5 cm long, blade linear-lanceolalé to oblanceolate, ca. '^ the length of claw, 2 mm wide, sparsely pilose externally, claw gla- brate; fertile stamen 1, slightly exceeding the pet als, filament slender, slightly arcuate, glabrous, connate with staminodes at base, anther oblong, 3-4 mm long, red, glabrate; staminodes 9, alit nate staminodes slightly shorter, 1.0-1.5 cm lone connate for ca. !^ their length, pilose on inne surface at connate portion, abortive anthers pre ent or absent; pollen peroblate, 3-colporoidat® | reticulate; gynoecium subequalling fertile sta- men, slightly arcuate, ovary densely strigoS sica di and style sparsely and loosely stri stigma slight) A Meetai from style. Fruit linear, apiculate with persistent style, 7-10 cm long, ca. 1.5 C? wide, light brown, strigose to glabrate, gynophort ca. 2 cm long, glabrate; seeds oblong, mm wide, dark brown, funicular aril-lobe scars sub- equal, ca. 1.5 mm long. Bauhinia erythrocalyx is a rare species occur ring in Campeche, Quintana Roo, and Yucatán. Mexico and adjacent Petén, Guatemala. It occurs | in dry, often rocky, semideciduous forests à! d evations from near sea level up to about meters Bauhinia erythrocalyx is most closely related to B. jenningsii, a more common and widesP species occurring in the same general area. It # readily distinguished from B. jenningsii by IP bilobate leaves. Neat E ve (US): Specimens examined: MEXICO. CAMPECHE: Xpujil, Shepherd 16 (LL, WIS). YUCATAN: 8 km Xpujil, Waide s.n. (WIS); Mérida, Novelo 404 QUINTANA ROO: 44 km ga-Chetumal £ way on road to Icaiché, Guero & Crake 909 (MOK. km from the Chetumal-San Carillo Puerto highway ^" road to El Placer, Téllez & Cabrero 1277 (MO) M TEMALA. PETÉN: Dos Lagunas, bordering the North Contreras 1660 (LL). —— — — high | Gute |] — 1983] WUNDERLIN—ARBORESCENT BAUHINIAS F IGURE ], Bauhinia fryxellii. A. Flowering branch. B. Flower (X1.3). 111 112 11. Bauhinia fryxellii Wunderlin, sp. nov., fig. 1. TYPE: Mexico. San Luis Potosí: Municipio Ciudad Valles, Rancho Pago Pago, 3.5 mi. by road W of Chantol, near juncture of Río Mesillas and Río Mico, Fryxell & Anderson 3427 (holotype, USF; isotypes, to be de- posited in CHAPA, ENCB, K, MO, MEXU, MICH, pf, TEX, US). (*pf" is the personal herbarium of Paul Fryxell.) Bauhinia jucunda Brandeg. affinis, a qua imprimis differt foliis majoribus magis Dipl lobatus, stylo breviori, et filamento staminis f: d b Shrub to 1 m tall; branches sparsely puberu- lent-strigose when young, soon glabrate. Leaves chartaceous, bilobate % or more their length, lobes narrowly elliptic-lanceolate, divergent, 8 cm wide, e, lo nerv long. d slightly catialiculite. sparsely pu- ru strigose; stipules ovate, aristate tipped, ng, caducous; adpetiolar intrastipular only 1 well developed, linear, greenish-white, up to 1 cm long, pilose on inner surface, the smaller one up to 5 mm long or reduced to a short fil- ament, rarely with a third vestigial petal; fertile stamen 1, 2.0-2.3 cm long, short-connate at base with staminodes, filament ca. 1.8 cm long, pilose on lower '^, anther linear-oblong, 7-8 mm long, sparsely pilose on ventral surface; staminodes 9, united by their flattened filaments % to 3⁄4 their length, longest adjacent to fertile stamen ca. 8 mm long, shortest ca. 5 mm long, all without vestigial anthers, outer surface of staminal tube and free filaments pilose nearly to base, inner surface pilose only on upper part; pollen sphe- roid, Fcolporoidate, sexine Rrato-teti Lh cm n long, linear; ovary short stipitate, appressed pubescent, stipe ca. 5 mm long, strigose; style ca. 5 mm long, sparsely pilose, stigma capitate. Fruit linear-oblong, apiculate with persistent style, 6— ANNALS OF THE MISSOURI BOTANICAL GARDEN " [Vou. 70 7 cm long, dark brown, sparsely strigose, gyno- phore ca. 1 cm long; seeds subquadrangular, 7- 8 mm long, 4—6 mm wide, surface shiny, reddish- brown, funicular aril-lobe scars subequal, ca. 1.5 | long mm ion Bauhinia fryxellii is a rare species known only from the type collection. It was collected at 110 meters elevation in a cut-over field. Collected in May, it was in flowering and fruiting condition. Bauhinia fryxellii belongs to an alliance of five species that includes B. chapulhuacania, B. de serti, B. dipetala, and B. jucunda. This group differs from all other monandrous Neotropical species of Bauhinia in that they have a reduced number of petals, usually 2, rarely 1, 3. or 4. Bauhinia fryxellii is most closely allied to B. Ji- cunda but differs in its larger, more deeply lo leaves, shorter style, and the filament of the fer tile stamen pilose near base. —— ——7À 7 The species is named in honor of one of il | collectors, the noted student of Malvaceae, Paul | A. Fryxell. 12. Bauhinia jenningsii P. Wilson in Britt., Bull. Torrey Bot. Club 43: 463. 1916. Caspari Jenningsii ‘2 ee Britt. & Rose, N. Amer. Fl. 23: 216. 1930. TYPE: Cuba. He vana: Isle de Poor Coe's Camp, Ensenada —— de Siguanea, Britton & Wilson 14851 (h€ lotype, NY; isotypes, CM, F, MO. NY, P. US). Bauhinia edet Standl., de Field Mus. Nat. e r22 E: Belize tann C 8. 1940. T e] aie maaie: Creek tt ideft 48 3 (holo f F). Slender shrub or small tree to 6 m tall; branché | glabrate. Leaves subcoriaceous, ovate to lance late, 7-14 cm long, 2.5-5.0 cm wide, apex minate to obtuse, base truncate to rounded. "P per surface bright green, glabrous, lower $ lighter in color, strigulose to glabrate, 5 petioles 1.5-2.0 cm long, terete or slightly s aliculate, glabrate; stipules lanceolate, ca. 1.5 long, caducous, adpetiolar intrastipular - cence enlarged and forming a subulate project up to 1.5 mm long, others minute. Inflor escent racemose, terminal or subterminal and or 10-30-flowered, rachis loosely strigose, red : buds linear, 1.5-2.0 cm long, loosely -— red, free tips minute; bract lanceolate, ca. 0 m long; bracteoles similar to bracts but " shorter; pedicels slender, 1.0-1.5 cm long. 100% ly strigose, reddish; hypanthium cyathifor™ J | | $ 1983] 3 mm long; calyx spathaceous at anthesis; petals 5, subequal, white or cream, 1.2-2.0 cm long, blade linear-lanceolate to oblanceolate, sub- equalling claw, 2-3 mm wide, sparsely pilose ex- ternally, claw glabrate; fertile stamen 1, slightly exceeding the petals, filament 1.5-2.0 cm long, slender, slightly arcuate, glabrous, short-connate With staminodes at base, anther linear-oblong, 4-5 mm long, glabrous, staminodes 9, 1.0-1.5 cm long, alternate slightly shorter, connate for cà. /^ their length, sparsely pilose on inner surface at connate portion, aborted anthers present; pol- len peroblate, 3-colporoidate, sexine reticulate; &ynoecium subequalling fertile stamen, slightly arcuate, ovary densely pilose, gynophore and style sparsely pilose, gynophore subequalling style, ot not much differentiated from style. Fruit near, apiculate with persistent style, 5-10 cm d 1-2 cm wide, light brown, strigulose, gyno- es. » 2 em long, glabrous; seeds oblong, 7- Ee uu 5-7 mm wide, dark brown, funicular scars equal, ca. 2.5 mm long. ru: common species occurring in Cuba, the a 2 Peninsula of Mexico, Belize, and north- d Sings It occurs in rocky, calcareous Bis By s edge of dry to mesic or sometimes Woods. It flowers fr : tember. om February to Sep BEES i 4 : ais a jenningsii 1$ a unique species, readily ite guished from other Bauhinia species with- is Pei by its entire, Smilax-like leaves. It Youmans! related to B. erythrocalyx of the are very m The flowers ofthe two species * Simi ar, but y . ie ` é * entire leaf B. jenningsii differs in its The "Cow ua name for this species in Belize is is *C ongue” and the Mayan name in Mexico hakts'ulubtoke NePfeSentive S mi. S of (MO, UsF) 12145 ( ). GUATEMALA. PETÉN: Vaxactun, Bartleti UC, US). Beze. cayo: Vaca, 13, i2 pina jucunda Brandeg., Univ. Calif. "oaa i 326. 1920. Casparia jucunda 213, T Britt. & Rose, N. Amer. Fl. 23: Ca de Pan, TYPE: Mexico. Veracruz: Barran- UC), ancava, Brandegee 8535 (holotype, Shru Young, ie 2 m tall; branches strigose when n glabrate. Leaves chartaceous, bilo- WUNDERLIN—ARBORESCENT BAUHINIAS 113 bate for 2 or more their length, lobes lanceolate or triangular, divaricate, 5-8 cm long, 3-5 cm wide at base, 5-8 cm wide between lobes at tips, base cordate to truncate, apex of lobes ob- tuse, margins moderately crisped, upper surface glabrous, lower surface strigulose to glabrate, 5-nerved; petiole 1-2 cm long, slightly canalicu- late, strigose; stipules lanceolate, ca. 1.5 mm long, caducous; adpetiolar intrastipular excrescence enlarged and forming | i to ca. 1.5 mm long, other minute. Inflorescences 1.5-2.0 cm long, strigose, free tips minute; bract lanceolate, ca. 1.5 mm long; bracteoles similar, but smaller than bract; pedicels 0.5—1.2 cm long, strigose; hypanthium cyathiform, 3-5 mm long; calyx spathaceous at anthesis; petals usually 3, rarely 2 or 4, sometimes third and/or fourth rep- resented only by a filament or obsolete, rarely a fifth vestigial petal present, white, linear, 1.0-1.5 cm long when fully developed, blade linear, 2—4 times longer than claw, densely pilose, ca. 1.5 mm wide, claw densely pilose; fertile stamen 1, ca. 1 cm longer than petals, filament slender, slightly arcuate, glabrous, slightly connate at base with staminodes, anther oblong, ca. 1 cm long, sparsely pilose; staminodes 9, 5-8 mm long, con- nate for !^ or more their length, pilose, abortive anthers absent; pollen spheroid, 3-colporoidate, sexine striato-reticulate with elongate supratectal processes; gynoecium subequalling fertile sta- men, slightly arcuate, ovary tomentose, gyno- phore and style tomentellous, gynophore sub- equalling style, stigma capitate, slightly oblique. Fruit linear, apiculate with persistent style, 8-10 cm long, ca. 1 cm wide, brown-strigose, gyno- phore ca. 1 cm long, glabrate; seeds oblong, 6-8 ong, 4-6 mm wide, surface dull, brown, funicular aril-lobe scars equal, ca. | mm long. Bauhinia jucunda is a rare species endemic to a small area in eastern Veracruz, Mexico, and known from only a few collections. It occurs in low, deciduous forests at 200—500 m. It appar- ently flowers from July to November, rarely as early as April. 1 11; € This species t g p that includes B. chapulhuacania, B. deserti, B. dipetala, and B. fryxellii. This species differ from other monandrous Neotropical species of Bau- hinia in having a reduced number of petals, usu- ally 2, rarely 1, 3, or 4. Bauhinia jucunda is most closely related to B. fryxellii, but differs in having 114 smaller, less deeply lobed leaves, a longer style, and a glabrous filament of the fertile stamen. Specimens examined: MEXICO. VERACRUZ: Road from Xalapa on detour to Los Banos de Carrizal Emiliano Zapata, Calzada 2043 (F, MEXU, USF); Plan del Rio, Municipio Dos Rios, Ventura 5787 (MEXU, NY); 8994 (MEXU); La Ceiba, Municipio Puente Nacional, zu tura 12061 (MEXUJ); Remudadera, Purpus 8833 (G NY); 8986 (F, GH, MO, NY, UC, US); without ae locality, Purpus 8940 (UC). 14. Bauhinia lunarioides A. Gray ex S. Wats., Bi ye ar . Amer. Bot. 205. 1878. TYPE: Mex Coban Rocky hills near Santa ig pbi. 2901A (holotype, GH). Casparia congesta Britt. & Rose, N. Amer. Fl. 23: 21 1930. Bauhinia congesta (Britt. & Rose) dite. Phytologia 1: 214. E . TYPE: Mexico. Coahuila: Mou s 24 E by N from Monclova, P 285 adn US: ae LE NY P Casparia. jermyana Britt. in Britt. & Rose, N. Amer Fle 25 211. 1930, iq jermyana (Britt.) Lundell, Phytologia I: 214 TYPE: United States. Texas: Gillespie Co.: musei. Rock, Jer- my s.n. (holotype, NY). [See discussion below.] Shrub up to 4 m tall; branches loosely strigose when young, soon glabrate. Leaves subcoria- ceous, broadly ovate to suborbicular, bilobate for 34 or more their length, usually bifoliolate, lobes suborbicular, slightly divergent, inner margins straight to slightly convex, (1.0—)1.5—2.5(—3.0) cm long, (1.5—)2.0-2.5(-3.0) cm wide, base cordate or rarely truncate, apex of lobes rounded, mar- gins slightly to moderately crisped, upper surface glabrous, lower surface strigose to glabrate, (5—) 1 1-nerved; petiole 0.5—1.0(—1.5) cm long, slightly canaliculate, pee strigose; stipules broadly ong, caducous; intrastipu- hairs, occasionally purplish-tinged, free tips ca. 2 mm long; bract lanceolate, ca. | mm long; brac- teoles similar but smaller than bract; pedicels slender, 0.5-1.0 cm long, strigose; hypanthium cyathiform, 1-2 mm long; calyx spathaceous at anthesis; petals 5, subequal, white or occasion- ally pinkish-tinged, 1.5-2.5 cm long, blade ellip- tic-lanceolate to ovate, ca. 3 times the length of claw, 0.7-1.2 cm wide, glabrous to sparsely pi- lose on inner surface, claw pilose to glabrate; fertile stamen 1, subequalling petals, filament slender, slightly arcuate, sparsely pilose at base, ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo 3 ) short-connate with staminodes, anther oblong, 5-7 mm long, glabrous; staminodes 9, subequal, e for ca. !^ their length, e anthers | usually pue polidi spheroid, 3-colpo r \ gynor ca. ¥ the " of style, gynophore and ovary pilose, stigma cap- itate, bilobate, dark in contrast with style. Fruit linear, apiculate with persistent style, 5-8 cm long, 1-2 cm wide, light brown, glabrate, gyno phore ca. 5 mm long, strigose; seeds 7-10 mm long, 6-8 mm wide, surface dull, dark brown. funicular aril-lobe scars unequal, one ca. 1.5 mm long, other ca. 0.5 mm long. petals, slightly arcuate h m Bauhinia lunarioides is restricted to Coahuila and Nuevo León, Mexico, and adjacent Texas It occurs in desert scrub, primarily on calcareous soils at elevations of 500 to 2,00 m. It flower primarily from March to June. ——À There is considerable debate concerning tht | distribution of this species in Texas. In addition | to the several collections from the Anacacho Hills | Kinney County, by various collectors, single col- | lections have been seen that, according 10 the ! abel data, are from Gillespie (type of Caspari | jermyana), Maverick, Presidio, Val Verde, ani | Webb counties. However, B. L. Turn er and M. | C. Johnston (pers. comm.) suspect a the label | information for these collections is incorrect l that all Texas material is actually from the ) Anacacho Hills. The occurrence of the species? | Texas in these counties, with the exception A Gillespie (see below), is logical and these A | barium records cannot be completely disrega ed. However, not having personally verified | occurrence of the species at the reported localities in southern Texas, I reserve judgement 0n this } oint. The type collection of Casparia jeri 24 is stated in the protolog and on the of 4 | holotype specimen to have come Ait Ancha | ed Rock, Gillespie County, Texas. However. ie | is probably an error since recent work has fal "T to produce any species of Bauhinia at Enchanie? | Rock. Furthermore, B. lunarioides grows w ! marily on calcareous soils and Enchanted Rod s is a granite outcrop. It is possible that the oo lection is actually from the Anacacho Moun™® in Kinney County where it is well known. " Bauhinia lunarioides is readily disting™ from the other white-flowered monandro™ | species of Bauhinia by its usually bifoli a 1983] leaves. Other species within the range of B. /u- narioides that may occasionally have bifoliolate leaves have purple to pink flowers. It is probably most closely related to B. divaricata but is more northern and xerophilous than that species. Representive specimens: UNITED STATES. TEXAS: Kinney Co.: Anacacho Ranch, SW slopes of Anacacho Mountains, Correll & Rollins 32537 (FSU, TEX, UC, US). Maverick Co.: N Devil's River, 0.25-0.5 mi. o confluence of the Devil’s River and Dolan Creek, Smith & Butterwick 166 (L L). Webb Co.: Near Laredo, War I EX). MEXICO. COAHUILA: San Lázaro, near the northern entrance of El Puerto de San Lázaro, Wynd & Mueller 109 (A, ILL, MICH, MO, MSC, NY, US). NUEVO LEÓN: 8 mi. W of Monterrey on way to Saltillo, Hitchcock & Stanford 6835 (NY; UG US). 15, Bauhinia macranthera Benth. ex Hemsl., Diag. Pl. Nov. 49. 1880. Casparia macran- thera (Benth. ex Hemsl.) Britt. & Rose, N. Amer. Fl. 23: 212. 1930. TYPE: Mexico. Hi- dalgo: Zimapán, Coulter s.n. (holotype, K). pe ie Standl., Contr. U.S. Nat. Herb. 23: 416. 1922. Casparia ih age (Standl.) Britt. & fou N Amer. : 1930. TYPE: Mexico LUIS Potosí: Rae pape m holotype i : 212. 1930. TYPE: Mexico. Nuevo León: AH de la Silla, Monterrey, Pringle 2529 (ho- otype, d x isotypes, A, CM, F, GH, MO, MSC, YT). i NY PU a thia macranthera var. pears Wunderlin, Phy- rio $a a 15: 53. 1967. Based on decus luna- Ides A. Gray ex Britt. & Ros tie tree up to 4(—7) m tall; branches ose when young, soon glabrate. their len ie ovate, bilobate for '^ to l^ their iei rarely bifoliolate or lobed less than cay. lobes parallel or slightly divergent, cordate, ape ) cm long, (3.5-)5-8(-9) cm wide, base gins slight} pex of lobes rounded to obtuse, mar- glabrous, y to strongly crisped, upper surface te,7 E surface strigose, pilose, or gla- long stign p, needs petiole 1.5-2.5(4.0) cm &abrate- NL Ce tomentose, strigose, Or -lanceolate, ca. 2 mm long, demi intrastipular excrescences minute; in- €s short-racemose, subterminal, axil- s Coto , rachis strigose or tomentose, Strigose y lanceolate, 2.0-2.5(-3.0) cm long, mice ae tomentellous, free tips setaceous, othe mm long; bract linear-lanceolate, than ong, bracteoles similar but smaller Ct; pedicels 3-6 mm long, strigose or WUNDERLIN — ARBORESCENT BAUHINIAS 115 tomentose; hypanthium cyathiform, 1-2 mm long; calyx spathaceous at anthesis; petals 5, sub- equal, pink to purple darker-veined, (2.5-)3.0- 3.5(-4.0) cm long, blade obovate to elliptic, sub- equalling length of the pilose claw, 0.5-1.0 cm wide, glabrous or sparsely pilose at base; fertile stamen 1, ca. 3⁄4 the length of petals, filament stout, arcuate, pilose at base, short-connate with staminodes, anther oblong, 7-9 mm long, sparse- ly pilose; staminodes 9, subequal, 6-8 mm long, connate ca. 2 their length, pilose below, abortive anthers present or absent; pollen spheroid, 3- colporoidate, sexine striato-reticulate; gynoe eci- slightly differentiated from apiculate with persistent style, (6—)8-12(-15) cm long, 1-2 cm wide, light brown, tomentose to glabrate, gynophore 1.0-1. 5 cm long, glabrate; seeds oblong, 8-12 mm long, 6-8 mm wide, sur- face uui brown, funicular aril-lobe scars equal, ca. 4 m Bauhinia macranthera occurs in central and northeastern Mexico in open, deciduous forests at elevations ranging from 600 to 2,500 m. It flowers from March to August. The closest relative to B. macranthera and that with which it has been confused is B. ramosis- sima. It is distinguished from that species by usually pink rather than purple flowers, larger, usually bilobate rather than bifoliolate leaves, and setaceous calyx tips. C, NY, US). HIDALGO: Puente de la Zorra, near Km 284 on highway NE oe Moore 1711 " UC, US bor € mr re enta bel a i SW of Galana, Mueller & Mueller 654 (A, F MICH. NY, P, TEX, US). s Tosi: Ca. 10 mi. NE of Ciudad del Maiz (17 mi. by ed McVaugh 10434 (DUKE, GH, LL, MICH, MO, EX, US). TAMAUL "I of Ciudad Victoria, King 4 : UC, US). VERACRUZ: Eof fabio Ga 760 (MO, Y 5. Pauletia multinervia H.B.K., 16. Bauhinia multinervia (H.B.K.) DC., Prodr. 2: 515. 18 6. bend TYPE: Ven- Nov. Gen. Sp. Pl. 6: 31 ezuela. “Prope Caracas, Caripe et Montana de Capaya." Humboldt r ye 576 (holotype, P; microfiche, MO; isotype B(W); photographs ofi isotype at B(W), F, MO, US). 116 Pauletia glaucescens H.B.K., Nov. Gen. Sp. Pl. 6: 317. 1823. Bauhinia glaucescens (H.B.K.) DC., Prodr. 182 V 2: m 5. TvPE: Venezuela. Sucre: Humboldt & Bonpland 2 ye otype, P; egi ine Mo: Bauhinia Sante dew dab Bri SAI: 1860. LECTOTYPE: ae nt: ot eo icd totype, K; ato Of | lectotype, F, IJ, US). Lectotype here designated. Large shrub or small tree up to 10 m n branches ferruginous puberulent when yo soon glabrate. Leaves chartaceous, H to oblong, bilobate for ca. 1⁄2 their length, 7-15 cm long, 6-14 cm wide, base cordate to rounded apex of lobes rounded, upper surface glabrous, lower surface glabrate or occasionally glaucous, ferruginous puberulent on nerves, (5-)7-9 (-11)-nerved; petioles 2-4 cm long, canaliculate, ferruginous pubescent or glabrate; stipules ovate, ca.2mm longe CACONI; adpetiolar intrastipular d forming a subulate pro- jection up to 1. p mm long, others minute. Inflo- rescences racemose or paniculate through reduc- tion of subterminal leaves opposite single flowers, up to 20-flowered, rachis ferruginous puberulent to velutinous, buds linear, 6-8 cm long, ca. 1 cm thick, ferruginous velutinous, apex rounded; bract triangular, ca. 1 mm long; bractioles similar to bract; peduncle 1—3 cm long, ferruginous puber- ulent or velutinous; hypanthium tubular, 2-3 cm long; ay splitting into 3 topes a lobes partly na pa ale green on inner tao petals 5, subequal, white, fili- form, 8-12 cm long, 3-5 mm wide; fertile sta- mens 10, equalling or exceeding petals, alternate ones slightly shorter, slightly spreading, filaments slender, white, short-connate at base for ca. 1 cm, puberulent toward base and on inner surface, anthers 1.5-2.0 cm long, glabrous; pollen sphe- roid, 3—6-poroidate, sexine reticulate with blunt, spinelike infratectal processes; gynoecium sub- equalling stamens, ovary and gynophore purple, puberulent, style glabrate in upper '2, greenish- white, gynophore subequalling style, stigma oblique. Fruit linear, apiculate with persistent n suborbicular, ca. 1.5 cm long, 1.0-1.5 cm wide surface dull, dark brown, funicular aril-lobe scars extending '^ way around edge of seed. Bauhinia multinervia occurs in our range only in the Lesser Antilles where it is found in mesic forests at elevations up to about 900 m. This species also occurs in Venezuela, Suriname, and Brazil. It is sometimes cultivated as a novelty ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 within its natural range and in tropical areas out- side it. Bauhinia multinervia flowers throughout the year, but most abundantly from October to March. This species is pollinated by bats. Bauhinia multinervia is distinguished from other Bauhinias in our range by its large flowers. It is a part of the B. ungulata complex, which consists of sea a PREEN centered in the Plan- alto region of B None of the dd of B. multinervia seen from the Lesser Antilles have been made sinc the late 1930's In MEE the species is known Y the local names Dieet and “Uri Specimens examined: ANTIGUA. cond st (volca- nic ec Fig Tre ree e Hill sg 1 Sy 9 (UC, US); without precise lo cality, Rose 1 (US). GUADELOUPE. ithout precise locali yo s.n. (P) pe id 690 (US); Richard s.n. (P). MARTINIQUE. St. Stehlé & Stehlé 3550. (US); without precise ical Belanger 490 (Py; Duss 1124 (NY); Terrasson 44 (P. St. VINCENT. Without precise locality, Smith & Smith 273 (NY). | | e | 17. Bauhinia pansamalana Donn. Sm., Bot. Gaz 13: 27. 1888. Casparia | na (Donn. Sm.) Britt. & Rose, N. Amer. PE 23: 216. 1930. TYPE: Guatemala. Alta Verapaz: Put samala, Tuerckheim 68 1 (holotype, US; 18% types, NY, Shrubs or small trees up to 5 m tall; brane | glabrous. Leaves chartaceous, ovate, entire % — bilobed up to their length, lobes slightly dive" gent, 12-22 cm long, 9-15 cm wide, base cordate to rounded, apex of entire leaf acuminate to cat- date, apex of lobes acuminate, margins smooth or nearly so, glabrous on both surfaces, 7-nerved: petiole 3-6 cm long, slightly canaliculate. gla- trastipula! terminal, or subterminal and axillary, 2 flowered, rachis and rachilla strigose to glabral® buds linear, ca. 2 cm long, strigose, free tips m nute; bract ovate, ca. | mm long; bracteoles sim- ilar to bract; pedicels ca. 1.5 cm long. strigose hypanthium cyathiform, 4—6 mm long tals 5 subequal, pu 203.5. cm long, blade narrowly elliptic. @ 3 times the length of the sparsely pilose claw: i mm wide, glabrate; fertile stamens 3, SU adnate with staminodes, subequalling P* adaxillary and each separated by staminode: aments slender, arcuate, pilose at base, an 4-5 mm long, glabrous, stamens and stamino“ sparsely pilose on inner surface; staminodes fil- ; | 1983] subequal, 7-9 mm long, connate for 2 or more their length, abortive anthers present or absent; pollen spheroid, inaperturate, sexine reticulate with blunt, spinelike infratectal processes; gy- noecium subequalling petals, arcuate, ovary stri Bauhinia pansamalana is a rare species native to southern Mexico and Guatemala. It occurs in moist deciduous or tropical rainforests up to ne meters in elevation. It flowers from April uly. Bauhinia pansamalana is the only Middle American species of Bauhinia, other than B. ess that has three fertile stamens. It occurs à a the south of the range of B. coulteri and can be further distinguished from that species by lts large leaves. Seo Ee MEXICO. CHIAPAS: Finca Ir- Cedi een (F. GH, MO, NY, UC, U xhil between Cedillo and La Escua- dra, V, i LM et al. V-338 (MEXU). GUATEMALA. ALTA Et T NY, US) a id ig 91659 (F, US); Standley 91858 (F); Pansa- Cook "pd d 1631 (US); near Finca Sepacuiter, 'gs : f : Stevermark 41738 ( Nee IZABAL: Along Rio Bonita, 18, ini ; Bauhinia pauletia Pers., Syn. Pl. 1: 455. 1805. on Pauletia aculeata Cav. Paule; Mid aculeata Cav., Icon. 5: 6. 1799. Type: Panama. ie "Near Panama City, Herb. Cavanilles Bauhinia on MA, not seen). 4 spinosa Poir., Encyc., Suppl. 1: 599. 1811. Bauhinia sa Paule tia aculeata Cav. sci namensis Spreng., Syst. Veg. 2: 334. 1825. Bauhinia 1, on Pauletia aculeata Cav. > Amphipodial rhizome system FiGureE 5. Rhizomes. TEM 1983] SODERSTROM & YOUNG—COLLECTING BAMBOOS 135 MEXICO Olmeca recta Soderstrom Veracruz: Mun. Catamaco, 10 Km N of Catamaco on road to Sontecomapan, 18*30' N / 95*01' W. Growing under tall trees in disturbed evergreen rain forest; common throughout the Tuxtla Range. Culms well separated from each other, 11-12 m tall, 4.5-5 cm diam., thin-walled, often filled with water, rotting easily. Rhizomes to 3 m long, running near surface of the soil. ulm leaf: sheath bluish-green, becoming pinkish-brown then stramineous; blade erect, deciduous, bright green. Branching: only from upper nodes, the branches intravaginal, arching out and becoming vinelike. Common name *Jimba." Elev. 435 m Thomas R. Soderstrom 2235 2 Oct. 1977 Collected under the auspices of the Smithsonian Institution and Colegio de Postgraduados, Chapingo, Mexico FIGURE 6. Herbarium label for bamboo collection. 3. Branches. e. Modifications: developing as spines a. General Characteristics. (Fig. 31). l) Occurrence: upper nodes only; all To Collect: Take a representative branch nodes. complement from mature or old culm, even 2) Habit and Length: main branches from dead culms that show the features ac- elongated and vinelike; upper curately. Cut culm ca. 5 cm above and below branches angled upward, lower re- the node and trim branches to ca. 5 cm, at flexed. least including the first node. Large culm sec- 3) Development: intravaginal (Fig. 11); tions may be split in half to save space. As extravaginal (Fig. 1h); both intra- with culm sections, allow to air-dry and re- and extravaginal. move all sheaths. b. Number and Arrangement: single 4. Branch Leaves (Fig. D branches (Fig. 3c); 2 subequal branches a. General Characteristics. = ig. 3d); 3 or more subequal branches 1) Condition After Cutting: wilting or (Figs. 3e, f, g); 1 dominant with fur- aT immediately; remaining ther branche igs. 3h, resh. — i); 1 Meo crop ci re- 2) uenis stiff, flexuous; erect; pen- E e d opes : b C La on both surfaces; lighter ranches below or around it (Fig. . ‘ 3b, shown in bud stage); or in apsidate on one surface than the other; varie- arrangement without a central branch gated. ; ; (Fig. 3k). To Collect: Leafy twigs, arranging the leaves c. Origin: produced at nodal line (Figs. to show both surfaces. Include leaves from 3c, g, h); produced above the nodal line young and old branches if there is a signifi- (Fig. 3f); produced from a specialized cant size difference. Press immediately to Process (Figs. 3e, j, k). avoid curling. If this is not possible the A d. Posture at Node: appressed (Figs. 3f, terial can be wetted and retained in a tightly me ; horizontal (Fig. 3g); angled upward an - 3c-f, h-k); angled downward (Fig. 1). secured plastic bag. Large leaves may require trimming and folding or dividing and press- ing in more than one sheet. 136 5. Inflorescences (Fig. 4). a. General Characteristics. 1) Habit: erect; lax; drooping. 2) Occurrence: terminating leafy branches; occurring throughout a leafless plant. b. Size: length and width if larger than an herbarium sheet. c. Color: green; stramineous; purple. To Collect: Flowering branches at all stages of development. When a flowering bamboo is snis in dn: SENE a and leafy non- 1 nearby, do not assume that the ie are the same species. If uncertain, collect each plant under a sep- arate number and cross-reference. 6. Rhizomes (Fig. 5). a. General Characteristics. 1) Length Between Culms and Di- ameter: short and thick; long and slender. 2) Habit: specialized as props for culm; running overground: running un- ound ergr : b. Occurrence of Buds on Rhizome Neck: present (Figs. 5c, d); or absent (Figs. 5a, b). c. Position of Roots: at the nodal e sd (Figs. 5c, d); at random (Fig. To Collect: For running types collect a sec- tion of the rhizome ca. half a meter long. For clump-forming types collect 2-several rhi- zomes and sever culms ca. 15 cm above the ground. Air-dry the sample and remove all scales (sheaths). C. COLLECTIONS IN LIQUID PRESERVATIVE. in the field and either stored this way or transferred to 70 percent ethyl alcohol, tOH. FAA is prepared by mixing 90 parts of 50 percent EtOH, 5 parts of glacial ace- tic acid, H(Ac), and 5 parts of formalin. To make 1 liter, use 900 ml of 50 percent EtOH, 50 ml of H(Ac), and 50 ml of for- malin. First add the H(Ac) to the EtOH ANNALS OF THE MISSOURI BOTANICAL GARDEN N [VoL. 70 and then mix in the formalin. After sev- eral days the solution undergoes esterifi- _ cation and emits a sweet smell, which in- dicates the solution is too old for fixing materia . Plant Parts for Preservation. | a. Branch Leaves: the blades and portion - containing the petiole and ligule are thè © important parts to collect. On sm | leaves cut the blade just above the bast, _ on larger ones cut a section 3-5 cm | long from the mid-portion of the blade, including the midvein. On exception- ally large blades cut a strip (3-5 cm wide) that includes the midvein and one side of the blade. For the petiole and ligule portion, cut the leaf st above the base of the he and just below the apex of the s | Buds at Different Sd E Deor Zi present (Fig. 3a) but sometimes moè — than one occurs at a node (Fig. 3): | The base of the sheath is often thick ened and remains as a ring, or girdle, around the bud (Fig. 3a). Any excess - culm tissue from around the bud €? | be trimmed away | Fleshy Fruits: iu may be collect | in a separate bottle and sliced into se | tions if unusually large. jon 9 & Roots: sections ca. 2 cm long be cut for preservation. et Seedlings: only young ones, with p seed still attached, should be collecteó. Culm Sections: some may be P^ served if they exhibit unusual colors patterns, such as mottling or Siri that would be lost in air-drying m Por . Cytological Studies: young spikelets pr be fixed in a 3:1 solution for 24 hours then transferred to 70 percent EtOH hes kept under refrigeration when poss" — The 3:1 solution is made by mixing 3 vé of absolute EtOH and 1 part of po Spikelets should be collected from 1? b rescences that are still covered by the 9! tending sheath or just beginning to emere (Fig. 4h). —— A GUIDE TO COLLECTING PANDANACEAE (PANDANUS, FREYCINETIA, AND SARARANGA) BENJAMIN C. STONE! The family Pandanaceae is paleotropical and consists of three genera, Pandanus L. ex Stickm. with the most species (over 600) and the widest distribution, from W. Africa eastward through- out the tropical areas to the Pacific Islands; Frey- cinetia, with about 180 species and the next larg- est distribution, from Ceylon (but not India, except the Andaman and Nicobar Islands) east- ward through the Malesian area into the Pacific A ae Zealand: and Sararanga, a restricted Ku. genus of two species, one in the Philippines the other in New Guinea and the Solomon Islands. hid with stems reaching perhaps 6—7 cm di- : yr leaves of perhaps 150 cm length and rather sm ie thaps 10 cm; but most species are Bd na er and some are really quite small ee extreme perhaps being Freycinetia species of ve leaves only 2-3 cm long. Bot with rather araranga are arboreous, branched, ànd erect t massive leaves 200 cm long or more, ns, uolui They have, unlike all other pan- are Lace paniculate inflorescences that The genus p and may be well over 100 cm long. ious spiral aa eadily recognized by its quadrifar- section P yllotaxy, and concomitant square- fruits are me rachises. The flowers and bein also unique in the family, the latter er wi $ : boreous E range of form and size, with tall ar- is mir ants such as Pandanus antaresensis of nea, which may reach 33 m height and vem : thick. assive proproots 10 m long and 20 cm developed into epiphytic shrubs, some with a pseudo-lianous type of growth, others cespitose and *stemless" and most of these are small shrubs, although one species (P. epiphyticus) is, apart from its reduced stem, quite massive. The family is important in several regions wherein it has developed a high degree of ende- mism and contributes to the fundamental struc- ture and physiognomy of the vegetation. In other regions its presence is limited to coastal areas. Only in the Philippines, New Guinea, and the Solomon Islands can all three genera be found occurring together. Outlying posts of the family include New Zealand, which has only Freyci- netia, and only one species; and Sao Thomé Is- land (off Angola, W. Africa), where one species of Pandanus occurs. Central areas of endemism are Madagascar, which has only the genus Pan- danus but about 100 endemic species; Thailand, which has both Freycinetia (but only two or three species) and Pandanus, the latter with perhaps over two dozen species; Malaya, with eight species of Freycinetia and about fifty of Pandanus; Bor- neo, with two dozen species of Freycinetia and more than fifty of Pandanus; the Philippine ar- chipelago, with about the same numbers of both genera as Borneo; New Guinea, with probably about 60 species of Freycinetia and 70 or more of Pandanus; the Solomon Islands, with about 23 species of Freycinetia and 28 species of Pan- danus; and New Caledonia, with about 14 species of Freycinetia and 21 species of Pandanus. Some important secondary centers include Australia, especially Queensland, with only four species of Freycinetia but a somewhat richer Pandanus- flora (possibly 15-20 species); Fiji; Mauritius; the Seychelles; Burma and the Himalayan foot- hills; Sumatra; and East Africa. Besides these, several small regions have one or a few local endemic species, usually of Pandanus to the ex- clusion of the other genera; examples are Lord Howe Island, Christmas Island (Indian Ocean), Hainan, etc. L De i partment of Botany, University of Malaya, Kuala Lumpur, Malaysia. Ann, Missouni Bor, Garp. 70: 137-145. 1983. 138 MORPHOLOGY Basic habit has been described above. In all species, the leaf is simple, usually more or less tice often long linear-attenuate, but (most- ly in Fi t The margins of he leaf are usually denticulate, as is the midrib on the undersurface. In Pan- danus (only) many species also have teeth (spi- nules, aculeations, denticuli, prickles) on the up- per (ventral) leaf apex, along the two pleats; their presence is a major taxonomic character and is a principal reason why good, intact material of leaf apices must be present in all pandan collec- tions. The size, spacing, form, color, and other details of the leaf teeth are also of taxonomic importance. In Freycinetia, the leaf-sheath is elaborated by the presence of a pair of auricles which are membranous flanges of the leaf- base; these are more or less distinct and taxonomically significant, and careful attention to obtaining good specimens is required. In Pandanus, such auri- cles are almost always lacking but a few species possess them and in any case good complete leaf- bases are a requisite for adequate representation in the herbarium. As the leaves are often very large (in a few cases, up to 9 m long and 21 cm wide), it is often impossible to collect more than a few. In fact, most herbarium material consists of a single leaf, or less, except in the case of very small plants such as Pandanus herbaceus. The collector must attempt to represent leaf variation in his collection. (Methods are described below.) e trunk and proproots may furnish some taxonomically useful data, but are often either not collected, or may be difficult and bulky. Slabs of wood, with bark, from the main trunk, can be obtained, and segments of proproots. Certainly bi E Norare there proproots in the Sararanga species. Epiphytes commonly produce feeding axillary roots. The inflorescences are invariably terminal. However, they may terminate a normal leafy shoot, or a specialized lateral or axillary shoot. The latter differ from normal leafy shoots in pos- sessing a reduced number of ordinary leaves, or 2-ribbed prophyll at the base. Species of Pan- ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 | danus and Freycinetia may have such lateral in- florescences (examples include P. danckelman- nianus and P. halleorum, and F. funicularis). Therefore it is very important for the collector to represent, prefer ably by means of the collec- | | ocument it. The inflorescences of Sararangaare always terminal. | The construction of the inflorescence varits with the species. However, all may be descri as a branched or, rarely, an unbranched, spud | The most important consideration is the strict unisexuality of the plants; a given individual 5 male or female and can therefore bear only om sexual kind of inflorescence. One of the mos important tasks of the collector, therefore, is 0 attempt to obtain both male and female inflores cences of any particular species. Unfortunately. | this is by no means easy. Staminate material is poorly represented in herbaria; but correlat staminate and pistillate materials judged 9 proven by the collector to represent the sexes? a single species are even rarer. Isolated pn] tem is based upon characters of t dE plants, such species are often fascinating bu d sedis" or ev iis The role of the staminate material i$ sd | er becoming increasingly important as M° : is Wer ofthe micromorphology and a and in some cases can already be of critical pi nificance in taxonomy; but good co of loca still rare. One reason is the difficulty of loca staminate inflorescences. As flowering 1$ ts al in most pandans, while fruit developme? b a lengthy process, there is a much higher ability of finding pistillate trees with partly \ ‘ only one to three days; while the flowering may only be a week or two. In contras development may take several months. P ogy of flowering in the pandans is still " known. However, there are certainly dn | terns already in evidence. In some cases (Part ely danus fragrans) flowering may be pue | every other month. In other cases, flower! | 1983] STONE—COLLECTING PANDANACEAE 139 be at intervals longer than one year. Many species are probably gregarious with synchrony in flow- ering, but the period may be very short. Robert Tucker in Queensland has reported that annual flowering in Pandanus conicus and P. tectorius is highly con lina two-week period, with in fact most anthesis taking place in the same week, Much of the rest of the year, the plants appear to be sterile, except those that are ges- tating fruits. Such patterns of flowering mean that collectors often encounter flowering pandans at random and by chance, and find ripe flowering materials very much less ly than unripe, or even ripe, fruits. However, it should be re- membered that if staminate individuals are seen to be in flower (they are usually the more ob- vious, as the infi lax and pen- dulous, and the white or colored spathes are no- ticeable), then it is highly likely that searching will reveal pistillate plants in the same locality also in flower. Such searches are of the greatest hice in linking staminate with pistillate ns to establish full knowledge ofa species. ut a caution must be noted: there are many Pio In which several Species of the same Mis sh sympatric and intermixed, and ex- EL a ves be taken to establish vegetative bes to of the sexes. The danger of attrib- Wit ering male and female plants to the Pecies, but in fact having material from male : s Nes of one species and females of another, is Quite serious. Mc apa inflorescences are spikelike. ., >? l€re are several spikes, each with an Cases eii from the main rachis. In a few aDparentl in orescence is reduced to a single, Y terminal spike, though there are al- Wa E] Y atn bracts. (The bracts are sometimes s from stami ®ation o Perianth, al nate flower consists of a small or large ag- variation Eres and there is a considerable SParately a " rm; the stamens may seem to be 'àchis, or a individually attached to the spike i borne in small or large clusters on ort ki or long axes (usually called columns or ste- Pet nores), Poll Nitiduia 1^ always and everywhere to include Toving, beetles and thrips, supplemented by Predatory earwigs. Pollen consumption is rapid and unless the collection is made early in anthesis and is quickly and adequately pre- served, most of the pollen will be turned into frass and larvae. Numerous herbarium speci- mens of staminate collections now serve only to show floral form (if that), and lack pollen. Pollen characters are proving to be taxonomically use- ful, so care should be taken to preserve pollen in the specimens. If the collector has the facilities, minate spikes should be obtained as supplemen- tary to the dried material. The same is true of early stages of anthesis of female inflorescences. The female inflorescences may also be simple or compound. If one applies this terminology to the gynoecial pattern of the flower, the same dif- ference obtains. Consequently one may recog- nize four types of pistillate inflorescences; (a) with a single spike (head, or cephalium) formed of l-carpellate flowers; (b) with several spikes formed of 1-carpellate flowers; (c) with a single spike formed of polycarpellate flowers; (d) with several spikes formed of polycarpellate flowers. In cases (a) and (c), the inflorescence is describe as “solitary.” This is the normal state in many, but not the majority, of Pandanus species, but is quite rare in Freycinetia. In cases (b) and (d), the inflorescence is “spicate”’ and this is the usual case in Freycinetia, and quite common in Pan- danus. In Sararanga, as already mentioned, the massive inflorescence is always a richly-branched panicle. In Freycinetia, the spikes (both male and fe- male) tend to be closely adjacent and often ter- nate, so that the open ripe inflorescence seems to be an umbel. In a few cases (F. angustifolia, F. jagorii) the inflorescence is racemiform, as it is in nearly all species of Pandanus in the males, and in many species also in the females. The bracts provide some useful characters, particularly as to color. On an inflorescence there is a spectrum of bracts, those at the base virtually leaflike, those subsequent each more altered, by the reduction of the length and the spread of the lower, colored part, until even the extreme apex is colored rather than green. The base is usually broadly expanded and boat-shaped, and the tex- ture thinner or, especially in Freycinetia, actually thicker but much softer. In Freycinetia, the up- permost bracts are quite modified, becoming (in the apt phrase of Paul Cox) "solid bat nectar, i.e. soft, sugary, and entirely palatable. This does not occur in Pandanus, in which all bracts still retain a certain (if thin) stiffness. (The bract tex- 140 ture in Sararanga is unknown.) As the inflores- cence ages, the bracts wither and many drop off or remain as tattered brown shreds and wisps. In Freycinetia, only a few basal bracts are re- tained and in ripening fruit these finally fall; most of the upper (visually, inner) bracts by this time have long since been eaten (around anthesis) or rotted away (later). But in Pandanus, the inter- mediate bracts usually wither and dry (turning brown) but remain in place on the ripening fruits; ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo 7 ' of young flowering pistillate material are highly | desirable. However, when dried, this is some ' times so altered in aspect, lacking in useful marks, | and in interior features, that it may be uniden- tifiable. Just as it is important to correlate male and female specimens, it is important to correlale fruiting with flowering pistillate collections. In the drupes or phalanges, the style and stigma , is of crucial importance taxonomically. In some | species, tl truct ither fragile (in many, | or, in the staminate inflorescence, dry up. The d: = ee 5 pd +. A 1 1 TT jo 2.9 E so that the inflorescence tends to break up as, or before, it falls. As already stated, this process is very rapid, taking only a few days, hastened by rapid chewing up of the softer parts by the local fauna of the inflorescence. In short, the collector usually only sees the full array of bracts in ma- terial at or before anthesis; and the colors are best, or only, visible then. The color, size, and details of the bracts afford a rather important means of correlating male and female of the same species; not that bracts of the two sexes are ex- actly alike, but they will be substantially similar. Naturally, collectors’ observations on fauna, in- cluding possible pollen or seed vectors, will be of great value. The fruits provide the most depended upon, readily observable, and preservable characters for identification. The components of the in- fructescence, besides the bracts and the rachises, are the heads (cephalia), or syncarps as they are sometimes called; this term however is rather ambiguous, as it can also be applied to the ag- gregate drupes, also termed polydrupes or pha- langes, which are the main component units of the cephalium in those species which have them. In other species, with simple (1-celled) drupes, the term is not ambiguous but as it has been used for both cases, I prefer to abandon it and use the y; ee. phalange dimensions. Unripe fruits also lack the critical mature coloration which can prove valu- able taxonomically. On the other hand, well-pre- served (particularly liquid-preserved) collections fragile ( y | they can be easily abraded or broken off), or the) are severely altered by the drying process. In such cases close observation and sketches or sharp close-up photographs may preserve information — later lost. | The leaves, apart from gross morphological data, provide a wealth of useful taxonomic In- — formation in their micromorphology. Data from | anatomical and scanning electron microscop? | studies have shown that highly significant ad- | ditional and correlative information can be &* - tracted from a small sample of a leaf. Howl this has to be based on a standardization. = | such studies the sample must be: a ee | the leaf, about 1-3 cm long (i.e. along the | | long axis), and including both margins, L6 5 | complete median segment; or in large leaves 0 with minimal materials, a segment including one margin and the entire midrib, but omitting the other margin, will ordinarily suffice. This -— | must be taken from a fully adult leaf, mide | the prophyll-leaf spectrum (i.e. representing i | series of longest leaves). It must also be from: | adult shoot (not a *sucker-shoot" or à flowering from herbarium specimens, but it is more r cult in such cases to be sure that the standa pi ization criteria have been met; hence, — J leaves taken in the field might well be ma™ m specially “for anatomical study.” (Anao using herbarium material would appreciate Re | Separate pollen collections (preserved of ^. 4 oughly dried and disinfected) are highly d?" able, as a great deal of pollen is lost through | usual drying techniques when making herba specimens. meni Correctly labelled liquid-preserved spec " of heads or partial heads at anthesis are h | useful in the interpretation cf stigmatic pir and this supplementary collecting iS $ | 1983] urged upon collectors who have the means avail- able. (If FAA is lacking, strong alcohol or just formalin may suffice.) FIELD COLLECTING METHODS AND SAMPLING Small pandans of either Pandanus or Freyci- netia can often be treated as any other plants, as whole branches with attached leaves, inflores- cences or infructescences, and sometimes part of the root system, may be small enough to fit in standard size drying papers. Where inflores- cences and infructescences are concerned, how- ever, the advisability of providing supplemen- lary liquid-preserved collections must be emphasized, For bigger plants of any of the three genera, some careful sampling has to be done. Where the leafy stems are more than 2-3 cm in diameter, Aa split longitudinally. Such a method bit ib a reducing the thickness of specimens dide " disadvantage of cutting through the sedia rede in such a way that useful char- Theses e leaf-bases may be damaged or lost. is thee €, what is always required in such cases kei mm preservation of carefully detached io thee in separately. They should be removed striding a sheath is intact (especially with and forth Fi eycinetias); they can be folded back » accordion-style, to fit into the drying er ones es away a series of leaves (old- tesh (ie until the leaf-bases of leaves with view: igs dried or withered) auricles are in : eaves should then be removed by Circumfere i a individually. Shallow slicing, and preserved o iq and Sararanga the adult leaf- large: they n very rigid and inconveniently do n pe flattened readily if they are Ins 3e cis (l git di lly divided). Bid E ry long leaves, leaf-length mea- ents: ; n be supplemented by standard leaf > IN such cases the segments have to be STONE—COLLECTING PANDANACEAE FIGURE 1. Types of leaf apices in Pandanaceae. A. Upper (ventral) surface of a species of Pandanus with denticulate apical-ventral pleats (avp). B. Upper i f Pa cal-ventral pleats. C. Lower (dor: apex showing the armed midrib. All the denticulations are antrorse. numbered or lettered sequentially so they can be “reassembled” later for study. The collector has to make a pragmatic decision about how many leaves to collect; since each herbarium specimen may consist of one leaf, or less, and still be so bulky as to spill over onto two or more sheets, it is especially important that the collector pro- vide careful and logical sampling techniques in the field. A choice of organs [prophyll, some scale leaves, some transitional bracts, fully-formed bracts, etc., as well as fully adult leaves, imma- ture leaves, “sucker-shoot” leaves, and (when possible) seedlings] must be made so that a fuller nowledge of both leaf structural and dimen- sional variation, and age and developmental variation, is preserved inherently in the speci- mens. It cannot be overemphasized that such careful sampling techniques can be disrupted CSS. as : ead Ant aridi die tributed; if at all possible, the bulk of the collec- tion, or the most representative, should be re- 142 FIGURE 2. Types of leaf-bases in Pandanaceae. D. Pandanus and Sararanga. Leaf-base wi o (or usually no) auricles; midrib near the base often (not E. Freycinetia. Leaf-base with auricles (a), and un- armed basal portion of midrib. tained until taxonomic study is completed. Duplicate specimens often appear which contain (for example) one prophyll, three phalanges, and a single leaf apex; needless to say, the form and the important diagnostic features of the plant cannot possibly be represented by such an ag- gregate of parts The goal of sampling should be to represent all possible organs and structures in such a way that a part, supplemented by notes, drawings, and photographs, can provide an accurate notion of the entirety. But this goal is more difficult when it comes to the whole habit of the plant, more so when variation in a population is found. However, thoughtful consideration of this prob- lem usually results in some functional ad hoc technique that works fairly well. The question of intra-population variability is, at least in some species or complexes, a serious matter, which needs intensive study, in view of the sometimes drastic differences in taxonomic points of view expressed by different botanists concerned with the same species or species-groups. Thus, col- lection of a series of individuals (all, of course, carefully correlated and separately numbered) ANNALS OF THE MISSOURI BOTANICAL GARDEN Tm. \ X i Á i fine UN ) NN e FIGURE 3. Habitin risate P- erect MR proproots (pr). S = Sararanga; palmoid roots at base (par). F — Fi wishes rect trunk. ec stem with clasping roots (cr). l is within a pragmatically determined ps i and certain groups of Freycinetia. COLLECTING TOOLS AND EQUIPMENT | (1) A strong, sharp parang or machete of - | knife; for very large specimens, a hatchet 7 also be useful. ees Pruning shears or Sequitents E d | ready to attack). | (4) Field notebook. (5) String tags (jewellers’ tags). These 26^ necessary especially when leaves etc. are mented sequentially. A small stapler is also = | kd (6) Pencils and (preferably) ladini apne «4 latter are extremely useful for labels to be place in liquid-preserved material. | (7) Plant presses— the usual materials. | 1983] Reina leaf. ray 2 = aod cales and normal folia age). E : 7e lind normal foliage leaves. tb — transitional (infor (8) Liquid ; i iini preservative supply. and vials for liquid-preserved spec- "n 10 oon fungicide (especially for pol- (11) Ca and film su (12) Skee i D. hts ok (if additional to field note- (13) Plast (14) Corda Le (15) Final f ede LIVING COLLECTIONS for collecti of almost any species are desirable 9n as ripe seed, small germinating dens with Or cuttings. There are few botanic gar- 800d collections of the family [among i ; STONE—COLLECTING PANDANACEAE 143 FIGURE 5. Types of inflorescence and fruit for- mation in pane: eae. A. S . B. Spike of cephalia. C. Pseudumbel of (ternate) cephalia (Frey- cinetia). D. Raceme of cephalia (Freycinetia). E. Poly- rmous berry of Sararanga (black dots are stigmas). rupe — ph a- la of Panda : (bicarpellate, multi- ovulate) of Freycinetia. I. Pyrene of Pandanus. J. Seed of Freycinetia. K. Seed of Sararanga these may be mentioned first those tropical gar- dens that have outdoor collections, e. g. Bogor, and those which have above average indoor col- lections, examples being Kew, Munich, and per- haps some others]. Even old established gardens tend to have very few species and frequently only one individual of a species (which, as they are unisexual, tends to be lost in due course). Pan- danus seeds tend to be fairly long-lasting and, wrapped in moist sphagnum, can be ai irmailed almost anywhere and survive; germination tends to be a long affair. Freycinetia seeds tend to ger- minate more rapidly; little is known of their hor- ticultural requirements. Except for the Lae Bo- tanic Gardens, the genus Sararanga is (I think) quite unknown in cultivation; further distribu- 144 ANNALS OF THE MISSOURI BOTANICAL GARDEN FUNGICIDE INSECTICIDE POWDER poe FIGURE 6. Some suggestions for collecting specimens of Pandanaceae. Above: — sir segmented, ^" segments num and some of the phalanges or dru note-takin tion, specifically of S. philippinensis, and of both sexes of both species, is highly desirable. The pandans in general tend to be horticulturally in- teresting and often curious or unusually elegant, and tend to cause considerable public interest. The species that are already comparatively well known in cultivation are few and can be con- veniently listed here so collectors can (if they wish) avoid the slightly more elaborate task of collecting and distributing living material for propagation: Pandanus tectorius, P. odoratissi- mus, and most of the “horticultural” varieties of mbered sequentially with both the collector's number and, the reat eran leaf folded back and forth à la saisis dean to fit on the rved; staminate spike similar div h perapi in "m with nicis: nd fungicide powders. All treatments sequential to photograP y r. Below this t | ! heet. Below: fruit longitudinal dive | len ided, parts also preserv' a ] both of these (especially the striped, variegated leaf forms); P. utilis, P. pygmaeus; and p. dubi The | (known in horticulture as “P. pacific ificus”) a only freycinetias in cultivation tend to be F nicularis, F. multiflora (of F. cumingiana) #7 A F. sumatrana, and even these tend to n ed to tropical gardens; nearly any Freycineti can only be deduced from natural habita ditions, so the seed distribution should include useful notes on these. Many 1983] species are rather indifferent to a number of pa- rameters, but others tend to be very finicky. Crit- ical factors tend to be soil pH and soil moisture retention (swamp pandans tend, as might be ex- pected, to be ecologically very narrow-minded, While savanna species are pretty tolerant). At- mospheric humidity and nocturnal temperature regimes may also be very important. So little is known in detail, however, that most attempts will have to be regarded as experimental, partic- ularly in habitats fairly unlike the origin. Palm species tend to be fairly good comparisons; if a pangan from the same region and basic habitat 1S grown under the same conditions as the palm, there is probably a good chance of succeeding with the pandan as well. IDENTIFICATION OF PANDANACEAE part from a few major herbaria, advice and entiation can usually be obtained from the airly small number of specialists on the family, STONE—COLLECTING PANDANACEAE 145 of whom (I believe) I can append the following virtually complete list: the author (B. C. Stone, Herbarium, Botany Dept., University of Malaya, Kuala Lumpur 22-11, Malaysia); Dr. H. St. John, Robert Tucker, Anderso C Townsville, Te Australia (especially for Australian species), D .-L. Huynh, Institute Botanique, Université B8 Neuchatel, Switzerland (particularly for determinations based on sterile or staminate collections of unknown correlation, through microtechnical methods). Personally I am always ready to receive and identify specimens of Pandanaceae from any part of the world. GUIDE TO THE ILLUSTRATED SECTION A number of sketches are attached with the idea of familiarizing the reader with the basic parts of pandans, some of their gross variability, and some special collecting techniques. STUDIES ON THE REPRODUCTIVE SYSTEM OF NIVENIA CORYMBOSA (IRIDACEAE), AN APPARENTLY ANDRODIOECIOUS SPECIES! ROBERT ORNDUFF? ABSTRACT zo ia COME has flowers Mis a stamen- and style-length dimorphism but pollen mono- Results o co ne lon The small South African genus Nivenia is of interest because it is one of two genera of Iri- daceae in which heterostyly has been reported (Ornduff, 1974). Three species of this small woody genus are reportedly distylous with one cahy, 1965). At least three species of Nivenia are omorphic with a floral morphology resem- bling that of the long-styled form of the distylous species (Ornduff, 1974). This paper presents the results of an artificial pollination program using the dimorphic N. corymbosa (Ker) Baker that was designed to determine the presence and na- ture of incompatibility in this species. MATERIALS AND METHODS The crossing program utilized four field-col- lected short-styled plants of N. corymbosa and one short- and four long-styled plants derived by selfing à long-styled plant. The terms “long” and “short” will be used in this paper to refer to these two morphs, but in view of the apparent andro- dioecism of this species, they are not equivalent to the usage of “pin” and “thrum” of distylous species. The material orginated from Bains Kloof, Cape Province, South Africa (Ornduff 7666, UC). Plants were grown in an insect-free greenhouse in Berkeley and pollinations were made usin fine-tipped forceps. Self-, intramorph, and inter- morph pollinations were carried out, and as con- trols some flowers were left unpollinated and rmorph crossing dd m ie Hae this species is andro- dioecious rather than covet oiially distylous A single pr ned 1:1 :sho a selfed long-styled plant rts, suggesting that the longs of. hd Adasen y ede condition are the ka morph rather than the AE on others were emasculated. One long-styled plant was selfed and the style length of the progeny scored. As capsules matured, they were placed in seed packets and the number of seeds was counted at the end of the program. Pollen via- bility, size, and wall sculpturing were determined using pollen grains from two plants of each morph mounted in aniline blue-lactophenol or Jacto- phenol. RESULTS The pollen viabilities of the field- collected plants ranged from 92 to 100 percent and of the progeny grown from seed from 56 to 95 percent. Pollen size was monomorphic: for two longs 37.7 um (e = 1.8) and 37.3 um (c = 2.2) two shorts X 37.7 um (c = 1.8) and 41.4 um (0 * 2.1). Wall sculpturing was also monomorphic. None ofthe emasculated or intact control p ers produced seeds (Table 1). Self-pollinations 0 shorts produced no seeds and of longs produ an average of 0.04 seeds per pollination. Intra" morph pollinations of shorts and longs pr an average of 0.2 seeds per pollination. me morph pollinations of shorts as seed paren duced an average of 0.3 seeds per poll of longs as seed parents 3.4 seeds per pollin The single progeny grown from a selfed long OF tained 6 longs and 6 shorts. DISCUSSION UR in The presence of an incompatibility pase Nivenia corymbosa is evidenced by the lo set following intramorph pollinations es ; esearch, pported in part by grants from the National Science Foundation and the Committee on R rk ! Su University of California, Be - Pepa ment of Botany, University of California, Berkeley, California 94720. he che er genus for whic which plants of some popula of A heterostyla L. Bolus e dimorphic flowers that are not conventionally distylous (Ornduff, 1974). ANN. MissouRi Bor. GARD. 70: 146-148. 1983. | | | 1983] TABLE 1. Results of nom and intermorph pollinations of Nivenia corymbos - Num- Number Number be r Seeds Flowers Cap- Seeds p Uti- sules TO- olli- Cross! lized Produced duced nation Controls: SP1 (intact, unpoll.) 11 0 0 0 LP2 (intact un 21 0 0 0 LP2 (emasc., unpo 18 0 0 0 LP3 (intact, un 16 0 0 0 Self-pollinations Fl 49 0 0 0 SF2 69 0 0 0 SF3 92 0 0 0 SF4 56 0 0 0 LP2 52 l 3 0.06 LP4 15 0 0 0 Intramorph cross-pollinations: SFI x m 14 5 26 1.9 23 0 0 0 SF2 x SF 26 0 0 0 SF4 x SP] 55 5 5 0.09 SPI x SF4 28 5 2 0.07 LP3 x Lp? 53 3 9 0.2 Intermorph pollinations: ore x LP2 8 22 0.37 3 XLP2 29 4 3 0.1 = X SF4 56 21 230 4.1 P2 x SF] l 1 4 4.0 X SF4 41 20 115 2.8 LP3 x SF4 31 12 87 2.8 - ie e nn opea of individual plants are used: F €d plant, P = progeny of selfed field-col- lected plant: t L= long, "i : bus y with that following intermorph pollinations of Proximate] ing intermorph pollinations is ap- Pollinati Y €qual to that following intramorph tons and does not exceed 5 percent of the vty on seed set of longs. The pollen int one short used as a seed parent in an T Eo of these crosses probably 9t be attributed to gametic sterility fac- ORNDUFF-— NIVENIA 147 tors. It is also possible that this difference in seed set between the two morpha : is attributable to mechanical probl pol- len grains on short stigmas. This does not seem to be a likely explanation for these differences, since observations of pollen loads on naturally- pollinated stigmas of N. binata (a species with a visually between intra- and intermorph pollen grains; Ornduff, pers. obs.). If such figures are a measure of normal stigmatic pollen loads carried by fecund individuals of the distylous N. binata in the field, the number of compatible pollen grains per stigma required for modal seed set of both longs and shorts is very low and would likely be deposited on stigmas via artificial pol- linapony. 5 third possible explanation of the ob- tin N. corymbosa is that the species is is essentially ‘androdioecious, that m mue in- duce staminate flowers. The data at hand, in- cluding the occurrence of pollen monomorphism in this species, are compatible with this sugges- tion. That hort ionally prod ll numbers of seeds suggests that androdioecism is not fully developed in this species. Various types of reproductive systems involv- ing a sexual separation have evolved from het- erostyly in other groups. These include full dioe- cism in some members of the Rubiaceae and Menyanthaceae and androdioecism in Oxalis suksdorfii (see Ganders, 1979). That various modifications of the distylous breeding system have evolved elsewhere in Nivenia is evidenced by the occurrence of long-styled monomo orphism in at least three species with apparent associated self-compatibility in N. stokoei (Guthrie) N. E. Brown (Ornduff, 1974). The 2:1 rather than 1:1 long: short morph ratio of a field population of the distylous N. binata also suggests that the breeding system of this species may deviate from that expected for a conventional distylous species (Ornduff, 1974). It is of interest to note that the progeny of a selfed long-styled plant of Nivenia corymbosa produced a 1:1 ratio of longs to shorts. Assuming that the androdioecism of this species is derived from a more conventional distylous ancestral condition, this suggests the possibility that in Nivenia longs are the heterozygous genotype, a 148 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo.. 70 condition otherwise known for certain in the di- stylous Hypericum aegypticum (Ornduff, 1979) and also possibly in the dimorphic Armeria mar- itima (Baker, 1966). For those interested in the comparative evolution of breeding systems and in the selective forces that lead to the breakdown of heterostyly, Nivenia clearly merits further at- tention and field studies are particularly needed. LITERATURE CITED BAKER, H. G. 1966. The evolution, functioning and breakdown of heteromorphic incompatibility sys- tems. I. The Plumbaginaceae. Evolution 20: 349- 368. | Ganpers, F. R. 1979. The biology of heterostyly. New Zealand J. Bot. 17: 607-635. Mutcany, D. 1965. Heterostyly within Nivenia (Ir -351. OnNDurr, R. 1974. Heterostyly in South African flowering plants: a conspectus. J. S. African Bot. 40: 169-187. 1979. The genetics of dd in Hypericum SOSIMcHN. Heredity 42: 271-272. a ————— EE ——— T —— A TAXONOMIC REVISION OF CONDYLOCARPON (APOCYNACEAB)! MARY E. FALLEN? ABSTRACT Condylocarpon, which has a center of distrib l Here the number of species is reduced from 17 t Condylocarpon intermedium Muell.-Arg. subsp. laxum (Muell.-Arg.) Fallen is made. collected and thus is not well known. Peculiar petal appendages, fruit type, and the shape of the style-head are shown to d the ution in Brazil and the Guianas, has been infrequently 0/8 be important taxonomic characters that can be used to define two basic groups within the genus. Condylocarpon is a small neotropical genus of Apocynaceae occurring in disjunct habitats from Nicaragua to Brazil. Five of the species are en- demic to Amazonian Brazil, the Guianas, and adjacent regions, suggesting an Amazonian ori- gin for the genus. Although some of the species are geographically wide-ranging, all tend to be ecologically restricted to special life zones. Most are members of the lowland wet or moist forest community. One occurs in the more xeric cer- rado community; another, the only species of Condylocarpon to occur outside continental South America, has fruits that show adaptations for salt s matio than 8 Alth collecting along rivers that are important trans- the Toutes, it suggests that these indehis- veer. ^ god be broken apart and dispersed cab rends of diversification are illustrated ration of the corolla lobes and modifi- Scri H M IE Since 1860, five are reduced to synony- Em. — EE 1 rich for their assistance, and in parti T constant support and uidance. TAXONOMIC HISTORY Condylocarpon was validly published by Des- fontaines (1822), who described the species C. guyanense based on a fruiting specimen sent to him by Joseph Martin, director of the Jardin de Botanique de Cayenne, French Guiana. The de- scription is quite accurate with the exception of his interpretation of the areolate tertiary vena- tion on the lower leaf surface as groups of scales. In the Prodromus, DeCandolle (1844) recog- nized two species of Condylocarpon. In addition to C. guyanense, he correctly assigned Echites isthmica (Vellozo, 1829) to Condylocarpon and made the new combination. He also described Maycockia, honoring Dr. James Maycock, a physician and naturalist, who in the early part of the 19th century compiled a catalogue of the appeared in Vellozo's (1829) Flora Fluminensis, the poor quality of which was noted by - Candolle (1844) in the new combination, as well as by Mueller-Argoviensis in his later treatment ofthe group in Martius' F/ora Brasiliensis (1860). Although the plate of Echites isthmica includes flowers as well as fruits, the flowers are so am- biguously depicted that little more can be dis- cerned than that the corolla is 5-parted. The type, therefore, must be considered functionally as a fruiting specimen. Thus at the time DeCandolle described the genus Maycockia, the only two species known to belong to the genus C. ondylo- Supported b ; ; EESE Institut für Systematisch : y a Fulbright/Swiss Universities Grant. I thank the staff of the Institut für Systematische potanik-Zü raka ; icular, I thank Professors Friedrich Markgraf and Peter Endress, ai g ; Present aut für Systematische Botanik der Universität Zürich, Zollikerstrasse 107, 8008 Zürich, Switzerland. per Institut für Allgemeine Botanik, Ohnhorstrasse 18, Ann. MOURI Bot. GARD. 70: 149-169, 1983. D-2000 Hamburg 52, Federal Republic of 150 carpon— C. guyanense and C. isthmicum—were known only from fruiting specimens. Since the type of Maycockia was based on a flowering spec- imen, it is understandable that DeCandolle failed to recognize the two genera as synonymous. In 1851, Miquel published the description of an apocynaceous liana collected in Suriname as the type of a new genus. He noted that it was a peculiar genus showing affinities to Tabernae- montana and Ochrosia and he named it after Nicolas Hortsman, an early explorer in the Guianas and Pará, Brazil. Hortsmania is syn- onymous with Condylocarpon. It is possible that Miquel failed to recognize this specimen as be- longing to the genus Condylocarpon due to the unusual fruit development of this species, in which only one of the paired carpels develops, with only the lowermost seed in that carpel ma- turing, so that instead of the usual paired, multi- articulate fruits, which are characteristic for most species in the genus, C. myrtifolium has a soli- tary, single-seeded fruit. Mueller-Argoviensis’ monograph of the genus (1860) appeared in his treatment of the Apocy- naceae in Martius’ Flora Brasiliensis. He was the first to connect Maycockia and Hortsmania with fruiting specimen collected in Para, as Anechites (?) amazonica, it was raised by him to generic rank when he received flowering material in 1930. Markgraf realized that there was a close affinity between Rhipidia and Condylocarpon, but he recognized the former as distinct on the basis of its short corolla lobes, high placement of the sta- mens on the corolla tube, ovarian disc, and hairy, filiform fruit. Rhipidia was reduced to synonymy under Condylocarpon by Ducke (1943), who at that time described two new Amazonian species of Condylocarpon, C. hirtellum (= C. pubiflo- rum) and C. reticulatum (= C. amazonicum). GENERIC RELATIONSHIPS There is, unfortunately, no standard intrafa- milial classification system for the Apocynaceae. Schumann's treatment of the family in Engler and Prantl’s “Die natürlichen Pflanzenfamilien" (1897), and the studies put forward in a number of papers by Pichon (19482, 1948b, 1951) are ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo.. 70 the most recent comprehensive studies of the family. Condylocarpon is a member of the large, het- erogeneous subfamily Plumerioideae by virtue of its completely fertile, basally unappendaged anthers, which are not adherent to the style-head. Following Schumann, the apocarpous, bifid fruit of Condylocarpon put it in the tribe Plumerieae. He splits the Plumerieae into four subtribes: Al- stoniinae, Tabernaemontaninae, Rauwolfiinae, and Cerberinae, based, in part, on number 0 seeds per carpel. Segregation based on 2, 4, or 6 seeds per carpe (Rauwolfiinae and Cerberinae) versus many seeds per carpel (Alstoniinae and Tabernaemontanr nae) is problematic. Markgraf (1930) criticized Schumann’s placement of the “typically few- seeded" Condylocarpon in the Alstoniinae. He felt that Rhipidia (= C. amazonicum) sho affinity to Anechites by the ovarian disc and the structure of the fruit, and that these two genera, dylocarpon shows a fourfold variation in the number of ovules per carpel. Therefore, Schu- mann's system is not useful for the classification of Condylocarpon beyond tribal rank. i In his classification of the Plumerioideae, Pi- chon (1948b) split the subfamily into seve tribes, based, for the most part, on whether the fruits were dehiscent or indehiscent, and dry 0 fleshy. He maintained Rhipidia and Condylo- carpon as separate genera within his tribe p volfiae, due to differences of the "disc" of R ‘ pidia, and in their fruit structure (apparently no having seen Ducke's (1943) publication). Te gether these two genera comprise his subir! Condylocarpinae, which he notes as having problematic affinities. d I agree with Pichon that Condylocarpon p member of the Rauvolfieae. However, I do de support his system beyond tribal rank. His r^ tribes are often too narrowly circumscribed suggest insufficient study. I feel that it 1$ " mature at this time to assign subtribes within Rauvolfieae. A more in-depth study of the e involved is necessary in order to recognize di evaluate phylogenetic relationships within x tribe. I disagree with Markgraf (1930) that CO” dylocarpon and Anechites are closely related. studies of Condylocarpon have failed to € the disc that he describes. In Anechites, 0? other hand, the disc is well-developed and "i tinct from the ovary. The resemblance of the dis | | 1983] fruits of Condylocarpon amazonicum to those of Anechites is probably best explained by conver- gence. Both genera are forest lianas with inde- hiscent fruits, and it is logical that a light, one- seeded mericarp covered with hairs has evolved independently in both genera as an aid to dis- persal. The structure and differentiation of the style- head is an extremely valuable taxonomic char- acter in the Apocynaceae (Schick, 1980; Fallen, 1983). The style-head has, for the most part, been relatively little affected by habitat conditions, and, together with the structure of the anthers, it pro- vides one of the most reliable means of recog- nizing natural groupings of taxa. The style-head of Condylocarpon is unusually simple and un- differentiated. This type of style-head, the in- dehiscent fruits, and the longitudinally folded seeds suggests a close relationship with Alyxia, à paleotropical genus of about 120 species, and to a somewhat lesser extent, with the pantropical gnus Rauvolfia. MORPHOLOGY Leaves : Of the seven species of Condylocarpon rec- COT in this treatment, two (C. guyanense and : Isthmicum) have three leaves per node; the tremel | C. myrtifolium, which has ex- i ei ae leaf characteristics, the leaves ne Show considerable variability within a of all s in both their size and shape. The leaves Pecies are entire and glabrous to subgla- i T ramps The abaxial surface is glabrous los m rum, C. guyanense, and C. myrtifo- ,, and pilose only along the midvein in C. e variability. Therefore, leaf charac- Value E to be of relatively little taxonomic ployed i Ough they have frequently been em- delimitat; Specific and particularly supraspecific equate “i in earlier treatments. The more ad- vides a be rium material now available pro- ability w; Wer perspective of the range of vari- Y Within the genus, Inflorescence Thei i ie scence in Condylocarpon is a many- Only or à yrse. These thyrses may be terminal terminal and axillary. They often exhibit FALLEN —CONDYLOCARPON 151 or, leas | P 1 cyme- 4-4. or even umbellike (Fig. 1). Flowers The calyx of Condylocarpon is regular, pen- tamerous, and without squamellae on the adaxial surface of the sepals. Aestivation is sinistrorse and the corolla is glo- bose in bud in most species. The latter aspect may be partly due to the bulk of the inwardly rolled appendages of the corolla lobes. Although the corollas of all species are less than 1 cm in diameter, they exhibit variety in form as well as in relative size. A very useful character for spe- cific delimitations within the genus is the struc- ture of the corolla lobes. In some species (i.e., C. glabrum, C. guyanense, C. intermedium, and C. isthmicum) the corolla lobes are equipped with lorate appendages that are speckled or streaked with reddish brown to red-violet markings on the abaxial surface. The corollas of these species may be up to 7 mm across, but greater than half of this distance is accounted for by the append- ages themselves; the diameter without the ap- pendages is no more than 2.0-2.5 mm (Fig. 2A). The corolla lobes of the other species lack lorate appendages. In C. myrtifolium they are shortly auriculate and ascending. This is the smallest- flowered species, with corollas about 3.5 mm in diameter, including their short, acuminate petal tips (Fig. 2B). The corolla lobes of C. amazon- icum and C. pubiflorum are oblique and spread- ing to slightly ascending. The corollas of C. pubi- florum, 9 mm in diameter, are the largest in the genus (Fig. 2C). The stamens of Condylocarpon have filaments about 0.2 mm long and are inserted midway or slightly above in the corolla tube. The anthers are ovate to lanceolate and range from 0.3 to 0.5 mm in length. The thecae are fertile completely to the base. The anthers are without sterile basal appendages and are not adherent to the style- ead. The style-head of Condylocarpon is one of the least specialized that occurs within the Apocy- naceae. It is orbicular in most species (Fig. 2D), but turbinate in C. amazonicum and C. pubiflo- rum (Fig. 2E), and has two free apical append- ages. It is covered with secretory papillae and shows no regional differentiation of the epider- mal cells (Fig. 3). The style elongates very little during development, so that at anthesis the style- head is subsessile or with a style no more than 152 ANNALS OF THE MISSOURI BOTANICAL GARDEN FIGURE 1. Condylocarpon guyanense. 0.4 mm long. The gynoecium is apocarpous, being composed of two carpels that are united at their apices by the common style-head. The Ovary is conical, glabrous, and ranges from 0.3 to 0.8 mm in height. [VoL. 70 shed 3$ Condylocarpon has usually been acsi ud lacking nectaries or a disc (Miquel, 1851; ler-Argoviensis, 1860; Miers, 1878). Markgraf (1930) described Rhipidia (= E lararnan . ` K : dice a Howeve® 0 ysed 1983] FALLEN—CONDYLOCARPON 153 C. pubi m a Basic structural types of Condylocarpon. A-C, corolla: A, C. isthmicum; B, C. myrtifolium; C, endo m. D-E, style-head: D, C. guyanense, E, C. amazonicum. F-H, fruit: F, C. soa G, icu H, C. myrtifolium. LI phon f pubiflorum: 1, from the side; J, in transverse sectio FiGURE 3. Scanning electron ig of the nee nf Condylocarpon guyane this as one of the criteria to distinguish Rhipidia from Condylocarpon. Ducke’s (1943) description of C. reticulatum included a disc. He stated at the end of the description that the study of the ovary of dried material was very difficult, and that the data used in the diagnosis were provided by Dr. Kuhlmann. I have studied the gynoecia of all species in serial sections and have found no distinct disc. Nectar seems to be secreted from the ovary wall. Markgraf's and Ducke’s mistake may be due to the lobed appearance of the ovary base in desiccated material. The number of ovules per carpel is usually four, with the ovules arranged biserially (C. ama- zonicum, C. glabrum, C. guyanense, and C. myr- winds In both C. amazonicum and C. myr- tifolium all but one of the ovules in each carpel degenerate during development, resulting in a degenerates as well. Both C. intermedium and C. isthmicum usually have five or six ovules per carpel, and the specimens that I have examined usually had four or five ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo.. 70 seeds in each mature carpel. The ovules of C pubiflorum are arranged in four or five rows rath- er than biserially, and have a greater number of ovules per row than in the other species, so that the usual number of ovules per carpel is ca. six- teen. Only three of the fifteen fruiting specimens of C. pubiflorum that were available to me for study had intact, mature fruits. These had 12, 15, and 16 developed seeds per carpel, respec- tively; I surmise that most of the ovules in this species usually reach maturity. Fruit The fruit of Condylocarpon is usually of posed of two pendulous, woody, indehiscent mericarps. In those species that have lorate ¢0- | rolla appendages, the mericarps are glabrous and articulated into several one-seeded, somewhat flattened or angled segments (Fig. 2F). In C. 8 anense and C. isthmicum the carpels are flat tened. The seed-containing segments are broader than the sterile regions between them. The tissue of the ovary wall at maturity contains many air spaces, making the fruits lightweight and buoy ant. Fruits of C. isthmicum remained afloat i? the lab for 16 days. The segments of C. intet- medium are 3-angled in transverse section, and at maturity the seed-containing segments are in ated, the inner ovary wall being composed of? light, corky tissue about 3-4 mm thick. In the regions linking the seed-containing segments the carpel is only a few millimeters in diameter 2? d | lacks a corky layer. The breaking up of the carp® | at the constrictions would result in four or AY¢ individual corky diaspores per carpel. It is pry” ected for as a protection to the embryo inc the shift to salt water dispersal (cf. Steba Those species that lack appendages on the T rolla lobes have mericarps that are terete covered by an indument. The paired merica ring to . V Curae extensions at both ‘ends, and are co , 20) is glabrous and flattened, and degeneration à ing development results in a one-carpellate: gle-seeded fruit at maturity (Fig. 2H). idee: The seeds are fusiform, longitudinally fo p. The and measure from 8.0 to 10.0 mm in length. i 1983] testa is verrucose (Fig. 21, J). The embryo is straight; the cotyledons are narrowly elliptic and constitute slightly more than half of the entire length of the embryo. Pollen Condylocarpon is one of the eight genera of Apocynaceae that are known to have pollen in tetrads (Erdtman, 1952). The tetrads are sur- rounded by a substance that stains deep blue in safranin and astra blue. INFRAGENERIC RELATIONSHIPS : The species of Condylocarponcan be separated Into two basic groups. The first group is char- acterized by having the left margin of the corolla lobe elaborated into a lorate appendage with red- ish brown markings on its abaxial surface. All of the species in this group have a simple orbic- ular style-head and glabrous, distinctly flattened or angled, articulated fruits. This is the larger group and includes C. glabrum, C. guyanense, C intermedium, and C. isthmicum. The second group is comprised of the two Am- Pis endemics, C. amazonicum and C. pub- our In this group the corolla lobes are au- a E or oblique and lack lorate appendages, Meus seem is turbinate. The fruits are Nod Y an indumentum and the segments Ke €. It may seem that the single-seeded is iis i C. amazonicum show little resem- fen 2 T e many-seeded mericarps of C. pubi- Mens ut the similarity of their floral structure : aan very close relationship between these hen €s, and at the same time distinguishes rom all other species in the genus. a al myrtifolium shows character- £roups. The corolla lobes lack lorate ough the NEL So that the fruit is a single, one-seed- E. composed of one glabrous, flattened segm METHODS Mo ; of a Wile logical studies were done with the aid With Slereoscope and a Wild microscope, trations, drawing-mirror attachments for illus- Only dri able "dipon herbarium specimens were avail- Study. Material to be studied in serial FALLEN—CONDYLOCARPON r55 sections was first boiled for 3—4 minutes, to re- store pliability and form, and then put into a dilute KOH solution and kept at 62°C for 1 hour. To remove the KOH the material was then placed in water and kept at 62°C at least 4 hours; this step was repeated. Specimens were then dehy- drated in an alcohol-xylol series, embedded in paraplast, cut with a rotatory microtome at 10 micrometers, and stained with safranin and astra ue. Specimens for study were provided on loan by the following herbaria: A, BM, BR, CAY, F, FI, f EL KE M. MO NY. P RB, S, U, UC, US, W, and Z. Unless otherwise stated, all type specimens were seen. TAXONOMIC TREATMENT Condylocarpon Desf. Mém. Mus. Hist. Nat. 8: 119. 1822. TYPE: C. guyanense Desf. Maycockia A. DC., Prod. 8: 380. 1844. TYPE: M. rau- wolfiae A. DC. = Condylocarpon rauwolfiae (A. DC.) Muell.-Arg. = Echites isthmica Vell. = Con- dylocarpon isthmicum (Vell.) A. DC. Hortsmania Miq., Natuurk. Verh. Holl. Maatsch. We- ch. Haarlem ser. 2, 7: 167. type: H. myrtifolia Condylocary ifolium (Miq.) Muell.- Arg. Rhipidia Mgf., Notizbl. Bot. Gart. Dahlem 10: 1033. TYPE: Anechites (?) amazonica Mgf. = Condylo- j ke. carpon amazonicum (Mgf.) Duc Woody lianas with slender, lenticellate twigs with copious, milky latex. Leaves entire, mem- branous to subcoriaceous, decussate or verticil- late, glabrous to subglabrous on the upper sur- face, glabrous to tomentose on the lower surface; petioles glabrous or with an indumentum. /n/lo- rescence a many-flowered thyrse, becoming con- gested in the terminal branchlets; | l ll ovate to triangular, glabrous to tomentose, usu- ally ciliate. Flowers 2.5-5.0 mm long; calyx 5- b bt totri , ciliate, often lobed, th to g hyaline, inner surface without glands; corolla white, yellow or orange, funnelform or salver- form, often globose in bud, aestivation sinis- trorse, lobes oblique or with red-brown streaked, lorate appendages on the left margin; stamens 5, lanceolate to ovate, without sterile basal ap- pendages and not adherent to the style-head, in- serted near mid-tube or above, filaments short; ovary apocarpous, bicarpellate; style short; style- head orbicular, or turbinate, shortly bilobed at the apex. Fruit bifid, glabrous or with an indu- mentum, articulated into several one-seeded, in- dehiscent segments (in C. myrtifolium, reduced by degeneration during development to a soli- 156 tary, one-seeded segment), or filiform and not conspicuously articulated, and with only a single seed developing per carpel; seed 8.0-10.0 mm long, fusiform, longitudinally per testa ver- e; embryo straight, cotyledons narrowly el- liptic. Pollen in tetrads Distribution: Condylocarpon comprises 7 species that occur principally in Brazil and the Guianas, with one species reaching Central America. KEY TO THE SPECIES OF CONDYLOCARPON la. Corolla lobes with red-brown markings and t ndages; fruiting carpels distinctly articulated, flattened or angled in transverse section, glabrous; leaves opposite or verticil- late. 2a. Leaves 3 (occasionally 4) per node. secondary veins 6—9 p. Ke E ‘isthmicum 2b. Leaves opposi 4a. Corolla eie darker from base to . C. glabrum terete and w ind n , if gla- brous, then fruit consisting of a single one- eded ; leaves oppos Fruit comprised of two c a with indumentum; leaves sparsely to densely to dense golden-brown praedas 6a. Petioles 2.0-6.0 mm long; flowers 4.0-5.0 mm long, corolla lobes 2.0— . C. pubiflorum . Petioles 0.5-1.0 mm long; flowers 3.5-4.0 mm long, corolla lobes 1.0- (en g g per carpel, covered by long, ‘peendine ferruginous- giten hairs . C. amazonicum 5b. Fruit consisting of one sec. one- ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo.. 70 seeded, glabrous carpel; leaves entirely | glabrous; branches glabrous or puberu- | lous in younger parts ........... 6. C. myrtifolium 1. Condylocarpon amazonicum (Mgf.) Ducke, Trop. Woods 76: 28. 1943. Anechites () | amazonica Mgf., Notizbl. Bot. Gart. Dah- | 926. Rhipidia amazonica (Mgt) tasto Ln ei i a M totype; MO, isolectotype). The holotype Mis | destroyed at Berlin in 1943; the isotype i5 here designated as the lectotype.—Fi& 4A-E ep arg iata Ducke, Trop. Woods 76: j 1943. razil, Amazonas, Esperança, pE of Río pae Igarapé, puoi 1172 (flow | ering) (MG, holotype; MO, NY, isotypes). i OEE MEA Mgf., iade 3: P30. 198 Brazil, Amazonas, Porto Velho- o-Guijat mirim, jan tier Hp eo ys 828 (flowering) Z holotype; INPA, isotype, not seen). High climbing, woody /ianas; branches terete, j reddish brown infused with black, glistening | golden verrucose or lenticellate, opposi di | tic to broadly lanceolate, subcoriaceous, | cm long, 2.5-5.0 cm broad, acute to shortly acu minate at the apex, obtuse to truncate at the base. the upper surface glabrous to subglabrous. the lower surface subglabrous to densely tomentos | secondary veins 12-15 pairs, prominent 0n the lower surface; midvein plane to shallowly «d aliculate above, prominent € o E. to- thyrses, puberulous to densely reddish bro cs | ntose or velutinous: bracteoles lance? “40 rs 3.5 brous to tomentose, lobes ovate to tf! ciliate; corolla greenish white to creamy Y° salverform, tube 2.0 mm long, 0.5-1.0 mm e constricted at the base, with an annnm e t | tion at the throat, lobes 1.0-1.8 mm mm broad, ascending to spreading, oblique o neither lorate appendages nor red-brown ot j sessile, turbinate; ovules four per ca ally arranged. Fruit bifid, woody, densel 1983] 0.5mm i FIGURE 4, Pper surf; ace. F-H, Condylocarpon glabrum: F, flower; G, gynoecium; H, leaf, lower surface. tino : Tiras odis ferruginous-brown hairs litorm, tere On each carpel 10.5-15.0 cm long, carpel de with only one seed developing 3.0 mm b €d region enlarged, ca. 1.5 cm long, Dion nan: Secondary and riverine non-in- undated : razil eR 1n central and western Amazonian adjacent regions in Venezuela, Suri- Na me, and Bolivia (Fig. 5). m : = ens Mon ne AMAZONAS: Bre- cnn us , INAME: Heyde & Linde- (U); 190 (Uy; Irwin et al. 55022 (K, NY, U). Speci teler 475 3 (F, M, MO, Z). Sur FALLEN—CONDYLOCARPON 157 NN EET 2T ^ e AUN EB E Ze Ex? 2 E Re: ZAR Ek ^ v4 z 2 —E eA = ZHE IT. JE ZATA AS Ae SPN d ue E 3 ai. B 4 x = = = las Sa SHN PA 25 AAA z = hh. dL E /. E ES i / ths , = «Tl JA E A lk T ANSA ANE, T [XX T A-E, Condylocarpon amazonicum: A, corolla; B, flower in bud; C, fruit; D, gynoecium; E, leaf, BRAZIL. AMAZONAS: Chagas 416 (MO); Cid et al. 825 NY., S, US); AcRE: Forero 6317 (NY, Z). PARA: Fróes 30423 (MO, US); Oliveira 3610 (NY); Silva 936 (F, NY, Zy; 1419 WAG). RONDONIA: Cordeiro 136 (MO); 594 (Z); 828 { (Z); Prance et al. 5627 (F, K, NY, U, US, Z). BoLIvIA. PANDO: Rusby 2599 (NY, US). This species is distinct from all other species 158 Lbs i-i | FIGUR =< E 5. Geog ANNALS OF THE MISSOURI BOTANICAL GARDEN | (0 Me aS ie Li ti | i} i | | I if || ALE I] 100 200 300 400 500 600 MILES | | FOOT CONTOUR SINUSOIDAL EQUAL-AREA PROJECTION | [| IE 1 | A ui raphical distribution bed Y Condvlocarpon Square = C. amazonicum; star = C. 8! thm triangle — C. won iyi circle — C. ist MÀ ——— — AA. 1983] FALLEN—CONDYLOCARPON 159 of Condylocarpon by the fruit, which is bicar- pellate, but with only one seed developing per carpel, and by the filiform carpels, which are covered with long, stiff, ferruginous-brown hairs. 2. Condylocarpon glabrum Muell.-Arg. in Mar- tius, Fl. Bras. 6(1): 66. 1860. TYPE: Brazil, Rio de Janeiro/Espirito Santo, between Campos and Vitória, Sellow 304 (flowering) (K, lectotype; F, photo of isotype at B that was destroyed in 1943). A lectotype is cho- sen here because Muell.-Arg. did not choose a holotype.— Fig. 4F-H. glistening, Striate, lenticellate, becoming gla- brous, densely lenticellate and streaked with golden-brown. Leaves opposite, becoming crowded at branch apex, elliptic to oblong-ellip- m membranous, 9.0—11.0 cm long, 3.0-5.0 cm ke. apex abruptly acuminate with acumen 5.0 ba n acute to obtuse at base, glabrous on aight urlaces, opaque on the upper surface, Uy paler on the lower surface, margins rev- Y, arching and becoming less distinct at Margin, prominent abaxially; midvein im- "s the upper surface, prominent on the E ace; tertiary venation indistinct on the Pper surface reticnlat 4 1 lly areolate Mh ru surface; petioles 1.0-1.7 cm long, a termin “ae blackened, glabrous. Inflorescence a thyrse, primary branches lax, 3.0-4.0 nutily vate reddish brown, lenticellate, mi- ed: So. ous, terminal branchlets congest- ciliate | COP ifie ovate, acute, densely tated s s 2 mm long. Flowers densely aggre- mm long “ ca of the terminal branchlets, 2.5 itels up ds Á : Se in bud, subsessile or with ped- dense] . mm long; calyx glabrous, tube .- y Mottled ES red-brown markings, lobes ins €. broadly orbicular, margins cil- Ys M pale, glabrous, fünneléusn Tb ca. ing i 1.0 mm broad, lobes with spread- 6 mm Steg appendages, ca. 1.0 mm long, deep S has that are densely speckled with abaxial aig id to red-brown markings on the brown, redo stamens lanceolate, deep red- "arpellate 5 e at mid-tube; ovary conical, bi- ' ^7 mm long; style-head subsessile, "ound; oy > OVules fi ese Fruit Nae per carpel, biserially arranged. Pres: Distribution: Known only from the type, col- lected from the Atlantic coastal forest of Brazil, between Campos and Vitória (Fig. 5). Condylocarpon glabrum shows affinities to C. intermedium, but differs by the pallid, broadly orbicular calyx lobes, in not having the lower portion of the corolla tube darkened, and with leaves which lack inter-secondary veins. 3. Condylocarpon guyanense Desf., Mém. Mus. Hist. Nat. 8: 119. 1822. type: French Guiana, Cayenne, Martin s.n. (fruiting) (FI, holo- type; P, isotype; F, NY, photo of isotype at B (destroyed in 1943), P, photo of holo- type). — Fig. 6A-D Climbing, woody /ianas; branches terete, red- dish brown, lenticellate, at least the younger por- tions puberulous. Leaves three per node, elliptic to oblong-elliptic, membranous to subcoria- ceous, 9.0—13.0 cm long, 3.0—4.5 cm broad, sub- caudate to narrowly inate at apex, with acu- men 1.0 cm long, 1.5-2.5 mm broad, acute at the base, glabrous on both surfaces; secondary veins 9-11 pairs, impressed on the upper surface, plane on the lower surface; midvein canaliculate on the upper surface, prominent on the lower surface, often with inter-secondary veins that di- verge from the midvein but anastomose with tertiary venation; lower surface conspicuously reticulate with well developed, polygonal areoles with branched veinlets; petioles 1.2—1.8 cm long, canaliculate, puberulous. Inflorescence lax ter- minal and axillary thyrses, with primary branch- lets 4.0-9.0 cm long, spreading, puberulous to pubescent; bracteoles ca. 1.0 mm long, acute, puberulous, ciliate. Flowers ca. 3.0 mm long, subsessile or with a slender pedicel up to 3.0 mm long; calyx pubescent, lobes ovate, densely cil- iate; corolla creamy yellow to orange, salver- form, glabrous, tube 1.5 mm long, 0.6-0.8 mm broad, lobes ca. 1.0 mm long, 0.5 mm broad, with lorate appendages, brown markings restrict- ed to bases of lobes; stamens lanceolate, inserted at mid-tube; ovary conical, bicarpellate, 0.5 mm high; style-head orbicular, subsessile, ovules four or five per carpel, biserially arranged. Fruit apo- carpous, reddish brown, glabrous, composed of each two woody carpels, y as five one-seeded, indehiscent segments, each segment 1.5-2.5 cm long, 7.0 mm broad, some- what flattened, striate, with a longitudinal streak over the seed area. 160 I A H, Condylocarp ANNALS OF THE MISSOURI BOTANICAL GARDEN Condylocarpon guyanense: A, flower: B, gynoecium; C, fruit; D, leaf, lower E } 1 A-D —L, on intermedium subsp. intermedium: E, fruit; F, flower; G, gynoecium; H, leaf, upper | 1983] FALLEN—CONDYLOCARPON Fic me dium, subsp. intermedium; white Distribution. w “ot Lowland forest in the Guianas tr. Amapá in Brazil (Fig. 7). Specime k S seen. GUYA K) Maguire 22890 (MO NA. VE Dept. Br. Gui. 83 (CAY). (CAY, P); Oldeman I 125 (CAY); Prevost 514 Y) ; 4c E. intermedium Muell.-Arg. in i na Fl. Bras. 6(1): 65. 1860. TYPE: Guy- » Danks of Pomaroon River, Schomburgk is. at was destroyed in 1943).—Fig. "dca, in rhe intermedium var. brevifolia Muell.-Arg. Us Fl. Bras. 6(1): 65. 1860. TyPE: Brazil, RE 7. Geographical Sat ee of Condylocarpon. Star = C. guyanense; black square = C. inter- = C. intermedium subsp. laxum; triangle = C. pubiflorum. Minas Gerais ?, St. Hilaire s.n. (flowering) (P, lectotype; F, isolectotype). A lectotype is chosen sei because Muell.-Arg. did not designate a ho- 1860. Type: Brazil, Bahia, Blan- chet 1 15; 7 : (flowering) (G, lectotype here designat- ed). without locality, Sellow s.n. (fruiting) (BM, Scan: Pro parte quoad plantam yntypicam, non Silva da Manso s.n Condrcarpn obtusiusculum Muell.-Arg. in Martius Fl. Bras. 6(1): Condylocarpon ipic print 2 -— L-Arg e sow tius Fl. Bras. 6(1): e p pepe Sow 57 7 1 Viroiting) K y solectotypes, here designated). E: Nicaragua, Dept. Zelaya, Bluefields, Long 1 3 (flowering) (ENAG, holotype not mer Condyiocarpon p Standl. & Wms., Ceiba 3: 34. 9057 T E: A , isoty , NY, photo of isotype a Tu ua, Dept. Zelaya, "Bluefields, Rio ^ condi, “Molin 1806, 1814 (ENAG paratypes seen); Nicaragua, along Rio Tanten, Molina 2081 odas & fruiting) (BM, ENAG not seen, F, parat 162 Slender, woody /ianas; branches subverrucose or lenticellate, greyish, glabrous or puberulent in ceous, 6.5-10.0 cm long, 2.8-4.5 cm broad, acute to obtuse at apex, obtuse to rounded at base, glabrous on the upper surface, glabrous on the lower surface except along the midvein, which is sparsely pilose; midvein impressed to subcan- aliculate on upper surface, prominent on lower surface; secondary veins 9-13 pairs, plane and somewhat indistinct on upper surface, promi- nent on lower surface, often difficult to distin- guish from inter-secondary veins; petioles 0.5— 1.5 cm long, puberulous to glabrous, shallowly canaliculate to terete. Inflorescence rather lax ter- minal and axillary thyrses, branchlets puberu- lous to glabrous; bracteoles ovate to acute, hya- line, ciliate. Flowers 2.5-3.0 mm long, 3.5-4.0 mm broad, globose in bud, subsessile or with slender pedicels up to 2.0 mm long; calyx pu- berulous to glabrous, brown-speckled, lobes ovate to acute, hyaline, ciliate; corolla white, yellow or orange, with a darker region from the base ex- tending to ca. the level of stamen attachment; tube 1.0-1.5 mm long, funnelform, base con- stricted, throat 1.5 mm broad; lobes 1.0-1.5 mm long, 0.6-0.8 mm broad, with spreading lorate appendages with reddish brown markings on their abaxial surface; stamens lanceolate, inserted at mid-tube; ovary apocarpous, conical; style-head orbicular, subsessile, or with a very short style, ovules five or six per carpel, biserially arranged. Fruit glabrous, composed of two woody carpels, each articulated into up to five one- seeded, in- dehiscent segments, each segment ovoid, 1.0 cm long, 0.7 cm broad, with a corky inner layer ca. 3-4 mm thick. KEY TO THE SUBSPECIES la. Vineis ciere about the xem length as the subsp. intermedium Ib. seriem D times het chet the leaves subsp. laxum a. Condylocarpon intermedium subsp. interme- dium Inflorescence about the same length as the leaves; branchlets brownish grey. Distribution: Coastal forests in Nicaragua, Trinidad, Venezuela, Guyana, and in Brazil, from Bahia to Rio de Janeiro (Fig. 7). s seen. VENEZUELA. DELTA AMACURO: Stey- mk p a 114448 (MO, Z). TRINIDAD. Baker TRIN ANNALS OF THE MISSOURI BOTANICAL GARDEN 14568 (K); LE s.n. - (Ky Finlay 2767 om | GUYANA. Becket .R Martinelli 4184 (Z), Mello RB 49349 (MO, D; Schwache 3100 (GOET). b. Condylocarpon intermedium subsp. laxum ( [Vo.. 70 | uell.-Arg.) Fallen, stat. nov.—Condylo- | carpon laxum Muell.-Arg. in Martius, Fl. Bras. 6(1): 66. 1860. TYPE: Brazil, Rio de Janeiro, Serra Tingua, Schott 5478 (flow stroyed in 1943). A lectotype is chosen here from among the syntypes, since Muell.-Ar£ did not designate a holotype. er. 28. qu gracile Miers, Apocyn. S. Amer. ee 78. TYPE: Brazil, Rio de Janeiro, ee 4019/ 086 (flo taces fruiting) (BM photo of isotype at B that was M RD in 1943). Inflorescence two to three times as long ke leaves, lax and spreading; branchlets glossy : Distribution: known only from Rio de Jan (Fig. 7). Specimens seen. BRAzit. RIO DE JE er chaud s.n. (F, NY, photo of specimen a ped stroyed); Kuhlmann 3788 (RB); RB 39290 (Z). ^! out exact locality: Raddi s.n. (G). 39.1831 (1827); reprinted in Arq. Mus. Rio de Janeiro 5: 106. 1881. TYPE: p Santa Crucis and Fharmacopoli Hi l tration in Fl. Flum. Icon. 3: A-D Cond "pe on — e (A. DC.) Muell.-Ar£ M n s Fl. m ip. 1860. as bester dupl 8: 324. TYPE: B ed de Janeiro ? gis. n. Metti rin holotype; P, isotype; F, NY, pho reser ead hoe: rauwolfiae var. kien Mue ering) (BR, paralec Viel sn Riedel 85 (G. paralectotypes); Langsdorff s.n. (B. paralect estroyed in 1943). Brazil, Prov. Sào " Sa Hilaire 1227 (flowering) (P, paralectotyP® ——— 1983] Ficur FALLEN—CONDYLOCARPON _ 1cm Cond, EA. A-D, C Condylocarpon isthmicum: A, gynoecium; B, fruit; C, leaf, lower mee De D, flower. E-I, n myrtifolium: E, flower, opened; G, gynoecium; H, fruit, top view; I, fruit, 163 164 w s.n. (flowering) (BR, K, paralectotypes); Gau- dichaud = (flowering) (P, paralectotype), all here signa Condyiocarpon rauwolfiae var. aean Muell. E rtius Fl. Bras. 6(1): 64. 1860. TYPE: Bra São Pa ulo, São Carlos, Manso da Silva 335 e ing) (BR, lectotype); Brazil, Mato Grosso, Martius s.n. (frui iaba, anso da Silva s.n. (flowering) (BR, lectotype, here designated). Large, woody /ianas; branches terete, lenticel- late, glabrous or pubescent in younger portions. Leaves verticillate (mostly three per node), ellip- tic to ovate-elliptic, membranous to subcoria- ceous, 7.0-11.5 cm long, 2.5—4.0 cm broad, acute to acuminate at apex, acute to obtuse at base, upper surface glabrous to subglabrous, lower sur- face glabrous with patches of pubescence in the axils of the secondary veins or tomentose; sec- ondary veins 6—9 pairs, impressed on the upper surface, prominent on the lower surface; midvein impressed or canaliculate on the upper surface, prominent on the lower surface; tertiary venation well developed on lower surface, reticulate, po- ee areolate with branched veinlets; peti- oles 0.9-1.5 cm long, glabrous to tomentose. In- eee terminal thyrse, branches puberulous to tomentose; bracteoles ovate to triangular, hya- line, ciliate. Flowers 3.5 mm long, 6.0-7.0 mm broad, pedicels 1.0-4.0 mm long, glabrous to tomentose; flower buds globose; calyx glabrous to tomentose, tube with red-brown markings, lobes ovate, hyaline, ciliate; corolla cream, yel- low, or pale orange, funnelform, tube 1.5-1.8 mm long, throat 1.0-1.5 mm broad, lobes with hpc py appendages on the left margin; appen 3.0 mm long, 0.5-0.8 mm broad, with ip, s brown. markings on the abaxial sur- face; stamens ovate to lanceolate, inserted at mid- tube; ovary apocarpous; style-head orbicular, subsessile; ovules five or six per carpel, biserially arranged. Fruit glabrous, pendent, composed of two woody carpels, each articulated into several (usually four or five) one-seeded, indehiscent seg- ments, each segment ellipsoid, 1.5-2.5 cm long, 1.0-1.5 cm broad. Distribution: Secondary, gallery, and period- ically inundated forest and campos in the Bra- zilian Planalto and margins, from Ceará to Ar- gentina (Fig. 5). ANNALS OF THE MISSOURI BOTANICAL GARDEN Representative collections. BRAZIL. CEARÁ: Eugenio 980 (RB); 981 (RB). coiás: Anderson 7822 (NY, D); 8265 (F, K, MO, NY, RB); Glaziou ws G,K P); Heringer 9915 (Z); Irwin et al. 18542 31786 (F, NY, Z). DIST. FEDERAL: NY, U); Duarte 10185 (Z), Heringer 9189 (Z); Irwin et al. 8 PE F; , : F, NY); (NY, Z). BAHI ley et al. 19907 (Z), Mori et al. 12327 (Z). MINAS GERAIS: Barreto 914 (F); , RB); Davidse & Ramamoorthy 5074 (K); Irwin et al. 20764 (K, MO, RB); 2 Krieger et al. 10846 (RB); 11373 (RB); Lindberg NY); 192 s eye g^ 1226 (MO); Mexia 4680 (F, G, Y, P U3 1736 (RB); St. Hilaire 172 (P. Teodoro 750 (RB), Warming s.n. (K, NY). ESPÍRITO (Z); Kuhlmann RB 19123 (G, K, P, RB, U); te Brade 13242 (RB); Marquete et al. 123 (Z), Mar e 10050 (RB); Martinelli 1327 (Zy, 3006 Oo [Vo 70 Z); p pea a (BR, F, GOET, M, — ren cam, 10701 (BR, K, P, Z); 14851 (MO); 14873 (BR, F, K); | 15716 (BM, F, K, MO, P, ( MO); Dn 2842 (Z); RU (Z); ! e 12543 (Z); 1315 5, 2 14181 (Zy; 17604 (F. 17869 (NY, P, Z); 2 222 (NY, Z); 25308 (NY. f^ Z); 26460 (Z); 26591 c 39849 (MO, NY); 417 m" Hoehne RB 23421 (RB); Joensson 461a (M G); (G); Kuniyoshi 4091 (Z), Lindeman 3306 MO. 5288 (K, NY, U); 5350 (F, U); Pedersen I or ost NY, Z); Pereira 5385 yz: Reitz & Klein 1743 P); Smith et al. 14514 (F). SANTA CATARINA: 1 in 130 (Ky; Smith & ies 11685 (MO, ONS bo ( T 6 & GENT man 1587 (NY); Meyer (em (U); [ibi E w Z); parasite 4215 (MO). ENTRE RIOS: Tro (F, MO). The type specimen of Echites isthmica, was published in Vellozo’s Flora Flu (1829) was lost. Therefore, the type of the SP". becomes the plate in the Flora Fluminenst a the case of Echites isthmica the plate is. at inadequate, and it is not without some 4 hension that I say that Condylocarpon foe which plant depicted in plate 39 of the Flora nensis (1831) clearly has three leaves Pet minenss | agi rrt 1983] This, together with the fact that the specimen was collected in Rio de Janeiro, near Santa Cruz, almost certainly make it the same species as the type specimen of C. rauwolfiae. The two varieties of C. rauwolfiae, acuminata and tomentosa, are not recognized in C. isth- micum. These varieties were circumscribed by Mueller-Argoviensis (1860) by leaf shape and the indumentum of the lower leaf surface, both of which show considerable variability, not only among individuals, but also from the same in- dividual. There may be a two- or three-fold dif- ference in leaf size, depending upon which part of the liana was used for the specimen, as well as upon the maturity of the leaves. In addition, there is a negative correlation between leaf size and latitude, from Ceará at about 5°S, to Rio Grande do Sul at about 27°S. Condylocarpon isthmicum is similar to C. in- lermedium and C. glabrum in having the lorate, red-brown streaked petal appendages and gla- brous, articulated fruit. 6. Condyloc. so ye RA: 1.) Muell.-Arg. in Martius FI. Bras. 6(1): 65. 1860. — Horts- mania myrtifolia Miq., Natuurk. Verh. Holl. Maatsch. Wetensch. Haarlem. ser. L1. IL 1851. TYPE: Suriname, without exact local- ity, Hostmann 1196 (flowering) (U, lecto- type; F, G, GOET, K, MO, NY, P, isolec- totypes; F, photos of isolectotypes at G and : NY, photo of isolectotype at G); 462 (lowering) (P, U, paralectotypes). One of the SYntypes is here designated as a lectotype, use a er was not designated by Miquel.— Fig. 8E] com ing silvery grey and glabrous with age. Leaves 0 PPosite, oblong- -elliptic, membranous, 6.0-10.0 ™ long, 2.0- se > Blossy, glabrous on the upper surface, paler, Upper surfa inter. -Sec veins 11-13 pairs, prominent on the ace, with numerous equally prominent Ondary veins diverging from the mid- Tving and anastomosing with sub-prom- inent, oblique tertiary venation or occasionally id- FALLEN —CONDYLOCARPON 165 surface; petioles 2.0—4.0 mm long, glabrous, can- aliculate for about half this length. Inflorescence a short, very congested thyrse, glabrous or pu- Flowers 3.0—3.5 mm long, subsessile or with ped- icels ca. 1.0 mm long; calyx iate; corolla salverform, greenish white to creamy yellow, darker near the base; tube 2.0 mm long, 6.0 mm broad, minutely puberulous, constricted at the base, ca. 1.5 mm broad at the throat; lobes 1.0 mm long, 0.6 mm broad, shortly auriculate, ascending and slightly spreading, without lorate appendages or red-brown markings; stamens lan- ceolate, inserted at mid-tube; ovary conical, 0.3 mm long, darker than style; style 0.3 mm long, topped by an orbicular style-head; ovules four or five per carpel, biserially arranged. Fruit con- sisting of only one carpel (the second degener- ating during development), a pendulous one- seeded, flattened, ellipsoidal disc with an apical notch, 4.0—5.0 cm long, 1.5-2.0 cm broad, green- ish en, leathery, with lateral veins diverging from the canaliculate central suture; seed area longitudinally striate, raised on the abaxial sur- ace. Distribution: Non-inundated forest in Suri- name, Venezuela, and Amazonia (Fig. 5). imens examined. VENEZUELA. DELTA AMACURO Dept. Br. G Jenman 88 (P); 243 (NY, P); Tutin 162 (BM, RB, U). SuRINAME. Focke 958 (U); Gonggrijp 120 (U); Kegel 1009 (GOET); Stahel & Gonggrijp 3000 (U); Went 551 (U); Wullschlaegel 1654 (BR). BRAZIL. PARA: Black 48- 3006 (RB); Ducke 7382 (RB, U); RB 15816 (RB); RB 21592 (RB); RB 21618 (RB); RB 21620 (RB). TERR. : 66 S RE. U); Maguire et 2). A. PANDO: Prance et al. 8654 (F, NY, P, U, Z); Rusby 2391 (F, G, K, NY). This species is easily recognized by the oblong, caudate-tipped leaves with numerous inter-sec- ondary veins, and glossy upper surface, and the one-carpellate, inarticulated fruit. 7. Condylocarpon pubiflorum Muell.-Arg. in Martius Fl. Bras. 6(1): 67. 1860. TYPE: Bra- zil, Amazonas, mouth of Río Negro, Spruce 1564 (flowering) (BR, lectotype; F, G, GOET, K, M, NY, P, isolectotypes; F, NY, photo of isolectotype at B that was destroyed in 1943).— Fig. 9. 166 ANNALS OF THE MISSOURI BOTANICAL GARDEN i SC FIGURE 9, Condyocarpn p ee A, corolla, top view; B, opened corolla, side view; C. fruit: E Pores E, leaf, upper surf: Su 1983] Condylocarpon ciliatum Muell.-Arg. in Martius Fl. Bras. 6(1): 66. 1860. TYPE: Amazonas, n to ofisolectotype at B that was destroyed in 1943). ~~ hirtellum Ducke d ips pe 76: 28. E: Brazil, Amazonas, mouth of Río. Joi Ducke s^ Pri (MG. holo- type, not seen; NY, isotype). Large, high-climbing, woody /ianas; branches terete, reddish brown, somewhat blackened at nodes, golden verrucose or lenticellate, with a sparse indumentum of golden-brown hirsute hairs that is denser at the nodes. Leaves opposite, el- liptic to oblong-elliptic, membranous, 6.5-15.0 cm long, 2.5-5.0 cm broad, narrowly acuminate to rod caudate at apex, acumen 5.0-10.0 mm long, acute to rounded at the base, upper surface i n the lower surface, tomentose tó velu- uec in th surfaces; petioles 2.0-6.0 mm long, a , darkened, golden-brown tomentose Flowers 4.0-5.0 mm long, 6.0-9.0 mm broad, oe or with glabrous pedicels up to 2.0 mm ; Calyx glabrous, lobes ovate, membranous, "mea wo dd white to creamy yellow, salver- dena: a a 5-2.0 mm long, base constricted, striction, | K mm broad, with an annular con- Li a S spreading, 2.0-5.0 mm long, 1.0- nor fet Toad, with neither lorate appendages "d -drown markings. Stamens lanceolate, in- mm | near top of tube; ovary conical, ca. “Ong, style 0.3 mm long; style-head turbi- nate: i do Ca. 16 per carpel, arranged in four sé. m Fruit apocarpous, densely golden tose, composed of two woody carpels, each I indistinctly articulated into several (up to 16), "d *d, indehiscent cylindrical segments, ee .0-2.5 cm long, 2.0-3.5 mm broad. ribution: Non. inundated or periodically in nun R a dated forests in central and western Ama- nia (Fig. 7). 10823 pe Seen. COLOMBIA. PUTUMAYO: Cuatrecasas AMAZONAS: Schultes 6677 (US). VENEZUELA. FALLEN—CONDYLOCARPON 167 AMAZONAS: Liesner 3603 (MO); Maguire et al. 36752 (F, MO, NY); Morillo & Hasegawa 5165 (F). Peru. LORETO: Gentry 21821 (F, MO, Z); Revilla 321 (F, MO, Z); 1820 (MO); Williams 2011 (F); 3770 (F). BRAZIL. AMAZONAS: Ducke RB 22434 (G, P, RB, U); RB 23938 (RB, U); Fróes 21195 (NY); Krukoff 6760 (NY); 8265 (NY); Martius (M); Spruce 3417 (K). The nearest relative of C. pubiflorum is C. amazonicum. The two species most closely re- semble each other in the westernmost part of their ranges in Amazonian Peru, Colómbia, and Venezuela. Without mature flowers or fruit, specimens of these two species from this region can be very difficult to distinguish from each other. NUMERICAL LisT OF TAXA Condylocarpon amazonicum (Mgf.) Ducke Condylocarpon glabrum Muell.-Arg. Condylocarpon guyanense Desf. Condylocarpon intermedium Muell.-Arg. a. subsp. intermedium b. subsp. /axum (Muell.-Arg.) Fallen 5. Condylocarpon isthmicum (Vell.) A. DC. 6. Condylocarpon myrtifolium (Miq.) Muell.- A AYN > rg. 7. Condylocarpon pubiflorum Muell.-Arg. List OF EXSICCATAE i TI e as given above i in the numerical list of taxa. Collection year follows s.n., when available. Anderson, W. R., 7822 (5); 8265 (5) Baker, R. E. iy TRIN 14568 (4a) Ball, J., s.n., (5) Barreto, M., B (5); dd Aes 5177 (5) et al., 218 (6) xo Dot (5) Cid, C. et rt 835 b 925 (1) , 136 (1); 594 (1); 828 (1) (42) eg Cuatrecasas, J., Cunningham, A., 1867 (5) La Cruz, J. S., "0850 (6); 4117 Duarte, A. P. et al., 1749 (5); 3538 o. 39900 (8) 10185 redd A., 353 (1); 699 (1); oa 1171 (7); 1172 (1); 2344 (1); 7382 (6); RB 15816 (6); RB 17480 (1); RB 21592 (6); RB 21618 (6); RB 21620 (6); RB 21697 (1); RB 22434 (7); RB 23876 (1); RB 23938 (7) 168 n, P. K. H., 3410 (5); 7063 (5); 7798 (5); 9094 (5); 10701 9: 14851 (5); 14873 (5). 15716 (5); 15911 (5); 16455 (5) Edwall, Pt RB 119753 (5) Ekman, E. L., 1587 Engler, 65 (5); s.n., 1864 (5) Eugenio, J., 980 (5); 981 (5) Fiebrig, K., 5838 (5 Finlay, Focke, 6) Fontella et al., 1118 (5) Forero et al 3 ES (1 For. Dept. Br. 1. (Guyana), 83 (3); 315 (4a); 4146 (6); 5066 odes 5193 (4 (4a) Frasão, A., RB 86684 (5 Fróes, R. i 21195 2a 30423 (1) Gardner, G., 5074 asiad E ae (5); s.n. (4b) Gent (H et al. uio (7) Ginzberges, A. et al. Glaziou, A. F. © 15223 (5); 31940 S, O. C. et al., 262 9 435 (5); 541 (5); 992 (5); 1032 (5) Gonggrijp, J. W., ie ©) Graham, Bro. Wm., n. (5) Hanbury, G., 2 (5) 4 (5); Sn. (SX s0. 1868 (5 Harley, R. M., 08 a 7758 (5); 7886 (5); 13470 1866 (5); s.n., , 2842 (5); (5); 13155 (5); 14181 (5); (5); 25308 (5); 26460 (5); (5) 12138 (5); 12473 (5); 12543 17604 (5); 17869 (5); 24222 26591 (5); 39849 (5); 41759 Heringer, E. P., 9189 2 9915 (5) Heyde et al., 65 (1); 190 (1) Hitchcock, A S. D (4a) Hoehne, F. C., RB 23421 (5); RB 28439 (5) Hoshimoto, A. 52 (5 Hostmann, W. R. et al., 264 Im T Thur, E, ae n 1879 (6 in, 8339 (5); 9539 (5); 11141 (5); 11420 (5); 1535905. 15657 (5); 15913 (5); rn 20764 2 (1) (6); 1196 (6) nt 27417 (5); 31786 (5); 48 ipis E enman, G. S., 88 (6); 243 Ae n nsson, y" , 461a (5); 1168a (5 e Kegel, H., 9 (6 Krieger, P. T et al., 10846 (5); 11373 ry Krukoff, B. A., 6760 (7); 8007 (1); 8265 (7) Kuhlmann, J. D 460 (5); RB 3788 e 6411 (4a); RB 19123 VI RB pes (6); RB 39290 (4b) Kuniyoshi, Y. S 1 (5) Lescure, 531 b , 13242 (5) Lindberg, G. A. 92 Lindeman, 2 C. d (5); Long, L. E., 4a) Magalhàes, M., 1226 (5) Maguire, B. et al., 20800 (3); 36752 Manso, A. L. P. da Silva 50( 5288 (5); 5350 (5) (7); 56763 (6) , 335 (5); s.n., 1834 (5) , 123 (5) s.n. 3 Martinelli, G. et al., 1327 (5); 3006 (5); 4184 (4a); 8450 (4a) ANNALS OF THE MISSOURI BOTANICAL GARDEN Martius, K. F. P. von, s.n., 1819 (5); s.n. (7) Mello, M., s.n. (4a » 3573 (S) Miers, J., preh Mer 8086 (4b) Mo lina, R. A., 1 (4a) (7) Mosén, C. W. H., 621 (5); 3653 (5) Mueller, E. 130 (5 e "dui (8); 11931 (5) 56 7 (4a) Pires, J. M. E. x yv (3); 58060 (5) Pohl, J. E., Prance, G. Er i al., 5627 (1); 8654 (6); 14230 (1); 14708 a | Prévost, M. a 514 (3) Raddi, G., s.n. (4b Re; egnell, A. ye II- "E n eitz, P. R. et al. iS 0 (5) Revilla, 3:327 d 128b AV 1820 (7) Riedel, sa , 85 (5); s.n. (4b Rodrigu . H., 2391 (6); 2599 (1) r Saint Hilaire, E oe (5); ae 1404 (5); s.n., 18] 4a 821 Sastre, C., 4701 (3); 5470 (3); 5906 (3) 4a); 1426 burgk, re (4a) Schott, H. W., 5478 Sf 479 (5) Schwacke, C. due Schwarz, G. J., 5(5 w 1 Sellow, Fa 5 ^ ps (2); 571 (4a); s.n., 1815-18! (4a); s Silva, N. i Rig a) $ Smith, L. B. et al, 11685 (5); 13126 (5); 13162 a r 1564 (T 2475 (T); 3417 () al 'G. et al., 3000 (6 rmark, T et al., 114448 (4a) 111 (5) (6 "s E F.. 64 (5); he i 1236 (5) Williams, Hr , 2011 (7); 3770 (7) Wullschl laegel, H. R., 165 $40 ) LITERATURE CITED ys De CANDOLLE, A. . Apocynaceae, in 31748. systematis naturalis regni v—— 8 veau € diss 22. re. Mém. Mus. Hist. Mat. 8: 119-121. [Vo 70 — } —À l ʻ 1983] Ducke, A. 1943. New forest trees and climbers of the Brazilian Amazon. Tr 6: 15-32. Ee T 1952. Pollen morphology and plant my. VY vocali Almqvist & Wiksell, uU. 539 p FALLEN, M. E. 1983. 5. aug revision of Ane- chites (Apocynaceae). bei uai 35: (in press). Besta! i ADOOS aus Sued- erika IL Notizbl. Bot. Gart. Dahlem 9: 959- — —. 1930. Neue Apocynaceen aus Suedamerika III. Notizbl. Bot. Gart. Dahlem 10: 1033-1039. — 22 0. Algumas Apocináceas novas do Brasil. 23 77-87. Mass y "1878. On the Apocynaceae of South Amer- ms. & Norgate, Edinburgh. 277 1851. i surinamenses. Na- tuurk, Verh. Holl. Maatsch. Wetensch. Haarlem, ser. MURLER-ARGOvIENSS, J. 1860. Apocynaceae, in Martius, n Bras. 6(1): 63-67. PICHON, M. 948a. Classification des Apocynacées. VOU: et Ambélaniées. Mém. Mus. Natl. Hist. Nat. 24: 111-181. FALLEN —CONDYLOCARPON 169 1948b. Classification des Apocynacées. IX. Rauwolfiées, Alstoniées, Allamandées et Taber- naémontanoidées. Mém. Mus. Natl. Hist. Nat. 27: 153-252. . 1951. Classification des Apocynacées. XXV. Echitoidées et suppl. Mus. Natl. Hist. eis Sér. B, Bot. i: 1-174. , A. d . Prantl, Die natürlichen Pinani 4(2): 109- 189. Leipzig. STEBBINS, G. L. 1971. Adaptive radiation of repro- ductive characteristics in angiosperms, = fa = seedlings. Annual Rev. Ecol. Syst. 2: 2 VELLOZO, & M. pA C. “1825 [1829]. " ane i ra Fluminensis 3: 112; Icones 3, t. 39 “1827 [1831]. " Text and plate NDA. n in Arq. Mus. Nac. Rio de Janeiro 5: 106. 1881. SUPPLEMENT TO THE CONTRIBUTIONS TO THE SYNANTHROPIC (ADVENTIVE) FLORA OF THE RAILROADS IN ST. LOUIS, MISSOURI, U.S.A.! VIKTOR MÜHLENBACH? ABSTRACT Based on 93 excursions from 1972-1980, 22 additional th the railroad network in St. Louis. Included were 13 new speci Missouri. -M addition 37 taxa, native to Missouri but ite rag introduced into the St. Louis region, sed. are disc My summary paper on synanthropes of the railroads of St. Louis (Miihlenbach, 1979) was based only on fieldwork from 1954 through 1971. From 1972 through 1980 93 additional excur- sions were undertaken on the Missouri side of the Mississippi River. In addition, from 1978 through 1980 railroad premises and freight yards on the Illinois side of the Mississippi River were visited 38 times and the results were published n Shildneck, Jones, and Mühlenbach (1981). ery similar on both dd of the Mississippi River. Only three species were found in Illinois that were not encountered species— Mirabilis linearis (Pursh) Heimerl, syn- anthropic in pee d. native in Missouri — was found only in Illino Far this supplement only 22 "ew gredi ompared to the 393 of the first repert t (Mühlenbach. 1979). Included among these new discoveries are plants collected during the earlier years but not fully identified until recently. Although the low per- centage of new discoveries for the latter period G. Eiten, TF. W. Gould (deceased), T. D. Jacobsen, A G. Jon L. Reveal, C. M. era Wilso nally, du wre this second period I en also helped to covered on njoyed the company of many persons, mostly members of the Botanical Gaiden n or students of Washington University, ' who does me bog the freight yards in their coll k^ s and 3 AW vant for d flora of can be expected after the initial 17 years of con- centrated fieldwork, an additional factor also ) plays a part. This was the much greater use o. herbicides. Traditionally many synanthropes r have been discovered in the classification 0f switching tracks, areas where freight trains afe rearranged for further movement. Most such are are now devoid of vegetation throughout the year because of herbicide spraying. During the years of my fieldwork the vegetation in the EC sification tracks was frequently not sprayed un the plants were SERE Then one 16 wo n seen places. Such abundant plant growth is gi only in exceptional cases. One case is rev! later in detail. RESULTS Novelties for the state of Missouri are Ma first. They include 13 species and 3 new varie of known species. The names of the cited companies are substituted by their O ronyms: BN (Burlington Northern); C&EI cago & Eastern Illinois), MKT (Missouri- EI (Chi- nes, R. Kral, J. McNeil, A. Melderis, | E d i) s, J. M. Rominger, N. H. Russell, R. M. Tryon, R. E. Weaver, L. C. Whee "e Boufforó Jim Conrad, Sous Co Lewis, Ilze Muehlenbachs, Karlis Muehlenbachs, Dunca in Edmon souri Botanical Garden, P.O. Box 299, St. Louis, Mie 63166, U.S.A: ANN. Missouri Bor. GARD. 70: 170-178. 1983. ese pe: nrad, Dorothy England, Mary Fallen Peter Hoch, Jak Humbles, Randy Hym ns are David n. I express my gratitude to e of them es ficial * ' : RS sas-Texas); MP (Missouri Pacific): M p i n the second piede von D. Clayton, A. F. Clewell, T. ‘Creo! G. Davidse, J. VA aia 1983] (Manufacturers Railway); RI (Rock Island, now in the possession of St. Louis Southwestern Rail- way = Cotton Belt Route (SSW)); SL-SF (St. Louis & San Francisco Railway, now in the possession of BN); TRRA (Terminal Railroad Association of St. Louis). Bromus japonicus Thunb. var. grandis Vel. M 3830A, Carrie Ave. Fr. Yd. (RI), 28 June 1973, one specimen amidst a colony of the typ- ical specimen of this species. Determined by H. Scholz. Bulgaria (Velenovsky, 1898). Bromus arvensis L. M 4169, Carrie Ave. Fr. Yd. (RI), 4 July 1978, one colony. M 4172, Bulwer Yd. (TRRA), 4 July 1978, two specimens. Both collections deter- mined by H. Scholz. Eurasia. Echinochloa utilis Ohwi & Yabuno. M 1577, Bremen Fr. Yd. (TRRA), 7 Sept. 1959, one specimen. M 3349, Lesperance St. Fr. Yd. (MP), 25 July 1970, one colony. Both collections were previously identified as E. crusgalli (L.) Beauv. The new identification follows the an- notations of P. Michael (University of Sydney, NSW.) This species was first discovered in 1961 In Sakai, Province Izumi, Japan, and described as à new species by Ohwi and Yabuno in 1962 (Ohwi, 1962). A fuller morphological and geo- sraphical description is given by Yabuno (1966) a especially by Vickery (1975: 197), who noted at Prior to 1962 this species appears to have iain widely confused with E. frumentacea of Puede : - the two are totally distinct... . E. utilis ue ess originated in eastern Asia, including Oria and is considered by Yabuno . . . and galli 9 have originated from Echinochloa crus- may b Cultivated for forage and grain and „ “an occasional weed of cultivated areas Bs trawfordii Fern. (C. scoparia Schkuhr : Minor Boott). REA 1L former Railway Exchange Agency par ik Clark Ave.) leased from TRRA, 24 F » One large, dense tuft. Determined by ge It was most unusual and per- orthern « find C. crawfordii. It is basically a and in he frequent in Canada (Nfld.-B.C.) S.A o the 13 northern border states of the barium cording to the literature and the her- cited MAN terial at MO. Only a few states are àccordin E The nearest state to Missouri it occu des Fernald (1950) is Tennessee where | 75 In the high mountains. In the principal MÜHLENBACH —SYNANTHROPIC FLORA 171 area of its distribution, Gleason (1968) charac- terizes the habitats as wet soil, meadows, swamps, and shores, places that very seldom produce syn- anthropes. Northern introductions are excep- tional events, in Europe as well as in the U.S.A. The assumption that we are dealing with a native plant cannot be completely neglected, but the probability seems low. I (Mühlenbach, 1979) previously cited several synanthropes for which native status could not be completely ruled out. Nevertheless, I classified them as synanthropes because they grew in association with other rare synanthropes. I treat C. crawfordii likewise. It grew on the track along one of the loading plat- forms of the former REA, again in company of interesting and rare synanthropes. See Hermann (1970) for further details. Salix alba L. X S. fragilis L. (S. X rubens Schrann). M 2068 (flowers) and M 2183 (leaves), from a unicate shrub, 7 April 1963 and 28 Aug. 1963, Harlem Fr. Yd. (TRRA). Determined by G. W. Argus. Eurasia. Rumex cristatus DC. (R. graecus Boiss. & Heldr.). M 725, SL-SF, west of Knox Ave., 14 Aug. 1955, one specimen. M 3967, SL-SF, west of Macklind Ave., near the Des Peres sewer, 5 July 1976, one specimen. M 4028, MP, at the branch- ing of the siding from the trunk line to the RF Macaroni and Spaghetti Works, 30 May 1977, one small colony. M 4060, Lindenwood Fr. Yd. (SL-SF), 17 Sept. 1977, three small colonies. M 4188, Baden Fr. Yd. (MKT), 15 July 1978, one specimen. All collections determined by K. H. Rechinger. To the best of Rechinger's knowledge R. cristatus has not been found in the U.S.A. as an adventive plant. This species was treated by him in Hegi (1953) and in Jalas & Suominen (1979). Eastern Mediterranean. I succeeded also in discovering this plant in Illinois (M 4224, 4 Aug. 1978, along the Eastern Connection be- tween the Madison Fr. Yd. of TRRA and that of the Chicago and North Western Railway of the same name). Rumex orbiculatus Gray, cited in Steyermark (1963: 580), turned out to be the first specimen of R. cristatus found in St. Louis (M 725). Polygonum cuspidatum Sieb. & Zucc. var. compactum (Hook. f.) Bailey. M 3356, MP, south of Zepp St., 30 July 1970, two specimens. M 3476, TRRA, west of Union 172 Blvd., 7 Nov. 1970, one colony. In both cases the plants were sterile. Determined by H. A. Wahl. Japan. Brassica napus L. var. napus (B. napus L. var. arvensis (Lam.) Thellung). M 100, between the right-of-ways of MP and SL-SF, west of the Kingshighway overpass, 29 May 1954, one specimen. Determined by H. Scholz. Known only as a cultigen (rape, an oil and forage plant). Brassica oleracea L. M 1565, North St. Louis Fr. Yd. (BN), 23 Aug. 1959, one specimen (B). M 2088, Carrie Ave. Fr. Yd. (RI) or Bulwer Yd. (TRRA), 30 May 1963, one or two? specimens. Both determined by H. Scholz. Widely cultivated vegetable, orig- inating in the Mediterranean. Reseda lutea L. M 3931, Carrie Ave. Fr. Yd. (RI), 12 June 1976, one specimen. Determined by H. Scholz. Eurasia. This species is missing in Steyermark’s Flora of Missouri, while Fernald (1950) cites Missouri pud sues in which R. lutea is in- troduced. As I st , 1979), Steyermark’s (1963) Flora of Missouri was used as the starting point for the decision as to what is native and introduced in Missouri. Besides, Gleason (1968) explicitly omits Missouri from the list of states where this species has been found. Viola arvensis Murray. M 3663, North St. Louis Fr. Yd. (BN), 27 June 1974, one small colony. Determined by N. H. Russell and H. Scholz. Eurasia. Viola arvensis belongs to the group of plants which Steyermark (1963) excluded from his Flora as erroneously determined. Oenothera heterophylla Spach subsp. hetero- phylla. M 942, O'Fallon Fr. Yd. (TRRA, now dis- mantled), 16 June 1956, two large specimens, from which only parts were taken; M 1018, 14 July 1956, the same two plants. This species was annotated by W. Dietrich (University of Düs- seldorf). It was previously determined as the na- tive O. rhombipetala Nutt. According to Correll & Johnston (1970) a plant of eastern and north- central Texas and western Louisiana. Plantago indica L. (P. ramosa (Gilib.) Asch- rs). M 3977, Carrie Ave. Fr. Yd. (RD), 17 July ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 1976, one specimen. Determined by H. Scholz. Eurasia. Lonicera maackii Maxim. M 3985 (flowers) and M 4001 (fruits) from the same plant (?), MP between Shaw Blvd. and Kingshighway Blvd., 14 Aug. 1976 and 18 Sept. 1976, one or several shrubs. M 3997 (flower) — and M 4004 (fruits), from the same shrub, 28 ` Aug. 1976 and 4 Oct. 1976, Inbound Fr. Yd. (RI, SL-SF, C&EI), one shrub. Determined by R. E. Weaver. Asia. Symphoricarpos albus (L.) Blake. M 3485, MP, between Tower Grove Ave. and Chouteau Ave., 8 Nov. 1970, one shrub. Deter mined by R. E. Weaver. North America, missing in Missouri. Lactuca sativa L. LH M 4002, Lesperance St. Fr. Yd. (MP), 18 Sept. | 1976, one specimen. It lay uprooted on ^ ground, but it was seen in living condition on? Aug. when it was too juvenile for collecting. De termined by R. P. Wunderlin. Widely cultivated vegetable. Origin uncertain. The following eight synanthropes were m ported by Steyermark (1963) to occur in x souri, but were not previously ob bserved by " in the railroad network from 1954 through 19 or were not surely recognized at that time. Festuca ovina L. " M 2963, South 7th St. Fr. Station (MP). July 1968, three specimens. Determined by R W. Pohl. Eurasia. Rorippa sylvestris (L.) Bess M 3822, dead end of the MP’s Ch along Landsdowne Ave., 15 May 1973; 30 June 1973, the same small colony. mined by R. L. Stuckey. Eurasia. M 3832 Syringa vulgaris L. 974, M 3856, Pickrel Fr. Yd. (TRRA), Ec 4 one sterile shrub. Determined by R. E urope. Physalis ixocarpa Brotero. M 2358 and M 2388, South Ranke (TRRA), 14 June and 19 July 1964, in both one specimen (UMO). Determined by D- and and W. G. D’Arcy. Southwestern U. S. A. Mexico; aei angulata 196 2493A, River A Yd. (MRS), 10 July Dete- | i risty Lead, ‘ $ | 1983] : FIGURE 1. - Greenhouse-grown progeny of Mühlenbach 4074 with normal growth form Pe: specimen. Determined by W. G. D'Arcy. Peis U.S.A., West Indies, South America, Tagetes erecta L. Pe = pn & Conrad 3899, Baden Fr. Yd. ), 6 Sept. 1975, one specimen. Determined d R. P. Wunderlin. Discovered by J. Conrad. exico. ME maritima L. var. agrestis (Knaf) Mott (M. inodora L.). a 3891, North St. Louis Fr. Yd. (BN), south- reserve tracks of the grain elevator, 24 June 19 d 75, one colony. Determined by R. P. Wun- throspermum Á MP) tad 3916, Lesperance St. Fr. Yd. p A af 1975, one specimen. Determined lenbach pun. As mentioned by me (Müh- Was carrie d ) a constant lookout for this species ünicate wa yi t from 1956, but in vain. The first the last sn Y located in 1975. It was also of St, Tum z 1s odd indeed for on the streets an extreme rari laevigatum does not seem to be rarity, The fol the list o Taraxacum laevigatum (Willd.) DC. (T. ery- ndrz. “ht species may also be added to railroad synanthropes in St. Louis. zo undulatum (Nutt.) Spreng. arlem eae 1955; M 1254, 4 July 1957, ing to Ste : e d. (TRRA), one specimen. Accord- live counti rmark (1963) it is rare in Missouri Arizona €s) and ranges from British Columbia » €ast to Manitoba, North Dakota, Ne- MUHLENBACH—SYNANTHROPIC FLORA 173 Aristida oligantha Michx.—a. Densely tufted, highly branched form (Mühlenbach 4074, MO). — h form. braska, Missouri and Oklahoma; in other words, it must be considered to be native. The same judgement on native status in Missouri and the general range of this species is given by Fernald (1950). But G. B. Ownbey, who annotated all my Cirsium collections, wrote to me on 9 December 1971: “You also have three additional American species: C. discolor, C. altissimum are native in your area, C. undulatum is adventive from the West, the Great Plains, to be exact.” The total number of species recorded from the railroads was 901, 414 being synanthropes and 487 native. About 20 additional species have withstood all efforts at determination. Some are too juvenile, some o i some damaged. Many of them are Amaranthus hybrids in dif- ferent combinations. As previously, I have tried to avoid publication of uncertainly determined material. Of the native species, one group merits a closer look, namely native Missouri plants that were collected in St. Louis County for the first time. (Steyermark (1963) combined the city of St. Louis with St. Louis County for this purpose, as do I.) Altogether there were 37 taxa (23 species, 14 intraspecific taxa) that were not previously re- ported from this area. Most of them were seldom encountered, and they were usually not numerous. Likewise, the majority tend to be rare in Missouri. According to Steyermark (1963) eighteen of them were re- ported in five or fewer counties, six even being found only in one county. Twelve were known from six to twenty counties, and only seven were found in twenty-one and more counties. Stey- 174 ermark mentions further that among these thir- ty-seven taxa, eight grow along railroads. At least half of the thirty-seven were originally restricted to the western half of the state. Quite a few (sev- en) were known only from the southeastern low- lands (Missouri’s Bootheel). It is of course of interest to learn the status of these plants. Are they autochthonic elements of St. Louis County’s flora or have they been introduced by one means or another? It likely that tl ity, the great majority, havei in fact been introduced. even one or two of three counties (St. Charles, Franklin. Pide bordering St. Louis County. Surely many of these native plants that appeared in St. Louis did not originate in Missouri itself but came from far away. It is easy to prove such assertions when varieties and forms are found that are missing in the native flora. Such obser- vations have been made in Europe, but not in Missouri. From 1954 through 1971 the richest freight yards with rare synanthropes were Carrie Ave. (RI) and Baden (MKT). Of interest is the fact that half of the introduced natives were also found in both these yards. This is an additional indication that they are indeed introductions. Five taxa were found along the so-called car cleanout tracks, where heavily polluted freight cars are swept and tidied up before loading. These Finally the remainder is hauled away. It is as- tonishing to see that in spite of such harsh con- ditions many plants manage to survive. Such sites are the most promising places in the freight yards for discovering synanthropes. These tracks are quite confined in size and length and were visited and meticulously inspected at least three times in a season. In such places almost every plant, be it synanthrope or native, must be introduced by the railroad operations. It is almost incon- ceivable that these native plants had been over- looked for years before they were noticed. Recently I became aware of a paper by Schultz (1976) on the synanthropic flora of the railroad monton, Canada, it is quite evident that the role of refuse, dirt, and debris, which normally ac- ANNALS OF THE MISSOURI BOTANICAL GARDEN companies freight, is more important in the in- [Vo 70 . troduction of synanthropes than the role of the ' freight itself. This conclusion has apparently nev- er been explicitly stated in the literature, but its importance is underscored by comparing the way refuse is handled in St. Louis and Riga compared with Edmonton. In the huge Calder Yard of the Canadian National Railways in Edmonton, freight cars are cleaned in an elaborate wast , disposal system in which all solids are removed by washing or a vacuum system and taken away | by truck to garbage dumps or an incinerator. In i | both St. Louis and Riga the cleaning system is | much less elaborate, and solid refuse ends up 0? the car cleanout tracks and nearby areas. A5? consequence, the car washing areas of the rail- road yards in St. Louis and Riga were the richest sites for synanthropes, whereas very few syi- anthropes have ever been found in the vicinity of the Edmonton facilities. The remainder of these thirty-seven native plants are most probably also introduced. Nt- merous observations in European yards have often demonstrated the introduction of p plants into them. One may even speak abou secondary or even triple introduction, aS si in souri, Chenopodium desiccatum var. E^ um, has until now only been reported from county, Jackson. Du other phenomenon should not be o looked. As experience in Europe has taught. d pt a few plants are able to spread spontaneo along trunk lines by ferroviatic migratio sud | one station or yard to a neighboring One wt evidence is missing in St. Louis. ibo. taxa indeed found on the right-of-w trunk lines or along some elit: of them was also found in a yard. ig ` rarities. The Participants ud a ferroviatic m Ee tion usually shov in St lines, something: which was totally on d ouis. They are o the followinE var. pseudohye met Th Pe N 4 x 1 Equisetum ONT L: (Farw.) Mort how? M 121,5 deis 1954, one colony that has alate intention to persist. This species was located n from | j | | | | | | l í 1983] in several other places. Determined by R. M. Tryon. Typha angustifolia L. M 2472, 31 May 1965, one large colony, per- sisting for many years. Another colony was later discovered, but not collected. Muhlenbergia frondosa (Poir.) Fern. f. com- mutata (Scribn.) Fern. M 2228, 21 Sept. 1963, one small colony. De- termined by R. W. Pohl. Brachiaria platyphylla (Griseb.) Nash. M 2245, 29 Sept. 1963, two specimens. M 3069, 30 Oct. 1968, one large tuft. Determined by J. B. Van Schaack and R. W. Pohl. Paspalum setaceum Michx. M 3614, 12 July 1971, one colony. Deter- mined by R. W. Pohl. Cyperus odoratus L. (C. ferax L. C. Richard). M 1862, 16 Sept. 1961, one specimen. M 3427, 13 Sept. 1970, scattered. Determined by R. Kral and T. M. Koyama. Scirpus atrovirens Willd. var. georgianus (Har- per) Fern. k M 3290, 9 June 1970, one dense colony or a uge tuft. Determined by T. M. Koyama. i VA austrina (Small) Mackenz. (C. muh- enbergii Schkuhr var. australis Olney). É MS. 28 May 1961, one tuft (US). M 2334, PB ay 1964, one huge tuft. Determined by F. - Hermann. Tradescantia ohiensis Raf. X T. virginiana L. i 1763, 30 May 1961, several specimens. De- ermined by E, Anderson. E. olygonum coccineum Muhl. var. pratincola b. cene) Stanford. M 2984, 15 Sept. 1968, one ony. Determined by H. A. Wahl. P, olygonum pennsylvanicum L. var. laevigatum > f. albineum Farw. ij 2578, 9 Oct. 1965, one specimen. M 2814, vd 1967, one colony. Determined by H. A. C | ice henop odium desiccatum A. Nelson var. des- atum. M 1794 coll » 1 July 1961, five specimens. Eleven ac followed. Eight of them came from is din 7 yard. The impression is that this plant * widespread in St. Louis. It was known MÜHLENBACH — SYNANTHROPIC FLORA 175 until now only from Jackson County. All deter- mined by H. A. Wahl. Chenopodium strictum Roth var. glaucophyl- lum (Aellen) Wahl. M 1860C, 16 Sept. 1961, four specimens. M 18844A, 6 Oct. 1961. Determined by H. A. Wahl. Mirabilis hirsuta (Pursh) MacM. M 2373, 21 June 1964, one small colony (about eight specimens) (UMO). The plant grew on the same place the next year (M 2458, 22 May 1965), but then disappeared. Determined by D. B. Dunn. Polanisia dodecandra (L.) DC. var. trachy- sperma (T. & G.) Iltis. M 1807, 29 July 1961, two specimens. Deter- mined by H. H. Iltis. Draba reptans (Lam.) Fern. var. micrantha (Nutt.) Fern. M 3498, 20 April 1971, one large colony. De- termined by R. L. Stuckey. Fragaria virginiana Duchesne var. virginiana. M 3225, 25 April 1970, few specimens. De- termined by F. G. Meyer. Rubus frondosus Bigel. (R. pennsilvanicus Poir.). M 3965, 5 July 1976, a small colony. Deter- mined by R. E. Weaver. Rosa arkansana Porter. M & K. Kramer 556, 7 May 1955. Determined by W. H. Lewis. Schrankia uncinata Willd. (S. nuttallii (DC.) Standl.). M 2462, 22 May 1965, one specimen. Deter- mined by D. Isely. Lotus americanus (Nutt.) Bisch. (L. purshi- anus Clements & Clements). M 1770, 17 June 1961, five or six sterile spec- imens (only one taken). M 1789, 1 July 1961, three specimens at the same place. Determined by J. D. Dwyer. Psoralea tenuiflora Pursh var. floribunda (Nutt.) Rydb. M 1945, 20 May 1962, one specimen. M 1949, 20 May 1962, one splendid specimen surround- ed by many small sterile ones. Both collections gathered in the same freight yard. Determined by D. Isely. Petalostemon multiflorum Nutt. M 1978, 8 July 1962, one colony. Determined by L. H. Shinners. 176 Parthenocissus inserta (Kerner) K. Fritsch. M 2930, 19 May 1968, one colony. Deter- mined by K. R. Robertson. Chaerophyllum tainturieri Hook. M 1635, 11 June 1960, two specimens. Six further collections followed. This species is not a rarity in St. Louis. Determined by M. E. Ma- thias. Spermolepis echinata (Nutt.) Heller. M 1729, 13 May 1961, one colony, observed on the same spot several times later, also in 1962. Determined by M. E. Mathias. Ptilimnium capillaceum (Michx.) Raf. M & White 4045, 23 July 1977, two dwarf ~ specimens. Determined by M. E. Mathias. * Monarda punctata L. M 1785, 25 June 1961, few specimens. M 3938, 12 June 1976, one specimen. Determined by J. A. Steyermark and A. G. Jones. Penstemon tubaeflorus Nutt. M 2098, 2 June 1963, one small colony. De- termined by A. G. Jones. Solanum carolinense L. f. albiflorum Benke. M 3086, 4 June 1969, five specimens. M 3089, 4 June 1969. In two different places. Astranthium integrifolium (Michx.) Nutt. M 1822, 19 Aug. 1961, one specimen. Deter- mined by W. H. Lewis. Aster parviceps (Burgess) Mackenz. & Bush. M 2262, 6 Oct. 1963, one small colony or a large tuft, also M 2305, 26 Oct. 1963. Deter- mined by R. P. Wunderlin. Aster dumosus L. M 2568, 30 Oct. 1965, one huge specimen. More flowers than leaves, a “white cloud" in- deed. Determined by R. P. Wunderlin. Rudbeckia amplexicaulis Vahl (Dracopsis am- plexicaulis (Vahl) Cass.). M 2111, 22 June 1963, one specimen. M 3938, 12 June 1976, one specimen. Determined by R. P. Wunderlin. Echinacea angustifolia L. M 2119, 25 June 1963, one specimen. Deter- mined by R. P. Wunderlin. Helianthus rigidus (Cass.) Desf. ssp. subrhom- boides Heiser (H. laetiflorus Pers. var. rigidus (Cass.) Fern.). ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 M 2745, 9 Oct. 1966, one colony. Determined by C. B. Heiser, Jr. Aster pilosus Willd. var. pringlei (Gray) Blake. M 2769, 30 Oct. 1966, few specimens. Deter- mined by R. P. Wunderlin. This taxon can even be considered as a novelty for Missouri, because it was excluded in Steyermark (1963) as erro- neously determined. Finally three species will be mentioned that — were not found along the railroads, but that arè all very rare synanthropes, previously observed | in Missouri in only one or two counties of Mis- i souri outside St. Louis. ! Stellaria graminea L. É | M 3929, 6 June 1976, lawn along the house, 5201 Landsdowne Ave., one very dense colon persisting in the following years. The residents of this house were unable to explain how this | plant could have arrived here. Determined by). | McNeil. Eurasia. Papaver somniferum L. É M 1667, 16 July 1960, between the side | and the house, 6422 N. Broadway, one spe | Unknown in the wild, probably originating " - the Mediterranean region. Coronopus didymus (L.) Sm. (ny M 3851, 1 June 1974, in the backyard ©! house, 4984 Neosho St., one colony. pier ie there up to now. Determined by R. L. ee Europe. I suspect that this synanthropic plan s been "introduced" by me, like a number o | mentioned earlier (Mühlenbach, 1979: 10 1). suddenly appeared on my property One mb another. They are all frequent plants on ther i roads in St. Louis. For this reason I 2 | them as introductions from the rail iy backyard. The seeds were picked uP por stat during my excursions by my clothing 4” stripped off. Especially the cuffs of my pee were frequently full of seeds of different P db Unfortunately, I have never seen Cor onop e » ymus on the railroads. However, this P ovt. | very inconspicuous and might have a pot looked. One more new introduction, ai rium rugosum Houtt., was discovered » | backyard in 1979. It is common along roads. The last arrival was Carduus nuta this case I do not blame the railroads. Most P" e ably the seed was transported by wind from vacant lot in the city where this plant has | considerably in the last few decades. 4 m Y 1983] It is also worthwhile to mention a weed that unexpectedly appeared in the Missouri Botanical Garden. Chenopodium ficifolium Sm. was dis- covered by the Director of the Kew Royal Bo- tanic Garden, J. P. M. Brenan, on his visit to the Missouri Botanical Garden, on the lot behind the newly erected Lehmann Building (W. G. D'Arcy 5963, 27 June 1972). Chenopodium fi- cifolium was found at several other places in the bes in the later years. It was last observed in 6. Finally, a most unusual collection of Aristida (M 4074, former RI, SL-SF and C&EI, Inbound Fr. Yd., 22 Oct. 1977) may be mentioned. It was originally collected as a unicate in a normal po ulation of A. oligantha Michx. As illustrated Figure la, which represents only one-third of the plant, it was extremely densely tufted and highly branched. The size of the spikelet parts Were as follows: lower glume 0.9-1.2 cm long including the awn 0.1-0.2 mm long, 5-7-nerved, upper glume 1.4—1.6 cm long including the awn 0.4-0.5 mm long; lemma 1.1-1.6 cm long, the central awns 1.6-1.8 cm long, the lateral awns 1.3-1.8 cm long. This specimen seems not to have àn exact parallel with any described species In this large genus. A number of prominent 3érostologists studied it but were unable to name it. It was, however, generally thought to be closely related to 4. oligantha Michx., a species with re (1.5-)2-3 cm long, the lower 3-5-nerved, pe n (1.5-)1.8-2.2 cm long and the awns we cm long. Subsequently seeds of M 4074 bi Lee (by Dr. P. Hoch) in the greenhouse. "a fell within the normal range of A. oli- on es pranching was much less dense (Fig. si " Spikelet parts range as follows: lower 207i -8-2.0 cm long, 5-nerved, upper glume y. cm long, lemma 1.4-1.5 cm long, awns Ries. cm long. The unusual plant may well dii ss àn extreme recombinant or a mutant, ig: unusual growth effects due to chemicals bbs. nvironment cannot be completely disre- De in EUN beyond any doubt that modern weed Vegetati xerts a disastrous effect on the railroad that in E was of special interest to observe rapidly Spite of this the vegetation may quite sudd Eve when the herbicide spraying is Ranken.. IScontinued. This was observed in both as Uta (TRRA). These yards were used number = yards for Passenger trains. As the Yards w Of these trains constantly dwindled, the i ere abandoned and later dismantled. Of MÜHLENBACH —SYNANTHROPIC FLORA 177 course, herbicide spraying was discontinued. A nice vegetation reappeared, demonstrating at the same time the phenomenon of trivialization, the process whereby native species take over syn- anthropes. th 4l qs d 1 1 KINU LOOK p yard, namely, Carrie Ave. freight yard (RI). This yard had the highest yield of rare synanthropes (Mühlenbach, 1979: 92). Here, as everywhere, the effects of weed killing increased each year. But in 1974 weed killing was suddenly stopped for two years, while the railroad company con- tinued normal operations, contrary to the situ- ation in the Ranken yards. When I visited Carrie unicates). One frequently hears about vanishing prairie, rain forests, and so on, but the same thing is now happening to the railroad flora in St. Louis and, of course, everywhere else as well. Only very few have realized that. Later in 1976 the freight yard was thoroughly sprayed with herbicides and the whole splendor was annihilated. In the autumn this yard looked like all other freight yards in St. Louis, a desolate, sterile landscape. I revisited it again in 1980, and to my surprise the vegetation was flourishing again, but no rarities could be detected. As it turned out, this freight yard was taken over by another railroad company (St. Louis and South- western Railway or Cotton Belt Route), which stopped all regular activities in the yard. Only one track was kept in operation for through-trains of other companies. As a consequence, the vege- tation was thoroughly trivial. On the basis of these observations, it is pos- sible to enumerate a rule concerning the weed killing process. If herbicide spraying in a yard is discontinued, vegetation will promptly reappear. In case the usual operations of a yard are not continued, trivialization will follow, but when normal operations of a yard are continued, syn- anthropes appear. How many and which kind will depend on the volume and character of the cargo. But natives may also appear for the first time. My conviction that native plants are also spread by railroads was strengthened by finding Rudbeckia amplexicaulis Vahl in 1976. It is a very conspicuous plant and it is inconceivable that I had overlooked it for 22 years. As in other botanical endeavors, railroad 178 botany published determinations of synan- thropes must occasionally be changed. In Miih- lenbach (1979) Kochia iranica (Hausskn. & Bornm.) Litw. turned out to be K. sieversiana (Poll.) C. A. Mey. according to W. A. Weber. Kochia sieversiana has been found repeatedly in Missouri, but in only one place, Stockyard Switch, Joplin, Jasper County. Centaurea stoebe L. subsp. stoebe and C. maculosa Lam. in Mühlenbach (1979) turned out to be the same species. Ac- cording to G. Wagenitz (Góttingen), who has de- termined my Centaurea material, plants of the C. stoebe group are not rare in North America, but have mostly been determined as C. macu- losa. A mistake in the citation for Sauer (1967) was also made in Miihlenbach (1979) and is here corrected. R. L. Stuckey has kindly pointed out to me that there is an older American paper (Stair, 1900) on railroad weeds, that was overlooked by me earlier. Besides this he mentioned two other more recent papers (Catling & McKay, 1974; Thomp- son & Heineke, 1977) not cited by me. An ad- ditional paper that has recently been published is Arnold (1981). Mrs. Jean Warholic (Freeville, N. Y.) brought to my attention Ross (1943). This paper treats the weeds at the Knapp farm of George Peabody College for Teachers in 1923. Among other things, the plants on the right-of-way of the Nashville, Chattanooga & St. Louis Railroad that trans- verses the farm were enumerated and evaluat- ed—altogether 82 pin The great majority were also found in St. L : It also seems iac to mention here a corrected determination for Arctium tomento- sum Mill., mentioned as a synanthropic species in Stevani (1963). Both collections at MO are not A. tomentosum but A. minus (Hill.) Bernh. LITERATURE CITED NOLD, R. M. 1981. Weeds that ride the rails. Nat. Hist. 90(8): 58-65. CATLING, P. M. & S. M. McKay. 1974. An interesting association of plants along a railway track at West Hill, Ontario. Ontario Field Biol. 28: 49-51. ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo.. 70 CorRELL, D. S. & M. C. JOHNSTON. 1970. Manual of the Vascular Plants of Texas. Texas Research Foundation, Renner, Texas. FERNALD, M. L. Gray’s M of Botany. Ed. 8. American Book Co., New GLEASON, A. H. 1968. The New Britton and Brown Illustrated Flora of the Northeastern United States and Adjacent Canada. Hafner Publ. Company, Inc., odes I. M. & G. C. Druce. 1919. The Adven- ve Flora of Tweedside. T. Buncle & Co., At- ath. 1935-1974. Illustrierte Flora Ma MS a, Ed. 2. Carl Hanser Verlag, Mün Hid. " J. 1970. Manual of the cin " the Rocky Mountains and Colorado Basin. je vice, U.S.D.A. Agriculture Handbook No. 3 Washington, D.C. ; JALAS, J. & J. SUOMINEN. 1979. Atlas florae EUM HEGI, G. MUHLENBACH, V. Contributions P. d anthropic ETA A flora of the railroa e Louis, Missouri, U. S. A. Ann. Missouri Onwi,J. 1962. On Japanese Echinochloa. Actà Phy- : 3: fields at s in Wire T s and 1967. The grain. amaranths be E relatives: a i E . Ann. ergy Bot. Gard. 54: 105" Ms | on : 76. Adventive ne (Moscov " Leningrad) 61: 1445-1454. 1981. SHILDNECK, P., A. G. Jo NES & V. MÜHLENBACH: ois plants — atus : and a note on the occurrence of RE Mig | in North America. Phytologia 47: 265- vM | STAIR, L “PD 0. rt on railroad W' ual Rep. Ohio Acad. Sci. 44-50. The lows STEYERMARK, J. A. 1963. die Missouri. | State Univ. Press, Ames, Iow: asculat — Diu RILATE PME 975. prairie | ra of the Depots ae rai um Jackson County, Illinois. Acad. Sci. 70: 114—127. jement VELENOVSKY, J. 1898. Flora Bulgarica. SupP I. Fr. Rivnaé, Prague. LM VICKERY, J. W. 1975. - Echinochloa e p. EB Nato | Gramineae. Flora of New Sou er rans. Illinois Stale Herb. New South Wales, Sas ege YABUNO, T. 1966. Biosystematic study of € Echinochloa. Jap. J. Bot. 19: 277-3 l —— ——— Ó,ÓQHRRIRAEREEDUEP P REVISION OF THE GENUS STERIGMAPETALUM (RHIZOPHORACEAE JULIAN A. STEYERMARK ! AND RONALD LIESNER? ABSTRACT enus TATE mdpetatun consists of 7 species, with 4 newly described species (S. exappen- eo xum, eg ps S. het subsp. Macs O S. resinosum, and S. . tachirense), and 1 new subspecies (S. guianense e basis of the presence or absence of a resinous exudate, Betonet or reduction of sn NM of the petals, and opposite or Msc leaves, two ner Aided are rec- ognized: pu and Balsam ocaulon. On the bas e species is provided, followed by a description of the spe asis of the yp or ypogyneae, within the F Rhi- cies and their ot geographical distribution. Newly discovered anatomical evidence, which reveals the presence of a -type sieve element in the Myrtales and is additional SPEA ce supporting t details that the family is atypical in the Myrtales, Sterigmapetalum was established in 1925 by uh!mann with a single species, S. obovatum, collected from Amazonian Brazil. It was differ- , by the Possession of dioecious flowers, of Sal) seg- ments and stamens, absence of a disc, and a greater number of cells of the ovary. Since that i. two additional species, S. colombianum j onachino ( 1944) from Colombia, and S. gui- nense Steyermark (1952) from the Venezuelan uayana, have been published. is recent explorations in Venezuela near ain border, a fourth species has been : ao i red. In an attempt to bring our present w dai of the genus up-to-date, material has us E wed from the herbaria of F, NY, and Ee the dioecious character of the us, iol with the paucity of specimens n herbaria, have added to the diffi- vision. However, as a result of this additional taxa have become manifest and Scribed below. We wish to thank the cu- rato : Ts of the Roc ea herbaria for the loan their mate eria GROss MORPHOLOGY FE exudate covers the young buds, k aay upper internodes of the stem in s. not found mand S. exappendiculatum, but is This Ene other members of the genus. er, associated with the reduced lat- ' Her ? Misso ANN. Missouri Bor, GARD. 70: 179-193. 1983. izophoraceae, indicates that the family should be excluded from the the view provided by the gross eM floral and should be removed from that order. eral flanges of the petals, and opposite leaves, form the basis for the establishment of a subge- neric category within the genus. Stipules. The stipules are uniformly interpet- iolar throughout the taxa. They are glabrous in . colombianum, S. tachirense, S. heterodoxum, and S. Aa einan. n neate at the base and broadest above the middle, resulting usually in an obovate shape. In S. tach- irense they are rounded or broadened at the base. In S. exappendiculatum they are elliptic-oblong and broadest toward the middle. The leaf mar- gins are mainly entire, but show varying degrees of crenulation, rather markedly so in S. exap- pendiculatum. Inflorescence. The inflorescence is usually sub- terminal and arises in the axils of the uppermost leaves. Two or more peduncles near the apex of the stem in the axils of the uppermost leaves are mocaulon the peduncles arise lower on the stem between the second and sixth internodes below the apex, and are solitary in the leaf axils. The bario Nacional de Venezuela, Caracas, Venezuela. Ssouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166. 180 flowers are arranged in unbranched or 1-4- branched dichotomous cymes, and are sessile to shortly pedicellate. In subgenus Ba/samocaulon the peduncles are -a with a resinous exu- date, especially near the Calyx. The calyx is ia Janae 4—7-toothed. Sterigmapetalum tachirense is the only species thus far known in the genus with four calyx lobes. The calyx tube encloses the floral parts, but is hypogynous and not coherent either with the ova ary Or other €— kd —- got The outer surf: tate y Ser iceous. In subgenus Balsamocaulon a resinous exudate covers the young calyx and margins of the lobes, which are manifestly elevated into thickened, linear ridges. The inner surface of the calyx lobes is usually densely strigose or sericeous. Petals. The petals are free and hypogynous, inserted at the base of the flower between the base of the calyx tube and the disclike membrane to which the st hed (Fig. 1A). They vary in number from four to six and differ greatly in ornamentation (Fig. 2). In shape they vary from linear, oblanceolate, to panduriform, with a usually 3-lobed or 3-parted apex and two lateral flanges or lobes bearing 2-18 short to elongated, fimbriate, ofte (Fig. 1A). Be- tween the lateral flanges is a central, hooded, incurved portion with three to many crowded, involute appendages. This central apical part is often surmounted by a penicillate tuft of hairs, which are also inflexed. Hairs may also be pres- ent on the dorsal part of the petal, and occa- sionally elsewhere. Sometimes, as in S. exap- pendiculatum, the apical tuft of hairs, as well as the fimbrillate lateral flanges, are lacking (Fig. 2: 2A, 2B). Androecium. The stamens vary in number from 8 to 12 in the staminate flowers, and are usually unequal in an alternating arrangement (Fig. IC). The anthers are versatile, dorsifixed, broadly ob- long, and rounded at both ends. The filaments are attached to the border of an annular mem- branous structure, which has been referred to as a disc (Fig. 1C). This structure could be equally interpreted as the fused bases of the filaments. It is inserted at the base of the flowers between the base of the petals and the ovary or ovary rudi- ment. Gynoecium. The ovary is completely superior in the pistillate flower (Fig. 1A), as is the ovary rudiment in the staminate flower (Fig. 1 B). Both are inserted on the receptacle at the base of the flower. The ovary is depressed-globose to sub- ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 obovate, and 5-6-locular with two pendulous ovules in each cell. In the few species in which - pistillate flowers are known, it is covered by a dense tomentum. The style is very short or ob- solete. The very short stylar branches are covered by a crowded mass of incurved stigmatic threads, which form a radiate structure depressed cen- trally. At the base of the ovary but not adherent to it is the disc. In the pistillate flowers, minute projecting staminodia, usually 10-12 in number, are attached to the border of the disc (Fig. 1B). Anatomical features. As a result of his recent studies of sieve-element plastids in various fam- ilies usually included within the Myrtales, Behnke (1983) found the so-called (S-type) plastid, those containing several spherical starch grains, to characteristic and constant within the order. of great significance is his discovery that the sieve- element plastids of the Rhizophoraceae are com- PSY unlike those of any of the other pe | the Myrtales in BOEDE, s Nu instead . contain whi ch rea- son these plastids ai are placed i in the E pes group. Although the genera Sterigmapetalum and T sipourea were not studied by him, four oth genera (Rhizophora, Bruguiera, Carallia, 29 Crossostylis) were investigated. GENERIC AFFINITIES Rhizophoraceae. In a revision of d pre species of the family, Prance et al. 0r sai entiated these two genera on the basis ° Cass cillate (Sterigmapetalum rea), and pedunculate inflorescences petalum op sessile, axil = ae iir (Cassipourea). In light of our present kno it should be modified to state that the Pe n | Suriginopen usually have two lateral smb | ach w ith 2-18, short to elongated, fi abide ic borne on either side o or inflexed, central lobe furnished cil dense tuft of hairs or with three a ppendage> aun When Kuhlmann described Sterigma er (1925), he stated that the genus, as then kn only from Brazil, comprised v was habiting high land, whereas Cassipou" = considered to be a genus of small trees. H y related erior | men 181 STEYERMARK & LIESNER—STERIGMAPETALUM SAA SINN NN segments of petal. | | stigma. calyx lobe. ovary. Style. calyx tube. ovule. basal part of petal. Staminode. A Pistillate flower. C | calyx lobe. petal. Calyx tube. anther. filament, ovary rudiment. B ~ j tds e ee Staminate flower. en QURE 1, A. St terigmapetalum obovatum, pistillate flower, semi-diagrammatic. B. Sterigmapetalum gui- longer’ Staminate flower, semi-diagrammatic. C. Portion of androecium with annular disk, sh and shorter filaments, Sterigmapetalum guianense. owing alternating 182 ANNALS OF THE MISSOURI BOTANICAL GARDEN famy Ue Beiceso- l, ventral view. Sterigmapetalum exappendiculatum, 2A. Petal, ventral view; 2B. Petal, a Sterigmapetalum cade m, 3A. Petal, dorsal view; 3B. Petal, ventral view; 3C. Petal, kein view. e: FiGURE 2. Vd iid guianense subsp. ichunense, 1A. Petal, lateral view: 1B. Petal, dorsal = | petalum heterodoxum, 4A. Petal, dorsal view; 4B. Petal, ventral vi iew; 4C. Petal, lateral vi l 1983] this latter observation does not stand the test of observations made on Cassipourea in other countries. In Panama, for example, Croat (1978) reported Cassipourea elliptica as a tree “13-17 m tall,” and refers to a report by Allen to a tree as much as “30 m" tall. In Venezuela Cassipou- rea may vary from a shrub only 1.8—4 m tall to trees 8-15 m in height. As a result of a study of these two genera, we find that the chief differences existing between them may be summarized in Table 1. TAXONOMIC POSITION Since both Sterigmapetalum and Cassipourea have hypogynous flowers, it is instructive to note how Previous workers have characterized the family Rhizophoraceae in their respective treat- menis, At the time Kuhlmann (1925) published Sterigmapetalum, he considered it closest to the a Blepharistemma, as a member of the *Pi&ynous, rarely perigynous families. The Rhi- "ophoraceae were described as having the ovary more or less united with the floral axis. oe and Hooker (1862) divided the Rhi- ies. iia (by them designated Rhizophoreae) ni tribes. The genera Cassipourea, Ble- "A emma of the West Indies, and Dactylo- des s of Africa and Madagascar, were placed are be Legnotideae and separated from oth- thera that tribe by their superior ovary. an 76) characterized the Rhizophoraceae MER ng an ovary “saepissime infero, raro su- : Aat Cassipourea is described as having sia "y liberum." Schimper (1893) recog- ed two subfamilies ofthe Rhizophoraceae: (1) we and (2) Anisophylloideae, the os Isunguished by stipulate, opposite leaves, US Or perigynous flowers, single style, en- while ^ Present, and the fruit a berry or capsule, the latter was differentiated on the basis of i oo calcium oxalate crystals, and placen- With ; ‘uous with the style, and (2) Macariseae Perigynous flowers, seed with appendages, STEYERMARK & LIESNER—STERIGMAPETALUM 183 TABLE 1. Comparison of Sterigmapetalum and Cassipourea. Sterigmapetalum Cassipourea Flowers dioecious. Flowers perfect. Style absent with sessile Style elongated and large stigma. exserted. Stamens included. Stamens at least equal- ing calyx and slightly exserted. Stamens 20-25. Inflorescence sessile. Stamens 8-12. Inflorescence peduncu- Appendages of lateral flanges glabrous. Lateral flanges of petals with 2-18 appendages. Appendages of petals pi- losulous. Lateral flanges of petals with numerous ap- ndages. Petals with 2 lateral Petals simple, not lobed. flanges (lobes) and a central incurved lobe. Ovary 5-6-celled. Ovary 2-4-(usually 3)- celled. Leaves 3-4-verticillate, Leaves opposite. more rarely opposite. simple crystals together with pairs of calcium oxalate in the leaf blades, and the placenta dis- continuous with the style. This treatment was modified by Melchior (1964) by treating the above categories as four equal tribes: (1) Macariseae, (2) Gynotrocheae, (3) Anisophylleae, and (4) Rhizophoreae, the first two characterized by the possession of stipulate opposite leaves, one style, and endosperm, the Anisophylleae by its exstipulate leaves, three to five styles, and no endosperm, and the Rhi- zophoreae by being mangrove plants with seeds producing an elongated [ ling hypocotyl that germinates within the fruit; the other tribes con- sist of inland plants whose seeds germinate in the soil outside the fruit. Many authors have characterized the Rhi- zophoraceae as having perigynous or epigynous flowers with the ovary stated to be mostly infe- rior. Lawrence (1963) described the stamens as “situated on the outer edge of a lobed perigynous or epigynous disk,” but further characterized the family as having the “ovary variable in position depending on degree of adnation of perianth (i.e., superior, half inferior, or incompletely inferior)," while Standley and Williams (1962) described the ovary as “usually inferior,” but also stated 7 184 that the “‘calyx-tube is more or less adnate to the ovary, rarely free, the limb produced beyond the ovary," but, in their description of the genus Cassipourea indicated the ovary as free. Hey- wood (1978) stated that the flowers are "hypog- ynous to epigynous," and the ovary "inferior or superior." Aristeguieta (1973) described the ovary as varying from superior to semi-inferior, and correctly characterized Sterigmapetalum and Cassipourea as genera having superior ovaries. Croat (1978), likewise, correctly stated that the zophoraceae Hutchinson (1959) stated that the calyx tube is “adnate to the ovary or free.” Cronquist (1968) placed the Rhizophoraceae in the order Cornales, characterized by him part- ly as having epigynous flowers and indehiscent fleshy fruits. He also Considered that. its posses- sion of well-devel unitegmic ovules was consistent with its inclusion as a fam- ily within the Cornales. At that time he believed it to be “most useful to include the Rhizopho- raceae in the Cornales as a near-basal side-branch not far distant from the Myrtales. If one prefers small orders and does not object to monotypes, the establishment of a separate order might per- haps be defended." This concept of a distinct order for the Rhi- zophoraceae was subsequently defended by Cronquist (1981: 655-659). In this latest work (Cronquist, 1981), he recognized the family as a monotypic, separate order, Rhizophorales, which he placed between the Myrtales and Cornales. Principally on the basis of its absence of internal phloem, he separated the Rhizophorales from the Myrtales, while from the Cornales the sep- aration is based mainly on the possession of bi- tegmic instead of unitegmic ovules, although he previously (Cronquist, 1968) stated that uni- tegmic ovules were common to both Rhizo- phoraceae and Cornales, as well as by additional characters of stipulate leaves and absence of ir- idoid compounds. The flowers of the Rhizophoraceae are stated by Cronquist (1981) to be perigynous or epigy- nous, be he —— the ‘genera regione and A a superior ovary. Actually, Cronquist (1981 De ex- pressed uncertainty in the proper disposition of the Rhizophoraceae, and suggested various pos- sibilities for its placement, not only in the Cor- nales and Myrtales, both of which he believed ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 unjustified, but also in the order Rosales. The latter order he considered as having characters too primitive to encompass the Rhizophoraceae, although he suggested that affinities with the Grossulariaceae and Hydrangeaceae might be conceivable. Since the Rhizophoraceae are not compatible with any of these three orders, Cron- quist's alternative was to treat the family as 4 distinct, monotypic order. ecent anatomical investigations on the plas- tids of sieve elements support the view that the Rhizophoraceae should be excluded from the Myrtales. The previous work on nearly 400 fam- ilies of flowering plants by different workers - (Behnke, 1974, 1976, 1981a, 1981b, Ee 1982b; Behnke & Pop, 1981; Dahlgren, 19 Dahlgren & Thorne, 1983; Mabry, 1977) hav shown that the plastids of the sieve element may _ be referred to either the S-type with starc e or the P-type with protein crystals. Based on à type plastids in the Rhizophoraceae, as m ed with S-type in other families of the M e (Melastomataceae, Crypteroniaceae, Wes aceae, Oliniaceae, Combretaceae, Lythraceat Sonneratiaceae, Punicaceae, Myrtaceae, Trapt ceae, and — Behnke ( 19820) er ed that th be excluded Myrtalee? These new anatomical data provide evidence to support the previously noted trè in the family from hypogyny to epigyny. ma- ndition of Sterig pisei the atypical gh inconsistent further families of the Myrtales. The combi | | | trend i atomical evidence with that of the diverge” well | ral structures of the Rhizophoraceae the Myt justify the exclusion of the family from : ids alliance of families having P-type plast! compatible floral features. TAXONOMY Since both Sterigmapetalum and Cara " have hypogynous flowers with complete : oral perior ovaries that are not united to n axis (Fig. 1A, B), they may be conside tuni as constituting a new tribe within the t described as follows: TRIBE HYPOGYNEAE Steyermark & D: Type: Cassipourea Aubl., Fl. Guiane Lt 211.1775. | 1983] Flores itn Folia opposita vel verticillata. Ovarium superu Sterigmapetalum Kuhlmann, Arch. Jard. Bot. Rio de Janeiro 4: 359. t. 32. eee TYPE: EET mapetalum obovatum Kuhlm buquerque in Prance et al., Acta Amaz. 5: 1975. Trees with interpetiolar stipules. Leaves op- posite or 3-4-verticillate, petiolate, entire or crenulate-serrulate. Inflorescence axillary, sub- terminal or in the upper axils, pedunculate, cy- mose, the cymes congested, unbranched or di- chotomously 1-4-branched. Flowers unisexual (the plants dioecious), actinomorphic, few to nu- merous, sessile or pedicellate. Calyx campanu- late, Nha or -lobed, the teeth or lobes °r panduriform, 3-lobed or 3- -parted at the apex With 2 lateral flanges bearing 2-18, short to elon- - poe appendages on either side of a a tral, hooded, inflexed portion bearing 3 to any crowded, incurved appendages, this cen- tral portion often surmounted by an inflexed Penicillate Pubescent tuft. Stamens 8-12, usually hequal. Anthers versatile, dorsifixed, broadly STEYERMARK & LIESNER—STERIGMAPETALUM 185 oblong, rounded at the base and apex. Filaments attached to the border of or fused with a basal annular membrane which is hypogynously in- serted at the base of the flower. Ovary rudiment present in the staminate flowers. Ovary threads. Fruit a septicidal, 5—6-celled capsule; seeds collateral, pendulous, oblong, winged; wings oblong-falciform, appendiculate or elongated, caducous. Endosperm oleaginous, slightly thick; radicle cylindrical; cotyledons foliaceous, flat, about equaling the length of the radicle. Geographical distribution: Seven species are known, distributed in Amazonian Brazil north to the Guianas, Venezuela, and eastern Colom- bia (Fig. 3). The genus Sterigmapetalum may be divided into two subgenera: la. No resinous exudate present; leaves usually 3-4-verticillate, rarely opposite; lateral flanges of the petals well developed with several to many fimbrillate Bis ceca cadit Sterigmapetalum pe- duncles, and upper cauline oxic Ja S Op- posite; lateral — of the petals recs ed to nearly obsolete .. Subgenus Balsamocaulon a - i-a KEY TO THE SUBGENERA AND SPECIES OF STERIGMAPETALUM la, Leaves always opposite; resinous exudate present on apical portion of stem, — and buds; Peduncle solitary, uiae between the 2nd and 6th nodes below the apex of the st mieten: 2a. Leaves a acute | or subacute s > < P an [x — e BE =p D =] go oO a e Lad Ras “p Laud O A g E [e e = c Qa [c] HH 36 Leave rarely Dm 4a. ER 23 cm 1 21 mm long .. Peduncles 0. flow Wers terminating 4 S u nd narrowed at the apex, “broadest at the middle; stipules acu - 6. Salis rounded at the base; staminate flowers 4- mer “saga Hea acute or subobtuse at the i ud or r pistillate flowers 56-mero 3-3c cm long; inflorescence either 1 the axes; lateral flanges of the peta sigg mm long, scarcely equaling or only slightly exceeding dede calyx : J a. s - enus s Balsamocaulon S. hurts irense manifestly 2-4-dichotomously branched; lateral flanges of the petals riches mm “rahe including the appendages greatly exceeding the an. tube, 14- i uS ee ie . S. obovatum “dichotomous c or with E rue crowded ls, pcd the appendages, at most more or less denticulate-crenulate; petals lacking fimbrillate * lateral vada 5 3. € ee iai Sb. Leaves ‘broadest above the middle, ‘runcate-rounded or 'emarginate at the apex; Fo entire; petals with fimbrillate lateral 6a. Upper leaf surface with conspicuously event tertiary venation. 186 ANNALS OF THE MISSOURI BOTANICAL GARDEN 8 5. OBOVATUM O S. GUIANENSE subsp. GUIANENSE à ndi 2 A S. GUIANENSE subsp. ICHUNENSE LX : DICULATUM = S. EXAPPEN * | y 5 t | : 2 BIANUM d: VENEZUELA e». 3X- S. COLOM M e fj s. TACHIRENSE / s" O s. REsINOSUM | BUE ys Jauvana ` Q s. HETERODOXUM | cordura A , summan ere \ | A 0 :GUANA, A> E Bi | y L] Fd BRAZIL F | / fp | E] | | BOLIVIA f FiGURE 3. Geographical distribution of the genus Sterigmapetalum. 7a. Stipules parim without; calyx exterior glabrous or at most "ena appre pubesce 7b. Sipai Sud puescon without calyx exterior aah seri Sterigmapetalum subgenus Sterigmapetalum. TYPE: Sterigmapetalum obovatum Kuhl- mann. sie resinosum nullum. Folia vulgo 3-4-ver- ticillata, ra l bene evi cum appendicibus fimbriatis.- 1. Sterigmapetalum obovatum Kuhlmann, Arch. Jard. Bot. Rio de Janeiro 4: 360. t. 32. 1925. Fig. 1A, Fig. 3 Trees 10—30 m tall; branches 3—5-verticillate, quadrangulate; upper internodes densely buff-to- mentellose, the lower internodes glabrescent. aves 3—5-verticillate, petiolate, the petioles 5— 15(-20) mm long, densely buff tomentellose; leaf blades coriaceous, obovate, rounded to sub- emarginate at the apex, obtuse to acute at the ssed- e colombianum™ | | bsp. guianense | | ceo . guianense pire: p- sp. chu b. S. guianense subs Í L| base, 7-17 cm long, 3.5-11 cm wide, pel and prominently nerved above, tome - on the nerves, less densely pubescent 0 nerves | faces, entire, S MEUS revolute, the hic: 11-19 on each side, ascending, S ideda near the margins, subsulcate evated below, the tertiary veinlets reti prominent on both sides. Infloresc e t | =| | chotomously 3-forked. Flowers | each axis. Calyx deeply campanulate, to \ long, 3-4 mm wide, densely buff-app the lal | mentellose without. Petals 15 mm long, "m | inate portion linear-oblanceolate to ligula lanceolate, 5 mm long, 1 mm wide, e 4 } apically 3-lobed, the apical central portio 9-10 0.8 mm long, bordered by 2 lateral — je | mm long, each lateral flange bearing a a mentous appendages basally attached tO | 1983] inar portion 2 mm long; apical margin of the central portion bearing 3 short appendages 0.5- 0.8 mm long on one side and 1 elongated ap- pendage 9-10 mm long on other side. Stamens 10 in one equal series; filaments free in the upper 0.5 mm, united into a membrane at the base 1.5 mm long; anthers narrowly oblong, subemargin- ate, 1.2 mm long. Ovary rudiment conic, 1.7 mm long, 0.8 mm wide, densely buff tomentose or sericeous. Pistillate inflorescence: forked cyme with 2-3 flowers borne on each axis. Flowers subpedicellate; pedicels 0.5—1.5 mm long, dense- ly buff-tomentellose. Calyx deeply campanulate, narrowed basally, 7-8 mm long, 4-5 mm wide above the base, 1.5 mm wide at the base; calyx lobes 5-7, triangular, acute, 1.5-2.7 mm long, 1.5-1.7 mm i | mm long, and 2 lateral flanges bearing 4- ME appendages, the united portion at the hes 2.5-3 mm long, 0.7-1 mm wide. Ovary die. depressed-subglobose, subtruncate at the Pm sn 5-angled, 2 mm long, 3 mm wide, dé d tomentose. Staminodes 10, 0.8 mm long, ad rted anthers globose, 0.1 mm long and mia the filamentous portion 0.6—0.7 mm long, "s a than half-way down, the united por- d "s long at the base. Fruit obovate or indi 95ovate, 4 cm long, 2-2.5 cm wide, he E sericeous. Seed oblong, plano-convex, 8 ng, 3.5-4 mm wide. Speci : pality Huan examined: BRAZIL. AMAZONAS: Munici- sin rahe near Livramento, rio Livramento, ba- Mandos Madeira, Krukoff6788 (NY), 7021 (NY, US) Es ies lower rio M (P; "uere "ed 75 (NY, US); Manáos, Ducke R p ‘gues, Coêlho & Coélh à m, Kill; i o 8468 (NY); Manáos, 25 » Killip & Smith 30197 (NY, US). PARA: Rio Tapajoz, s PEU 31 Aug. 1916 (fl.), 15 Dec. J. TER Way, 215 k Maguire m W o m, Magui ls rcipmapetalum guianense Steyermark, b diana Bot. 28(2): 422. 1952. TYPE: Ven- ela, Estado Bolívar, dense forest at base STEYERMARK & LIESNER—STERIGMAPETALUM 187 of cerro along río Karuai, Ptari-tepui, 1,220 m, 27 Nov. 1944, Steyermark 60658 (F, ho- lotype).— Fig. 1B, C. Trees 15-20 m tall; branches 3-4-verticillate, terete; upper internodes sericeous, the lower in- ternodes glabrous. Stipules ovate-triangular, ob- tuse, 6-8 mm long, 2-2.5 mm wide, densely se- riceous both sides. Leaves 3—4-verticillate, shortly petiolate, the petioles 5-9 mm long, glabrous to densely sericeous, leaf blades coriaceous, broadly cuneately obovate to obovate- or oval-oblong, subtruncate to slightly emarginate or rounded at the apex, subobtuse to acute at the base, 3.5-8 cm long, 2-7 cm wide, the margins remotely and inconspicuously minutely glandular-denticulate to entire, mainly glabrous except the midnerve beneath sparsely to more densely appressed-pu- bescent, the lateral nerves 9-12 on each side, arcuate-ascending, anastomosing before the margins, prominently impressed above in subsp. guianense, elevated below, the tertiary veinlets conspicuously impressed above in subsp. gui- anense, absent or obsolete in subsp. ichunense, conspicuously reticulate below. Inflorescence 1 to 3, terminal or in the axils of the uppermost leaves, compactly dichotomously cymose. Sta- minate inflorescence: peduncles 0.3-3 cm long with two main axes 1.5-2 mm long, moderately to densely appressed-pubescent or pilosulous, 12- 15-flowered. Flowers white, on short pedicels 1.5— 2 mm long or subsessile, 6-8 mm long, 5 mm wide, 5-merous. Bracteoles narrowly lanceolate, subobtuse, 1.5-2 mm long, 0.5 mm wide, pilo- sulous both sides, involucrate. Calyx 5-dentate, urceolate, 4-5.5 mm long, 3-3.5 mm wide, gla- brous to sparsely pilosulous without, densely se- riceous within, the teeth triangular-ovate, sub- acute, 1.5 mm long. Petals 5, spatulate, dorsally densely pilose, ventrally sparsely pilose, subun- guiculate toward the base, 6 mm long at matu- rity, 1.8 mm wide apically, the central cucullate portion shorter than the lateral flanges, provided with an incurved penicillate tuft; lateral flanges exceeding the central portion, 1.5-2.5 mm long, 3-4-fimbrillate, the appendages glabrous. Sta- mens 10; filaments unequal, 3-4 mm long, united in the basal 1 mm, glabrous; anthers broadly oblong, 1 mm long. Staminodes squamelliform, 1 mm long, bordering the basal sides of the fil- aments, united with the base of the filaments. Ovary rudiment densely hirsute. Pistillate inflo- rescence: Fruit obovate, 16 mm long, 8 mm broad, densely appressed-pubescent. 188 Y TO THE SUBSPECIES OF STERIGMAPETALUM GUIANENSE la. Upper leaf surface lo m and sub- reticulate vue | venation .. guianense € subsp. guianense lb. Upper leaf nu ace lacking A m" subreticulate tertiary venatio 20S. tiaina ‘subsp. ichunense 2a. Sterigmapetalum guianense Steyerm. subsp. guianense opip distribution: Talus slopes and dw forest on sandstone table mountains at 1 ce 1,615 m altitude in southeastern Estado Bolivar, Venezuela, and dwarf forest bordering Savanna at 550-565 m altitude at base of Ta- felberg Mountain, Surinam (Fig. 3). Specimens examined: VENEZUELA. BOLIVAR: Ptari- tepui, Steyermark 60658 (pistillate plant, type of S. radas at base of slopes, Stey- ermark 60042 (F, US, VEN); vicinity of “Misia Kathy dier No. I, Tafelberg, 565 m, Maguire 2 24737 e NY); 200 m, south Savanna No. IV, Tafelberg, 550 m, Maguire 24780 (NY). 2b. Sterigmapetalum guianense subsp. ichu- nense Steyermark & Liesner, subsp. nov. TYPE: Venezuela, Bolívar, Sierra Ichün, la- ss boscosas Ly filas s sur del he d tributario del río pue 4°46'N, 6318" W, 500—625 m, 29 Dec. 1961, Julian A. Stey- ermark p Mon holotype: NY; US. isotype). — 1:4r. 1: b er supra haut manifestis vel obsoletis. Geographical distribution. Sandstone forested areas in southern Venezuela at 110-500 m (Fig. 3). AOL specimens examined: VENEZUELA. BOLI- AR: Sierra Ic soup Ping ised 90428 (US, VEN). TERR. FED. AMAZON arlos de Rio Negro, dense forest alrededores del Faster se 125 m, 17-18 Aug. 1970, Steyermark & Bunting 102747 (US, VEN); road from to Santa Barbara, 12—40 km rom ando, 110 m, 24 Mar. 1974, Gentry & Tillett 10868 ( (MO, VEN). The specimens cited from Terr. Fed. Ama- zonas tend to have longer petioles and shorter pubescence. Additional collections with flowers from that region may eventually show that an additional subspecies or species is represented. ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot 70 3. Sterigmapetalum exappendiculatum Steyer- mark & Liesner, sp. nov. TyPE: Venezuela, Bolivar, montane forest, cumbre of La Es- calera, Alto río Cuyuni, rio Uiri-Yuk, 1,000 m, Aug. 20-21 1962, Maguire, Steyermark & Maguire 46876 (VEN, holotype; NY, iso type). — Fig. 2: 2. . Arbor 12-20-metralis, ramulis juvenilibus intera ovato-lanceolatis subobtusis 7-10 mm longis, E] mm latis dense sericeis; foliis 3-verticillatis, D 5 11 mm longis sericeis; laminis subcoriaceis one | vel elliptico-oblongis apice rotundatis vel late 0 raro acutis, basi obtusis vel subacutis 7-14 us] | 3.7-8 cm latis supra praeter costam sp: M P n glabris, subtus modice vel dense adpreso-pilos! ense sericea, marginibus utroque latere Ii bs et | ulato-dent 12 d ante marginem 4-10 m ems a | mosantibus supra conspicue impressis subtus vat- venulis tertiariis utrinque manifeste reticulatis oo | isque; inflorescentiis solitariis pedunculatis ' foliorum supremorum insidentibus, pedum ^d longis dense breviter puberulis; infloresceat i: cin s longis 1-1.5 mm latis: calyce 5-denta bus iceo; petalis juvenilibus obovato- “subdolabrifornib» | dorsum superneque papillatis ventraliter 1n bes pi | pilosulis, apice breviter pilosulis, parte m » incu rvata tribus appendicibus brevibus sica juveni o basi voideo apice obtuse dense hirsutulo 5-lobat csineltis circumcincto. | Trees 12-20 m tall; young branches Wee internodes densely tomentellous. Stipules pe lanceolate, subobtuse, 7-10 mm subcoriaceous, oblong or elliptic ed or broadly obtuse, rarely acute, obtuse or subacute at the base, 7-14 cm Young pistillate angen congest 5mm flowered, the bracts 3-5 mm long. !- wide. Calyx iia sericeous man without. Young petals obovate-subdom y dorsally and above papillate, v ventrally and 1983] pilosulous, the apex shortly pilosulous, the me- dian portion incurved at the apex bearing 3 short appendages, the appendages none. Young ovary ovoid, obtuse at the apex, densely hirsutulous, 5-lobed, surrounded at the base by squamellae. .PARATYPES: VENEZUELA. BOLIVAR: Chimantá Massif, vicinity of Bluff Camp, at base of west-facing sandstone bluffs of Chimantá-tepui (Torono-tepui), 1,700 m, 2- 4 June 1954, Steyermark 75634 (NY Geographic distribution: Montane forests overlying the Roraima sandstone formation, southeastern Estado Bolívar, Venezuela, at 1,000- 1,700 meters (Fig. 3). This species is well-marked by the absence of lateral appendages of the petals, sparse indument of the petals, crenulate-denticulate leaves, and rather densely tomentose stems, peduncles, and lower surface of leaf blades. 4. Sterigmapetalum colombianum Monachino, Trop. Woods 77: 10. 1944. TYPE: Colombia. Magdalena: Río Jabalí region (La Victoria), 1,000-1,200 m, Dec. 1931-Feb. 1932, Es- pina & Giacometto A31 (NY, holotype; F, Isotype). Tree 10-25 m tall; uppermost internodes ap- Pika glabrous but covered by a minute, pie: perenne indument. Stipules ovate, sub- wae i. mm long, 1.2 mm wide, appearing ue M at resinous in bud. Leaves 3—4-verticil- Riis PY petiolate, the petioles 2-6 mm long, E ud y glabrous but basally with minute hairs; thes €s subcoriaceous, obovate, rounded at D ru acute at the base, 4.5—6.7 em E. €m wide, glabrous on both sides (mi- ent ben tered, pale, appressed hairs may be pres- each oer entire, the lateral nerves 8-12 on the m 6 ascending, anastomosing 3-5 mm from rubus o tertiary veins finely reticulate and On both surfaces. Pistillate inflores- lary, shortly pedunculate in the axils long. dp Tmost leaves; peduncles 8-10 mm bitin: Owered, densely minutely appressed- only). Cal Flowers umbellate, sessile (in bud rely a yx (in bud) apparently 5-dentate, brous wi Bo Pabescent or practically gla- Within, p met, densely appressed-pubescent ; Petals (immature) subobovate, glabrous hai “Ana Slightly sericeous in the lower ‘wo lateral flanges each bearing 4—6 glabrous STEYERMARK & LIESNER—STERIGMAPETALUM 189 fimbrillae, the apex of the middle portion be- tween the lateral flanges bearing a tuft of short hairs. Staminodia 10, scalelike and united at the base into an annulus (according to Monachino). Ovary spheroid, pubescent, 5-angulate. Geographical distribution: Forests of north- eastern Colombia, in Department of Magdalena, at 1,000-1,200 m (Fig. 3). Two collections, in addition to the type, from the Department of Magdalena, Colombia, both sterile, were identified by Standley as S. colom- bianum and cited by Monachino (1944). The data for these collections are: Cincinnati region, 1,300-1,500 m, Dec. 1931-Feb. 1932, Espina & Giacometto A142, and Río Toribio region, slope of San Lorenzo, 1,000-1,200 m, Espina & Gia- cometto A90. The description of this species is incomplete and unsatisfactory, due to the 1) immature de- long," but an examination of the F isotype shows peduncles 8-10 mm long. The peduncles and uppermost part of the stems in the NY holotype appear to be covered by a microscopically dense appressed indument, but the F isotype is covered by minute hyphal threads that complicate the interpretation of the presence or absence of in- dument. Similarly, the calyx tube on the F iso- type is covered by fungal hyphae to the extent that uncertainty remains as to whether the ex- face i letely glab bears some terior E y glabro minute appressed indument. Monachino described the NY holotype as hav- ing an inflorescence composed only of the bud of female flowers, “single on end of peduncle” (1944), but the F isotype shows three peduncles arising just below the apical portion of the twigs, each peduncle bearing 3 sessile flowers. In the publication of the Amazonian species of Rhi- zophoraceae (Albuquerque in Prance et al., 1975), the pedicels of S. colombianum are stated to be 3 mm long, whereas this length refers ac- tually to the peduncles of the NY holotype as described by Monachino, since the flower buds themselves are strictly sessile. The two specimens cited by Albuquerque un- der S. colombianum for Venezuela and Surinam respectively (Steyermark 90428 and Maguire 24780) are treated in the present paper as S. 190 guianense subsp. ichunense and S. guianense subsp. guianense. Additionally, a specimen col- lected by de la Cruz 2170 from Guyana (NY) and identified as S. colombianum, proves to be- long to the Malpighiaceae. 5. Sterigmapetalum tachirense Steyermark & Liesner, sp. nov. TYPE: Venezuela. Tachira: cloud forested ridge top, along steep slopes 72°13'W, 1,200-1,380 m, 11 Nov. 1979, tree 20 m tall, leaves coriaceous, spreading, rich green, Julian A. Steyermark, Ronald Liesner & Angel González 120032 (VEN, holotype; MO, isotype). — Fig. 4. Arbor 20-metralis, ramulis juvenilibus dense mi- nuteque papillato-pubescentibus indumento adpre munitis; stipulis non visis; foliis 3-verticillatis raro op- positis brevipetiolatis, petiolis 2-2.5 mm lon: nute papillato- puberulentibus vel glabrescentibus; laminis 6-8.5 cm longis 5.3-7 cm latis utrinque glabris inte- imis; nervis lateralibus utroque latere 12-13 di- varicate patentibus ante margin: em 2- T mm anasto- Tua supremorum cheers pedunculis 1.2- I paberalennbur pilis papillatis instructis, cymoso- ra- mosis » axibus secundarii a mm longis minute ad- masculinis breviter puma a pedicellis 0.5 mm lon- gis; db ies nde o 3.5-4 mm longo 2- ene s 5 on i extus sparsim ad- wake ito ii AAAS ti intus spar sim dese So- piloso; pera lobis 4 deltoideo-ovatis obtus mm longis 1.3 esso-pubescent talis 4 anguste ee 5 mm longis basi m is, supe mm latis basim versus un- guiculatis, ue | ato; petalis apice 3-lobatis, parte centrali cucullata se ntis aggregatis incurvatis munita, utroque latere extensionibus duobus laterali- bus in 15-18 bri as glabras fissis, infra staminibus 8, filamentis 3. 5 mm lo basi 0.5 mm | blongi m longis 0.4 mm latis; staminodiis subulatis apice ce akala angustatis | mm longis; ovarii rudimento conico | mm longo basi 0.7 mm lato adpresso-piloso. ongis 0.2 mm latis Trees 20 m tall; i Juvente bises covered with dense, minute, apr Stipules not seen. Leaves 3-verticillate, rarely op- posite, short- petioled, the petioies 2-2. 5mm long, minutely papillat blades suborbiculas d rounded at the apex; rounded at the base, 6-8.5 cm long, 5.3-7 cm wide, entire, both sides glabrous, the lateral nerves ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 12-13 each side, — spreading, anas | tomosing 2-7 m margin. Staminate in- florescence: maine in axils of the uppéme | leaves; cymosely branched, the secondary axes 4-5 mm long, minutely appressed-puberulent peduncles 1.2-1.7 cm long, sparsely or moder- ately appressed-puberulent, the hairs papillate flowers in close succession. Flowers with pedicels 0.5 mm long. Calyx deeply campanulate, 3.54 mm long, 2-2.8 mm wide, the tube 2.5 mm long, sparsely appressed-puberulent, glabrescent with- — out, sparsely appressed-pilose within. Calyx lobes 4, deltoid-ovate, obtuse, 1.2-1.5 mm long, 13! mm wide, glabrous without except for the ciliate apex, within densely appressed-pubescent. Petals | 4, narrowly oblanceolate-ligulate, 5 mm long, 0.5 | mm wide at the base, upwards 1.5 mm wide, unguiculate toward the base, the claw 1.5 mm ! wide; petal apices 3-lobed, the central part CU- | cullate, provided with incurved, clustered sêg- ments, on each side with two lateral flanges part- | ed into 15-18 elongate, glabrous imbit ] lateral segments below and the margins dense? hispid-pubescent. Stamens 8; filaments 3. 5 p | long, 0.2 mm wide at the base, connate in a | basal 0.5 mm; anthers narrowly oblong, L. A mm long, 0.4 mm wide; staminodes po , | Ovary rudiment conical, 1 mm long, wide at the base, appressed-pilose. e forests of Geographical distribution: Montan M at |, Estado Táchira, western Venezuela, 1,380 m elevation (Fig. 3). This species differs from the other d members of the genus in its tetramerous flo dd with 4 calyx lobes, 4 petals, and 8 stamens. il in the suborbicular-oval leaves rounded at | ends. | Ste. ! Sterigmapetalum subgenus Balsamocaulon ermark & Liesner, subgenus novum. a s | Sterigmapetalum resinosum Steyerm | Liesner. sita. Ext tum resinosum adest. Folia oppo obsoletàt Exsudat siones laterales petalorum deminutae vel o Venamo y Cuyuni), km 121. | Dorado, 1,050 m, 27-28 Dec. 1970 b ) m, leaves coriaceous, deep green abov® | 1983] STEYERMARK & LIESNER—STERIGMAPETALUM 191 C D tover FIGURE 4, Sterigmapetalum tachirense. A. Habit, flowering branch, X1. B. Ovary rudiment, staminate Mid Staminate flower, lateral view. D. Peta Androec oecium, staminate flower, showing one c ied attached a 1, ventral view, staminate flower. E. Peta l, dorsal view, staminate omplete stamen, the other stamens with euam of lyx lobe. to basal membrane. G. Inner upper surface of ca v below, Julian A. D ee G.C.K. E. Dunsterville 104569 (US, eic dn die No. 2581558-A; US, isotype, shee 581559-A). Fig 2: 3. ties Centibus, ioi = ramulis glabris vel glabres- dem S itis petiolatis, petiolis 10-15 mm longis glabris hl ellipticis ptici5 vel i oblanceoato-oblongis apice abrupt e angustatis use acutis basi acutis 5-10.5 €; m longis -5 cm latis arp glabris subtus obsolete minu- teque glanduloso-punctatis, marginibus | utroque latere en dem utroque latere 11-14 ad angulum 20-25* pa- tentibus supra impressis subtus paullo elevatis ante margine m 2-4 mm anastomosantibus, venulis tertia- pra vix manifestis; inflorescentiis masculinis (imma- NC Dre ETE secundorum usque ad sextos breviter pedunculatis in- sidentibus, pedunculis (immaturis) 1-1.7 cm longis mi- Is ca. 6-floris; floribus breviter gl longo subcoriaceo 5-angulato, lobis 5 deltoideo-ovatis subacutis marginibus elevatis apice pilosis extus aliter ala eio int d 24 1 1s 1 bris talis 5 (immaturis) panduriformibus apice rotundatis cucullatis, appendices tres subulatas i renti i is 0 i utrinque gla dorsum obtuse bus lateralibus nullis; staminibus 10 (immaturis), fi- lamentis brevissimis basi dilatatis connatis; ovarii ru- dimento ovoideo-conico apice rotundato glabro i 5-angulato; stigmate sessili. acute or acuminate, glabrous, 4-5 mm long, ca- ducous. Leaves opposite, petiolate, the petioles 10-15 mm long, glabrous or glabrescent; leaf blades coriaceous, oblong-elliptic or oblanceo- late-oblong, the apex abruptly narrowly acute or obtusely acute, the base acute, 5-10.5 cm 2-5 cm wide, both surfaces glabrous, abaxially , , margins elevated late, incurved, the base rounded, 0.8 mm long, . mm wide, both sides glabrous, dorsally ob- PARATYPE: VENEZUELA. BOLÍVAR: En el drenaje del río Cuyuni, a lo largo del río Anawaray-parü, vecin- ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo.. 70 dades del km 134 y campamento 134 al sur de El Dorado, 1,300-1,350 m, 25 Dec. 1970, J. Steyermark, G.C.K. & E. Dunsterville 104463 (US, VEN). Ge 2 E EDUR | + AA + f; n Roraima dst f ti f Estado Bolívar, southeastern Venezuela, at 1,050-1,300 m ele- vations. This species is distinguished by the resinifer- ous exudate on the apical portion of the stem, buds, and peduncles, the opposite leaves with abruptly acute apex, and glabrous petals which lack lateral flanges. 7. Steri TET - 1 Q4avormart B. i Liesner, sp. nov. TYPE: Venezuela. Falcón: Sierra de San Luis, Montaña de Paraguariba, forested area bordering savanna, entre € Hotel Parador y Curimagua, 1,300 m, 2! July 1967, tree 3 m, leaves coriaceous, deep green above, paler green below, flower buds green with 5 valvate sepals, Julian A. Siéy- ermark 99409 (VEN, holotype; US, 18° type). — Fig. 2: 4 rioribus resino lateralibus utroque latere 10-11 a ange divaricate patentibus ante marginem 2-5 si tomosantibus, supra subimpressis inconspic su brevis nibus 10 inaequalibus in alabastro, flame ae simis 0.6 mm longis basi connatis; 0 columnari-subconico 0.6 mm longo. : upper 1? Trees 3 m tall; branches glabrous: ‘ ternodes resinous. Stipules suborbicular-O obtuse, glabrous, resinous. Leaves oppo pé petioles 8-11 mm long, glabrous; leaf pe riaceous, obovate, the apex truncate-round base cuneate, 5.5-9 cm long, 3.5-5 cm W! 3 l- aad - yously ! sides glabrous, the margins inconspicuous? BR a i ainena leah f icuis subtus 1 o e o gy 1983] 12 crenulate-toothed, the sinuses minutely glan- dular, the lateral nerves on both sides 10-11, divaricately spreading at a 15—20° angle, anas- tomosing 2-5 mm from the margin, above in- conspicuously subimpressed, below subelevated, the tertiary veins below minutely reticulate, above obsolete. Male inflorescence: (immature) resin- ous, solitary, in the axils of the second to sixth riceous within, 5-dentate, the teeth pilose at the apex, the margins elevated, indurated. Petals 5 (immature); dorsally pilose above, the margins ciliate, otherwise glabrous; central part of the apex cucullate, hispid, bearing 3 subulate, in- curved appendages, each side bearing 2 lateral, unequal, erect, subulate appendages. Stamens 10, unequal in bud; filaments short, 0.6 mm long, basally connate. bini rudiment conical-sub- conical, 0.6 mm This species, like the foregoing S. resinosum, is well-marked by the resiniferous exudate on the Ni ii leaves. It differs from S. resinosum in Bic a leaves truncate-rounded at the apex, ud aii ee and petals densely pilose at the ateral flanges bearing two unequal fimbrillate appendages LITERATURE CITED AR Arbo d L. 1973. Families y Géneros de los €s de Venezuela. Edición Especial del Insti- Deli ee: Caracas. 605 pp. Pasi 1974. und S-Typ Siebelement- 1 "rs bei Rhan Beitr. Biol. Pflanzen 50: Evoluti . Behnke ini PNE 126 if Centrospermous Families, Pl. Sys E eve-element characters. Jn H. D e ‘ior, Ultrastructure and Systematics E. 981b. a "s pee tein w b. Siebel a MEN Phloem-Pro- koryiedone Evolution der Blütenpflanzen: II. Mon- oe r. Deutsch. Bot. Ges. 94 (in press). t eve-element plastids, exine sculp- wring and the ME afinities of the Buxa- L Syst. Evol. (in press). STEYERMARK & LIESNER—STERIGMAPETALUM 193 1982b. onl eens pe: plastids of T S acea on and Rhizophoraceae. Presen tion and ig iiis ance of Subtype-PV Plastids. PI. Syst. Evol. (in p 1983. Ultrastructure of sieve-element plas- tids of uo hase allied groups. Ann. Missouri Bot. Gard. im 1981 Sieve-element dope and crystalline Sabine Ga Legum e: Mi- ical characters as an sr 5 e cir- tors), Advance in Legume Systematics. Academic Press, London. BENTHAM, G. & W. D. HOOKER. tarum 1: 677-678. CROAT. T. 1978. Flora of Barro yid Island. Stanford Univ. Press, Stanford. 648 p CRONQUIST, A. 1968. Evolution and CUM EG i owering Plants. Houghton Mifflin Co., Bosto 1862. Genera Plan- pp ——. 1981. An Integrated System of Classification owering Plants. pp. 655-659. Columbia Uni- versity Press, ^ REN, R. M. T. 1980. A revised system of clas- sification of the angiosperms. J. Linn. Soc. Bot. 24. & R. THoRNE. 1983. Circumscription and the variation of the order Myrtales, Ann. Missouri Bot. Gard. (in press). ENGLER, A. 1876. In Martius, Flora Brasiliensis 12(2): Hrvwoop, V 1978. Flowering Plants of pa. World. ord University zo Oxford. 158 Horciraow, T 1959. milies of Flowering Plants, : 304. QUUM Cease? Jonxes, $ s 1942. In A. Pulle, Flora Suns 3(2): Ee J. G. 1925. Plantas novas. Arch. Jard. Bot. Rio de Janeiro 4: 359—361. t. LAWRENCE, G. H. M. 19 axonomy of Vascular Plants, 8th printing. The Macmillan Co., New York. 629 ae . The order Centrospermae. Ann. 220. In A. Engler’s Syllabus der Pflanzenfamilien, "12th ed. 358-3 A tei species of Ste- 0-12. PRANCE, G. T. ETAL. 1975. Ad taxonomica das espéci _amazonicas de Meo Acta Amaz SCHIMPER, Pi F W. A. Engler & K. Prantl, Die Natürlich. Pflanzenfam. III Teil. 7 Abteil. 43-56. STANDLEY, P. C. & L. O. WiLLiAMs. 1962. Rhizopho- raceae, in Fl f Guatemala. Fieldiana, Bot. 24(7): 1892-1893. Rhizophoraceae, in 263-264. STEYERMARK, J. A. 1952. Contributions to the flora of Venezuela. Fieldiana, Bot. 28(2): 422-423. NEW SPECIES AND COMBINATIONS IN THE GENUS OENOTHERA (ONAGRACEAE)! WARREN L. WAGNER? ABSTRACT r new Ripe in Oenothera sect. Megapterium and one each in ns Pachylophus and Fou othera are made her the Gre Oen to make them available for use in regional flo en Pachylophus) from hac us eater un. 0. Reseed eile a Hebi era harringtonii donee subi t. Raimannia) Illinois, and O. het erophylla ert: orientalis ipsius from fam i them availat A series of revisionary papers on Oenothera sects. Megapterium (Wagner, in prep.), Pachy- lophus (Wagner et al., in press), and Oenothera subsects. Oenothera (Dietrich & Wagner, in prep.) and Raimannia (Dietrich et al., in prep.) that contain new taxa and combinations are cur- rently in press or in the late stages of preparation. The publication of one or more regional floras that contain information resulting from these de- tailed studies on Oenothera, however, will likely precede that of the revisions. Therefore, in order to make these new names and combinations available for the Great Plains Flora, the Vascular Flora of the Southeastern United States, vol. 3 and the Michigan Flora, vol. 2, they are pub- lished here in advance of the viene Detailed discussion of these changes and new taxa will be included in the revisions. Oenothera sect. Megapterium (Spach) Endl. Oenothera macrocarpa Nutt. subsp. incana (A. ray) W. L. Wagner, comb. et stat. nov. Oenothera missouriensis Sims var. incana A. Gra Boston J. Nat. Hist. 6: 189. 1850. T TYPE: "Cana- y in the Texas Panhandle in one of the following counties— Oldham Potter, Hutchinson, Roberts or Hempshill — Texas. April photo MO; BOR etl Munz, Amer. J. Bot. 17: 366. E N ` Q S $2 = w k Q X o e > o > = Oenothera macrocarpa Nutt. subsp. oklahomen- sis (Nort.) W. L. Wagner, comb. et stat. nov. jin oklahomense Nort., Annu Rep. Mis- ri Bot. Gard. 9: 153. 1898. ry pr: Moe Payne ! This guine is based u Foundatio Depattison of Botany, Bernice P. Bishop Museum, P.O. Box ANN. Missouri Bor. Garp. 70: 194—196. 1983. o disjunct areas in Alabama and us are Press described bett in order to also make le. County, Oklahoma, 1893, F. A. thos 183 (MO- 5 ted; KSC, US, nn MO shee Oenothera macrocarpa Nutt. subsp. fremontii (S Wats.) W. L. Wagner, comb. et stat. nov. Oenothera fremontii S. Wats., Ped Am Acad. Hi 8: 587. 1873. Type: “White rock on Smoky Hif River,” Piwan: Russell, Ellis, E. Cove, Logan county, Kansas, 7-23 Jun e 1867, "n 1 Parry 79 (GH. lectotype here designated sheets, US, isolectotypes). z (Ame nt 17: 367. 1930) indirectly ni. the Fremo tson saw : : at- I have selected the other collection cited by Wi son, Parry 79, as the lectotype. that Oenothera macrocarpa is a variable mrs has differentiated extensively in the Great à region. Fach of the four subspecies oA | in general they are sharply distinct. e treated as subspecies primarily becau o complete interfertility and extensive o dation in any area of marginal contact. i mediates are known between O. mac'r 00D subsp. macrocarpa and subspp. fre montii men oklahomensis and between subspp. ok klahom sis and incana. There is also evidence that xd gests past hybridization between subspP- Ber and fremontii, although there is at prese? contact between them. pon research supported by a grant to Peter H. Raven from the National 19000-A, Honolulu, Hawai'i 96819. į | | | | | —— DRE: ctm qNNNNEPG | 1983] Oenothera howardii (A. Nels.) W. L. Wagner, comb. nov — howardii A. Nels., Bot. Gaz. cl 8. 1902, non M. E. Jones, Zoe 3: 301. 1893. TYPE: "In clay, Vermilion," Sevier Co., Utah, 1,60 m, 16 July 1894, M. E. Jones 5631c (RM- 13996, lectotype here designated; BM, G, MO 2 sheets, MSC, NY, POM, UC, US, isoleciotypesi: Jones’s publication of the name Oenothera howardii is eo a anical Nomenclature, 1978). He placed his ce of *Oenothera howandit" qe to Dan pcan for his helpful comments on uati Oenothera trachycarpa ig Munz, Amer. J. Bot. 17: o part Plants from Hamilton Co., Kansas, eastern Colorado, Utah, and Nevada formerly included E Oenothera brachycarpa are segregated here as O. howardii. This species is quite distinct from i O. brachycarpa in its brilliant yellow petals that a reddish brown, floral tubes 4.3-1 1(-12.5) cm us dab In that it consists of entirely tetraploid, Es aes Sud octoploid plants whereas O. bra- ies ien rom west Texas to southeastern Ari- dift ea rn Mexico has pale yellow petals le Ty lavender, floral tubes (10.5—512-22 cm ng and is entirely diploid, n = 7. Oenothera L. sect. Pachylophus (Spach) Endl. Oe nothera caespitosa Nutt. subsp. macroglottis ydb.) Mes Stockhouse & Klein, comb. et stat. no P, es 2591 Es Rydb., Bull. Torrey Bot. Club tarieg 1503. TYPE: "Turkey Creek and tribu- "Ws e Enc vial of La Veta Pass, Heur- 1900, P «Color ado, 2,800-2,900 m, 8 & 9 June (NY, ^ ydberg and F. K. Vreeland 5857 reco Tel photo MO; RM, isotype). Locality u-— n with the aid of R. Barneby (pers. £z E. ead treated here as Oenothera caespi- (1931. Bus acroglottis were included by Munz tificial O. ) within his broadly defined and ar- caespitosa var. montana (Nutt.) Du- Eee Grove. OR 9 fe Klein, Morris es W. Dietrich, WAGNER—NEW OENOTHERA "n his eae and O. harringtonii are contributed € wi d W. M Arboretum, 9414 Me 195 rand. Slightly curved capsules (1.7-)2.5-5.6 cm long with an undulate ridge along the valves, coarsely dentate spatulate to broadly oblanceo- late leaves and cool montane habitat clearly dis- tinguish this entity from the other subspecies of O. caespitosa. It ranges from southern Wyoming, Colorado to southeastern Utah and northern New Mexico YPE: United States: o he Colorado Springs, 2,000 m, 29 May 1939, J. H. Ehlers 7461 (MICH, holotype, photos MO, COLO; MICH 2 sheets, NA, isotypes). Oenothera harringtonii W. ie Wagner, Stock- S Oenothera caespitosa var. eximia sensu Munz, Am J. Bot. 731. 1931, pro parte (as to plants “he Colorado). Herba annua caulescens. Folia ane coe Yum oralis 3.1-6 c s. mum pa Capsula pose recta, (2.1-)2.5-3(- ga, valvularum nibus 6-8 tuberculis prominen- kis 2-3 mm Pm rote coalescentibus porcam si- formantibusque, per dimidium velduos lon- atoiak trientes deshiscens. Numerus chromosoma- ticus gameticus, This attractive species named in honor of the late Harold D. Harrington (1903-1981) is known only from grasslands in the southeastern Colo- rado counties of El Paso, Fremont, Otero, and Las Animas. Oenothera harringtonii is distinc- tive in the O. caespitosa complex in being a coarse annual with densely leafy, erect stems; capsules ovoid with 5-8 conspicuous tubercles 2- m high and five or more flowers opening at one time. Oenothera L. sect. Oenothera subsect. Oenothera Oenothera elata H.B.K. subsp. hirsutissima (A. Gray ex S. Wats.) Dietrich, comb. et stat. nov.* deed biennis L. var. hirsutissima A. hae m mer. Ac 43. 1849; ex S. oc. Amer. Acad. Arts 8: 579. ^de TYPE: : Valley of posi Fe Creek, Santa Fe , New Mexico, June 1847, A. Fendler 218 (GH, Siototype: MIN, MO, isolectotypes; Munz, Aliso 2: 18. 1949 Recent careful study of a large number of spec- ith R. Stockhouse, Pacific University, Forest adowbrook, Philadelphia, PA 1 19188. Botanisches Institut of Ae University of Düsseldorf, Germany. 196 imens of plants assigned to Oenothera hookeri Torr. & A. Gray and O. elata suggested that the two entities are extremely similar and should be recognized as one species (Raven et al., 1979; Dietrich & Wagner, unpubl.) Populations in the are here recognized as O. elata subsp. hirsutis- sima. Oenothera L. sect. Oenothera subsect. Raimannia (Rose) Dietrich Oenothera clelandii Dietrich, Raven & W. L. agner, sp. nov. Type: United States: Michigan. Mason Co., Amber Township, E half of sec. 23 NE '4, 30 September 1974, C. J. Barkley s.n. (MO-2383779, holotype; MO, isotype). Mer Hs rhombipetala sensu Munz, N. Amer. FI. II. 108. 1965, pro pa Herba biennis. Inflorescentia compacta, non inter- rupta ramis lateralibus; efflorescent flores complures per diem. Petala late elliptica vel ovata, acuta (raro rotundata), 0.5-1.6 em longa, flava vel pallide flava. Semina ambito elliptica, brunnea, 1-1.9 m Numerus chromosomaticus gameticus, n = 7; ae chromosomatice stein Mer ica This dict; tural heterozygote. A ring of 14 chromosomes is thus formed at meiotic metaphase I and the pol- len is only about 50% fertile. Populations for- merly included in Oenothera rhombipetala from southeastern Minnesota and Michigan south to Iowa, Illinois and scattered localities in Arkan- sas, Missouri, Kentucky, New Jersey, and New York are all autogamous complex heterozygotes and are here separated from the larger-flowered, outcrossing, fully fertile, pair-forming popula- struc- r ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo.. 70 tions from South Dakota south to Texas and central Arkansas. Oenothera clelandii is named in honor of Ralph E. Cleland (1892-1970), life- long student of Oenothera cytogenetics. Oenothera heterophylla Spach subsp. orientalis Dietrich, Raven & W. Wagner, subsp. nov. TYPE: United States: Alabama, Sumter Co., sandy clay field, 5.2 mi S of Dancy, 3 May 1972, R. Kral 46366 (MO- -2379457, nee DUKE, GA, GH, NCU, NY, USF, VDB, isotypes). Tubus floralis 3-4.7 cm longus, immaculatus, sint pilis pustulatis. Sepala eese sine pilis pustu- latis, apices sepalorum 1-3 m i, erecti. Ovariu et capsula saepe omninoglabrum. Mendi chromo- somaticus gameticus, n = Oenothera heterophylla subsp. orientalis is clearly distinguished from subsp. het erophylla, which occurs from central and eastern Texas t0 eastern Louisiana, in its lack of red pustulate hairs on any flower parts, sepal-tips erect in bud, 1-3 mm long and the ovary glabrous. It is known from only two disjunct areas: Greene, Pickens and Sumter counties in Alabama and Calhoun, Nevada, and Ouachita counties in Arkansas. LITERATURE CITED Munz, P. 1931. Studies in Onagrace genus Pachylophus of the genus Ochi J. Bot. 18: 72 ae VII. The sub- hera. Amer. RAVEN, P. H., W. DIETRICH An outline of the systematics of Qenot sect. Euoenothera (Onagraceae). Syst. Bot. 232. WAGNER, W. L., R. E. STOCKHOUSE (In press). P systematics an Oenothera caespitosa species C complex ( ceae). Ann. Missouri Bot. Gard. This species and the next subspecies contributed in collaboration with W. Dietrich and Peter H: Rave Mig a Botanical Garden, P.O. Box 2 99, St. Louis, MO 63166. g & W. M. KEN volution of the de E NOTES CONFIRMATION OF THE CHROMOSOME NUMBER IN CEPHALOTACEAE AND RORIDULACEAE Cephalotaceae (monotypic) and Roridulaceae (ditypic) are two Southern Hemisphere families of restricted distribution. Cephalotus is endemic to southwestern Western Australia and Roridula tothe Cape region of South Africa. The two fam- ilies share little in common apart from their being insectivorous in that they derive supplementary nitrogen from insects and other small organisms trapped by their leaves. Cephalotus is herbaceous and has relatively large leaves that produce pitch- ers, while the two species of Roridula are shrubby and have small leaves covered with sticky but evidently not glandular hairs. Because of the intrinsic interest of these fam- ilies as well as the taxonomic difficulties sur- rounding them, and because living material was available, we embarked on a cytological study and have determined the chromosome number m Sfhalotus and one species of Roridula. Ini- lally we believed that both families were un- p as cytologically (Raven, 1975) but the work ondo (1969), Keighery (1979), and Johnson (1980), in which the base number of x = 10 was established for Cephalotus has now come to our attention. In addition, Kress (1970), published chromosome numbers for both Cephalotus and Roridula. Kress found 2n = 12 in Roridula gor- gonias and also recorded 2n = 20 in Cephalotus. These numbers are here confirmed, meiotic ma- terial having been studied in the case of Ceph- alotus. Since no illustration of the chromosomes of Roridula has previously been published, a photograph (Fig. 1) of a metaphase spread is in- cluded here. Chromosome number and voucher information are as follows: Cephalotus follicularis Labill. n = 10. Western Australia, Flinders Park, Albany, Ornduff 8823 C UC). Roridula gorgonias Planch. n = 6. South Af- rica, Cape, Vogelgat, Hermanus, Williams 2790 MO). We thank Ion Williams, Vogelklip, Hermanus, South Africa for the seed of Roridula; and R. FIGURE i AN N. MISSOURI Bor. GaRp. 70: 197-208. 1983. Mitotic metaphase in Roridula gorgonias. 198 Ornduff, University of California, ahd for providing the fixed buds of Cephalot LITERATURE CITED JOHNSON, M. A. T. 1980. Chromosome numbers in Akania and Cephalotus. vg Bull. 34: 37-38. KEIGHERY, G. J. ome counts in Ceph- alotus (Cephalotaceae). Plant Syst. Evol. 133: 103- KoNpo, K. 1969. Chromosome numbers of carniv- orous plants. Bull. Torrey Bot. Club 96: 322-328. Kress, A. 1970. Zytotaxonomische Untersuchungen ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 an einigen Insectenfangern (Droseraceae, Bybli- daceae, Cephalotaceae, Roridula is Sarraceni- aceae). Ber. Deutsch. Bot. Ges. 83: 55-62. RAVEN, P. H. 1975. The bases of angiosperm phy- logeny: cytology. Ann. Missouri Bot. Gard. 62: 724—164. — Ching-I Peng, Institute of Botany, Academia Sinica, Taipei, Taiwan, R.O.C.; and Peter Gold- blatt, B. A. Krukoff Curator of African Botany, P.O. Box 299, St. Louis, Missouri 63166. A NEW HESPEROMANNIA (COMPOSITAE) FROM AUI ISLAND: HAWAIIAN PLANT STUDIES 116 Hesperomannia (Compositae) consists of three species and four subspecific entities, mostly en- demic to the single islands, Kauai, Oahu, Mo- lokai, and Maui. They are attractive trees, with ower heads like 1 pink thistles. Now, oo species has been discovered | in west Mau Hesperomannia mauiensis Sp. nov. HOLOTYPUS: Hawaiian Islands, west Maui Island, Iao alley, Makalaloa Stream, steep forested slope, July oe 1980, Robert Hobdy 859 (BISH).— Fig. Diagnosis Holotypi: Arbor 2.3-3.3 m alta est, pe- tiolis 17-30 mm longis in basi tdg. oo 9.5- 16 cm longis AT, 5 cm latis chartace ae is va florescentia terminali ascendente puberula « capitibus, involucro 30-32 mm alto dense puberulo, phyllariis superis lineari- lanceolatis, roca 30-40 luteis, corollis cum tubo 15-17 mm ongo lobis 12 mm longis 0.3 mm latis extra qs re 7- 8 mm longis, aculeis pappi 27 mm um 4-5 Tree 2.3—3.3 m tall; leafy branchlets 2. 5-5 mm in diameter, terete, brown, densely pale ascend- ing puberulous; leaves in a plume at the branchlet tips; internodes 2-8 mm long; nodes scarcely en- dulate, above dark green, glabrous, below green. glabrous, secondary veins 7-9 in each half, as- cending, the lower ones straight, the upper ar cuate; inflorescence terminal, racemose, W1 4- 5 heads, densely ascending en peduncle 2-5 mm long, 2.5-3 mm in diameter; pedicels 7-12 mm long; involucre 30-32 mm high, nar- rowly campanulate, with numerous imbricated phyllaries, these pinkish, but densely pale as- cending puberulous, the lowest ones 2-3 mm long, ovate, acute, the median ones lanceolate, 3.5 mm wide, the upper ones linear lanceolate, florets 30-40, canary yellow; ovary 5.5 mm 10 prismatic, puberulous; corolla tube 15-17 mm long, glabrous, the 5 lobes 12 mm long, 0.3 m wide, almost linear, but tapering to an acute b. oig pilosulous without, with a midrib; di -8 mm long; anthers 7-8 mm long. ai dcn thee finally splitting apart; style exsert i dark; pappus bristles 38, and 27 mm long, go mineous, mostly upwardly barbellate (mat achenes not seen). The closest relative of this novelty is H.E borescens Gray ssp. Swezeyi (Deg.) Carla.. er with the blades oblanceolate (or narrowly ye obovate, obtuse or subobtuse; all or at least inner phyllaries glabrous; corolla tube 20 E long, the lobes 18 mm long, 1.5 mm wide; £ thers 9 mm long; and the pappus bri isties iip Hesperomannia mauiensis has the blades © ap tic, acute to subacuminate; phyllaries all | 1983] NOTES wao Smm Ficure 1, Hesperomannia mauiensis St. John, from holotype. a, habit, X/; b, floret, X2. 199 200 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo. 70 pressed puberulous; corolla tube 16-17 mmlong, the type locality, Maui, and -ensis, the Latin ad- the lobes 12 mm long, 0.3 mm wide; anthers 8 jectival place ending. mm long; and the pappus bristles 38. d St. John, Bishop Mu : ; — Harol The new epithet is formed from the name of 19000A, Honolulu, Hawaii 96819, USA. TORTULA CHISOSA SP. NOV., A BISTRATOSE-LEAVED SPECIES FROM THE UNITED STATES, MEXICO, AND SOUTHERN AFRICA Tortula chisosa Magill, Delgadillo et Stark, sp. nov. is described here from material collected in North America and southern Africa. It is closely related to 7. ammonsiana and T. bartramii. In 1976, one of us reported the occurrence of an unusual specimen of Tortula from the moun- tains of west Texas (Magill, 1976). The specimen (Magill 1341) was of interest because of a unique combination of characters that separated it from all North American species of Tortula. In con- nection with a study of mosses of Zacatecas, Mexico (Delgadillo & Cárdenas, 1979), four specimens that are nearly identical with the Tex- as plants were obtained; L. Stark and R. C. Cas- tetter collected five others in New Mexico, and R. Magill secured two additional collections in southern Africa. Comparison of these specimens against specimens and descriptions of other taxa suggests that the first represent an undescribed species of Tortula. Tortula chisosa Magill, Delgadillo et Stark, sp. Habitu et forma foliorum et propagulorum T. am- monsiana Crum et Anderson simile, sed foliis bistra- tosis et apicibus sine denticulatis et costa dorsali supra papillosa differt. Plants with the habit of T. ammonsiana. Stem loosely appressed when dry, erect spreading when moist, 0.5-0.7(-0.8) mm wide, 1.2-2.3 mm long, smaller toward the base of the stem, crowded near the tip, oblong to spatulate. Leaf blade bi- stratose, i ] gins plane; upper and median leaf cells quadrate or hexagonal, firm-walled, 8— 13 um in diameter, with several hollow c-shaped papillae per lumen; basal leaf cells hyaline, quad- rate to rectangular, smaller toward margin, form- ing distinct fenestrations on lower third or fourth of the leaf. Costa reddish, excurrent as a short mucro, dorsally papillose on the upper third to halfway down, 58-103(-113) um wide at mid- leaf, in section with a strong dorsal stereid band of 4-7 layers of cells, 2—4 guide cells and 24 adaxial papillose cells; the hydroid group 1s 1- conspicuous. Propagulae in groups on stalks axils of upper leaves, stalked, costate, leaflike, with upper cells papillose, ending in a single hya- line smooth-walled apical cell, that may be sub- tended by 2 or 3 smooth-walled cells. Perichae- tial leaves not differentiated. Male inflorescentt and sporophyte unknown. TYPE: U.S.A. Texas: Big Bend National Park, Chisos Mountains, on soil in small canyon below Lost Mine Trail, W side Lost Mine Peak, && 1,600 m, 19 June 1973, Magill 1341 (MO. he lotype). Additional specimens examined: U.S.A. NEW MEXICO" Doña Ana Co., ca. .25 mi E of Dripping Springs ro W side Organ Mountains, 32°19’N, 106734 W, 2 O: m, 10 July 1980, Stark & Castetter 1154 (MEXU, P i ountains, 106°35’W, 1,813 m, Moe Stark & Castet 31024, 3105A, 3106A, 3107A (PAC). MEXICO. A TECAS: Cerro de la Bufa, 22°47’N, 102°34'W, 2.7 Junio 1979, Cárdenas 332b (MEXU); 2 km E 50m. rio, Mpio. Chalchihuites, 23°35'N, 103°49'W, 229: Marzo 1981, Cárdenas 1077c (MEXU); zo 1981, Cárdenas 1113 (MEXU); 8 km 22°22'N, 101°35'W, 2 n along ge River, sd gill 4242 (MEXU, PRE). SourH WEST AFRICA: berg summit, S Namib, Sept. 1977, Williamso" (MEXU, PRE) 2654 32°21'N | Marzo 1981. &' B 49 ape no NN Rt PH NOTES 201 FIGURES |. i section of da i Tortula chisosa.—1. Leaves.—2. Leaf apex. —3. Leaves and propagulae.— 4. Median cross- : , showing dorsal papillae.— 5. Median cross-section of leaf margin.— 6. Propagulae and stalk.— Showing costal de S, costa at upper right corner. — 8. Basal leaf cells and right basal leaf margin. —9. Propagulae Figures 4. Ed velopment (All Magill 1341). Bars equal 0.1 mm in Figures 1, 2, 3, 6, 8, 9 and 20 um in tered, costate, leaf-like propagulae; and a terricolous or saxicolous habitat. This combi- nation of character states is remarkably constant eav : 7 €s with reddish, papillose costae, bi- but not exempt from variation in the collections aminae- , minae; the production of stalked, clus- 202 thus far analyzed. For example, the leaf laminae, while consistently bistratose in all specimens, in several cases show unistratose patches in mid- leaf cross sections. Stalked rer have not been observed in three specimens: Cárdenas 1077c, Magill 4242, and Williamson: 2675d, even though other diagnostic features are present. The last specimen has additional variations that are of interest in the evaluation of the species; its leaf cells are consistently smaller (7-10 um) than those of the American collections and the angle of leaf blade insertion is steeper. Also, the leaves are keeled, a fact that is not evident on other specimens. Tortula chisosa is similar in many respects to T. caninervis (Mitt.) Broth. (syn. T. bistratosa Flowers); however, in the latter species the plants are larger, the leaf margins are recurved, and the leaves awned. Through the production of leaf- like propagulae, 7. chisosa is also related to the T. pagorum-laevipila complex. The similarities in habit and habitat, leaf form and method of propagule production clearly indicate relation- ship to T. ammonsiana. This last species was described from the eastern United States (Crum & Anderson, 1979) and would appear to be one of the closest relatives of T. chisosa. However, the differences between them are clear; the leaves of T. chisosa are bistratose, they lack minute apical teeth (also absent in African specimens of T. ammonsiana), and the costa is papillose dor- sally toward the apex n New Mexico Tortula chisosa has been found intermixed with plants of T. bartramii Steere; study of both plants from the same sites has shown that they are similar in cell size, stem, and costal structure. Leaves of 7. bartramii had occasional bistratose patches and, in one instance, one plant showed a few small apical leaves that resembled the propagulae of 7. chisosa. Despite this, the ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 combination of character states by which we de- fine our new species is maintained and allows easy separation of mixed samples. The nature of factors promoting appearance of leaf-like pro- pagulae and bistratosity are beyond the scope of this paper, but, undoubtedly, they should be ex- plored before a final evaluation of the relation- ships of 7. chisosa and T. bartramii. The present known distribution of T. chisosa may at first appear unusual. However, it appar- chisosa in South America or Australia to com- plement its occurrence in Africa. We would like to acknowledge the valuable comments of Mr. Brent Mishler, and the field assistance of Mr. R. C. Castetter; Miss A. Cár denas kindly provided her collections for anal- ysis. Contribution No. 220 from the Department of Biology, The Pennsylvania State University. LITERATURE CITED wan ban A LE ADEM. 1979. Tortula am- na, a new species from West t Virginia. Maoa 82: 469-47 ^u DELGADILLO M., ,C. & M. A. CARDENASS. 1979. Wer México. I. Boletín de la Soci nica de México g Botán 38: 1—6. MAGILL, R. don of m Bend National Park, Texas. Bryologist 79: 269-2 —R. E. Magill, Missouri Botanical Garden, P. Box 299, St. Louis, Missouri 63166; C. be dillo, Instituto de Biología, U.N.A.M., Ap. Posta 70-233, 04510 México, D.F.; L. R. Stark, De partment of Biology, 202 Buckhout Laborator, The Pennsylvania State University, U niversit) Park, Pennsylvania 16802. | | 1983] NOTES POLYGALA DUKEI (POLYGALACEAE), A NEW SPECIES FROM PANAMA Polygala dukei K. Barringer, sp. nov. TYPE: Pan- ama. Darien: the Bolimina, small hill, Duke & Brista 385 (US!).—Fig. 1 Ab P. scleroxylon Ducke ramulis inermis foliis al- ternis bracteis glabris setas paucis subtentis differt. Small tree with hard wood, white bark. Stems glabrous, without spines, with persistent leaf bases at the nodes. Leaves alternate, glabrous; petioles to 6 mm long, rugulose, drying darker than the blade; blade ovate-lanceolate, entire, subcoriaceous, shiny, 9—13 cm long, cm wide, the base cuneate, the apex acuminate, the margin jt veins prominent on both iubens irregularly re- ticulate. Inflorescences axillary, racemose; axis 5 mm long; bracts glabrous, rigid, 1 mm long, sub- tended by a few stiff bristles. Flowers unknown. Fig fruit wiih oe 4 Uh olygala dukei K. Barringe indicates habit ^ ger. a. portion of branch € removed to M aril and seed. Upper portion of scale indicates fruit size, lower portion size, units equal 1 c Calyx deciduous. Capsule green, glabrous, loc- ulicidal, 2-loculate, 1.5 cm long, 2.2 cm wide, conspicuously 2-lobed, the lobes ovoid, not winged; locules 1-seeded; seed smooth, glabrous, globose, pendulous, with an umbraculiform aril. Additional Collection: PANAMA. DARIEN: near Estero Grande off Río Marea, 13 May 1967, Duke 10965 ) Duke reports that the fruits are tasteless and odorless and are eaten by currasow, parrots, and parakeets. The type is without flowers, but the fruits, hab- it, and leaf venation place it in Polygala section Acanthocladus. Chodat (1891, 1896) defined the section by its usually thorny stems, flowers in axillary fascicles, and stiff bracts. He keyed the section on the basis of its deciduous calyx, keeled petals, undivided stigma, and glabrous, un- Polygala dukei with remnants of axillary inflorescence. b. 204 winged fruit. He described two species, P. k/otz- schii Chod. and P. bennettii Chod. [= P. albicans (Bennett) Grondona], both from Brazil. Later, two more Brazilian species were described, P. pulcherrima Kuhlm. and P. scleroxylon Ducke. All four species are trees or shrubs with opposite leaves and persistent leaf bases. Short lateral branches often become spines after the leaves fall. The inflorescences are very short racemes, so that the flowers are fasciculate in leaf axils. The calyx is deciduous and the fruit is conspic- uously 2-lobed, each lobe being ovoid and con- taining a single, pendulous, arillate seed. The leaves are subcoriaceous and have prominent looped secondary veins and irregularly reticulate tertiary veins. Polygala dukei shares all these characters except presence of spines and opposite leaves. Polygala dukei can be easily distinguished from all other Panamanian species by its arborescent ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 habit, its fasciculate, axillary flowers, and its 2-lobed fruits (Lewis & Herrera-MacBryde 1969). I would like to thank Robin Foster for pointing out the problem. William Burger, Michael Nee, and Michael Huft helped by critiquing the manu- script. Research was supported in part by NSF Grant DEB 8103184 to the Field Museum. LITERATURE CITED CHopar, R. 1891. Monee Polygalaceae. Au- bert-Schuchardt, Geneve . 1896. Polygalaceae, in ne Prantl, Nat. Pflanzenfam. III. 4: 323—345 Lewis, W. H. & O. HERRERA-MACBRYDE. 1969. Po- lygalaceae, in Woodson et. al. m i Panama. Ann. Missouri Bot. Gard. 56(1): 9 — Kerry Barringer, Department of Botany, Field — Museum of Natural History, Chicago, Illinois 60605-2496 A NEW VARIETY OF DECLIEUXIA CACUMINIS (RUBIACEAE) FROM BAHIA Among Raymond M. Harley’s Rubiaceae from Bahia, Brazil, are numerous collections of De- clieuxia H.B.K., of which there are two gather- ings of D. cacuminis. This species has been pre- viously reported from Minas Gerais (Kirkbride, 1976 One of these collections, Harley et al. 21216 from the Serra Geral de Caitité, is referrable to paratype collections, all from the vicinity of Grao Mogol (Kirkbride, 1976), approximately 350 air- kilometers south-southwest of the new station. In my discussion of the pubescence of var. de- currens (Kirkbride, 1976), it was described as less densely puberulous and with longer hairs on the mericarps than var. cacuminis. The pubescence of Harley et al. 21216 is denser with shorter hairs on the mericarps than the type collections. So the variation of pubescence in var. decurrens en- compasses that found in var. cacuminis. The other collection représents a new variety, which is presented here Declieuxia cacuminis Müller Argoviensis glabra Kirkbride, var. nov var. Frutex omnino glaber, stipulis trilobis, decurrenti- bus, -a caeruleis, fructibus apicis pe 044 mm, mericarpiis 2.2-2.4 x 2-2.2 mm, ca. 0.6 m crassis. TYPUS: Brazil, Bahia, Serra das Almas, ue and upper N.E. slopes of Pico das Almas, Cê. km W.N.W. of the Vila do Rio de Contas, @ i 41°57'W, 13°33'S, alt. 1,600-1,850 m, Hoa al. 19691 (holotypus, UB; isotypus, CEPEC. eri- The overall appearance of this variety po coid, and therefore similar to that of var. x: minis. It is easily separated from the other varieties by its totally glabrous condition other details that are set forth in key to the varieties of D. cacuminis: la. Shrubs completely PN mericarps with the emargination 0-0.4 eep ——— Pius. scd " cacuminis var. r. glabra » Shrubs or subshrubs puberulous except - calyx and sometimes gs core EF , mericarps with the emarginati tion 0.5-1 my dee Ib. 2a. Stipules ee or tridentate, not d urrent; toa blue, vs kee H ae ae ericarps 2.2-2 1.6- mm, 0.7-0.9 mm ae nis . D. cacuminis vat. T Á Stipules unidentate, decurrent, L5- white, externally glabrous; mericarps ! N c | the following i a PR 1983] 2.1 X 0.8-1.1 mm, 0.2-0.4 mm thick .. cc DÀ. Cacuminis var. decurrens The varieties of D. cacuminis are an example of taxa whose distribution at the highest eleva- tions in the Serra do Espinhaco and further north in Bahia is correlated with their differentiation. These may have resulted from either long dis- lance dispersal and random selection of geno- types or Pleistocene climatic change alterations in the distribution of the vegetation and resulting differentiation (Kirkbride, 1976). ith dractic til Ui ASUL NOTES 205 In my opinion, the latter is more important in this case and in the genus Declieuxia. LITERATURE CITED A revision of the genus Mem. New York Bot. KIRKBRIDE, J. H., JR. 1976. Declieuxia (Rubiaceae). Gard. 28(4): 1-87. —Joseph H. Kirkbride, Jr., Departamento de Biologia Vegetal, Caixa Postal 153081, Funda- ção Universidade de Brasilia, 70.910 Brasilia, DF, Brasil A NEW COMBINATION FOR A PROBLEMATIC CENTRAL AMERICAN APOCYNACEAE Prestonia woodsoniana (Monachino) A. Gentry, comb. nov. Echites woodsoniana Monachi- no, Bull. Torrey Bot. Club 86: 245 TYPE: Mexico: Michoacán: Hinton 15325 (holotype, NY, isotype, MO). Echites parviflora Sesse and Mog., Fl. Mex. 44. 1893. ie (1832) nec Afz. ex Ettingshausen (1861). ata Woods. Ann. Missouri Bot. Gard. 47:79. 1960. TYPE: Costa Rica: Puntarenas: Holm and Iltis 243 (MO). Bisi aceae, it may well be a survivor of the wu plexus from which Asclepiadaceae p Mably Woodson never saw material of ma ala Sesse and Mogifio. In his MR Echites and related genera (Ann. peque t. Gard. 23: 169-438. 1936) he listed oie hassigned rejected species “impossible Amen However, two years later (North nad ws Flora 29: 103-192. 1938) he resur- scription reproducing the Sesse and Moçiño de- (Bull, T and accepting it in Echites. Monachino ' *Orrey Bot. Club 86: 245-247. 1959) re- alized that E. parviflora Sesse and Mogifio is a later homonym of E. parviflora Roxb. as well as of E. parviflora Afz. ex Ettingshausen. From the description Monachino recognized the Sesse and Mocifio plant as being conspecific with a plant collected by Hinton in Michoacán and Guerrero, Mexico, and proposed the nomen novum E. woodsoniana for it, substituting one of the Hin- ton collections as type. Meanwhile, Woodson re- ceived a Costa Rican collection of the same species and described it as an unusual species of Prestonia, noting that its only close relative in Prestonia is West Indian P. agglutinata (Jacq.) n tube. Unfortunately, Woodson's description of the Costa Rican plant appeared one year later than Monachino's of the Mexican one, necessi- tating the new combination Prestonia woodson- iana (Monachino) A. Gentry, if Woodson's de- cision to include the species in Prestonia is accepted. This species turns out to be widespread in the dry forest area along the Central American Pa- cific coast, ranging from Guanacaste and adja- cent Puntarenas Province in Costa Rica to Guer- rero and Michoacán in Mexico. There are recent collections from both Honduras and Nicaragua, as well as Costa Rica at MO. Unfortunately, the recent collections of P. woodsoniana have all been misidentified as Echites tuxtlensis Standl., a somewhat similar-looking (except for the much longer corolla tube and non-caudate corolla lobes) species from the Caribbean side of northern Cen- tral America and southern Mexico. Echites 206 woodsoniana was apparently included in E. tux- tlensis in the Flora of Guatemala (Fieldiana, Bot. 24(8): 334-407. 1966) as well. Presumably all Pacific coast collections of E. “‘tuxtlensis’ are actually Prestonia woodsoniana. It should be noted that my transfer of Mon- achino's epithet to Prestonia reflects no firm opinion that this species really belongs in that genus. Rather, I am merely accepting Woodson's judgement to that effect. The definitive character ANNALS OF THE MISSOURI BOTANICAL GARDEN of Prestonia, the faucal annulus around the mouth of the corolla tube, is so weakly defined in 7. woodsoniana as to be virtually non-existent. Nevertheless, the specialist in Echitoideae con- sidered this plant best referred to Prestonia and I assume that that remains the best placement pending another revision of the genus. —Alwyn H. Gentry, Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166. ALSTONIA (APOCYNACEAE): ANOTHER PALAEOTROPICAL GENUS IN CENTRAL AMERICA Representatives of several palaeotropical gen- era have been discovered in Central America in recent years. So many examples of this pattern have now turned up (see Gentry, 1982: 124-125) that this disjunction is no longer surprising, if no less interesting. Presumably many of these gen- era with Central American range disjunctions, usually with relationships to Asian rather than to African taxa, reflect remnants ofa widespread tropical Laurasian Tertiary flora (cf. Raven & Axelrod, 1974; Gentry, 1982). While preparing a summary of Apocynaceae taxonomy for the Missouri Botanical Garden's weekly Floristic Taxonomy Seminar, I was star- tled to realize that the endemic Central American genus Tonduzia looks suspiciously similar to many species of the widespread palaeotropical genus Alstonia. Further study revealed that the striking resemblance is a real one and Tonduzia should be reduced to the synonymy of Alstonia. Indeed such a reduction was proposed long ago by Pichon (1947), who noted that there is nothing to distinguish Tonduzia from Alstonia and re- duced it to a section of that genus. Although Pichon treated Tonduzia as a distinct section of Alstonia, it constitutes a poorly demarcated one: every one of the distinguishing fi fsection Tonduzia are shared with one of the other two sections of A/stonia that he recognized. That even such an inveterate generic splitter as Pichon (cf. comments in Gentry & Tomb, 1979: 756-757; Gentry, 1980: 8), who proposed two new seg- regate genera from Alstonia and resurrected two others in the same paper in which he sank Ton- duzia, considered Tonduzia congeneric with A/- stonia should have rendered the merger defini- tive. However, Pichon's proposal appears to have been overlooked by all subsequent workers (6.£ Standley & Williams, 1969; Nowicke, 1970; Woodson, in herb.) and Tonduzia is still recog- nized in Willis’ Dictionary (Airy Shaw, 1973) and in all North American herbaria. Pichon (1947) was aware that several specie of Tonduzia had been described but had seen material of only the type species, Tonduzia lon- gifolia (DC.) Mgf. (T. parvifolia Donn. Sm. 5 sedis :ned from a synonym of T. longifolia) and refrained frc . ace . at ther species. pri tne 2 fara Neo- Thus the only ifi itl ) tropical species of A/stonia is A. longifolia ue Pichon. Three species of Tonduzia are pec cepted (Standley & Williams, 1969) necessita the following new combinations in Alstonia: Alstonia macrantha (Woods.) A. Gentry. comb. nov. Tonduzia macrantha Woods., Missouri Bot. Gard. 24: 12. 1937. Guatemala. Quezaltenango: Skutch (MO). T. longipedunculata Woods., Publ. Field Mus P Hist., Bot. Ser. 23: 78. 1944. TPE: Guate Sololá: Steyermark 47313 (MO). Alstonia pittieri (Donn. Sm.) A. Gentry, com nov. Tonduzia pittieri Donn. Sm., Bot. 49: 456. 1910. type: Guatemala. mala: Deam 6098 (MO). arat Recognition of A. pittieri as specifically tinct is on rather ounds as the D^, | er tenuous gr on cited leaves and relatively high stamen inserti [Vot. 70 | | ‘ j A Guate | ol d w a 1983] in the Flora of Guatemala as distinguishing it from A. longifolia are not always associated. While additional collections are needed to re- solve the status of A. pittieri, A. macrantha is easily recognized by its larger flowers. LITERATURE CITED AIRY SHAW, H. K. 1973. Willis’ A Dictionary of the Flowering Plants and Ferns, ed. 8, Cambridge University Press. Gentry, A. H. 1980. Bignoniaceae, Part I: Crescen- eae and Tourrettieae. Flora Neotropica Mono- graph 25: 1-130. 2. Phytogeographic patterns as evidence for a Choco refuge. Jn G. T. Prance (Editor), Bi- ological Diversification in the Tropics. Columbia University Press. NOTES 207 A. S. Toms. 1979. Taxonomic implica- tions of Bignoniaceae palynology. Ann. Missouri Bot. Gard. 66: 756-777. NowiCKE, J. W. 1970. Apocynaceae. In Flora of Pan- ama. Ann. Missouri Bot. Gard. 57: 59-130 PicHon, M. 1947. Classification des Apocynacées: IV. Genre “Alstonia” et genres voisins. Bull. Mus. Hist. Nat. (Paris), sér. 2, 19: 294-301. Raven, P. H. & D. I. AXELROD. 1974. Angiosperm biogeography and past continental movements. Ann. Missouri Bot. Gard. 61: 539-673. STANDLEY, P. & L. O. WiLLIAMs. 1969. Apocynaceae. In Flora of Guatemala. Fieldiana, Bot. 24(8): 334- —Alwyn H. Gentry, Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166. THE GENUS BOTRYARRHENA IN VENEZUELA Since the publication of the author's Rubi- aceae of Venezuela (Steyermark, 1974), a num- a new taxa of the family have been added k € Hora of Venezuela. In addition to various Mee new to the republic, two genera of Ru- reed ae and Botryarrhena, previously un- je Or Venezuela, are newly recorded. The thas a genus, Joosia, was collected by the au- Gaa, mied by Ronald Liesner and Angel arr Z, in the state of Táchira, while Botry- only E mi genus, previously known G mazonian Brazil, was discovered by Send vidse, accompanied by Otto Huber and Phen S, Tillett. A second, Venezuelan species tryarrhena is described here. lene *'egated to a separate tribe, Retiniphyl- but otherwi ng simply racemose inflorescences, se differing from the genus Stach- stead of qi the possession of hermaphroditic in- two ovule aous flowers, bilocular ovary with Ovary wi in each cell instead of a 4—5-celled elon "umcrods ovules in each cell, and an Pues bilobed stigma. biis Eun of the Venezuelan collection, Zilian ead "om fruiting material, with the Bra- es, indicates that the two represent different taxa, the Venezuelan plant having short, erect infructescences with shorter, more crowded pedicels, larger leaves, and more prominently re- ticulate tertiary venation. The Venezuelan collection, the second species known for the genus, may be described as fol- lows: "ib 4 "M 1 2 Ce J k, sp. nov. J bol 10-metralis; stipulis late deltoideis obtusis 4 is 6 mm latis; foliis oppositis petiolatis, lami- vatis, venulis tertiariis utrinque prominulis valde re- ticulatis; petiolis 2-3 cm longis; calyce hypanthioque m longo (hypanthio 2 mm longo 2.5 mm lato bro; calyce 1.5 mm longo 3 mm lato apice truncato paullo repando glabro); bracteola sub flore lanceolato acuto 2 mm; pedicellis fructiferis 1.5— ongis; infructescentia 3-5 cm longa si Ww mm simplice vel prope basin | ramosa; ovario 2-loculare, orulis in o culo 2; fructibus congestis subgl i n angustatis, apice paullo elevato truncato munitis, exo- io 0.5 mm crasso, endocarpio glabro; seminibus subgloboso lit itatil tundatis 10 mm DYL 5 s longis 6-8 mm latis glabris. Tree 10 meters tall with all parts glabrous. Stipules broadly deltoid, obtuse, 4 cm long, 6 mm wide. Leaves opposite, petiolate, broadly oblong-oblanceolate to oblong-elliptic, some- what abruptly narrowed to a broadly obtuse apex, cuneately narrowed at the base, 22-30 cm long, 9—13.5 cm wide; lateral nerves 12-13 on each 208 ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 side, strongly impressed above, elevated below; tertiary venation prominently strongly reticulate. Petioles 2-3 cm long. Infructescence simple or once branched near the base, 3-5 cm long; fruit- ing pedicels 1.5-2 mm long, crowded on the rachis. Fruiting calyx and hypanthium 3 mm long; hypanthium 2 mm long, 2.5 mm wide; calyx truncate with a slightly repand border, 1.5 mm high, 3 mm wide. Bracteole lanceolate, acute, 2 mm long. Fruit depressed-subglobose, apically narrowed to a slightly truncate summit, 17-20 mm long, 15-17 mm wide, the base articulate with the pedicel; exocarp thin, 0.5 mm thick; endocarp thin, glabrous. Seed subglobose-oval, rounded at the ends, 10 mm long, 6-8 mm wide, glabrous. Ovary 2-celled; ovules 2 in each cell, axile. TYPE: VENEZUELA: Terr. Fed. Amazonas: Dep- to. Atabapo: southeastern bank of the middle part of Cafio Yagua at Cucurbital de Yagua, Lat. 3?36'N, Long. 66?34'W, alt. ca. 120 m, 8 May 1979, G. Davidse, O. Huber & S. Tillett 17427 (VEN, holotype; MO, isotype). LITERATURE CITED Duckr, A. 1933. Plantes nouvelles ou peu connus — de la région amazonienne. Arch. Jard. Bot. Rio de — 7; pper J: A. 1974. Rubiaceae, Flora de Ven- uela. Ed. Esp. Inst. Bot., Rec. Nat. Renov. brise ies y Cría. 9(1—3): 1-2070. —Julian A. Steyermark, Herbario Nacional, Ca- racas, Venezuela. X" ng CORRIGENDA The following corrections should be made in the paper titled **Vicariance biogeography in Mesoamerica," by C. J. Humphries (1982. 69: 444—463). The second sentence on page 456 should read: “Under assumption one (Fig. 13 iii-v), combining the two cladograms would give two components 0 and 3, and an informative cladogram." The last sentence in the same paragraph should read: *Combining the three cladograms together under assumption one is partially informative." Figures 13v and 14vii shown below should replace the incorrect figures on pages 457 and 458, respectively. 13v Mvii i The previous issue of the ANNALS OF THE MISSOURI BOTANICAL GARDEN, Vol. 69, No. ' PP. 735-998, was published on August 8, 1983. Systematics Symposia Published In 1953 the Missouri Botanical Garden began holding a series of annual systematics symposia each fall. Since 1969, with the sixteenth symposium, the Garden has pub- lished the proceedings of most in its quarterly series, ANNALS OF THE MISSOURI BOTANICAL GARDEN. Those published symposia which are available are listed below, and the price of all except the twenty-fourth is $10.00. The twenty-fourth, which was reprinted separately from the ANNALS, is available for $6.00. Many of these symposium issues of the ANNALS contain other papers, but each issue listed below contains the complete proceedings of the particular symposium. The proceedings of the 1981 symposium will be published during the second half of 1982, and if you wish to purchase a copy of that symposium, you may subscribe at this time. Upon publication, your copy will be sent. Sixteenth Annual Symposium, 1969—TROPICAL ISLAND BIOGEOGRAPHY. 56(3). 127 pp. $10.00. Seventeenth Annual Symposium, 1970—HYBRIDIZATION, EVOLUTION, AND SYS- TEMATICS. 59(3). 142 pp. $10.00 Twentieth Annual Symposium, 1973—PLANT-ANIMAL COEVOLUTION. 61(3). 132 p. $10.00. Twenty-first Annual Symposium, 1974 —8IOGEOGRAPHY. 62(2). 143 pp. $10.00. Twenty-second Annual Symposium, 1975—EVOLUTION AT THE POPULATION LEVEL. 63(2). 100 pp. $10.00. Twenty-third Annual Symposium, 1976—CHEMOSYSTEMATICS. 64(2). 98 pp. $10.00. Twenty-fourth Annual Symposium, 1977—SYSTEMATICS STUDIES IN AFRICA. $6.00. 223 pp. $6. Twenty-fifth Annual Symposium, 1978—PALYNOLOGY AND SYSTEMATICS. 66(4). 248 pp. $10.00. Twenty-sixth Annual Symposium, 1979—EVOLUTION AND SYSTEMATICS OF THE GRAMINEAE. 68(1). 104 pp. $10.00. Twenty-seventh Annual Symposium, 1980—REPRODUCTIVE STRATEGIES IN PLANTS AND ANIMALS. 68(2). 146 pp. $10.00. Twenty-eighth Annual Symposium, 1981—BIOLOGICAL STUDIES IN CENTRAL AMERICA. 69(3). 304 pp. $15.00. To place an order use this form or a photocopy of it. Orders should be prepa a $1.00 fee will be added to orders requiring invoices. No shipments are made be payment is received. Mail form with your check or money order, payable to Missou Botanical Garden, to: Department Eleven issouri Botanical Garden P.O. Box 2 St. Louis, MO 63166-0299 USA. Please send the SYMPOSIA checked above to: MMC Name Address ————————— — ——..— D Payment enclosed. W ress : i io t. L] Send invoice ($1.00 fee will be add to total) ———— M Postal Code Country m— pes cime E n —MÓÓÓÓ üt i, c IRE African Botany from the Missouri Botanical Garden In addition dd its own titles, the Garden stocks certain rks concerning the botany of Africa. For those titles Hid p as individual issues of serial pub- lications, the volume or number of the original is cited at the end of these entries so users will be aware of carefully. BRYOPHYTES OF SOUTHERN AFRICA. AN ANNOTATED Cueckuist, R. E. Magill & E. A. Schelpe. 39 pp. Illus- trated. 1979, $8.50. 316 species of hepatics and 591 species of mosses are recognized for the E of Southern Africa area. € d ot. Surv. S; Afric 43. AFRICA. J.H hos, 142 pp. sae, 1977. $20 reats (E ais and 90 species. gut of Southern HE AFRICAN Tur SPECIES. J. H. Ross. 0. vi hs 155 Pp. rond 1979. SE 30 oo for all of ESS Mem. Bot. Surv. ig OF SOUTHERN AFRICAN FLOWERING PLANTS. A. vicies ol. 1. DicorvLEDONS. 756 pp. 1975. cia Vol. 2. GYMNOSPERMS AND MONOCOTYLEDONS. 284 pp. 1976. $10.00. Vol. 1 treats 1574 genera, and Vol. 2 treats 356 genera. The total ed about 1596 ofthe world's genera of flowering pla: d nts, and about 550 of the gen- to Afri F SOUTHERN AFRICA. J. H. Ross. 159 pp. Mieres, 1975. $16.7 Treats 13 genera and 106 species. Flora of Southern Africa 16(1 Mosses (1)— FLORA OF SOUTHERN AFRICA. R. E. Ma- gill. xv + 291 pp. 1982. $30.00. The first of four projected fascicles. One hundred eighty species in the families Sphagnaceae through Grimmiaceae are treated. bac ES c pp OF AFRICA. 223 pp. Illustrated. 1978(1979). $6.0 Proceedings. vn the 24th Systematics (Odor s a: northern, tropical, southern Africa, and Mad- scar. Reprinted from the Ann. Missouri Bot. Gard. ). REVIEW OF PERMIAN PALYNOLOGY WITH PARTICULAR wii TO THE NORTHERN KAROO BASIN, SOUTH AFRICA. J. M. Anderson. 67. * 1334 13 + 188 pp. tosta 1971. $6.50. Mem. Bot. Surv. S. Africa. 41. ORDER FORM i Please type or pint peas address. No shipments u ent received: Bod € E Missouri E el pograd ical Garden St. Louis, MO 63166-0299 U.S.A. Title BELL ease. NM o Date: Ship to: Qty. Total "NU o oU Add $1.00 invoicing fee, if payment not enclosed Add 4% shipping and handling Total for books 9 € a v6 Total order (To place an order, use this form or a photocopy of it.) PERSPECTIVES IN TROPICAL BOTANY Contents: 1. Perspectives in Tropical Botany: Introduction by P. B. Tomlinson & Peter H. aven 2. Plant Inventory in the Tropics by Ghillean T. Prance . Plant Morphology and Anatomy in the Tropics—The Need for Integrated Approaches by P. B. Tomlinson 4. A Contribution of Rain Forest Research to Evolutionary Theory by P. S. i shton 3. Promising Directions of Study in Tropical Animal-Plant Interactions by Daniel H. Janzen 6 . Ecosystem Research in the Tropics by Carl F. Jordan & Ernesto Medina . Perspectives in Tropical Botany: Concluding Remarks by Peter H. Raven N “Intensified studies of plants, vegetation, and ecosystems of the tropics is not some esoteric or arcane aspect of pure science, but essentially an area of applied biology which is much neglected” (Tomlinson & Raven). “In the American tropics alone, the richest in terms of species of plants and animals of the world’s three major tropical areas, there are probably over a million unnamed and mostly uncollected species of organisms. A majority of these will become extinct during the next 25 years, and most will never have been collected even once by the time they disappear from the face of the Earth” (Raven). This Symposium gives an overview of many important aspects of tropical botany. The urgent need for basic biological information is stressed in each of the papers. Published as Volume 64, Number 4 of the ANNALS OF THE MISSOURI BOTANICAL GARDEN. 92 pages, soft cover. 1977(1978). Price: $15.00, post- paid. The simplest way to order is to use the order form below or a photocopy of it. Please include information requested below, if you order on separate paper. Orders should be prepaid: a $1.00 fee will be added to orders requiring invoicing, and no shipments are made until payment is received. Please send copy(ies) of PERSPECTIVES IN TROPICAL BOTANY @ $15.00 each to: ua pu Iu O Payment enclosed. Name TUUM U O Send i i 1.00 fee will be porcis nd invoice ($ added to total). —————À—À—À Postal Code Country INFORMATION FOR AUTHORS The Annars publishes original manuscripts in systematic botany and related fields. Authors are asked to follow the suggestions below in order to expedite editing and publication. If an author feels that his manuscript presents special problems, he should write the editor concerning the best way to handle these before submitting the manuscript. Page costs are $55.00 per page; charges may be reduced under special circumstances, and the editor will help authors to seek additional funding if necessary. Manuscripts must be typewritten on one side of good quality (not erasable) 81⁄2” X 11" paper. The manuscript should have at least one inch margin all around and be double-spaced throughout, including the abstract, notes, legends, tables, lists of specimens, and literature cited. Figure legends and tables should be typed separately and placed at the end of the text. Legends for each group of figures should be typed as one paragraph, using the style found in the current issue of the ANNALS. Authors should indicate in the margins the approximate places for illustrations and tables. Include author’s name and page number in upper right hand corner of each page. Sub- mission of the original and two carbon or xerographic copies of the manuscript is required, and the author should retain a copy of the final, typed draft. Acknowledgements to granting agencies, herbaria, illustrators, and technical assistants may be conveniently placed as a footnote on page 1. The author’s full mailing address should appear as the second footnote. An abstract must accompany each paper other than “Notes.” The abstract should succinctly oa the findings and conclusions of the paper and should be completely comprehensible Itself, A brief Latin diagnosis for each new taxon is preferred to a complete Latin description. A Complete description should be given in English. Use one Paragraph per basionym as follows: taxon, author, literature citation, type citation, €g., Bauhinia dipetala Hemsl., Diag. Pl. Nov. 1880. Casparia dipetala (Hemsl.) Britt. & Rose, N. Amer. Fl. 23: 216. 1930. Type: Mexico. Veracruz: Valle de Cordoba, Bourgeau 1713 (holotype, 5 Isotypes, F, GH, P, US). (See also Regnum Veg. 58: 39-40, 1968 for examples of synonym tations.) Descriptions should follow style found in other ANNALS issues, e.g., in Wunderlin, 1983 (70: 95-127), Keys should be dichotomous, indented, with numbered couplets. Cite rep- “sentative specimens as follows: Mexico. Oaxaca: Sierra San Pedro Nolesco, Talea, Jurgensen ion G, K, US). Use herbarium acronyms as listed in Index Herbariorum (Regnum Veg. E Abbreviations should be checked for consistency and to make sure they are unambiguous. “nods are used after all abbreviations except measures, compass directions, and herbarium Prints, Abut edges of photographs in composite plates. Do not mix line copy and continuous entry in the bibliography is referred to in the text, and each reference to a paper in the text " PNE in the bibliography. Citations of periodical literature should appear as follows: author s "st name, initial(s). Year. Full title of article. Name of Journal. Vol: inclusive pagination. For e. "e than one author, follow the style: author's last name, initial(s), second author's initial(s), ‘name & third author’s initial(s), last name. Names of journals are abbreviated tag ae nana €. : me, Mitial(s). Year. Full title. Edition, if ap 13th cher aspects of style, consult a recent issue of the Rouen Hie “dition. Universit i Chicago; or write to the editor. y of Chicago Press. Chicago; : : : s ; ssouri Botanical Garden, Po 4 manuscripts should be addressed to the Editor, Annals of the Mi "BOX200. St. Louis: Missouri 63166, A Taxonomic Revision of Condylocarpon (Apocynaceae) Mary E. d Fallen Supplement to the Contributions to the Synanthropic (Adventive) Flora of the Railroads in St. Louis, Missouri, U.S.A. Viktor Mühlenbach .... Revision of the Genus Sterigmapetalum (Rhizophoraceae) Julian A. Z Steyermark & Ronald Liesner New Species and Combinations in the Genus Oenothera Ee ceae) ' Warren L. Wagner NOTES Confirmation of the Chromosome Number in Cephalotaceae and Roridula- is ceae Ching-I Peng & Peter Goldblatt. ee A New Hesperomannia (Compositae) from Maui Island: Hawaiian Plant y Studies 116 Harold St. John E Tortula chisosa sp. nov., a Bistratose- Leaved Species from the United States, Mexico, and Southern Africa R. E. Magill, C. Delgadillo, & L. R. Stark Polygala dukei (Polygalaceae), a New Species from Panama 3 Kerry Barringer e A New Variety of Declieuxia cacuminis (Rubiaceae) from Bahia ose H. Kirkbride, Jr. A New Combination for a Problematic Central Ana Ap cynaceae Alwyn H. Gentry A . 205 Alstonia (Apocynaceae): Another Palaeotropical Genus in Gam oe merica Alwyn H. Gentry The Genus Botryarrhena in Venezuela Julian A. Steyermark z ^ 207 ANNALS d THE HISSOURI BOTANICAL GARDEN | A Revision of the Genus Anthurium (Araceae) of Mexico and Central America. Part I: Mexico and Middle America Thoma S B. r VOLUME 70 1983 NUMBER 2 ANNALS OF THE MISSOURI BOTANICAL CARDEN The ANNALS contains papers, primarily in systematic botany, contributed from the Missouri Botanical Garden. Papers originating outside the Garden will also be accepted. Authors should write the Editor for information concerning arrangements for publishing in the ANNALS. EDITORIAL COMMITTEE NANCY Morin, Editor Missouri Botanical Garden CHERYL R. BAUER, Editorial Assistant Missouri Botanical Garden MARSHALL R. Cro Missouri ATE VINE ERRIT DAVIDSE Missouri Botanical Garden JOHN D. Dwyer Missouri Botanical Garden & St. Louis University PETER GOLDBLATT Missouri Botanical Garden Published four timesa year by the Missouri Botanical Garden. . Louis, Missouri 63110 ISSN 0026-6493 Forsu hserinti £ Office ofthe Annals, PO Box 368, 1041 New Hampshire L Lawrence, Kansas 66044. Airmail delivery charge, $30 per volume. Four issues per volume. Second class postage paid at Lawrence, Kansas 66044 © Missouri Botanical Garden 1984 Aia ANNALS MISSOURI BOTANICAL GARDEN VOLUME 70 1983 NUMBER 2 A REVISION OF THE GENUS ANTHURIUM (ARACEAE) OF MEXICO AND CENTRAL AMERICA. PART I: MEXICO AND MIDDLE AMERICA! THOMAS B. CROAT? ABSTRACT á Anthurium (Araceae), a distinct neotropical genus with more than 700 species, is a member of ut subfamily Pothoideae. The genus ranges from Northern Mexico and the Greater Antilles to uthern Brazil and Northern Argenti d Paraguay. Species di ity is greatest at lower to middle elevations of northern South America, Panama, and Costa Rica, while the upper Amazonian forests h xico and ower Amazonian forests are relatively poor in species. In this paper, 122 species from Me : nown species, Guatemala hav as 4] known tax 26 s were describe including A. chiapasense ssp. t pipe d in this revision: A. armeniense, A. chamulense Matuda ssp. oaxacanum, A. halmoorei, ^ ncetillense, A, nelsonii, and A. rzedowskii. In addition, fi binations! l proposed, laxiacense (Matuda) Croat, A. cuneatissimum (Engler) Croat, A. pe- datoradiatum ssp. helleborifolium (Schott) Croat, A. schlechtendalii ssp. jimenezii (Matuda) Croat, and A. subcordatum ssp. chlorocardium (Standl. & Anthurium "002 i à e : wide for the tribe Acoreae, which contains the Pread genus Acorus and the Australian en- —H € Finally. : Colletti L. O. Wms.) Croat. demic Gymnostachys. The tribes Culcasieae and Zamioculcaseae with 26 species are strictly Af- rican. Four Asian genera, Pothoidium, Anaden- drum, Pothos, and Pedicellarum, account for eight percent of the species of the subfamily. Only Po- thos, with about 50 species, is very large. Araceae contains 110 genera and more than Thi i ;., his study was completed with support from National Science Foundation grant DEB 77-14414. The in Mis d R, F, G em U, . P, UC, and US. I thank the curators of these herbaria for the loan of materials. acknowledge the technical assistance of Frances Mazanec, Patricia Croat, Ann Ruger, and Emily ! Mi i ; 'ssouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166. Ann, Missouri Bor. Garp. 70: 211-420. 1983. 212 2,500 species. It is worldwide in distribution but (Croat, 1979). There are 14 genera restricted to Africa, and a few genera restricted to temperate regions of tl tł hemisphere, includi gtl Mediterranean region. Important local centers of diversity include subtropical and warm temper- ate South America, with eleven endemic genera, and the Indomalayan region, with thirteen en- demic genera. At least 1,350 species, roughly 55 percent of the total, occur in the New World tropics and subtropics. Roughly half of these are Anthurium. MATERIALS AND ACKNOWLEDGMENTS The terminology and usage in the descriptions in this paper are defined in “Standardization of Anthurium Descriptions” (Croat & Bunting, 1979). Colors referenced in the descriptions that follow are taken from the color chart by Berlin and Kay (1969). This color chart, available from the University of California Press, is a repro- duction of the Munsell Color Array of 40 hues, at maximum saturation, with nine degrees of brightness. The B & K (Berlin & Kay) color chart, as it is referred to here, represents 40 hues in the vertical columns and 9 degrees of brightness in the horizontal rows. Colors are arranged in 10 basic clusters with 4 different hues per cluster, ranging from red through yellow, green, blue, purple, and finally red-purple. The four columns for each color cluster are numbered 2.5, 5, 7.5, and 10. These numbers are repeated for each basic color type. The colors from the B & K color chart are read by first reporting the color, then the row followed by the column. For example, the third color in the fifth row in the red area would be called Red 5/7.5. The second color in the eighth row would be called Red 8/5. I have preceded the color references by B & K to in- dicate which color chart is being used. Although a more complete discussion of flow- ering behavior of Anthurium has been published elsewhere (Croat, 1980), mention should be made here of the descriptions of flowering behavior, especially with regard to the development of stig- matic droplets and the progressive development of stamens. Most observations reported here are based on the study of cultivated material, and references to the degree of accumulation of stig- matic fluid might be inaccurate in the absence ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 of pollinators to remove it. Nevertheless, these Observations show that important taxonomic haracters are exhibited by the timing and amount of stigmatic fluid and particularly by the devel- opmental sequence of the stamens. They are therefore recorded here as part of the description of the species where they have been observed. All estimates of ecological zones given in this paper are based on either Holdridge Life Zone maps for most Central American countries and for Panama or on the “Mapa de tipos de vege- tacion de la Republica Mexicana" (Flores et al., 1971). As yet no study has been made to correlate the vegetation types represented on the Mexican map with those of the Holdridge Life Zone sys- tem. For an understanding of this system Se Holdridge, et al. (1971). o DISTRIBUTION OF ANTHURIUM Anthurium ranges from Mexico (Tamaulipas on the Caribbean slope and Nayarit on the Pacific slope) to northern Argentina and peo Species diversity is greatest in the low and mi elevations of northern South America, Panama, and Costa Rica, with less important conten the mountains of southern Mexico and sou : eastern Brazil. While the upper Amazon S of Bolivia and especially Peru, Ecuador, and s lombia are rich in species, the lower Amazon relatively poor in species. DISTRIBUTION OF CENTRAL AMERICAN ANTHURIUM Because of the large size of the genus B taxonomic difficulty, revisionary work on si thurium must progress in stages. There di xico proximately 219 species of Anthurium 1n This and Central America (including Panam 122 paper, the first of a series, will deal wi m Mexican and Middle American taxa. yer exclude a revision of the Panamanian $ vill which number 152 species alone. The E m be presented in Part II of this paper. t parts the great morphological variability in MS of Anthurium, it has been very difficult to oa : a key to large numbers of species. In ade! ently the Panamanian species are not yet e pe well known to complete a key to the ee that curring in Panama. This is due to the pre rest Panama has almost as many species as ®® —— of Central America and Mexico put ! Massif annn.. gpass, | 1983] CROAT— Most of these species have been discovered only in recent years and many are still poorly known. The Costa Rican Anthurium flora is more closely related to that of Panama than to that of the remainder of Central America, but it is rel- atively well known (Croat & Baker, 1979), and thus will be included here. Although Nicaragua is beginning to receive se- rious botanical attention as the result of the Flora of Nicaragua Project conducted by W. D. Stevens (Missouri Botanical Garden), its aroid flora is proving to be relatively poor. Most Anthurium species collected in Nicaragua represent species of Costa Rica and Panama that reach their limit in Nicaragua. These are found principally in the very wet southeastern corner in the Department of Zelaya. Examples of these are A. acutangulum Engl., 4. hacumense Engl., A. ochranthum C. Koch, A. ramonense Engl. ex K. Krause, A. spa- thiphyllum N. E. Brown, and A. upalaense Croat & Baker. Other species are basically Mexican Species that just reach Nicaragua from the north. Examples of these are A. huixtlense Matuda and A. schlechtendalii Kunth. Additional Nicaraguan Anthurium species in- clude wide ranging species that range from Mex- Ico or Middle America to South America, such ^" A. cubense Engl. (also in West Indies), A. c/a- Vigerum P “a “get it is certain that more species lecting s ed to the flora with the ongoing col- little a Stevens and his associates, there 1s is greater t that species diversity of Anthurium Ris hia of the San J uan depression, which although Osta Rica and Nicaragua. Costa Rica, 65 a e ciably sialler, has approximately anama S of Anthurium (Croat & Baker, 1979). Never E I$ even richer with 152 species. Nica- M E Orests that I have investigated were poor es. For example, the cloud forest on Vol- mbacho east of Lake Nicaragua had only Pecies of aroids. Even the much wetter cán Mo a few s OF MEXICO AND MIDDLE AMERICA 213 cloud forest regions between Jinotega and Ma- ama, many species can be encountered in a small The remainder of Middle America is, like Nic- aragua, relatively poor in species. Honduras has ly own species with an additional four species, namely A. friedrichsthalii Schott, A. mi- crospadix Schott, A. schlechtendalii Kunth, and A. trinerve Miq., expected there. Like Nicaragua, Honduras is expected to have only a single en- demic species, namely A. /ancetillense Croat. Belize has no more than eight to ten species (none of them endemic) but Guatemala, with about 25 species, is richer. Although the greater number of species in Guatemala can perhaps be attributed in part to the fact that it has been much better explored than the remainder of Middle America, the chief reason for its greater number of known species is that it shares with Mexico a number of typi- cally Mexican species. These include A. chia- pasense Standl., A. huixtlense Matuda, A. mon- tanum Hemsl., A. seleri Engl., and A. titanium from Chiapas are to be ex These include A. berriozabalense Matuda, A. chamulense Matuda, A. cordatotriangulum Ma- tuda, A. nakamurae Matuda, A. ovandense Ma- hott. e 3 m =j a Aa 3 Ey ~ 3 Q n. & ~ = 3 Un o + and A. retiferum Standl. & Steyerm. However, all of these might possibly also be discovered in Chiapas with further collecting there. Mexico, with 41 known taxa and an additional four or five sp p : y rich Anthurium flora with 26 endemic taxa. Some of the endemic taxa are relatively widespread within Mexico, such as A. andicola Liebm., A. chiapasense Standl., A. schlechtendalii ssp. ji- menezii (Matuda) Croat, A. nizandense Matuda, A. ovandense Matuda, A. pedatoradiatum Schott, and A. podophyllum (Cham. & Schlecht.) Kunth. The majority, however, are relatively narrow en- demics, mostly known only from northern Oa- xaca (seven taxa, A. cerropelonense Matuda, A. chamulense ssp. oaxacanum Croat, A. longipel- tatum Matuda, A. machetioides Matuda, A. nel- sonii Croat, A. subovatum Matuda, A. umbrosum 214 Liebm., A. yetlense Matuda); northern Chiapas (four taxa, A. berriozabalense Matuda, A. clari- nervium Matuda, A. leuconeurum Lem., A. le- zamae Matuda); or southeastern Chiapas (two taxa, A. cordatotriangulum Matuda, A. naka- murae Matuda) any of the Mexican species appear closely related, and their current isolation is perhaps due to changes in the environment brought about by increasingly arid conditions and the formation of broad expanses of intervening desert or ex- tremely arid areas. Thus, for example, A. hal- moorei Croat, A. schlechtendalii ssp. jimenezii (Matuda) Croat, and A. nizandense Matuda all appear to have been derived from A. schlechten- dalii Kunth, a common species from the Carib- bean slope (or from a common ancestor). Spe- ciation here would appear to have resulted from the onset of arid conditions with the resulting isolation of these species in specific parts of west- ern Mexico. Similar isolation of species has oc- curred in Chiapas, where relatively close species are now endemic to northern Chiapas or to southeastern and southwestern Chiapas. That Mexico and Middle America were iso- lated from Costa Rica and Panama during the evolution of many Mexican Anthurium species seems obvious from the fact that only six Mex- ican species, including A. flexile Schott, A. mi- crospadix Schott, A. pentaphyllum var. bomba- also to Panama and the latter two range through- out much of South America as well. The situation in Guatemala is little better with only ten species reaching Costa Rica. These include (in addition . f, A. cubense Engl, A. gracile (Rudge) Lindl., and A. interruptum Sodiro. The three latter species are also present in South America. Further evidence of this isolation can be seen in the distinctive nature of many Mexican species and the predominance of D-shaped or broadly sulcate petioles that are uncommon elsewhere in Anthurium. This feature is exhibited in several leaf forms, including species with oblong leaf blades and ovate-cordate blades, as well as in pedatisect species, such as 4. podophyllum (Cham. & Schlecht.) Kunth and A. pedatoradia- tum Schott. The Mexican species in this group will be referred to as the Anthurium andicola alliance. They may ultimately prove to be a sec- ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 tion in their own right, but final judgment on this will be deferred until further work can be done with the South American species of the genus. HisTORY OF WORK WITH CENTRAL AMERICAN ANTHURIUM The last revision of Anthurium was that of Engler (1905), who reported 58 species for Cen- tral America. The work was completed at a time before serious botanical activity had begun in many parts of Central America, and the number of species is now estimated at 219, almost a four- fold increase. Since Engler's time, most work conducted on Anthurium in Central America has been the piecemeal description of new species and the compilation of floras for Guatemala, Costa Rica, and Panama. Most of the treatments were pre pared by P. C. Standley, who also made à sub- stantial number of collections. The Flora of Gua- temala treatment was completed by P. C. Stan- dley and J. A. Steyermark (1958). Steyermark also made many collections, especially in ni temala, and must be mentioned as one of the 1 collectors who paid any attention to descriptive detail of the plants in his notes. he most serious student of Anthurium was e few mately 30 years (beginning about 1949), A: with the Araceae of Mexico. In all, Matu since. Matuda's most important work (7 summarized the Araceae flora of Mexico was known then, but he also described a num 5 of species after the publication of that reves E Also important in the study of Mexican on ceae was G. S. Bunting, whose “Comment Mexican Araceae,” published in 1965, pic rized his knowledge gained from four € 6, field trips to Mexico during 1959, 1961, and I^ Some of his commentary was useful in pr rihe my own field trips to Mexico. A revision " M. palmately-lobed species of Anthurium 3 : the Madison (1978) has also been useful during Preparation of this revision. is exicd Only 35 of the 122 species inhabiting M d and Middle America were described befor (19 gler’s time. Of these, the largest numi species) were described by Schott with EL CU xa 1983] numbers by Hemsley (three species), Poeppig and Endlicher (two species), Masters (two species), and Liebmann (two species). The remaining sev- en species were each published by a different author: Hooker, Rudge, Lemaire, Chamisso & Schlechtendal, Aublet, Kunth, and Miquel. Engler published an additional thirteen species near the turn of the century and Luis Sodiro, "onn in Ecuador, described two Central period. Matuda, leading the list of Anthurium workers subsequent to Engler's time, published 19 species. Standley and his co-workers, L. O. Williams and J. A. Steyermark, published five species as a re- sult of their work with the floristics of Central America. A single species each was described by J. F. Macbride, T. G. Yuncker, and L. H. Bailey during the same general period. Recently de- scribed material includes 20 species by Croat and Baker (1979) and a single taxon each by R. Shef- fer and G. M. Barroso. This paper will describe six additional new species. Several people, although not describers of An- thurium, nevertheless made significant contri- butions by means of their collecting efforts. These Include D. Breedlove (Chiapas), R. Baker and nne act a species in the same area. When he id, he frequently (if not always) assigned the a Plant the same number he had as- dnt ad > a using onlya different date to m Pa. different specimens. This practice has his eika cases to mixed collections when is iid collections were in fact a different species Curred in the same area. TAXONOMIC CHARACTERS ous P ological characters used in the taxon- the sec Anthur ium depend to some extent upon that h tion being dealt with, but the characters shape ave been used most frequently have been length and venation of the leaves, internode spadix i the color and shape of the spathe and cussed. any of these characters are already dis- elsewhere (Croat & Bunting, 1979). we Although roots have not been used !uonally as taxonomic characters, they dis- CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 215 play a wide variety of qualitative characters. Typically, roots develop one or more at each node. In section Polyphyllium Engl., the roots are numerous, short, and slender and are borne along the internodes. In most species of section Pachyneurium Schott, and in other short- stemmed species as well, the roots per node are much more numerous. Since the internodes are very short, the root mass can be large and dense, with the roots contiguous or nearly so. In many cases, there has been a divergence in the function of roots so that some are utilized exclusively for aay are directed upward where they collect debris, or they extend into the rosette of leaves and collect debris. In other cases, the support roots of the plant are negatively geotropic, such as in the case of pendent species (Madison, 1978). Commonly, roots appear smooth when fresh; when they are dry it can be seen that they actually consist ofa dense mass of woolly tricl that closely matted they look contiguous before shrinkage of the root. Anthurium gracile (Rudge) Lindl. has roots with a thick white layer of velamen useful in absorption of moisture from the air. Other species that have been investigated lack this ve- lamen layer. Stems and cataphylls. Although stem anat- omy has not been adequately studied, the out- ward appearance of the stem provides few good taxonomic characters except for a considerable range in length and width of the internodes. Stems are usually scurfy and brown in age. While there is a considerable difference in the size, shape, and degree of indentation of the petiole scars, these characters have not been used systemati- cally. Much more important are the cataphylls that frequently persist on the stem. The color, texture, d the degree of weathering they undergo are useful taxonomic characters. Growth in Anthu- rium is sympodial with each shoot bearing a bi- carinate prophyll and a larger, single-keeled cata- phyll, a foliage leaf, and terminating in an inflorescence. The inflorescence, although always potential, may abort (Engler, 1877). The cata- phylls, which are usually green, promptly turn brownish, reddish, or yellowish and may be de- ciduous. More frequently they persist, either re- maining intact, such as in section Calomystrium Schott and some other species, or they are mar- cescent, variously weathering into a network of y accumulat e pe» 216 color, degree of weathering, and degree of per- sistence of the cataphylls frequently provide use- ful taxonomic characters. Leaves. Among the more useful taxonomic characters that have been overlooked previously is the cross-sectional shape of petioles (Fig. 1). Most Anthurium species have petioles that are terete or subterete in cross-section and frequently significant. Many Mexican species have petioles that are D-shaped or broadly sulcate in cross- section. Petiole cross-sectional shape, within entirely. Some of the more variable, wide-rang- ing species, such as A. schlechtend. lii Kunth, have considerable variation in petiole cross-sectional shape even within a single population (see that species for a discussion of this). Some species of Anthurium have sharply triangular petioles, oth- ers trapezoidal, square, or terete with several to many sharp ridges. The degree of channelling of the petiole also provides useful taxonomic char- acters. The length of petioles, like overall size of any morphological part of Anthurium, is highly unreliable because plants often flower when quite young and thus the overall size of leaves, inflo- rescences, and other parts vary remarkably in ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 size over the course of the life of the plant. More useful is the relative length of the petiole in com- parison to the length of the inflorescence. The most widely used taxonomic characters in Anthurium are the shape and venation of the leaves, since these are p I p di e than in any other genus of plants (Croat & Bunting, 1979). In addition, glandular punctations on the diro consistently have glandular punctations on one or both surfaces. Still, some species in oth- erwise epunctate sections, such as section Cal- omystrium Schott, have evolved glandular punc- tations independently. While many members 0 section C. ium Schott have t sh spots resembling the glandular punctations of section Porphyrochitonium Schott and others, at least one unknown isolated Brazilian species (Rosa & Lira 2325) has definite glandular punctations 0? the lower blade surface. A group of coriaceous, cordate species from Mexico (A. verapazense al- t XL 1 A 1 TET dl punctations. Other useful features in the leaf include the presence of papillate epidermal cells, such as in section Cardiolonchium Schott, and the presence of easily visible raphide cells so prevalent in s tion Calomystrium Schott. Inflorescences. The inflorescence, i generally critical for identification of spec! a Anthurium, has been under utilized, perhaps aon e -— : : í : 1), 10 Cross-sectional petiole shapes in Anthurium.— A. Basically terete: ranging from esulcate (1) FIGURE |. shallowly and acutely sulcate (2 & 3), to obtusely and broadl cate (5), ally to narrowly and acutely sulcate (6), broadly and acutely sulcate (7), shallowly and broadly sulcate (8), he of i i B. D- (9), flat adaxially with marginal ribs (10), to flat adaxially wi broader than thick: 3-ribbed abaxially (2), broadly and obtusely sulcate a adaxially, narrowly and sharply ribbed around the rem the sulcus (5). * Not yet observed but to be expected. ed abaxially (1), to obtusely and broadly sulcate a ially 5-ribbed abaxially (3), broadly and s aining circumference (4), or with one or or quae daxially. harply sulca more r! 1983] > œ O = m CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 217 CROSS-SECTIONAL PETIOLE SHAPES IN ANTHURIUM Basically terete 1 2 3 4 5 6 7 8 9 10 11 D-shaped or broader than thick 1 2 3 4 5 6 7 U-shaped or thicker than broad 1 2 3 4 5 6 Markedly angular nHavyvyvyyvy - Markedly ribbed abaxially diii 218 cause of too much reliance on herbarium ma- terial, which greatly diminished the value of the inflorescence. Peduncle length has frequently been used as a taxonomic character, yet length alone is extremely variable on plants of different ages and even in populations of plants of the same age. However, the ratio of the peduncle length to the petiole length is usually relatively constant and often provides a useful character. Like most vegetative parts, the overall dimensions of the spathe and spadix are not very useful, but again the ratio of the length of the spathe to the length of the spadix is relatively reliable. The shape of both the spathe and spadix and the texture of the spathe are usually much more reliable than size. The spadix and the many characters it em- bodies is perhaps one of the least used taxonom- ically. This is no doubt due to the fact that too few species have been critically compared in live condition. The dried spadix is difficult to study and loses many of the characters that are appar- ent before drying. One of the most useful char- acters exhibited by the spadix is color. Spadix color when seen at anthesis is usually stable, and species vary considerably in the colors exhibited. In some species, however, the spadix varies from green to violet-purple. While many spadices are green before the spathe opens, most quickly as- sume the color they will have at anthesis. Some species may be green when stamens begin to emerge and later may turn violet-purple. Some species change colors after anthesis, so it is im- portant to make note of the color at anthesis. Anthurium huixtlense Matuda and A. formosum Schott, for example, have spadices that become noticeably darker after anthesis. ther characters exhibited by the spadix in- clude scent, size, shape, and texture of the flow- ers; the number of flower spirals and the number of flowers per spiral; the disposition of the stig- mas and timing of the stigmatic droplets; the rate of emergence and disposition of the stamens, as well as the shape of anthers and pollen color. Many of these characters are discussed in detail in Croat (1980), but a summary ofthe important points will be included here as well. Although there has been no success as yet in determining scents of flowering Anthuri um, there is little doubt that a large percentage of them have distinctive aromas. Those with attractive bright spadices have sweet aromas, such as of mint or lilac, and are no doubt bee-pollinated. Euglossine bees have been seen in several cases visiting Anthurium (Croat, 1980). Species with ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 purple or drab colored spadices generally have yeasty or rotting fruit aromas (the scent of one species is like that of fresh cucumbers) and are probably fly-pollinated. Scents are usually not evident at night, and some species are pheno- logically active, presenting their aromas only during midday. Aromas of some species a a distance of several meters, but most species give only faint but distinct aromas. Flowers are closely aggregated in spirals and are quite variable in size and shape. Their overall shape at the apex may be square, rhombic, or 4-lobed. The tepals themselves may be 3, 4, or 5-sided. The texture of the surface, presence of droplets, raphide cells, and other features, all may be taxonomically important. ; Stigmas may be slitlike or brushlike and exserted. The production of stigmatic droples varies considerably from species to species, Wi da mportant. c sequence and duration of stamens as well as method of presentation is similarly imponi the tepals, the stamens in some species are es ed and later the filament shrinks and retracts t anther to the surface of the tepals. sana other species do not retract and the filament anthers remain exposed. Stamens in still others retract completely below the surface ofthe tepals and thus out of sight. The sequence of gr emergence is also quite variable (Croat, 129. with the rate of emergence being an à ditio a variable. Most species with typically a pered spadices have emergence beginning # i base of the spadix and progressing toward ices apex. Species with cylindroid or clavate : a have staminal emergence generally scatle types even beginning at the apex or middle. In all s of spadices the rate of emergence may ves all different. For example some species dev of their lateral stamens before the quid nate) stamen and finally the fourth (alt all st stamen emerge. Still other species have nel mens of the flowers in any spiral fully xs : ;rals havé before many flowers in succeeding SP words their lateral stamens exserted. In other pmen! the leading spirals in the staminal develo progression are few. In contrast, other adi may have, for example, ten or more of the! with spirals with only lateral stamens exposed. . another ten or more of the ing $ the bearing three stamens (i.e., the laterals piel third, usually anterior, stamen) and finally | ; ————— —— M — P f "m—— 1983] the lower (basal) spirals with a full complement of four stamens exposed All of the characters mentioned here have proven to be useful taxonomic characters, yet most have never been mentioned previously by other workers. Anthurium fruits consist of 2-loculed tries. The berries provide a number of useful characters including size, shape, and color, the nature of the pericarp and mesocarp, the pres- ence or absence of raphide cells, as well as the number, size, and shape of seeds. Engler (1905) appeared to realize the significance of the berries in classification, but unfortunately he had not seen berries on the majority of the species he Studied. Because Anthurium has markedly pro- togynous flowers (Croat, 1980), they do not readily set fruit on greenhouse collections with- out manipulation. Often this is not possible with- Out a second individual of the species. A great deal of attention has been given to those species in the living collections for which fruits were not Own, in order to encourage them to set fruit. The number of seeds per locule has long been considered an important character, pas such groups as sections Tetraspermium Schott, Gym- nopodium Engl., and Porphyrochitonium Schott having more than one seed per locule, while the remaining sections usually have one per locule. Seeds themselves are as individually different from species to species as are the berries. Most have sticky, mucilaginous appendages firmly as- CM with them (Croat, 1980) often attached One or both ends. In some cases the seed is completely enclosed in a mucilaginous envelope. ni extremely sticky structures seem ideally A to being attached to bird beaks, and An- urium berries seem ideally suited for bird dis- Fruits. CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 219 persal. Many of the Mexican species of Anthu- rium lack a sticky appendage on the seeds. SECTIONAL CLASSIFICATION OF CENTRAL AMERICAN ANTHURIUM In many cases the existing sectional classifi- cation systems of Schott and of Engler are in- adequate. While some species assigned to var- ious sections by Engler (1905) may be related, most sectional descriptions have many overlap- ping characteristics, and I believe that many species assigned to a particular section by Engler bear little relationship to each other, e.g., A. gra- cile (Rudge) Lindl. and A. friedrichsthalii Schott in Schott's section Leptanthurium, and A. pro- tensum Schott and A. durandii Engl. in Engler's section Urospadix. Anthurium friedrichsthalii Schott is a typical member of section Porphy- rochitonium Schott, as is A. durandii Engl., while A. protensum Schott is a member of section Pa- chyneurium Schott. A review of the sectional classification of An- thurium will be published elsewhere and, al- though an attempt will eventually be made to revise the sectional classification of Anthurium, it is premature to arrange Central American species in any subgeneric classification system in this treatment. An attempt will made here, however, tentatively to assign all Central Amer- ican species to a section. In some cases species will be placed within an alliance of species that may later be ranked at the sectional level. It is hoped that ongoing studies with live ma- terial including cytological, anatomical, and chemical studies will prove adequate to provide a thorough analysis of the sectional classification of the genus. PROVISIONAL SECTIONAL PLACEMENT OF sea SPECIES OF MEXICO AND MIDDLE AM l * uf blades prominently 3-lobed or — x palmatisect. - Leaf blades with the lobes united at t Ja. Leaf pets with three lobes poder iei or — s lobes often smaller than the central lobe) _ Section Semaeophyllium Schott * Three sections do not occur in Central — section Gymn A. subsignatum Schott A. tilaranense Standl. opodium Engl. is geses by a single mde Koch nnopus Griseb. h epresented by a single species A. o, ri from Cu ba; sec on Chamae repiu um Schott, rep mi M Haage, i IS restricted n" Braz — "Urospadix ie may be beiee e South America Oncent 220 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor 70 3b. Leaf blades with five or more lobes. .. Section Schizoplacium Schott A pedatoradiatum Sch A. pedatoradiatum ssp. yu (Schott) Croat A. podophyllum (Cham. & Schle 2b. Leaf blades with the lobes free to the base. Section Dactylóphititn Schott A. clavigerum Poepp. A. kunthii Poe A. puema. (Aub. G. Don var. bombacifolium poe sarc A. tri Ni Sodio lb. Leaf blades entire or nearly so, sometimes with the anter pir markedly constricted at the base but never distinctly 3-lobed. 4a. Leaf blades Hy eme dark glandular- -punctate at least on the lower surfac 5a. Leaf blades cordate, hastate, or otherwise with usually definite posterior lobes most hs veins s usually n merging with the margin of the blade proximal to the middle of the bl . Section Eu dais Schott (Anthurium ub s alliance) chiapase 4 chiapasense ssp. tlaxiacense (Matuda) Croat M primary lateral veins extending between all the basal veins giving the blade the appearance of a melastomataceous leaf blade (Section Digitinervium). 6a. Leaf blades with several pairs of basal veins acd eane RR extending to or | almost to the apex; primary lateral distinct, parallel nding more | or less straight between all the basal v veins, giving the leaf blade the appearance of a | melastomataceous blade; pistils often prominently exserted and quadrangular; number E ovules and seeds per locule unknown Section Digitinervium Sodiro A. lentii Croat & Baker 6b. Leafblad tł pair of basal vein he e collective vein) eum cp along the margin to the apex; primary lateral veins ican jio numerous s rie parallel; pistils not quadrangular; ovules and seeds usually two or more per ocu 7a. Plants mostly scandent: stems slender, mostly less than 6 mm diam.; intern rnodes Schott somewhat elongate; acute at apex. ction TAN f A. scandens (Aubl.) Engl. A. scandens ssp. pusillum Sheffer A. (ME vod | A. tri i 7b. Jee Tey more or less caespitose, at giia lacking slender scandent stems; stems speia more than 1 cm diam.; berries often more or less jones and indented rU pe __.. Section Porphyrochitonium acutan, gulum Engl. alatipedunculatum Croat & Baker austinsmithii Croat & Baker cosi eri Hook. f. | andii Engl. fetch Schott hacumense Engl. patie op Schott louisii Croat & Baker esee] "3 S EE S SERE 8593 g w, as pidesi . wendlingeri G. M. n ast not PES dark glandular-puncta ped in cross-section or desee channelled with more éd aes acute latera Schot! d tion Belolonchium f (A nthurium andicola : mr ami | ebm 4b. Leaf blades epunctate or at le 8a. Petioles D-sha margins A. andicola Li A. berriozabalense Matuda ui | 1983] 9a. Vernation of leaf blades involute (i.e., with b 9b. Vernation of leaf blades supervolute (i.e e., with one margin CROAT —ANTHURIUM OF MEXICO AND MIDDLE AMERICA 221 cerrobaulense Matuda chamulense Matuda chamulense ssp. oaxacanum Croat cordatotriangulum Matuda cotobrusii Croat & Baker mea kii Croat Sahe ova Schot subcordatum ssp. E onn (Standl. & L. O. ph? TxAc E dae aes e 3 A D o 3 O -4 o m u m Matuda . titanium Sandi & Ste yerm. 8b. ee bue terete or subterete in cross-section often narrowly aaile but with the argins unded; sometimes square, triangular, trapezoidal, or otherwise, but not usually (PO. Na 4 A th 170215 tad 1 in bud); plants frequently with a “bird’s-nest” habit rt blades often thick, fre Pret e more or less tie a or obovate od cue the prim lateral veins (at least the lower ones) joining the ma ms R E manum Shon pata ale Schott rolatum Croat & Baker priate Schott Leid aceti Croat nth schlechtendalii ssp. jimenezii (Matuda) Croat schottianum Cro Baker seibertii Croat & gs spathiphyllum N. E. Brown ere IRA e $ = 8875 e but " the alternate margin rolled around t n, giving the “cut-away” view a cochleate appear: we epe habit; leaves various; = _ ry lateral veins a oeat joining int lective vein a pei ore reaching the m 10a. Stem scandent and wiry; internodes pongais, x usually less than 3 mm pen with i ng their length; flowering nches Section M dem Engl. ce); plants usually lacking a o a col- seus us, bearing one to six leaves; hemes cataphylls "akin. URBE a. Stand 4 . flexile exile ssp. ui (Macbr.) Croat & Baker 10b. Stems short or scandent but not asi and wiry; internodes short or long but not ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 slender and wiry, usually more than 8 mm diam.; the adventitious roots restricted to the nodes; flowering branches bearing a single leaf (branches are short and do not appear like branches); 1-ribbed cataphylls present and usually cones la. * Section Urospadix centered in Brazil and that also have this characteristic Roots white, bearing velamen; blades thin, narrow T veins numerous and scarcely more prominent than the ma * stems rarely more than 1 cm diam. with persistent reddish-brown, inni ahaavne spadix with no more than three flowers visible per spiral e any angle); chromosome numbers 27 = 10 or 20. s Section Leptanthurium Schott A. gracile (Rudge) Lindl. . Roots usually greenish, lacking velamen; blades various, usually with the pri mary lateral veins D more prominent than the interprimary veins; stems various; cataphylls usually n Calomys- trium with cordate "m number of flowers per r spiral various but usually more than three per 12a. Leaf blades ao.. usually metallic green or blackish-green in ee the epidermal cells papillate or otherwise gere petioles freque ribbed; B-chr eoma present. Section ees Sod A. clarinervium Matuda A euconeurum Lem . lezamae Matu 12b. Leaf blades not velvety, usually smooth and semiglossy to glossy; s. epidermal cells flat, not papillate or raised; petioles various but n commonly ribbed; B- pieni lacking. 13a. Stems bearing co us, thick, intact, persistent, reddish-brown cataphylls; leaf bider: aay thick; the minor veins usually not prominently raised; the upper blade surface usually with linear raphide cells visible; spadix often colored, often sweetly scent ted; i spathe usually pred often br — _ Section Calomystrium Scho A. armenie A. pilosi esie & L. O. Wms. A. clavatum Croat & Baker A. formosum Schott A. hoffmannii Schott A. huixtlense Matuda A. monteverdense Croat & Baker A. obtusilobum Croat & Baker A. ravenii Croat & Baker 13b. Stam PR ‘ntaa mass oF fibers praece persisting intact, as in n A, "eximium wa beers usually lacking phige cells; spadix various, less "he quently scented; spathe various, often lanceolate. 14a. Leaf blades much longer than broad, oblong to linear (rarely ovate » then thin, veiny, and weakly bullate as in A. da- vidsoniae). 15a. Seems elongate, erect to scandent 7-20 mm diam 16a. Leaf blades thin, veiny, and weakly bullate; the reticulate veins often pices ens ee commonly cylindrical; berries c ly pek tt wn ie P ECT Section X. gie vllum Scho ks Meets alliance) A. davidson e Standl. A. pallens Schott : har- lób. Leaf blades usually coriaceous or thickly € ot taceous, smooth; the reticulate veins imer i conspicuous; spathe and agr kowi 5c diu on Xi ialophyllium Se (A. caucanum alliance) apparently not represented in Central America have leaf blade ——— MÀ— QM —— l — pe 1983} CROAT OF MEXICO AND MIDDLE AMERICA 223 A. iios ed diuo & Baker roat & Baker A. testaceum Croat & Ba ker I5b. Stems short; the internodes short; stems mostly less dT H 1 duod 1 1 p and many times s longer than the stamens Spiel ocio DOON opum. Shoe P pittiei A. pittieri var. x $m 17b. uns not po neni exserted long before an- done than stamens. ... Section Xialophyllium ' Schott idend urgeri Croat & Baker (usually with longer stems) 14b. Leaf blades ouch longer than broad (or at least not onen ng or ere mostly more or less ovate with posterior Rie Blades moderately thin, often conspicuously veiny or even bullate; anterior lobe usually with numerous, Reni more or less parallel veins; berries frequently Section Polyneurium Schott luni ca peratum Croat & Baker ott 18b. Blad s coriaceous or nearly so, usually smooth and not noticeably veiny or ird anterior lobe lacking numerous close parallel ve jg ene Belolonchium Schott A. brownii Mas A. ipe gu Matuda oncinnatum Schott vandense Matuda A. pluricostatum Croat & Baker A. silvigaudens Standl. & Steyerm. A. umbrosum Liebm. A. yetlense Matuda Key TO ANTHURIUM OF MEXICO AND MIDDLE AMERICA là. Leaves palmately lobed or divided; leaf oe more than three or if three, then divided to the base. KEY A Ib. Leaves : simple, sometimes lobed. Leaf blades 3- lobed, the margins constricted near the base of the auierior lobe, . . —. KEY B 2b. Leaf blades not 3-lobed. 3a. Leaf td cordate or subcordate at base. E Leaf blades with prominent Doetbrior os conspicuously black punctate below. .. KEY c Leaf blades with posterior lobes, not p te below or not conspicuously so. ..... KEY D 3b. Dur blades not bate or pedes at ee 5a. Leaves tate on one or both surfaces. MET KEY F s pun 5b. Leaves not Ule on i eiker surface. mmm KEY A Ee palmately lobed or divided: leaf segments more than three or if three, then divided to the base. ^ © Leaves E foliolate with the segments free to the base; plants terrestrial. A. trisectum Sodiro deed 5-13- benc or lobed (if Misc then the peduncle less than 5 cm long). Blade € segmen all free to the 3a. aa pi d scandent, epiphytic leaflets lobed; peduncle elongate; spadix to uc RR € dark violet-purple. —_____- 3b. P Sed relatively small, bar with long "internodes; leaflets not | lobed (except. E basal leaflet); peduncle long or short; spadix less than 25 cm long; spathe green to purplis 224 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo. 70 4a. Leaf segments all more or less equal, petiolules 4—6 cm long; peduncles slender and long, usually more than 20 cm long, longer than spadix; spadix whitish to greenish at anthesis. . kunthii Poepp. 4b. Leaf segments unequal, the lowermost more or less auriculate and usually markedly inequilateral; petiolules usually short, less than 2 cm long with the leaf tissue decurrent, often almost to the base; peduncles short and stout, usually less than 6 cm long, usually shorter than the spadix; spadix lavender at anthesis. ss pentaphyllum var. bombacifo ent (Schott) Madison 2b. Blade segments united at the base; plants terrestrial or on rocks on steep slope 5a. Blade segments generally pinnatifid with 5-12 narrow ditione. the siete rarely entire but then subcoriaceous; the peduncle longer than the vi joles. .. ig i ar que (Cham. & Schlecht.) Kunth 5b. Blade sspe entire, the peduncle shorter than ine petioles. 6a. Blades evenly segmented. .. pedatoradiatum spei ssp. pedatoradiatum 6b. Blades unevenly divided, at least the pes ÉL nt free or nearly so. — = N sey inerti ssp. helleborifolium (Schott) Croat KEY B Leaves simple, 3-lobed, the margins constricted near the base of the anterior lobe. la. Posterior lobes usually more or less orbicular; lower blade surface glossy, yellow-green when dry; pri _— b. imary lateral veins numerous, prominent, more or less straight and parallel; tepals yellow. ~- 5 andu Schott Posterior lobes not as above: ‘lower ‘blade surface not glossy; pets lateral veins either not numerous or not prominent or neither; tepals purple or yellow to gree E 3a. Plants scandent with long internodes; spadix yellow-green to yellow; venie violet-purple. subsignatum Schott 3b. Plants n more c or yr less acaulescent with short internodes: spadix neither ee nor yellow; berries red or violet-purple. 4a. silo sas vun dde anterior lobe; spadix creamy-white or greenish to reddish wn; berries red, at least at apex. .. A. tilaranense Standl. 4b. Posterior jei extended dy from anterior lobe; spadix green tinged with vor e; a violet-Duml. OO 15 9 02 o oo co ee eri E 2b. Basal lobes us at more or less right angles to anterior lobe. 5a. Posterior lobes broadly rounded, the sinus usually spathulate; spadix violet-purple; blade margin often Syrien Costa Rica to Ecuador. ~ A. brownii ME b. Posterior ihes narrow, the sinus not spathulate; spadix usually greenish; blade margins not markedly paipak Mexico and Guatemala. 6a. Blades with the posterior lobes usually longer than the anterior lobe; the mas between " e anterior and posterior lobes stra raight to convex; the collective vein arisi usually da m the uppermost basal vein in the anterior lobe. . ud Den riozabalense Me : Blades with the posterior lobes usually shorter or no ) longer r than the anterior lo margin between the anterior and posterior lobes usually markedly concave; the collectiv ve vein arising from one of the lowermost basal veins and extending along the "E Engl. both the posterior and the anterior lobes. em Un a o BEY C Leaves simple, cordate, or subcordate at base; with prominent posterior lobes, black punctate below. la. Leaf blades usually subcordate, stiffly coriaceous wit l ri. ore basal veins strongly pue roughly parallel to the midrib to beyond the mid in cross-section; berries white; Costa Rica, Panama, Colombi A. len owe fbl oriaceous, the basal veins not as pepa — he uppermost extending above mid lish cross-section; berries red, pees je purp subco rming a collective vein near t the Margin, usually only t is — pistils not markedly early emergent, terete in atm 2a. Blades mostly 2-3 times longer than broad, the anterior lobe usually more or less oblong and sometimes constricted at the base or narrowly triangular. 3a. Posterior lobes mostly less than 1.25 times longer than broad, the sinus parabolic to spathulat? atuda often closed; northern Oaxaca in the pise Madre Oriental A. longipeltatum M 3b. Bim, ce mostly more than 1.5 ti longer than broad, usually : directed promine ae outward; the sinus sed rai = ‘ethan Chiapas Belize, Guston n ne €, not closed; southern Oaxaca and 4a. Inflorescences usually 1. 5-3 tines longer than the red-o tioles; be eue to orange-red; seeds 6-9 per berry, less than 3 m a oe verapazense vns En ———— 1983] CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 225 4b. Sppe usually less than 1.5 times longer than petioles 5 x olive green at anthesis; primary lateral veins prominently raised below (fresh jd diy; restricted to severis recen leon asd uus tlaxiacense (Matuda) Croat 5b. Spad ateral southe nerit ante and western Guatema e SA. chi neue ' Standl. ssp. chiapasense 2b. Blades mostly 1.25-1.5 times longer 1 than broad, rarely more "than 2 times longer than broad, t anterior lobe usually more or y with the anterior | g or narrowly tri , Blades conspicuously coriaceous, acute at the apex with a short "e Deua tern Oa- ograndicolum Matuda 6b. Blades usually only subcoriaceous (at least when dry), acuminate at E Ad from on : 7a. Inflorescences usually 1.5-3 times longer than petioles; spathe usually less than half as ong as the spa ei ‘berries subglobose, red-orange to orange-red; seeds 3-5 per locule (7- 9 per berry), less than 3 mm long. A. verapazense Engl. 7b. Inflorescence kae: less than 1.5 times as long as the petioles; caen usually more ies half as long as spadix; berries obovoid, bright red; seeds 1 per locule (2 per berry), mo than 4.5 mm long. y ke cens Standl. ex Pike KEY D Leaves simple, cordate or subcordate at base, with posterior lobes, epunctate below. la. Plants scandent and twining or at pie with long slender stems with elongate hase held erect usually by rooting onto or being supported by trees; blades mostly less than 27 cm long. 2a. Leaves bullate; peduncle Heiny eena obscured by the base of spathe, s rarely t to A Standl. 2b. Leaves : not bullate (except A. davidsoniae); peduncles conspicuous. 3a. Stems mostly more than 2 cm diam.; blades 16-48 cm long, 11-44 cm rai En A. subsignatum Schott 3b. Stems mostly less than 1 cm diam.; blades Lou = fct ai cm long, 16 cm wi —4 mm diam., usually Short roots; sea level to 1,700 m 5a. Leaves usually palmine nerved, lanceolate, cordate at base and broadest below the middle of the blade; spadix usually with a conspicuous stipe 1-2 cm long; Mexico to northern Guatemala, 1,000-1,700 m. .. +h WAIU 111dlly A flex ile ssp. muelleri (Macbr.) Croat & Baker 5b. Leaves plinerved, elliptic to ‘narrowly elliptic, narrowly subcordate at base, m ss : at or ipn the middle of the blade; spadix sessile or dice Be stipe to 2 m exile a aa exile Mex o Panama, sea level to 1,000 m 4b. Stems nece ps 52 m diam., usually 1 m or less long, me stiffly erect, usually rece d by a tree, not twining, the internodes lacking roots; roots long, few, restricted o the grid oe 1,200-2,300 m oa: Leaf blades all cordate or, rar rely, su sübcormddle. -nman T Oe S 6b. aed blades subcordate intermixed with blades that are rounded or truncate, r. SA mi oud tok = lb. Pla ordate. EEA nts not ligia t; internodes short. a a. Species of Costa Rica and pir ofte ling into South America t t g (A. ravenii as far north as Hondura 8a. Geniculum situated at least 10 c cm below the base of the blade; blades purse terrestrial. oerstedianum Schott 8b. Geniculum immediately subtending the leaf blade. 9a. Spadix more or less clavate, broader in rss at apex than at base; spathe quae hes hes a ong cospidaie x, long-dec ied base; blades pn € brow middle elevations, Cordillera Fh jm nca. um Croat k Ae 9b. Spedix idro or tapering from base to apex, aves clavate; scala ats oblong, lanceolate, or ovate. 10a. Spathe purple to dark maroon or mottled green es species with spathe predominantly green wi and purple throughout (excluded ith purple tinged margins and/or eins). 1 la. Spathe twisted and contorted; peduncle short yum ani lowland si E long); sinus narrow; — owlands near Bribrí iod a i ianum Croat & Baker llb. luda flat; bosse more ‘than 20 cm long, or if funes not dwarfed by the petioles; sinus narrow or broad; mostly 400-2,4 400 m ER 12a. Leaf blades shallowly cordate or subco I arnar ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 1 » | 1 *9laj5l. 1 } > 11 1 lat 7m P I > Jy 1a vv vial "y 5s FAM | 1 a | in outline. I meor veins of ical re reddish beneath on diving A pe Mast. 13b. I 3-25 mm long; berries orange; major veins of leaf not reddish beneath. e anchoanum 1 Engl. 12a. Leaf seen deeply cordate at base or shallowly cordate pa triangular in outl 14a. Sie usually dark purple vicies stipe 0.5-2.5 cm ong stamens long-exserted. watermaliense Hort. ex L. L- . 14b. — vidis um tinged i purple; ve 2-6 mm Tai ns weakly exserted. otobrusii pe 5 Baker 10b. Spathe white, ye ae: green, or with purple-tinged veins and/or ma 15a. Leaf blades generally bs atts or broadly lanceolate in ORE Eds or subcordate basally, the lobes never Sida each other; spathe green or green tinged with purple, never 16a. Petiole 4-sided; blades E peduncle less than 25 cm long, muc han the Se l3 - A. por Scot bte A m long, slight iir Daa to e than | petio ioles. 17a. Spathe o oblong-lanceolate; berries purple-violet to reddish-violet major veins of leaf maroon below; blades almost nd with a purplish tint or with purplish blotches; 400-1,70 es A. bo | Mast. 17b. Spathe lanceoiate to ovate; ; berries orange; ma veins of leaf not maroon be ying p purpli ish h blotch- 500)1, 000.2, ae A.r anchoarmti Engl. . Leaf T ici heart-shaped in outline, with conspicuous and dee basal sinus, the | qe often overlapping each other; spathe white, green or green tinged with purple. 18a. Blades agii the posterior lobes usually orbicular; pp golden yel- dur E Schott [o 18b. Rian i. purple, or occasionaliy yellow 19a. Plan wing at elevations above a 000 m; flowering S spadix 5- tt 6 cm oi on a st ipe 1-2.5 c _ A. concinnatum Scho 2 m lon 19b. udis usually growing at elevations S iy 2. 000 m; spadix var- 2 us lengths, usually sessile (4. standleyi with a stipe 0. 6-6 cm on 20a. Largest leaf blades less than 55 cm lon; adix dark purple; spathe pies green, sometimes tinged purple. 22a. rie subcoriaceous, the basal veins coalesced the sinus much broader than deep when Seated spathe ovate to narrow ; berries orange A 2 N o ws EB C m US [^2] + M B5 B. e "em É < Q, z [^7] & [5 e = deeper on broae, mes closed with the lobes overa i 21b. Spadix white to yellow or pale green, eee purplish. 23a. Collective vein continuous from n basal lobe to apex of blade. .. A. obtusi 23b. Collective vein formed only in u blade, arising from vepe bar cus veins, asal veins running to m the 24a. Blades diede MR cR the sides of ils anterior lobe Mes apeo cataphy weatherin bers the usually purple; at ie fons rial plants ter situm Sot the lower PURIS Aw termaliense Hort. ex LH ils 24b. Blades peus the sides convex; catap y a hA o—— hers dum. , 1983] CROAT OF MEXICO AND MIDDLE AMERICA 227 persisting more or less intact, reddish- brown; spathe pale, aed dee purplish plants epiphytic or terr 25a. Jedem —-6(8. $e M ida. broadly —. A. mont oe Croat & Baker 25b. Peach less than 3 cm wide, lanceo- late to joueur A. hoffmannii Schott 20b. Largest leaf blades more than 55 cm long. 26a. 2 a s Leaf blades with tertiary veins sunken above, giving the blade a bullate appearance, drying pale yellow- green with a wrin nkle d, chartaceous character; sub- L 1 11 * the base; n pale s ien - heittle: spadix pale green es pale white. ... A. caperatum Croat & Baker in color various, often tinged purplish; spadix white, low, purple, or green tinged with purple; berries 28a. Seige violet-purple, long tapered, to 28 m long; blade prominently » nd ze S nii Mast. 28b. Specs white, pale green, sac i hail violet, short and only slightly tapered, 5 18 cm long. 29a. Spathe soon reflexed; spadix cream, white or pale yellow-green; berries red, round at NS ravenii Croat & Baker to 29b. Spathe remaini io close e o c orange base, sharp əy Bry ei a Schott 27b. Plants terrestrial. 0a. Co llective vein arising well above the rries orarige. Pi ou acs IUNII standleyi Croat mi eee 30b. Collective vein dis arising from o ofthe basal veins; spathe white, pale piu or pale violet or tinged with pale violet; spadix white, yellow, lavender, or pale BICC Lite ange. 31a. — usually drying with large, picuous, purplish blotches; green becoming -0 on eut ofthe Atlantic slopes at middle elevations, some- times seen as the wet mre on disturbed steep EEE osum 1 Schott 31b. Blades lacking dark splotches: spathe and spadix esed colors; berries not colored a 32a. Spadix bright alien at anthe- sis, the anthers not visible (pollen oozing out in slender posterior lobes elongated, the 228 7b. Species of Mexico and northwestern middle America, northwe d of Río San Juan depression (Nicaragua), som ANNALS OF THE MISSOURI BOTANICAL GARDEN 3 N E metimes in El Salvador and Honduras, rarely in [Vo.. 70 anterior lobe us sually some- smooth = E M C. Koch Spadix pdt loved or green d ragua. 33a. densis es broadly sulcate or D-shaped in cross-section, the dissi margins often acute, some- tim Ma 35a. obtus af blades merely subcordate, lacking well developed posterior lobes. T f een with the basal veins and primary lateral veins € —— ca in northern Oaxaca. Sierra Su ubovi tum Matuda 35b. Leaf t blades with all veins greenish; oe and southern Oaxaca to Chiat G uatemala, El Salvador, or Hondur. 36a. Blades less than 12 cm long; uiros Chiapas near — with Oaxaca. ALG rrobaulense Matuda 36b. Blades usually more than 15 cm long; Oaxaca, Guatemala, Honduras, or El Salvador 37a. Collective vein arising from on northern Chiapas to Gua . Collective n wae nb the primary lateral veins, the poste broad and not directed piecinandy ou 38a. Blades more or less broad, often broader near the 3 ~ c Chiapas, Gus in Mexico, e of the lower basal veins; Le lobes usually wenns ed broad and directed prominently E ES mala. than er oblong, three à four — o: ae Matuds 38b. Blades triangular to ovate, bw: dim three times longer than road, broadest at ba 39a, pen blades with margins of the anterior E usually con- A. è rrobaulense Matuda 39b. 40a, Spadix pale green; blades ce the collective vein bs — from the first basal v frequ secondary see ws vein v— irregu- Nea ite the margin to the nd basal vein mergi = wit pex; seco Leaf blade with margins of the anterior wes s sl y con- vein. uently or nearly to the idi the margin in the upper half of the blade; Mexico, in ecc and southern Oaxaca and Chiapas. merging with th v 0 ; : e margin below the middle n blade; Oaxaca and Serine Guatemala, El Sal dor, and Ho dura: 4la. Leaf ‘blades v ak à tertiary veins obscit® she Sal- bove M1 v» vador and western Hondura: .. A. chamulense Matuda, bie Standl. “A. subcordatum ssp. chlorocardium ( . O. Wms.) Croat, A. subovatu 41b. Leaf blades with tertiary veins minutely sunk- um Matu- 2 ^ E ue ——— Á— —————— O————— —————— M À— 3 — vss coser ETHNIC MMC — "—-— ^ —-—— -E ee S 1983] CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 229 en and visible above when fresh; Oaxaca and 42a. Leaf blades with tertiary veins below pina hae’ hana subcori riace ous; occu “bosque pin-encino”’ (see Fons z = 197 D m A. chamule nse Matuda ssp. chamulense Leaf blades with the tertiary veins clear- ly visible below; blades of medium thickness; occurring in “‘selva alta per- rennifolia 4 ÑN d mulense ssp. oaxacanum Croat o s ora ? du Cerro Sale, Biber (Dept. fent Bárbara). .. A. subcordatum ssp. m ` (Standl. & L. O. Wms.) Cro 43b. Spathe lanceolate (rarely narrow- ly ovate); berries bright red; west- ern Guatemala to El Salvador and Honduras along the Continental Divide. .. A. subcordatum Sli SSp. subcordat 34b. "vd blades with conspicuous, well developed posterior lobes. 4a. Blades about as broad as long. 45a. Leaf blades whitish-matte moe north central Oaxaca. ..... A. nelsonii Croat 45b. Leaf blades not whitish bene 46a. Tertiary veins iei raised even on fresh leaves; leaf blades usually les s than 40 cm long; Veracruz and Oaxaca. - ES s andicola Liebm. 46b. Tertiary terial inulous pe n drying; : leaf blades variable in size; soulheuciaeh Chiapas and Gua 47a. n po^ than 25 cm long, thin ES io rionegrense Matuda 47b. Blades more than 40 cm ong» subcoriacec BONS EE o o. i prac Standl. & Steyerm. 44b. Blades s substantially longer than broa ad. 48a. Lower blade surface matte, drying whitish; north central Oaxaca. .......... A. nelsonii Croat 48b. Lower blade surface ‘not ‘conspicuously ‘matte and whitish. 9a. Blade with the anterior lo p y southwestern Chiap A: cerrobaulense Y Matuda 49b. pede with the ae ‘lobe ‘usually convex x along the margin. Blades with the tertiary veins prominently raised even on fres E ves; Veracruz and Oaxa . A. andicola 42088 SOb. Blades with the tertiary veins s not ‘prominently raised on fresh le pr g. la. Blades ovate-triangular (ovate in A. subcordatum), the i d e collective vein always arising from the first basal vein, the second basal vein only rarely ng. gearing to the pale vein, usually merging with the margin well be- middle of the blade, the sinus usually broad, low t ien arabolic or arcuate; northern and southeastern Chiapas and northern a. 52a. Blades ovate-triangular 53a. Major veins of the blade and apex of the pet- iole reddish; Sierra de ne in northern Oa- aca. subovatum -orai 53b. Major ‘veins of the blade = Pas apex of th petiole green; northern and Sadana Chiapas. .. 52b. Blades ovate. 54a grab ‘ovate to narrowly ovate; berries or- A. "chamulense Matuda ssp. “chamulense 230 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 33b. Petioles terete or subter: ange (?; Honduras on Cerro Santa Bárbara SUPE: Santa Bárbara). ~ A. subcordatum ssp. ee (Standl. & L. O. Wms.) Croat 54b. Spathe lanceolate (rarely narrowly ovate); berries bright red; western Guatemala to El Salvador and Honduras along the Continen- tal Divide. ^ — — — —— .. A. subcordatum Schott ssp. subcordatum 51b. sey ovate to narrowly ovate or ovate-triangular; pos- r lobes usually well developed, often longer than broad ca in A. rzedowskii), often curved inward; bei: veins, the second basal vein often loop-conneced argin dulate; petioles not glaucous; frequently with a sec- ondary collective vein extending irregularly along the margin to the apex or nearly to the apex; Guer- rero, Oaxaca, and Chiapas. ........ AF r wskii Croat p . Blades subcoriaceous to medium thick, usually eiie rosea s (except A. titanium); ge ually glauc 562. Blades "broadly ovate to rounded in in | UU usually m more than t cm long. — 9 titanium Sandi & Steyerm. . Blades ovate-trian gia to Meis er the margins of the an r lobe m gd straight; spadix nail pie than m em 57a. Blades ovate-triangular; major veil t 5 an c when dry; spathe lanceolate, us T about as long as the spadix; southeas em Chiapas: — — — .— — A. cordatotriangulum N n Matuda s southeastern go? e Guatemala. * tanum Hem Pe E p usually ies bine one-third as long as longest petioles. ix dark p | Devens soo esi glossy; endemic to Sierra de Juarez, ra 400-2,800 Matuda AÙ 60b. Leaves subcoriaceous, ‘semiglossy: Lancetilla Valley, ca. “300 m x it white or lavender: leaves subcoriace n- ix green; cataph ylls dilacerating; peduni w than 10 cm long; © zia to Central Guatemala, 1,500-2,000 m or to northern Oaxaca, z 59b. Spadi daidi ,30 rropelonense A. lancetillense lense Croa 62a. Spathe broadly lanceolate, more than 2.5 times longer than broad, unded to acute at the base; spadix more than 5 times lo ar b broad, markedly tapered toward the apex; endemic p^ north yetlense Ma . Spitts broadly ovate, ess than 1.5 times es longer th than dal Me se; spadix less than 4 times longer than broad, sar tapered i the apex; endemic to central Guatemala.—— m 61b. Spadix white t io havesulil: 'cataphylls remaining intact; peduncles 21-39 69... long; Chiapas to Nicaragua. tien Cg hub m 1983] CROAT OF MEXICO AND MIDDLE AMERICA 231 58b. Peduncles more than half as long as longest petioles 63a. Uppe r blade surface papillate (with - epidermal cells cone-like) or alveolate; fresh leaves usually con spicuous sly ve af blades with veins co nsicuosl ans than the surface; epson lateral veins not conspic cuously more prominent when dry than the interprimary fects v veins; epidermal cells of leaf blade drying paras M coinie Plebis amae Matuda Leaf Ja des with veins conspicuously paler than the blade surface; prima ry aa ondary veins; epidermal cells obviously papillate with individual cells obviously raised, cone-like. 65a. Leaf blades almost as broad as long, the sinus closed or "C narro basal veins free or, if united into a posterior rib, the rib not m arin 65b. Leaf blades conspicuously longer than broad, the sinus obovate or hippocrepiform; basal veins d nto a conspicuous, naked (i.e. marginal to the sinus) posteri . A. leuconeurum Lem 63b. Upper blade Bis smooth, not enn cd (the ‘epidermal cells not raised and cone-like) or alveolate (with a raised, net-like reticulum); fresh leaves glossy or matte but not velvety; Mexico and Middle America. 66a. Posterior lobes longer than the anterior lobe. ........ A. berriozabalense Matuda 66b. Posterior lobes not longer than the anterior lobe. 67a. Cataphylls remaining intact, usually reddish-brown, vie acta leaf S pale reddish-violet, the stipe igi conspicuous, often curved, Guatemal a. Ss T A E . armeniense Croat 68b. Spadix at anthesis with tepals not conspicuously upturned, the spadix mi smooth, , rose-violet to lavender or ae, the stipe absent o rt, from Chiapas ioe western Nicaragua. „u... A. hui len wind 5 Maküda j idt dins usually weathering into aber or aioe leaf blades usually lacking raphide cells. 69a. oot inant nodding with spathe hooding spadix; endemic o Nicaragua on limestone outcrops. _A. beltianum Standl. & L. O. Wms. 69b. Inflorescences not “typically ‘nodding with spathe hooding spa- dix n - Ss Second pair of basal veins onto aon the mid- 70a. pedes occurring above 2,000 m elev la se = leaf margins below the middle of the T Sinus of leaf broad, shallow, arcuate; petioles coriaceous; principally northern Chia EE E E ETA aes EO cha mulense Matuda 72b. ape usually narrow, deep, hippocrep o parabolic; petioles Suc Siem ner inis eous with undulate margins; south- eastern Chiapas to Oui ala. montanum n Hemsl. 71b. Second. pair of basal veins joining leaf margin well above middle of leaf blade. 73a. Leaf blades broadly ovate, nearly a: as broad: as io Sierra i de Juárez in northern qoas 7 400- ) m. _ cerropelonense Matuda 73b. Leaf ovate- hte longer than broad; spadix green or green tinged with pu 74a. petioles glaucous, sinus usually spathu- late to hippocrepiform, rera parabol- ic; pm Oaxaca to Guatemala, ik montanum Hemsl. ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 74b. tpe: not glaucous; sinus usually ar- o parabolic, rarely spathulate; Dose northern Oaxaca and Chia- pas, 1,400-2,900 m. ... A. rzedowskii Croat 70b. Plants occurring below 2,000 m. 75a. Leaf blades not coriaceous. 76a. Leaf blades ovate-triangular, nearly as broad as long; rare, known only from an area near the border of Oaxaca and Chiapas. .......... , rionegrense Matuda 76b. Leaf blades ovate, longer than broad. — 77a. Spadix green at anthesis, soma i i ; ioles we n rero, Oaxaca, and Chiapas. ....— pne Croat Blades lacking a a secondary col- lective vein heirs irregularly along the margin to the apex or nearly to the apex; Guerrero, Oa- xaca, and Chiapas. 79a. Leaf blades with margins 7 oo S southwestern Guatemala, 1,200-2,900 m. ——— .A. montanum Hemsl. 79b. Leaf ‘blades with margins minutely undis to sub- crispate; ~ to south- ern Oaxaca and Chiapas, vr E ovandense Matuda Spadix dark purple at anthesis; petioles usually not glaucous. 80a. In Raises noticeably shorter ie the lea T - d 80b. 1 c - 3 28 9 oO wn [e] oO 3 e e wn 2 e er w w an E or longer than leaves; Mexico ^s wise ges mbrosum m Lit 81b. Tue usuall edly undulate; pes 1,20029 montana Hes 75b. Leaf blades coriaceous. d 82a. Coll ective vein arising from one of bes lo er ce uatemala, 700-1, 82b. Collective vein arising from one of the We basal veins, usually extending along only | | 1983] CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 233 of the margin of the anterior lobe, canta Paule iis to lower basal ve . Sinus relatively shallow, Mira cu i Rs — lobes claivaly short, ow; southeastern Chus io rien 1 300- 3,000 m. mulense Matuda repi- 83b. Sinus sje spathulate to iri candi form o rabolic, the posterior lobes often ihe sometimes longer than 84a. Posterior lobes (at least on well ue [9] = S835 289 $8 a -g Ld gp an C : cous; southeastern Chiapas to south central sig ism XS . montanum Hemsl. KEY E Leaves simple, not lobed; blades neither cordate nor subcordate, punctate on one or both surfaces. la. Spadix coiled or spiralle 2a. Spathe d. showy, red to red-orange, ovate; leaf blades narrowly ellip scherzerianum Schott 2b. aie not showy, reddish or green tinged purplish, linear or eps. ae leaf blades oblong or ovate. 3a. Leaf blades oblong, to 80(150) cm Vni inn E 3b. Leaf blades ovate, te 20.cm long. e m Spadix not coiled or spira 4a. wn c A. wendlingeri G. M. Barroso A. louisii Croat & Baker tipe 3-20 cm long; leaf blades oblanceolate; Specs pale violet-purple. .. _... A. hacumense Engl. 4b. Stipe less than 1 cm long (usually the spadix sessile). 5a. Submarginal collective vein markedly i pate above, raised below and significantly more prominent than the primary soins veins; berries bright red; leaves narrowly gnus to nar- rowly oblanceolate, the apex acuminate. . - Submarginal collective vein pi oy as prominent ast violet, white, yello owrorange or rarely red; leaves of various shapes. 6a. Plants Sada. scan with persisting, dilacerating cataphylls enclosing the internodes duo akeri Hook. f. the primary lateral veins; ipsuni ale not dilacerating, deciduous); plants with long (except A. tonduzii with cataphylls entire, lucent white or pale violet or purple; Mie stems, often hanging from m s; berries trans blades usually much less than 14 cm long 7a. Internodes 2-21 cm long, not oth hed in persistent cataphylls; leaf blades usually rounded at base; primary lateral veins usually three to six per side. . tonduzii Engl. . h 2 em lon E rarely more than 4 cm lon ng, at least ru internodes uon it ges : boyi leaf blades usually obtuse x ez younger to acute at base; primary jea veins usually 8a. Flowering spadix purplish, the pistil ninh. gonnranel quadrangular, emerging prominently before stamens emerge; spathe persiste ntly erect and more or less roadly ovate to obovate, inc acute to —— 0 to 800 m. ESM edu A. trinerve ues oo o E oO d i=] go * © Qa — Ei 3 =. p e l^ oO B3 an E: s oO oO z o B — o [7] 5 growing at seer pst 0 to ae 700 a i - ii at semis blade punctate or no 9 Te flow aeo 11 d mm oa a A ANM teal Sep. scandens - d. 2-10 mm long is “anthesis blade punctate o > Sete 3-1216) ere ng candens ssp. pusillum Sheffer 234 6b. DI ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 1 HS e 1 + occa 10a A. friedrichsthalid — various colors; leaf blades em much longer than 15 cm, T. onally slightly s "Leaf blades oblong strap-like; berries pale rie: broader in direction of axis than perpen ar to axis, sunken at a lla. Spathe to 5 mm wide, early deciduous; leaf blades mostly P cm wide; s ile. sessi A. fr circha: Sehor 11b. Poe 1-1.5 cm Vis usually persisting; leaf blades — 4-7. x short-stipitat A. utleyi Gol id p . Leaf ones not oblong; “eat 12a. Peduncle filiform, long iter on fragile, the spadix often pendent; spathe tinged reddish-violet; berries peach-orange; all vegetative parts brittle; 0- 800(1,400) m. A. acutangulum Engl. 12b. Peduncle erect, elongate or short; spathe of various colors, itle green or purple; berries of various colors; vegetative parts not brittle; 01,700 m. x eee 3-sided and alate; petiole — in cross- ction; Osa Pen- nsula. tipedunculatum Croat & Baker 13b. Peduncle terete or nearly so (except A. pct sharply angulate); iole not triangular in cross-section; primarily Atlantic watersh v Leaf blades narrowly elliptic, the apex acuminate; spadix s sho rt m), on a stipe 1-6 mm long. A. utle yi Co 14b. Leaf blades lanceolate, oblanceolate, elliptic, or obovate; spa very long or short (A. /ancifolium, A. paludosum); eis sessile or nearly so. 15a. Leaf blades lanceolate to ovate-lanceolate or oblong-lan- ceolate, the apex long-acuminate; spadix grayi ish-white to m aie de le. . Peduncle terete; spadix grayish- or greenish-white, with four or five flowers visible in each spiral from any : . . P rity. den angie: berries violet-purple at matu dri i E jm Schott from any angle; berries Me tinged with violet-pur- ple at apex. aussi ; Croii & Baker 15b. Leaf blades | ad = ate } ong: acuminate at apex; spadix color v 17a. Cataphylls thick, ifie not dilacerating; pogod one sixth to one third as ms re blades; blades gue oblong, coriaceous drying y light green or yo" low-green on both surfaces “fresh mentite oS with a ipee Y m 17b. Cataphylls thin, soon weathering and deciduous; b sedit M arious; blades mostly elliptic to oblan wly divinse usually brownish or grayish to meer st ally paler on lower surface when fresh, the midri vli d paler; berries orange to red; elevation 18a. Leaves usually with moderately short her (one sixth to one third the length of the idol Muf oblanceolate to narrowly obovate Ens ually tapered to apex; berries pale ee flat at apex; 0-1,400 m NOM gue aM monense E Engl. € ex K. Ki third 18b. Mies re "i longer petioles -— o full P ugs 3. 5 abnuit soa sienna or stout, sensn or not, usually less tha diam. in A. acuta ngulum but sometime to 10 mm dain. at base in A. durandit, berries orange to peach-orange. em nt inte Roe e cane iai d HE AES sage ie caesi Lors esci ae ERR a AERA 1983] CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 235 19a. Blades short-acuminate, the acumen al- cataphylls usually pe deciduous Or with ing e fibro work. ........ Wut Engl. 19b. Blades noticeably acuminate, the acumen y persist- ing asa dense fibrous network. . A. acutangulum Engl. KEY F Leaves simple, not lobed; blades neither cordate nor subcordate at base, epunctate. la. Leon situated at least 7 cm below the base of the blade; plants terrestrial. ... A. oerstedianum Schott lb. Geniculum immediately Pacis the leaf blade. 2a. yes scandent, the stem elongate, with long internodes, the leaves dispersed throughout much of the length of the 3a. Spathe "ies on aped peduncle (0.5)1—1.5 cm; berries yellow to orange or white. .......... ind A. pittieri Engl. 3b. Spathe essentially not decurrent; berries green, yellow, orange or bright red to red-violet. with two or more successive, ahis internodes alternating with greatly = gated internodes, 8-33 cm long; leaves drying brownish-green or sie berri bright red or purple-red; below 1 1.000 mo s . interruptum odio . Stem with internodes nearly uniform length, decreasing slightly d the apex of th stem; leaves drying dark brown to blackish or pale green ; berries green, red or orange; usually above 1,200 m elevation, rarely below 1 000 m 5a. capa less than 3 cm long; peduncles less than 4 cm long, enclosed E pun in an ongate leaf sheath. — ________. tenerum Engl. 5b. ud more than 3 cm long, peduncles more than 5 cm long. 6a. Blades drying black or blue-black, dark k bro wn on the upper surface, medium TO AR o wn on the lower sur when fresh; e; 1,500-2,600 ot as PN IE E e A. carnosum Croat & Baker 6b. Blades drying dark green to eir green above, pale yellow-green or occasionally 700-2, 300 Y 6'a. a more than 7 cm long at anthesis, slender and narrowly tapered to x; berries red, early emergent, oblong. .. .. A. testaceum Pala & Baker and p rsistent, Il apical internodes not X Mu KM LL s oe Schott r less ‘acaulescent with the ‘stem short or the stem 2b. Plants not scandent, either appearing more or es short 8a. Petioles terete or subterete, ies quadrangular, recta rectangular, or triangular, often narrowly or i h the margins sharp or blunt but merely rounded on lower margins, sometimes weakly d don in apical half of blade, never with a unii, near the base i 10a. Peduncle usually shorter peat S auly short and thick, never "lon dumm , usually sho j - bane usually more or less oblong y _ A. cubense we pered; peduncle less than twice as long as spadix; berries Pe iae mostly two or more times longer than — ole; Mew etna never spat D (except A. consobrin peduncle more than twice as long as s spadix; berries red, white, cream, cid: - le. par or yellow, not violet purple Sn yellow, orange, sometimes LI A 1591110 QV : pets BE rae ^ia white tipped with orange, y yellow o slope Ts "Mature spadix cylindroid, not tapered, to ca. 9 cm long and 1.5 cm 10b. c ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 (including pistils) diam. at anthesis, the pistils exserted well above the stamens and emerging weeks ahead of the stamens; berries white at base, reddish at apex; midrib rounded on lower leaf a near the base of l co the blade. be Schott 12b. Mature spadix oo tapered, 12-15 om long and ca. 8 mm at base at an nthes pp , obsct d by the stamer rted; berries white, cream, yellowish, or orange, iE a solid color throughou peint usually more or less trapezoidal with two sharp ribs on the deeem leaf surface near the base of the blade. A. fatoense K. Krause llb. Pistils not emerging early, rounded at Mu berries greenish-white or red, usually blunt at apex; roots 3 mm or more thick, ines or not but not all uniformly turned upward, not iar cootaly acute 13a. Spathe ovate, pale green, erect, and enshrining xe Mats spadix pale green, moderately stubby, usually less than 6.5 cm long; plants usually terrestrial or on soil deposits on rocks; blades usually elliptic; roots i 1 cm diam. when fresh; eastern Guatemala -— western E on the Pacific slope. salvadorense Croat 13b. paei lanceolate (except A. halmoorei), usually green p violet-purple reflexed or spreading; spadix green to pale lavender or purplish, usually rather slender and long-tapered, usually more than 6.5 cm long; plants usually epiphytic or epipetric (except A. nizandense); blades usually oblanceolate-elliptic to oblanceolate (except sometimes elliptic on A. Petite eng roots mostly less than 6 mm diam. when fresh; Mexico to Colom 14a. Leaf blades obovate-elliptic, 1.3-2.5 times longer than petioles, ap — ie e s longer than broad; Mexico in soppaa rn Oaxaca andense Matuda 14b. Tsat sich PEE oblanceolate or EFSER 3 or more aean s longer than petioles; usually 3 or more times longer than 15a. Spat ovate to broadly ovate; berries pale greenish-yellow: stern Mexico in Nayarit, Jalisco, Colima, and Michoa i: Ai halmoorei Croat 15b. Spathe lanceolate to linear-lanceolate; berries s bright Mexico from Guerrero to Chiapas and Veracruz and t0 Colom ‘bia. 16a. Cataphylls cucullate (hooded), fist- shaped before opening; inflorescence frequently pendent at an e , spadix usually long-tapered, pale lavender, we A glaucous; spathe long and narrowly tapered to wir spreading, thin, almost as long as the spadix; Pact slope except from Nicaragua to Panama. -— €: m a "gis A. salviniae © niae Hem 16b. Cat lls 1 A : dix short- -tapered, green to violet-purple, no rÍ cous; spathe moderately short, reflexed, coriaceo u sually much shorter sni spadi cific Plants occ urring on rocky, dry slopes; i slope of Mexico in castes and sou sac, DEEE E OT Me 2) Croat uus sansa aneis ssp. jimenezii (Matuda) 17b. Med occu epiphytes or on TOC er usually moist nus — slope, princi from Verac nduras. ..—— ndalii . A. schlechtendalii Kunth ssp. schlec ger 9b. Major lateral veins mostly j joining into a ective vein, this icesases habit. 18a. Spadix with no more than ies to bee from any an f vie ME ris heres ow the middle, tapered to a long cuneate base; ete: red-violet; spadix purple-brown; berries red; roots usuall n —— 4 ng Tc strai ight o or conspicuously loop-conn : e s arising usually in the lower half of the blade; plants with or withou five flowers exposed across its diameter 1? any 19b. Leaf blades broadest. atc Or r below the middle, acute to — p e | 1983] CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 237 green; cones green, yellow or orange, sometimes tinged purple; berry color variou 20a. Saithe decurrent onto peduncle (0.5)1-1.5 cm; ig early sain nt many times longer than stamens at anthesis; berri ee wo 13 SSp. pittieri 20b. Spathe not rero dd onto peduncle; pistils Reus longer than stamens at anthesis; berries D en or red (not known in A. praa var. fogdenii but ea oran 21a. Spathe to 7 cm long; dh 4-12 cm "n jede: IE 1 testaceum Croat & Baker 21b. editi less than 3.5 cm long; spadix less ae 4.5 cm long; berries sh (A. pallens) or probably orange (A. pittieri var. fogdenii). riti reat blades less than 10 cm long, thick, veins sha and etched above, obscure below; ipod lateral veins t four per side. Pe var. Jotdeni SR Leaf blades more than 15 cm long, sid tely thin, tertiary vs Pierro below (dry), primary later p veins five A. pallens Schott 2 N S si 18b. Spadix with more than ie ‘sonar exposed across its diameter in each spiral from any angle of v 23a. Petioles mostly less than 10 cm long; blades usually six or more times longer than pe 24a. Collie vein dubiis from one of the primary lateral veins, well above base; berries white at base, reddish or purplish at yg Vosa early adipic pent longer than sai stamens. . A Sophi r . Collective vein arising from the base of the blade; ‘berries ied (not for A. machetioides); pistils not usd held at about the same level as stamens. 25a. Spadix stubby, creamy-white; spathe lanceolate-ovate, clasping at the base; —€— s thick, persisting intact; Costa Rica on the pe.. Pacific slo A coset (is 25b. © 1 » | Al or robtuse at the base; nios thin, inr) into fibers ARE A. mach ine Matuda N AR o axa i his usualy " cm or longer; blades no more than four times longer than prem Species of Mexico and Middle America, ranging as far south as northern eng ras; petioles broadly sulcate adaxially, the margins obtuse or 272. Leaf blades lacking any obvious — veins, the lowermost pri- mary lateral veins more or less e ual those higher up on the blade. 28a. Tertiary and higher order veins d elevated (at least when dry). ... .. A. retiferum Standl. & Steyerm. 28b. Tertiary and higher order v veins obscure, much less promi- nent than the primary lateral veins. .. A. vato ii age Leaf blades with — well developed basa ] veins which more promine ent t an the primary lateral veins or at least Puis to form a posterio ve rib 29a. Collective vein arising from f eins; primary ry late eral \ veins numerous or esas a so, siae more prominent than t the interprimary or tertiary veins; blade often a broadened at the base. . So COMES A. seleri Engl. 29b. Collective 1 vein "arising "from ‘the ‘upperm e vein or even from one of the primary lateral eae primary lateral veins usually ee ae | more prominent than the in- ry or tertiary veins broadened at the base (sometimes so in A. nakamurae). 30a. post broadest near or even above the middle, six or re times meer’ than broad; wap gene Chia nakamurae teda 30b. Blades brosdest | near rr the base, lei ca. three times longer than broad. 2 J d 238 ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 3la. Major veins of lower blade surface a (at least on younger leaves); northern Oax 31b. Major veins of lower blade surface gree temala to Honduras. . subc nz Schott 26b. Species of Costa sud and Panama or possibly a as is far Bes as Nicaragua; petiole 2a. es sulcate mundis est A veins obscure (prominulous or: hi but scarcely prominent than the tertiary venation); the midrib reddish re in ipie: half of blade (less noticeable after drying aap not 3 N io” Eg e < 13 M a = [o] z =] dilacerating, soon deciduous. ib & Baker jor | nt, much mo iis s than the ertiary veins; midrib not reddish; ipiis usually dilaceratin le tim at least at base, usually persisting for a considera 33a. Cataphylls less than 15 em long : and less than l cm wide a the middle, mostly ca. l cm diam., elongate, with ~ white to tan s usually erect at anthesis, ‘arching i in fruit; spadix usually ue than 6 mm diam. at base, berries pale green. „u angustispadix Croat & Baker 33b ~~ more than 15 cm — or more .5cm at the ften weathering into a net-like structure 8b. Petioles Aaa rectan lower (abaxial) si mo roots among the existing leaves; inflorescences usually spreading to pendent at anthesis, pendent in ore (except A. cuspidatum), spadix mostly more than 7 m iam. at base (except 4. protensum and A. cuspidatum); pr not pale green. 34a. Leaf blades ovate to oblong-ovate, mostly more than 20 cm wide, truncate to subcordate at base, the basal veins united into a short trunk vein. ....—— — uU EN o Soest — spin to oblong-elliptic, mostly! ne - 20 cm acute to attenua Mild a poste ib. Re 35a. Spathe si sii less than 8.5 cm long; spa ovate, cordate at base, ca. half as wide as long. . brenesii Croat & e Baker 35b. path and spadix more than 10cm ke pene lanceolate to linear, not cordate, much more an qure ce as long as wide. sty 36a. pma much paler and whitish benea wih spadix yellow-green; rosulate — - blades to 5 erect -spreading leaves blad l G des nud 36b. Leaves more or less oco M purplish or purplish tinged; plant an des leaves eed ndent; blades often m 37b. Spadix less han 2) HN mes longer thal wide at anthesis, to 20 cm long mel fruiting spadix usually less than cm long; apex of flowers less inf 1.5 mm wide; plants occur! 7, Schot a A. prote above 9 gular, decidedly triangular or prominently 1- or more rib subo parum ı Matuda | | | 1983] CROAT OF MEXICO AND MIDDLE AMERICA 239 38a. Major lateral veins mostly free to the margin, sometimes loop-connected i in apical half d oye never with a continuous, Wee straight collective vein arising from near the uall o ihe Deals berries pale olii: pee r red. 39a. Stem usually 1-1.5 cm diam., comparativ ely elongate, at least the lower internodes not completely obscured by the roots; inflorescences usually ar ion ati ding; infructescence pendent, the pert hes ish to pin ite or pale reddish to pale purple. 40a. sas ree thickly CAE berries cream to pale orange or ole: roots any, c e or less contiguous, tener. slender, Me pet- ioles Boe less iia 10 cm long. A. fatoense K. Krause 40b. Leaf blades subcoriaceous; destin red; roots Pre dense, thick. blunt, not ascending; petioles mostly more than 10 cm ipalaense Croat & Baker 39b. Stem usually more than 2.5 cm diam. (at least on mature in dividuals), the inter- nodes mostly completely obscured by the roots, the roots moderately stout, usually mo re than 3 mm diam., not all elongate tapered and ia directed upw ards; + let purple or green or green tinged purplish. 4la. Species occurring from ism to El n frequently on rocks; berries red or pale greenish-yellow or greenish-w 42a. Leaf blades boomen less hay dos times longer than broad, .3-2.5 times longer than petioles; southern Oaxaca and Guerrero. . A. n izandense Matuda 42b. I f blad lat three or usd times longer than broad; blades three o or more times longer than petioles 43a. Spate green, o SD ie vier pa enshrining the spadix; berries orange-red; tern Guatemala and western El vdd on the Pacific i auc Croat 44b. Spathe reflexed; berries pale greenish-yellow; petiole tusely 3- "ee western Mexico in Nayarit, Jalisco, Co lim and Michoa A. ha fono Croat 43b. Spathe green or ibd; T, lanceolate 45a. dorur peta the Pacific slope in Guerrero and Oaxac pes. . ocky A. schlechtendalii ssp. jimenezii ceni Croat 45b. Plans septo from vé sep slope; epiphytes or on ocks in Aii onde moist a hot dali Kunth ssp. schlechtendalii 41b. pakar of Costa Rica ‘and Panama in as far north as Nica . Peduncles usually shorter than petioles, rarely > i 5 times longer than d on lower side (abaxial), the sides unribbed; spathe sine spadix often short, 5-22 cm long; pistils emerging early; berries red, acute at apex. ...... MuR chai UENIRE uH UE ee d E purpureospathum Croat A. 46b. choirs eM. E 5-3 times auus than petioles; puso ubquad- a A area K. Krause 38b. Major lateral veins = mostly, joining "into : as “strai right c or r conspicuously loop-connected colle ve vein, this usually arising in the lower half of the blade; plants with or without he rosulate habit; berries green, red, violet-purple or orange 4 as broad as or broader toward oe s at middle; berries greenish to white 48a. Spadix usually clavate or subclavate; cataphylls soon seiten to inr tudinal, pers istent of the Atlantic slope from 0-650 m : 49a. Blades with 20—30 pairs of primary lateral veins, these e prominent sunken; spathe lanceolate to y lance “i a, Se about twice as long ly triangular, as the spadix; petiole common spt ng p iier sm oe 2 CM 49b. Blades with (5)10-15 pairs of primary incl se these weal wea eakly s nk- as the spadix; peace slc conne E cuneatissimum TA ) Croat : ie roid; spath Jess vola lanceolate-ovate; cata- p 48b. Spadi y Cy 240 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 phylls persistent, turning brown but not dilacerating; berries dn. hes the Pacific slope in tropical wet forest from 0-50 m ium Engl. 47b. Spadix usually conspicuously tapered toward the apex; spathe mis om to anceolate or oblong-lanceolate; berries greenish, violet-purple, orange or red. 50. Petioles triangular, acutely 1-ribbed on lower (abaxial side; blades markedly bicolorous, the lower surface e creamy-green, drying black to olive-brown, lower surface with iie major veins conspic cuously darker than the surface, —€— or not at all raised, the tertiary veins obscure; berries bright violet- A. michelii Guillaumin 50b. Poroka quadrangular r, usually 3-ribbed on lower (abaxial) na blades not markedly bicolorous but eae somewhat paler beneath, usually drying green, lower surface with the major lateral veins usually cine lighter than the lower surface, abe oi spicuously raised (scarcely raised in A. eiie ndi Vari with the tertiary veins prominently visible); berries greenish, orange o 51a. PN terrestrial; blades attenuate at base (broadly concave between lower third of blade and geniculum); berries poema acutifolium m Engl. 51b. Plants epiphytic; blades acute to truncate or rounded i base; berri r re 52a. Major lateral veins scarcely raised; blades acute at base, the tertiary veins clearly visible but not at all raised (except after drying), about as prominent as the collective vein and the dis parts of the major lateral veins; cataphylls turning brown, re- maining i es bright ri ng intact, not dilacerating; berrie "T c pe 5 Ñ e Mus anthers m lo nthe tepals, doe filaments usually not | . seibertii Croat ak Baker 53b. Ped ll than 20 cm wide: dbl ive xh more than 20 cm fon anthesis; stamens rominently exserted above kx tepal lon the filaments exposed; anthers ca. 0.9 ER ped "abile To Schott Anthurium Schott, Wiener Z. Kunst 3: 828. 182 the iced sheathed near the base, geniculate at ae very Conspicuous; cataphylls usually — lanceo- late (rarely cucullate), usually green, soon drying and then variously colored, with a medial rib cross-sectional shape variable, fr Mei terete and narrowly sulcate, sometim wl p to several ribs, the sulcus various, with M us 10 gins sharp or blunt; blades usually coni and subcoriaceous, very variable in form, sinp tely usually ovate, elliptic or lanceolate oF digita lobed, net-veined, the primary lateral veins vein. or basal veins often forming a collec e ing à the basal veins often united at base for IN- posterior rib on leaves with posterior sori or angular, usually elongate; Me zi usually not convolute at base, persistent, y free well before anthesis of the flowers: us 1983] spreading or strongly reflexed from the spadix, i sometimes hooding spadix, times cylindroid, clavate, or su or stipitate, many flowered, variously colored, elongated in fruit, usually flowering from the base upward in a slow or rapid progression, some- times beginning from the middle and developing rapidly throughout the spadix; flowers perfect, usually protogynous, closely aggregated in spi- rals, truncate and usually rhombic in outline at apex, sometimes the outline sub-4-lobed (i.e., margins ofapex) straight or gradually or jaggedly sigmoid; principal spirals with flowers aligned, their lateral margins forming a straight row; al- ternate spirals (opposite direction) with flowers aligned, their lateral margins not forming a straight row; tepals flattened throughout most of their length, broader and truncate at apex (some- what accrescent in fruit), usually triangular, the inner margin usually rounded, held above the IS of stamens emerging first one at a time, fol- *d by the anterior and posterior ones, the s flattened, somewhat rounded abaxially th : wi broadly ovate to oblong-ovate, opening il à longitudinal slit. INFRUCTESCENCE usu- Y pendent, rarely erect; berries ovoid, oblong- CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 241 ovoid, oblong, or obovoid, less frequently de- pressed-globose, usually acute to rounded or truncate at apex, sometimes with a depression at apex, exserted and dangling on 4 threads at ma- turity, succulent and juicy, variously colored, 2- celled, usually 2-seeded, sometimes 4- or more seede DISCUSSION OF MORPHOLOGICAL CHARACTERS PRESENTED OR OMITTED FROM SPECIES DESCRIPTIONS sce Aatailed n inti d p fthe species presented here and at the same time to provide relatively short descriptions, a number of assumed character states will be listed for the genus. While these might be apparent upon re- reading the generic description, this list should more readily elucidate assumed characters. Stems: Usually assumed short unless a specific length is given. Internodes: As above. Cataphylls: Assumed 1-ribbed. Petiole: Color mentioned only if other than green. Geniculum: Shape and color given only if dif- ferent from petiole. Leaf Color: Mentioned only if other than green or if lower surface is substantially paler than upper surface. Leaf Texture: Usually mentioned only if other than semiglossy. Venation: Basal veins and midrib assumed raised unless otherwise indicated. Inflorescence: All colors given are assumed to 1 i 1 i 1 +} i tatar ^4 assumed to be cylindroid-tapered unless otherwise stated. Flower size: “Length” is actually the width of the apical, thickened part of the flower in the direction of the axis; ^width" is the width of the apical part of the flower, perpendic- ular to the axis (only the visible portion is measured); “sides of the flower" refers to the line formed by the 4 outer margins of the apex of the flower (see Croat & Bunting, 1979). Filaments: Length and width is the measure- ment of the area exposed. 242 See Croat and Bunting (1979) and Croat (1980) for details of flowering behavior to understand the necessarily abbreviated discussion of stam- inal behavior in Anthurium. The discussion of flowering behavior under the section of this pa- per entitled “Taxonomic Characters" (Inflores- cences) gives information essential to under- standing the descriptions of this behavior in the following treatment of species. Anthurium acutangulum Engl., Bot. Jahrb. Syst. 25: 371. 1898. Type: Costa Rica. San José: near San José, 1,200 m., Tonduz 10360 (B, hololectotype; CR, isolectotype; designated Croat & Baker, 1979). Anthurium porschianum K. Krause, Ann. Naturhist. Mus. Wie dontis 611 (W, holotype; B, isotype). Epiphyte or rarely terrestrial; roots numerous; cataphylls subcoriaceous, 2.5-5(11) cm long, acuminate (the acumen apiculate), drying tan to light brown, persisting as linear fibers. LEAVES erect to spreading; petioles narrowly sulcate, rounded to sometimes 1-ribbed abaxially, 3-22 cm long, 4-7 mm diam.; geniculum 1-1.5 cm long; blades elliptic to narrowly obovate, sub- coriaceous, short-acuminate at apex, acute to ob- tuse at base, 15-34 cm long, 5-14 cm wide, broadest at middle or slightly above, both sur- faces semiglossy, the lower surface punctate; midrib convexly raised above and below; pri- mary lateral veins 8-10 per side, departing mid- rib at 55°-60° angle, + straight to collective vein, loop-connected from base; collective vein rising from base, sunken on upper surface, 3-8 mm from margin. INFLORESCENCE spreading to pendent, equal to or longer than leaves; peduncle 12-61 cm long, 4-4.5 mm diam., terete or strongly ribbed abaxially, sometimes 3-ribbed adaxially, rarely 4-6 ribbed, the ribs irregularly spaced, tinged with red-violet near base, much longer than petioles; spathe green, tinged with red-violet (B & K Yellow-green 8/ 10), oblanceo- late, 7-11.5 cm long, 1.2-1.5 cm wide, broadest at base, inserted at 40? angle on peduncle, some- times slightly twisted; stipe 6-7 mm long in front, 2-3 mm long in back; spadix yellow-green (B & K Yellow-green 6/7.5), 8.5-21 cm long, 4-6 mm diam. at base, 3-4 mm diam. at apex; flowers rhombic, 4.2-6 mm long, 2.5-3.5 mm wide, the sides + straight; ca. 5 flowers visible in the prin- cipal spiral and ca. 4-6 flowers visible in the ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 alternate spiral; tepals glossy, weakly punctate, papillate; lateral tepals 3—3.3 mm wide, theinner ~ margins straight and flat; pistils emergent, green; | stigma brushlike with droplets, dry and brown | as stamens emerge; stamens emerging rapidly from base or middle, the first lateral stamens emerging at apex as third and fourth emerge at base, exserted on transparent filaments, 0.3-0.5 mm long, 0.7-1 mm wide, retracting and holding anthers at sides of pistil, not contiguous; anthers creamy white, 0.5 mm long, 0.7-0.8 mm wide; thecae ellipsoid, divaricate; pollen white, abun- dant. INFRUCTESCENCE pendent; berries pale orange, + obovoid, ca. 5 mm long. Figs. 2 and 3. The species is known from Honduras (la Mos- quitia) to Panama from sea level to 1,400 m. lt is found only on the Atlantic slope in Honduras and Nicaragua, but on both slopes in Costa Rica and Panama. In Panama it occurs on the Pacific slope near the Continental Divide. Anthurium acutangulum is known from wetter parts of trop- ical moist, premontane wet, and tropical wet for- est. Anthurium acutangulum is recognized by its long-petiolate, more or less elliptic, pandit punctate leaves, which are abruptly ac E at apex; its slender, pendulous inflorescence: ! : peduncle and spathe that are usually tinged violet, and its pale orange berries. j is species is a typical member of — Porphyrochitonium despite its erroneous p ment in section Leptanthurium by Engler (1 iA The species is most easily confused with sm specimens of A. ramonense but that gp a much stouter peduncle and the inflori either held semierect or, if it is pendent, 1015 7 dition, the pistils of A. ramonense are a raised, the exposed portion being m gulum squared and violet-purple. In A. acuta" | the pistils are green, appear somewhat oy pals, and are promptly raised well above the giving the spadix a weakly knobby ap Costa RICA. ALAJUELA: between Cañas an N of Bijagua, Croat 3648 1 (MO); Santa Croat 4 3581 Park, Liesner 5130 (MO), NW of Zarcero, 7 (BR). (MO). carTAGo: valley of Río Tuis, Tonduz ò+ HONDURAS. GRACIAS A DIOS: near Rio ell & Cruz 4121 (MO) 1983] CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 243 N? 2385492 A. oo , ! he Ee | E jui E a EN eu ean. Ae S. FIGURES 2-5. 2. Anthurium acutangulum Engl., Croat 43340. Anthurium acutangulum En — oe ^4, thurin s acutifolium Engl., Pittier 6539 (Lectotype). gna T ersed alatipedunculatum Croat 4 Iker, Utley 1030 (Ty ype). f R De 244 NICARAGUA. RÍO SAN JUAN: near Cano Chontaleno, NE of El Castillo, She 3422 (MO). ZELAYA: Cano acis mete 8834 (MO); Stevens et al. 8862 dem i a America, N of El Empalme, Stevens 8335, 8412 (MO): E recen Rosita and Puerto, SW of Río Kukalaya, Stevens 8506, 8645 (MO); W of Rosita Empalme, Neill 4111 (MO); Cano at Salto La Oropéndola, Stevens 8947 (MO); along road from Siuna to Matagalpa, Stevens 7523 (MO); near Tala Has and Puente Mango, Stevens 7608 (MO); N of Limbaika, near Río Prinzapolka, Stevens 8253 MO). Anthurium acutifolium Engl., Bot. Jahrb. Syst. : 365. 1898. Type: Costa Rica. Puntar- enas: along the Río Hakum, Buenos Aires (SE of San Isidro del General), elev. 250 m, Pittier 6539 (B, hololectotype; BR, CR, iso- lectotypes; designated Croat & Baker, 1979). Anthurium scopulicola Standl. & L. O. Wms., Ceiba 3: 5 pou Rica. Puntarenas: vicinity of Palmar Norte, elev. near sea level, P. H. Allen 5788 (EAP, bolak F, isotype). Terrestrial or rarely epiphytic; roots numer- ous, thick, velutinous; cataphylls persisting as fibrous newe subcoriacnn, 6-10 cm long, apicula te) , drying dark tan (B & K Yellow 3/5). LEAVES erect to oe pd (2)6-22 cm long, (3)6-9 mm jam., ened to broadly sulcate adaxially, sharply 3- eec abaxially; blades epunctate, ob- lanceolate to broadly oblanceolate, gradually acuminate at apex, long attenuate at base, (1 1)25- 62 cm long, (3.5)5.5-27 cm wide, broadest at middle; midrib raised above and below, paler than surface, sometimes yellow; primary lateral veins 8-12 per side, departing the midrib at 409— 45* angle; collective vein arising from the third to fifth primary lateral vein, 3-5 mm from the margin. INFLORESCENCE erect, usually short- er than leaves; peduncle 24-53 cm long, 3-5 mm diam., much longer than petioles; spathe green, DE E 5-12 cm long, 0.7-1.1 cm wide; spadix sessile, green to white, sometimes tinged Ns red-violet, 7-16 cm long, 6-8 mm diam. at se, 3-4 mm diam. at apex; flowers square to hone ca. 2 mm in both directions, the sides t straight; 5-6 flowers visible in the principal spiral, 5-7 flowers visible in the alternate spiral; ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo. 70 lateral tepals 0.5 mm wide, the inner margins turned up. INFRUCTESCENCE with greenish- yellow, obovoid berries often not developing in apical one quarter to one half of spadix. Fig. 4. The species is known from Costa Rica and Panama in tropical moist, premontane wet, and tropical wet forest life zones at elevations of sea. level to 900 m. The species has been collected once on the Atlantic slope and lowlands in San José and Puntarenas Provinces of Costa Rica, including the Osa Peninsula. It has been collected in Panama only on the Burica Peninsula in Chi- riquí Province. Anthurium acutifolium is a member of section Pachyneurium and is distinguished by being terrestrial plant (rarely epiphytic) with thin, ob- lanceolate to broadly oblanceolate leaf blades that are attenuate at the base. The petiole : broadly sulcate to flattened on the upper side an sharply 3-ribbed on the lower side. The syntype Pittier 4099 at MO is unusually small s- the species but appears to differ in no other i Anthurium acutifolium is similar ue to A. consobrinum but that species is an e (apparently restricted to the Atlantic É early emergent, narrowly acute pistils, an ob ie spadix, and a petiole that is rounded on thelo Es side. Pistils of A. acutifolium are not early e gent and are rounded at the apex. 560 SA RICA. ALAJUELA: El Rodeo, Hunnewel IS ema GUANACASTE: Cabecera, Pittier 11 1 s: Boruca, Pittier 4656 (BR); of Pue ipo Burica Peninsula, S of ui F uentes 6210 (UMO); basin of El Gen (MO); Pacaca, doppie i near Santa Ar er & Liesner 7171 (MO); NW of Santa 17387 (NY, US). & Bake — Anthurium alatipedunculatum Croat Brenesia 16 des 1): 19. 1979. TYPE: E Rica. Puntarenas: Rincón de Osa 1974. ofair strip, 20—60 m elevation, 20 J Me Utley & Utley 1030 (F, holotype ^ type) Epiphyte; stems 1-1.5 cm diam erm sn decks 2-3 mm diam.; canaphyll ately thin, 2-3 cm long, acuminate " Rh istent rown, weathering into fibers, 1983] CROAT LEAVES erect; petioles 4-26 cm long, 5-6 mm wide, p sulcate, drying with a narrow, membranous wing; geniculum 0.5-1.2 cm long; cuim lanceolate-elliptic, acuminate at apex, obtuse to rounded at base, 15—23.5 cm long, 6- 10 cm wide, broadest below the middle, obscure- ly and sparsely punctate above, densely punctate below, the punctations reddish-brown, the mar- gin drying conspicuously undulate; lower surface paler than upper; primary lateral veins 10—20 per side, departing midrib at 55? angle, almost straight to collective vein; interprimary veins numerous; collective vein arising from the base, 3-8 mm from margin. INFLORESCENCE shorter than leaves; peduncle 12-19 cm long, ca. 3 mm wide, winged, 0.5 to 5 times longer than petioles; spathe green suffused with purple-red at margin, lan- ceolate-linear, 3-4.5 cm long, 6-8 mm wide, broadest at base, short-acuminate at apex (the acumen apiculate), inserted at ca. 30° angle on peduncle; spadix sessile, green, 10-12 cm long, -5 mm diam. at t base, 2-3 mm diam. at apex; flowers rhombic, 2.3-2.8 mm long (dry), 1.8-2.3 mm wide (dry), the sides + straight to weakly sigmoid; 4—5 flowers visible in the principal spi- ral, 3-5 flowers visible in the alternate spiral; g, 0.9 mm wide (dry); thecae Bin d un INFRUCT ESCENCE not en The species is known only from the Osa Pen- i nsula in tropical wet forest at less than 100 m elevati tion. Anthurium alatipedunculatum is most closely allied to 4. acuta his angulate peduncle and a triangular pet- ole dien Is sharply sulcate adaxially. The species member of section Porphyrochitonium. is ^ Rica. PUNTARENAs: Rin e Osa, region bags Strip, Utley & Utley 1 030 ( (F, cen MO, : olcán Orizaba, 2,500 m. Liebmann (K, ololectotype iere. designated). A RAM cucullatum C. Koch, Index Sem. Hort ol. App. 6. 1853. Anthurium andicola var. "Cue OF MEXICO AND MIDDLE AMERICA 245 cullatum (C. Koch) Engl., Monogr. Phan. 2: 169, N. 97. 1879. Type: Origin unknown (cultivated). Anthurium spicata Matuda, Anales Inst. Biol. om 216, fig. 55, 1954. TYPE: enango, d ,300-1,700 m, posco S.T n. (MEXU, holotype). Anthurium oaxacamonticolum uu Anales Inst. Nac. Dm 27: 344. 1957. TYPE: Mexico. Oaxaca: vic. La Gloria, near the Oaxaca- Chiapas border, Maebousall s.n. March 5, 1953 . (MEXU, ssa pe). Mex rte near Ayutla, is 2362 (NY). Anthurium chochotlenis Pob: Cact. Suc. Mex. 20: Mexico. Oaxaca: Chochotla (probably a Me ‘Chilchotla at 18°14'N; 9'W), Huatla de Jiménez, Matuda 38620 Ads holotype). Epiphytic or on rocks, sometimes in soil on rocky cliffs; stems less than 20 cm long; leaf scars ca. 1.7 cm wide; roots thick, descending; cata- phylls subcoriaceous, 2—4.5 cm long, tinged red at margins and apex, the apex round, minutely apiculate, drying medium brown (B & K Yellow 5/5), weathering to fibrous network at base. LEAVES with petioles erect to spreading, 18—47 cm long, 4-8 mm diam., sharply sulcate, weakly glaucous; geniculum 1.5-2 cm long, faintly tinged red; blades broadly hie TO thick, narrow- ly acuminate at apex (rarely rounded), usually prominently lobed at base are merely obtuse), 20—37.5 cm long, 13-27 cm e, broadest at point of petiole attachment; sieur lobe 19-2 cm long, the margins broadly rounded; posterior lobes 7.5-13 cm long; sinus arcuate with de- current petiole to parabolic to sometimes hippocrepiform; both surfaces matte to semi- glossy, the lower surface weakly glaucous; mid- rib convexly raised above, diminished and flat at apex, more acutely raised below; basal veins 3- 6 pairs, the first free to base, those remaining coalesced 1-3 cm, posterior rib naked, ned curved, the outer margin scarcely turned u mary lateral veins 3-5 per side, departing midrib at 55° angle, + straight to collective vein, sunken above, raised below, interprimary veins sunken above, raised below, tertiary vein prominulous above, below (drying p inent on both surfaces); collective vein arising from the first basal vein or from one of the lowermost primary lateral veins, sunken above, raised below, 4-8 m from margin, sometimes the second basal vein running almost to apex. INFLORES- CENCE erect, shorter than or equal to leaves; peduncle 16-40 cm long, 4-6 mm diam., terete, sometimes with a single ridge tinged reddish; 1 SCUIC 246 spathe thin, medium green (B & K Yellow-green 5/5), narrowly ovate to ovate-lanceolate, 5.5-7.5 cm long, 2-2.8 cm wide, narrowly acuminate at apex, rounded to subcordate at base, inserted at 45° angle on peduncle; spadix green heavily tinged purple or dark purple (B & K Blue-purple 2/10), 4-10.5 cm long, 5-9 mm diam. at base, 2-5 mm diam. at apex; flowers square 3-3.6 mm in both directions, the sides straight parallel to spirals, weakly to jaggedly sigmoid perpendicular to spi- rals; 5-6 flowers visible in the principal spiral, 6-7 flowers visible in the alternate spiral; tepals matte, weakly punctate, very minutely papillate, the lateral tepals 2.1-2.2 mm wide, the inner margin broadly rounded; pistils weakly emer- gent, green to purplish (paler than tepals); stigma linear, ca. 7 mm long, with large, clear droplets ca. 1 month before stamens emerge; lateral sta- mens emerging from the base, followed by al- ternates in rapid progression, leading stamens preceding the next in the series by 2 or 3 spirals; anthers white to pale yellow, ca. 0.7 mm long, 0.8 mm wide, held over pistil then retracting and held at sides of pistil; thecae ellipsoid, somewhat divaricate; pollen yellow, fading to white. IN- FRUCTESCENCE pendent, spathe deciduous; berries orange, obovoid, rounded at apex, 1.5cm long, 8 mm diam., 6 mm thick; mesocarp fleshy, orange; seeds 2, greenish orange, flattened, ca. 9 mm long, 5 mm wide. Figs. 6, 7 and 10. Anthurium andicola is endemic to Mexico, at 1,000 to 2,500 m elevation, occurring in both the Sierra Madre Oriental where it ranges from central Veracruz to northern Oaxaca and in the Sierra Madre del Sur south of Oaxaca City. One collection by MacDougall supposedly from La Gloria on the Chiapas-Oaxaca border is worthy PP OEE ieee dela Vil. Oaxaca and Chiapas is near the Pacific Coast at an elevation of near sea level. This locality is out of place for A. andicola both geographically (being ca. 270 km east of the nearest other collecting locality) and altitudinally (all other collections rom above 1,000 m). Yet the specimen differs little from collections made in the Sierra Madre del Sur (especially those from Pochutla). A col- lection from La Gloria (MacDougall s.n., March 5, 1953) was designated the type of A. oaxaca- monticolum. A MacDougall s.n. collection from La Gloria was studied that bears the date “Mar./ 53” (MEXU 110308) but it is uncertain whether ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 this is the type because the type description de- ) scribes a plant with larger leaves and inflores- cences. Anthurium andicola is recognized by its usu- | ally ovate, reniform or ovate-triangular leaves | with prominently raised tertiary venation (es- | pecially on drying), and D-shaped or broadly sul- , cate petiole with sharp margins and green to | violet-purple spadix. Both the petiole and the | lower leaf blade surface are often covered with | a thin waxy bloom. The species is among the most variable in Mexico and even within a single | population variation in leaf shape can be great, | varying from conspicuously lobed at the base t0 not at all lobed. Plants from central Veracruz n the type locality are uniformly ovate to broadly ovate with rounded lobes and an arcuate to pat | abolic or horse-shoe shaped sinus. Further south | at the type locality of A. macdougallii the pos- | terior lobes are often directed inward with ; sinus almost reniform. The same variation 0€ curs in the Sierra Madre del Sur but leaves there are often more diminutive with less pronounced posterior lobes and more frequently an — : | sinus. Sometimes the anterior lobe 1s Fo ably reduced and almost rounded at the such as in the case of Reko 6064 (Oaxaca) and MacDougall s.n. from La - (Chiapas), which is the type of 4. -— | ticolum. It has been placed in section Be | chium. í The species is not easily confused wil a other although some specimens, espec! f d de | with poorly developed posterior lobes It southern cordillera, are similar to chamulense. . One specimen collected at an unknown loci l in the Sierra de Oaxaca (September, po m | assigned a Matuda collecting number ( This col | sterile but probably is also A. andicola. : lection bears the herbarium name A. & ' tanioides Matuda. The blade is thinner be of and does not have the pronounced vena” ^ typical A. andicola. d | The herbarium name A. campi! based on Camp 2762 represents from andicola not much unlike the typical i wel Veracruz, even though it was collect near the type of A4. macdougallii. ed nett Matuda 38721, presumably collec and with | Mitla, SW of Oaxaca, has blades thinner ‘nicl | less conspicuous venation than mos A. | those of A 1983] Fic Cus > CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 247 Anthurium angustispadix Croat & Baker, Eso- Anthurium andicola Liebm., Croat 482 : A 26600 (Typo) ur Anthurium armeniense Croat, en roat 4090: ( ype). ANNALS OF THE MISSOURI BOTANICAL GARDEN Anthurium "m i v Q © Y ~ 3 3 y A U ns 9 X 3 m a " g [e] b d wv - ~ 3 = S " S © È p bd A = PA 9 3 3 M ^ È 3 x Š 3 = = - f v Croat 48215.—11—12. Anthurium andicola Liebm.. YO. Fioures 10-13. 1983] collections. The region was visited recently and turned up no obvious habitat for Anthurium so the live plants could not be investigated further. They are presumably only unusually thin-leaved collections of A. andicola, since that species oc- curs on both the Atlantic and Pacific slopes, but no other collections have been made in the vi- cinity of Mitla. Mexico. No other location, Galeotti 6096 (BR); r ménez, Croat 48215 (MO); Matuda 38620 (MEXU); ESE of Ixtlan de Juárez, Graham & Frolich 1046 (MICH); La Gloria, MacDougall s.n. (MEXU); near Mitla, SW of Oaxaca, Matuda 20321, 38721 (MEXU); District Pochutla, Cerro del Machete, Reko 6064 (GH); IR. ENCB); Munc. Juchique de Ferrer, Hernandez 85, 1596 (F); Orizaba, Botterie 85, 135 (GH); Ma- tuda 1540 (MICH). Anthurium angustispadix Croat & Baker, Bre- nesia 16 (Supl. 1): 23. 1979. Type: Costa ica. Puntarenas: primary forest along Rio Coto Brus in the vicinity of Alturas, 23 km N of La Unión (on Panamanian border), elev. ca. 1,500 m, Croat 26600 (MO- 2251852, holotype; CR, F, K, PMA, SEL, US, isotypes; Live at MO). Epiphyte, sometimes secondarily terrestrial; stems 18-30 cm long, 1-1.5 cm diam., branch- he creeping, elongate; internodes 0.8—2 cm long; ots numerous, dense, 3-4 mm thick, directed downward; cataphylls thin, 6-11 cm long, grad- "ally long-acuminate at apex (the acumen 2-4 cm long and inrolled) drying tan, persisting, di- peerating. LEAVES spreading: petioles 5-22 cm Pi 4-7.5 mm diam., narrowly, deeply, and yes Sulcate, rounded abaxially; geniculum a em ong; blades lanceolate to oblong-lan- Nic thin to medium thick, gradually long- "ilb. at apex (the acumen acute or very y inrolled), acute to attenuate at base, 13- 0 cm ng = with age, lower surface glossy or semiglossy; idrib raised, triangular above, diminishing to- CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 249 ward apex, convexly raised below; primary lat- eral veins 11-16 per side, departing midrib at 55°-65° angle, sunken above, raised below straight to arcuate-ascending, loop-connecting; collective vein arising from the base or the first or second pair of lateral veins, 3-4 mm from margin, sunken above, raised below. INFLO- CENCE erect-spreading; shorter to longer than leaves; peduncles 20-40 cm long, 1.5-3 mm diam., terete, equalling or longer than petioles; spathe medium green (B & K Yellow-green 5/ 10), linear-lanceolate, 4.5-8.5 cm long, 4-9 mm wide, broadest near the base, long-acuminate at apex, acute at base inserted at a 50-70? angle on peduncle; stipe 3 mm long in front, 1 mm long in back, 4-5 mm diam., or spadix sessile; spadix greenish-yellow at base, paler at apex (B & K Yellow-green 8/5 to 6/7.5), 4.5-10 cm long, 3- 6 mm diam. at base, 2-4 mm diam. at apex; flowers rhombic, 1.7-2.8 mm wide in both di- rections, the sides sigmoid; 5-6 flowers visible in the principal spiral, 5-10 flowers visible in the alternate spiral; tepals glossy to matte, papillate, green to greenish-yellow, brown in fruit, the lat- eral tepals 0.8-1 mm wide, the inner margin thin, turned up against pistil; pistils slightly emergent, green, at first covered by tepals and flat, rapidly emergent into a pointed mound; stigma 0.3-0.5 mm long, ellipsoid, brushlike, white, exserted, small droplet persisting for 1—2 days before sta- mens emerge, dry and gray when stamens emerge, usually not emerging in the apical half to third of spadix; stamens emerging rapidly from the base in a regular sequence, laterals first followed by alternates, held at the side of pistil, sometimes contiguous; anthers pale yellow, 0.2-0.3 mm long, mm wide; thecae ellipsoid, prominently divaricate, opening bowl-shaped; pollen orange- tact; spadix 5-9 cm long; berries obovoid to depressed-globose, almost flat at apex, pale yel- low-green, to 5 mm long; mesocarp watery, clear, sweet; seeds 2, pale green, flattened, to 2.8 mm long and 2.5 mm wide, each enveloped in a clear, gelatinous substance to 4 mm long. Fig. 8. Anthurium angustispadix is known from east- ern Costa Rica and western Panama at elevations from 100 to 1,500 m in premontane wet and tropical wet forest life zones. The species is perhaps best placed in section 250 Xialophyllium and is recognized by its slender elongate stems with weathered, tan to pale brown cataphylls; its lanceolate leaves with a narrowly cuneate or attenuate leaf base; and its greenish- yellow, relatively long and slender, tapering spa- dix with pale yellow-green berries. It is not easily confused with any other species. The species often branches from near the base in pots in the greenhouse. Plants sometimes set seed in the greenhouse without apparent manip- ulation. A RICA. PUNTARENAS: 10-15 km ENE of La (F); Finca Alturas, NE of Sabilito, Croat ET (MO); along Río Coto Brus, 23 km N , Croat 26600 (CR, F, K, MO, PMA, SEL, US), 26650, 26690 MO); between Río Cotón and Rio Negro (ca. 15 km from Sabalito) near Las Alturas lumber camp, Burger & Matta 4553 (CR, F, NY); E of Las Cruces, 5-6 km S of San Vito on and around the property of Robert Wilson, Burger & Matta 4437, 4455 (F); near Cafias Gordas, Pittier 11127 (US Anthurium armeniense Croat, sp. nov. TYPE: Guatemala. San Marcos: ca. 2 km from Fin- ca Armenia above San Rafael, 1,100-1,250 m, Croat 40905 (MO-2599479, holotype; Live at MO). Planta epiphytica; cataphyllum persistens intactum, brunneum; petiolus 19-28 cm longus; lamina ovata, 20-40 cm longa, 1. 5-21 cm 1 lata, basi cordata pro- dix caesius aut roseus, 3-5.7 cm longus, fragrans; tepala marginibus elevatis; baccae aurantiaca Epiphyte; stems 1.5 cm diam.; leaf scars ca. 1.5 cm wide; roots 4-5 mm disni: descending; cataphylls subcoriaceous, 7 cm long, rounded to emarginate and cuspidate at apex, drying red- dish-brown and persisting intact. LEAVES erect- spreading; petiole 19-28 cm long, 4-6 mm diam., terete; geniculum 2-3 cm long; blades narrowly 15 cm long, the margins broadly rounded; pos- terior lobes 5-12 cm long, directed downward to inward; sinus usually spathulate sometimes hip- pocrepiform, acute to rounded at apex; both sur- aces semiglossy, the upper surface with + con- the second to third pairs coalesced ca. 1.5 cm, the fourth to sixth coalesced ca. 2.8 cm, raised ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 above and below, diminishing and flat at collec- tive vein; primary lateral veins 2—4 per side, flat above, sunken below, departing midrib at 55°- 60° angle, curved to margin, loop-connected; col- lective vein arising from the first basal vein, 4- 15 mm from margin. INFLORESCENCE erect or spreading; peduncle 1 1.5—23 cm long, 2-3 mm diam., terete, as long as or longer than petioles; spathe white, narrowly ovate-elliptic or lanceo- late-oblong, 4.5-9(12) cm long, 1.9-3 cm wide, broadest near base, long acuminate, green at apex, obtuse to acute at base, decurrent ca. 2 mm, inserted on peduncle at 70° angle; the stipe 2 cm long in front, 3-5 mm long in back, usually curved; spadix pale lavender (B & K Purple 6/ 5), fading to pink (B & K Red-purple 6/2.5) at anthesis, 35.7 cm long, 7-9 mm diam. at bast, 6—7 mm diam. at apex, the flowers rhombic t0 weakly 4-lobed, 2.1—2.9 mm long, 2.8-3.6 mm wide, the sides straight to weakly and irregularly ! sigmoid, ca. 6 flowers visible in the principal i spiral, 9 flowers visible in the alternate spiral; the tepals matte; lateral tepals 1.6-1.9 mm wide, the inner margins straight to convex; pistils white, emergent ca. 0.5 mm, the stigma broadly elliptic, ergen whitish, A darcupi with short papillae exserted sion; anthers white, broadly ovate, long, 0.8-1.4 mm wide; bee broadly 9 0 an wly c upular; pollen abun e iam., wit anys stages); berries obovoid, style dark. t apex, pale orange (B & K K Yellow-re eG cm long, 6-8 mm wide; m ‘very ange, juicy, gelatinous with n umerous 5! palé raphide cells; seeds (1)2, ovoid-elipsid Pit whitish-green, darkergrecn aM reddish at apex, weakly flattened, 4.5 m sticky, gelatinous, translucent substance i gm prominent at lateral margins, extending ends. Figs. 9 and 14. | lo- e species is known only from the typ? } The o wet cality from 1,100 to 1,600 m in pe forests on steep slopes in southwestern mala in the Department of San Marcos The species is a member of sectio trium and is recognized by its et v moderately coriaceous leaves with rap 1983] CROAT evident on upper surface and by its thick, per- sistent cataphylls that dry reddish-brown, but es- pecially by its pale lavender to pink fragrant spa- dix that has a distinctly bumpy surface owing to the slightly upturned edges of the tepals. The species also is characterized by its white spathe. It is confused with A. Auixtlense, which has similar leaves and cataphylls (both are in the nearly level with the surface of the spadix so that it does not have the bumpy appearance. In ad- dition, A. huixtlense lacks the strong lilac scent of A. armeniense. GUATEMALA. S Finca A Rafael, Croat 40905, 40923 (MO). Rica. Alajuela: San Luis de Zarcero, elev. 1,450 m, Austin Smith 637 (NY, holotype; F, isotype). naming Epiphyt , Slender, gn ying grayish; stem moderately elongate, ca. 1 cm diam., 16-20 cm long, scurfy; cataphylls 2-2.5 cm long, drying dark reddish-brown, persisting as linear fibers. LEAVES spreading; petioles subterete, narrowly sulcate, 4-22 cm long, 2-4 mm di &niculum 7-20 mm long; blades subcoriaceous, bs long-lanceolate, 11-21 cm long, 3-6.5 cm wide, ee Me at apex (the acumen flat), obtuse ? acute (rarely attenuate) at base, broadest at or below the middle, the margin revolute on drying; both surfaces semiglossy, lower surface dark Pandular-punctate; midrib acutely raised above, aL and sunken toward apex, convexly pononga primary lateral veins 7-15 per side, ci midrib at 45° angle, + straight to the ad lve vein, slightly sunken above, promi- - ew. interprimary veins sunken above, V VISI leb 1 11 a: . za T, H we reg 4-8 mm from the margin, about equally i luas the primary lateral veins, sunken > Prominulous and darker below. INFLO- ceed; ENCE erect-spreading, equalling or ex- pis ila leaves; peduncle terete to sharply an- 21-25 ~3-ribbed, sometimes with only one rib, ibn long, 2-3 mm diam.; spathe pale green, est ju ate, 2.5-3 cm long, 4-8 mm wide, broad- et St above the base, acuminate at apex, nded at base, inserted at ca. 30° angle on the uncle: spadix sessile, pale greenish-yellow, 5— OF MEXICO AND MIDDLE AMERICA 251 7.5 cm long, 3-4 mm diam. at base, 2-2.5 mm diam. at apex, green at anthesis, becoming violet- purple; flowers rhombic, 3.3-3.5 mm in the alternate spiral; tepals semiglossy, the lat- eral tepals 2-2.5 mm long, the inner margin broadly rounded to straight; pistil green soon turning purple, emergent to 0.5-0.6 mm; stigma vides a d : L ined apex, with a circular depression around the style ca. 2 mm diam.; locules 2, with 3 seeds per locule; flattened to subterete in cross-section, smooth. Figs. 11 and 12 Anthurium austinsmithii is known only from Costa Rica in lower montane rain, lower mon- tane wet, lower montane moist, and premontane wet forest, primarily on the Atlantic slope at el- evations of 1,200 to 1,560 m. The species is a member of section PorpAy- rochitonium and can be distinguished by its acu- minate leaf blades, conspicuously punctate on the lower surface which dry a dirty gray, slender spadix with ca. 3 flowers visible in the principal spiral and by its whitish berries tinged with pur- ple-violet bearing a circular depression at the x Anthurium austinsmithii is similar to A. acu- tangulum, A. durandii, and the species allied to them. It is also similar to A. alatipedunculatum because of its angulate peduncle, but the latter species also has a triangular petiole and is known currently only from the Osa Peninsula at much lower elevations than A. austinsmithii. OSTA RICA. ALAJUELA: San Luis de Zarcero, Austin 36046 (MO) 232 Anthurium bakeri Hook. f., Bot. Mag. pl. 6261. 1876. Tyre: Costa Rica. Bull s.n. (K). } li l., Bot. Jahrb. Syst. 25: 406. 1898. Type: Costa Rica. Cartago: Turrialba, 500 m, Donnell Smith 4978 (B). Epiphyte, stems less than 10 cm long, 1.5 cm diam.; roots moderately slender; leaf scars ca. 1.7 cm wide; cataphylls coriaceous, 3—6 cm long, acuminate at apex (the acumen apiculate to 2 mm), drying brown, persisting as fibers. LEAVES erect to spreading; petioles (3)1 1-17 cm long, to 5 mm diam., subterete to sharply sulcate; ge- niculum 1-1.5 cm long; blades narrowly elliptic- lanceolate to narrowly oblanceolate, 19-55 cm primary lateral veins nearly obscure, flat, departing midrib at 45° an- gle, straight to collective vein; collective vein arising from base, sunken, significantly more prominent than the primary lateral veins, con- tinuous to apex of leaf, 2-4 mm from margin. INFLORESCENCE erect-spreading, much shorter than leaves; peduncle 5.5—30 cm long, 3- 5 mm diam., half as long or as long as petioles; spathe coriaceous, pale yellow-green (B & K Yel- low-green 9/10), the edges sometimes tinged with purple, 2-5.5 cm long, 7-28 mm wide, oblong- lanceolate, inserted at ca. 60° angle on peduncle, obtuse at apex and base; stipe 3 mm long, 2 mm diam.; spadix creamy white (B & K Yellow-red 9/5), 2-11 cm long, 5-15 mm diam. at base, 4.5— 12 mm diam. at apex; flowers square to 4-lobed, 1.6-3.5 mm long, 1.7-3.2 mm wide, the sides jaggedly sigmoid, 7-8 flowers visible in the prin- cipal spiral, 5-6 flowers visible in the alternate spiral; tepals glossy, minutely papillate, lateral tepals 1.5 mm wide, the inner margin convex; pistils not emergent; stigma minute, slitlike, ca. 0.1 mm long, droplets apparent before stamens emerge; stamens emerging in an irregular pat- tern, usually lateral stamens first followed by al- erect, spathe reflexed against peduncle; spadix 7.5-8.5 cm long, 1.2-1.4 cm diam .; berries red, pointed at apex, obovoid ca. 6 mm long; me- socarp juicy, pulpy, transparent with numerous oblong raphide cells; seeds 2, Ovoid, cream-col- ored, flattened, 3.2 mm long, 2 mm wide, 1.2 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 mm thick, enveloped by a gelatinous substance extending below the base. Figs. 13 and 15. The species is known from Guatemala to Co- lombia in wetter parts of tropical moist, pre- montane wet, and tropical wet forest at eleva- tions from sea level to 660 m. In Panama, the species ranges to 1,000 m elevation. In Colombia it is known from a single collection from Norte de Santander (Camp 84 on Pipeline, 600 m), M. & R. Foster 1716 (A). Anthurium bakeri can be distinguished by its moderately thin, elliptic-lanceolate blades that are reddish-brown punctate on the lower surface and have many weak primary lateral veins and conspicuously sunken collective veins. The lower leaf surface usually dries much paler than upper surface, sometimes yellowish, contrasting sharp- ly with the green upper surface. Other guishing features include the creamy white, al- most cylindrical spadix and bright red berries. The species is a member of section Porphyro- chitonium. BELIZE. BELIZE DISTRICT: near Quamina Creek, 3426 (MICH). STANN CREEK DISTRICT: Mid pp S-16 (MIC (MO); Maya Mountains, Bouti (MO); San Jose, Croat Mes 24359 (M CostTA RICA. AL aun N of Ángeles Note 3381 (DUKE), NNE of B 36430 (MO); PUE. & miy 533 Canalete, Burger & Baker 9 cienda Santa María, Liesner 5131 (MO); Wye olcan Arenal, Taylor 11588 (NY); NE 0 Croat 994 Croat 4 (MO). CARTAGO: Turrialba, GUANACASTE: NW o Lake HEREDIA: near Puerto ^ 4 72 (MO); Jiménez 105 (MO); Taylor 454 ja Vara Blanca, Skutch 3651 hamo LIMON: H Donahue Tapezco-Hacienda La Sue paige & Go rger e | = (MO); Finca Castilia "Dodge MO). PUNTARENAS: Osa paste (MO), Rift 1 10021, 10022, 10150 (MO): Liesner 3097 ( : r 32 (DUKE) D DOR cón de Osa, Utley & Utley from San Isidro del Gene Baker 10106 (MO Gua dia rapaz: vicinity cai ATEMALA. ALTA VE a Steyermark 44382 (NY); Mai Nae 41606. Steyermark 44917 (MO); O die bees Fat ` 41653 (MO). izaBAL: vicinity Exmibal, 4548] 32804 (NY). PETE of Chinajá, Steyermark (MICH). Hon —- ATLANTIDA: La Ce iba, n 8739 (M ; Lancetilla, Chickering 16 (MICH); L Lancetilla Valley, Croat 42638 V7 6, 104. (MO): we o 2898 LOO Silk Grass Creek, Gentle 2626 Kl ci | lum- 4 i Luteyn 283, 36326, | CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA Anthurium -— Hook. f., Croat 77-1275.— Anthurium i. D Drum 14-17. 14. Anthurium armeniense Croat, Croat 40923.—15. j —16. Anthurium berriozabalense Matuda, Honolulu Botanic Garden “dDalense Matuda, Croat 254 ro = s 2704 (DAV); ut 4960 (MICH, MO). A: NW of Siguatepeque, Molina 26062 (NY). pcs A DIOS: vicinity ity Rio P Plátano, Clewell & Cruz 4061 (MO). vono: Coyolito, Carleton 552 (US); Yuncker et al. 8187 (NY, S). EXICO. CHIAPAS: Mun. of Las Margaritas, Breedlove & McClintock 34076 (CAS). NICARAGUA t evy 1420 (P); N of hoe a 7471 ved MATAGALPA: Cerro Musün, Neill 1808 (MO); E of Matagalpa, Neill 2353 (MO). RIVAS: n Maderas, Neill & Vincelli 3247 (MO); Stevens 6507 (MO). ZELAYA: Caño Zamora, Ste- vens 8824, 8826 (MO); Cerro El Inocente, Neill 1906 (MO); Stevens 6714 (MO); Cerro Waylawás, Stevens 8756 (MO); SW of Colonia Naciones Unida, Stevens 5023 (MO); trail from Colonia San José, NNE to Par- cela, Stevens 9008 (MO); Colonia Agricola Yolaina, Se evens 4827 (MO); pe = al. 6437 (MO); Rio L, NY); Stevens 8909 927 (MO); vicinity of 9 (MO); Jae 4225 (MO); near Tala s and Pu Mango, Stevens 7620 (MO). Anthurium beltianum Standl. & L. O. Wms., Cei- ba 3: 103. 1952. Type: Nicaragua. Dept. of Jinotega: mountains east of Jinotega, along trail to Cerro la Cruz, elev. 1,050-1,350 m. Standley 10255 (F, holotype; EAP, isotype). Terrestrial; stems thick, to 4 cm diam., leaf scars 4 cm wide; roots sparse, descending, tan; cataphylls coriaceous, 4-12 cm long, caudate and erect-spreading; petioles 13-73 cm long, 5-8 mm diam., subterete, shal- lowly and bluntly sulcate, puc merely flat- tened adaxially; geniculum 1.5-2.5 cm long; blades subcoriaceous, triangular-ovate to ovate, acuminate at apex, deeply lobed at base, 14—51 cm long, 8-36 cm wide, broadest at point of petiole attachment, the margin undulate; pos- terior lobes 5-18 cm long from sinus to tip, the sinus hippocrepiform to triangular, acute to round at apex; both surfaces matte to semiglossy, some- times with punctiform or linear raphide cells (more conspicuous on dry specimens); midrib raised above and below; basal veins 5-6 pairs, the first and second free, the remaining coalesced 2.7-4 cm, raised above and below; posterior rib turned up along the outer margin; primary lateral veins 4-8 per side, departing midrib at 409-55» angle, flat to weakly raised above, raised below; interprimary veins flat above, raised below; less- er veins scarcely visible; collective vein arising from the first basal vein or one of the primary lateral veins, 1-2.5 cm from margin, flat above, raised below. INFLORESCENCE erect-spread- ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 ing; peduncle 20-67 cm long, 3-6 mm diam., slightly curved near base of spadix, shorter to current and clasping at base, inserted at 65°-8 angle on peduncle, held behind and hooding spa- Es pom pale green (B & K Green 7/2.5), c and a large crepes for 3 or 4 days before i stamens emerge; stamens emerging very E in a complete sequence from base, held in ca. 0.6 m tone: 1.1 mm wide; thecae ellipsoid, — somewhat divaricate; pollen pale white. I UCTESCENCE pend persisting; berries obovoid-ellipsoid, yellow E: yellow-orange, 13-15 mm long, 9-11 m e pericarp thick, darker colored and e rap ide cells near apex, numerous granular raphide cells in the lower half; mesocarp juicy, i sweet; seeds 1-2, semicircular to broadly ¢ "d p 1 soid, brown with nu 1 jections on surface, ‘flattened, 8-9 mm long s 9 mm wide, 3 mm thick, encased in à thin velope. Figs. 18 and 19. Nicaragua in the departments of Jino Matagalpa at elevations of 1, 000 to 1. premontane wet forest life zones, ON lim outcrops. The species was named in honor of the ? f P English naturalist Thomas Belt, author ^ “The Naturalist in Nicaragua." m is a m | of section Calomystrium and is with moderately thick, ovate-cordate iit E obovate or hippocrepiform sinus, ci ae : dix à ard FiGURES 18-21. (Type). 18-19. Inthurium beltianum Standl. & L. O. Wms., Croat 43003.—20-21. Anthurium brenesii Croat & Baker. Croat 46923 [£861 IVO3IO VOLIINWV TYIGGINW ANV OOIXAN AO WATWH.ILNF 256 Anthurium beltianum is similar to A. cotobru- sii, which has generally similar leaves and inflo- rescences but does not have the hooded inflo- rescences of A. beltianum. NICARAGUA. JINOTEGA: vicinity of Jinotega, Croat 43003 (MO); Standley 10255, 10359, 10883 (F); Wil- liams et al. 24733 (F); MATAGALPA: vicinity of Aran- juez, Neill 3020, 3452 (MO); Finca Santa Maria de Ostuma, Williams et al. 27887 (MO). A nthuri I iozabal Matuda, Anales Inst Biol. Univ. Nac. México 25: 214. 1954. TYPE: Mexico. Chiapas: Las Vistas north of Ber- riozabal, 1,250 m elev., Miranda 6707 (MEXU, holotype). Terrestrial on rocky, steep slopes to 0.6 m tall; stems 10 cm long, 1.5-3 cm diam.; leaf scars ca. 1.5 cm wide; ro ie ca. 5 mm dick. directed sowie cataphylls subcoriaceous, 3-5 cm long, acute at apex, the acumen minutely apic- ulate, drying light brown (B & K Yellow 4/5), persisting intact at apex, splitting into linear fi- bers at base. LEAVES erect to spreading; petioles 13-56 cm long, 3-7 mm diam., subterete, weakly flattened adaxially; geniculum 1.5-2 cm long broadly and shallowly sulcate; blades triangular, acuminate at apex, broadly lobed at base, 12-33 cm long, 11-23 cm wide, broadest at base; an- terior lobe 9-16 cm long; posterior lobes 12-18 cm long, directed outward; sinus triangular to parabolic to sometimes hippocrepiform; upper surface semiglossy to glossy, lower surface semi- glossy, the midrib convexly raised above and below, diminishing at apex above; basal veins 4—5 pairs, the first and sometimes the second free, second to fourth coalesced ca. 1.5 cm, third to fifth coalesced 2.3—4 cm, raised in weak valleys above, raised below; posterior ribs curved, na- ked, the outer margin upturned; primary lateral veins 4-5 per side, departing midrib at 459-50? angle, raised or flat above, raised below; collec- tive vein arising from uppermost basal vein (sometimes with lower basal veins loop-con- necting to the collective vein), sunken above, raised below, 6-10 mm from margin. INFLO- RESCENCE spreading to erect, shorter than the leaves; peduncle 16-40 cm long, 2-4 mm diam., terete, sometimes tinged with violet- -purple, longer than the petioles; spathe medium thick, green tinged with purple at margins, peus to oblong-lanceolate, 4.5-9 cm long, 0.9-2.7 wide, broadest at base, acuminate at apex, dum ing to truncate or rounded at base, inserted at 45°-60° angle on peduncle; spadix green (B&K ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 Yellow-green 5/7.5), 5-10 cm long, 7-9 mm diam. at base, 2-3 mm diam. at apex; flowers square or rhombic, 3-5 mm long, 3-4 mm wide, the ; sides + straight to jaggedly sigmoid; 3-5 flowers __ visible in either spiral; tepals matte, green, - | inconspicuously punctate, minutely papillate, the stigma oblong, shallow, green, 0.8-1 mm opening with minutely exserted papillae, droplets apparent for 7-10 days, dry and b for 2 or 3 days before first stamens emerge; mens emerging + rapidly from the base, the sec ond, third, and fourth stamens Ne 30 E iiaee that quickly retract, anthers held a of pistil but not contiguous; thecae ellipsoid, d- varicate; pollen yellow, fading to white. IN- FRUCTESCENCE pendent; spathe withel spadix to 10 cm long, 20 mm diam.; berries of ange (B & K Yellow-red 7/5), broadly obov round at apex with a conspicuous indentat 8-11 mm long, 7-10 mm wide; pericarp parent, somewhat thickened, lacking raP cells; mesocarp fleshy, juicy, with num raphide cells; seeds 1 or 2, flattened on one if 2, ellipsoid in cross-section if only one, y green, 6-8 mm long, 4.5-5(9) mm wide, 3.5(6.5) mm thick. Figs. 16, 17, and 26. 1 to Anthurium berriozabalense is restricted 10 — Mexico and is known for o Eu from 1 northeastern Chiapas at 1,000 to a n collection from eastern Oaxaca e val d | from Rio Grande (presumably a: i a Romero) is probably also this species. ected the type near El Bosque where I have s, p The species is tentatively placed i lolonchium and is recognized by its erit sagittate leaf blades with long. slender. pos! a lobes that are usually as long or longer i: anterior lobes, and also by its der spa spadix and subglobose orange berri Anthurium desee is pe to A. seleri, a more wide ra g species v o, Forms s range go a that of A. perriozabalenst es have can be distinguished from A. berrioz GURES 22-25. Baker, Croat 43358. — 2] El p E. Anthurium brownii Mast., Croat 37840.—2 radi brownii Mast., Anthurium caperatum Croat & Baker, C roat gi Croat 10389.—24. Anthurium caperatum Croat & [£861 VOLI3JNWV TIGGIN ANV OOIXJN JO WAIYAHINK —- LVOSIO ESC ANNALS OF THE MISSOURI BOTANICAL GARDEN Á + ( dg JRES 26-29. 26. j ji Croa oat Å Bake i Ant gee berriozabalense Matuda, Croat 47715.—27. Anthurium wt 99, Ar aker, Croat 46923 (Type).—28. nthurium Pur geri Croat & Baker, Burger & Baker 10096 (TY maed: carnosum Croat & Baker, C roat 486 8 CROAT LARVAL 1983] the much longer (or at least equally as long) an- terior lobes and by the anterior lobe and poste- rior lobes forming a broad concave edge. In A. berriozabalense, the margin is either straight or convex and the posterior lobes are usually as long or longer than the anterior lobes. In addition, the leaves of A. berriozabalense are thin whereas the leaves of A j 1 i Still ther difference is that the collective vein of A. ber- riozabalense usually arises from the uppermost basal veins or has the lower basal veins con- spicuously loop-connected with the collective vein. In A. seleri the collective vein arises from one of the lowermost basal veins and extends rather uniformly along the margin of both the posterior lobes and the anterior lobe to the apex. In the type description Matuda discussed the similarity of A. berriozabalense to A. umbrosum but these two species are not at all similar except that both have orange berries. MEXICO. CHIAPAS: vicinity Berriozábal, MacDougall 445 (MEXU); Miranda 6707 (MEXU); vicinity Si- mojovel de Allende, Croat 47715 (B, CAS, F, K, M, M O, NY, RSA, SEL, US, VEN). oaxaca: Rio Grande, MacDougall 311 (MEXU). Anthurium brenesii Croat & Baker, Brenesia 16 (Supl. 1): 28. 1979. Type: Costa Rica. Ala- Juela: along Hwy. 15 between Naranjo and Quesada, 3.2 miles N of Zapote, 1,560 m, Croat 46923 (MO-2682420, holotype; COL, R, DUKE, F, GH, K, NY, PMA, RSA, SEL, US, VEN, isotypes; Live at MO). Epiphyte; stems short, 1.5-2.5 cm diam.; roots moderately few, to ca. 5 mm diam.; cataphylls moderately thick, 10-25 cm long, acuminate at apex, tinged red at margins, weathering into fi- iol » drying light brown. LEAVES pendent; pet- oles 14-45 cm long, ca. 6 mm diam., terete; &niculum to 1.5 cm long; blades oblong, mod- — coriaceous, abruptly to gradually short- acuminate at apex, acute to weakly attenuate at base, 32-68 cm long, 6-10.3 cm wide, broadest d the middle, the midrib raised above and ON the primary lateral veins 13-15 per side, Parting midrib at 55° angle, sunken above, Prominulous below; collective vein arising from rai ae 2-5 mm from the margin, sunken above, P below. INFLORESCENCE erect or pen- nt, shorter than the leaves; peduncle (10)28- hice long, 4-5 mm diam., terete, shorter or mE. than petioles; spathe moderately thick, ium green sometimes tinged reddish at an- OF MEXICO AND MIDDLE AMERICA 259 thesis, ovate to ovate-elliptic, 5.5-8.5 cm long, 3.5-5.5 cm wide, broadest from 1 cm above the base to about the middle, acute to rounded and at base, inserted at 25?—40? angle on peduncle, + erect and held close to spadix or at ca. 45° angle to peduncle; spadix sessile or subsessile, green, reddish-brown to red-violet, 7-8 cm long, to 8 mm diam., at base, to 6 mm diam. at apex; flowers square to weakly 4-lobed, 1.7-2.5 mm long in both directions; 6—10 flowers visible in the principal spiral, 6-9 flowers visible in the alternate spiral; the lateral tepals to 1.2 mm wide, the inner margin concave, pistil green scarcely emergent, the stigma oblong, 0.5-0.6 mm long; sta ging rapidly throughout, lateral sta- mens first, quickly followed by alternate sta- mens, exserted on short, translucent filaments, retracting to hold anthers just above the margins of the tepals; anthers orange, ca. 0.6 mm long, 0.8 mm wide; thecae ovate-ellipsoid, slightly di- varicate; pollen orange, faintly spicy scented at anthesis. INFRUCTESCENCE spreading to pendent, the spathe usually withered, the spadix to 9 cm long, 2 cm diam.; berries exserted ca. 1 mm, before maturity, obovoid, flattened, orange (B & K Yellow-red 7/5), 9-11 mm long, 5-7.5 cats WIU ? 28 mm diam., p pj sparse, thick raphide cells; seeds 2 (rarely 1), tan, 3-4.5 mm long, 2.2-3 mm wide, 1.8-2.2 mm thick, enveloped in almost translucent (tinged orange) casing, ca. 10 mm long and as wide as the seed. Figs. 20, 21, and 27. J Anthurium brenesii is presently known only from the Cordillera Central in the Province of Alajuela between San Ramón and Quesada at 1,000 to 1,560 m elevation. The species is related to A. protensum with which it shares similar leaf texture and cata- phylls, and similar altitudinal range. It differs from that species, however, by its abruptly acu- minate leaf blades, its short, broad spathe, which is subcordate at the base, and its shorter, broader spadix. It is also similar, especially vegetatively, to A. prolatum, but differs in having much short- er, broader leaves. It is placed in section Pachy- neurium. Costa RICA. ALAJUELA: La Palma de San Ramón, Brenes 5738 (F); Cordillera Central ca. 15 km N of Zarcero, Williams et al. 29034 (F); along Hwy. 15 be- tween Naranjo and Quesada, 3.2 miles N of Zapote, Croat 46923 (COL, CR, DUKE, F, GH, MO, NY, PMA, RSA, SEL, US, VEN). 260 Anthurium brownii Mast., Gard. Chron. 6: 744, Figs. 139, 140. 1876. Type: Colombia (Nueva Granada). Wallis s.n. (K, exact sta- tus unknown). Epiphyte; stems short, 2.5-3 cm diam.; roots few, thick, puberulent to densely velutinous; cataphylls coriaceous, 9-15 cm long, caudate- acuminate at apex, drying reddish-brown (B & K Yellow 4/2.5), persisting — intact, ultimately theri twork offil EAVES übers at base erect to spreading; petioles 13-100 cm long, 3- 10 mm diam., subterete, shallowly and bluntly sulcate; geniculum 2-2.5 cm long, sometimes tinged red-violet (B & K Red-purple 4/10); blades narrowly ovate-triangular, moderately thick, 15— 66 cm long, 11-48 cm wide, broadest at base, gradually acuminate at apex (the acumen flat, apiculate), prominently lobed at base; anterior lobe 11—40 cm long, the margins straight to con- vex; posterior lobes 4-13 cm long, directed sharply upward; the upper surface matte to semi- glossy, lower surface paler; midrib convexly raised above, sometimes tinged reddish, sunken at apex, acutely raised below; basal veins 5-6 pairs, the first to fourth coalesced 2—6 cm, the posterior rib naked, upturned along outer margin; primary lat- eral veins 4-8 per side, departing midrib at 25° 35° angle, raised in valleys at base above, sunken at apex, raised below; lesser veins visible but less prominent; collecti in arising fi rst basal vein or lowest primary lateral vein, 2-7 mm from margin. INFLORESCENCE erect, longer than leaves; peduncle 51—65.6 cm long, 4-6 mm diam., terete, green mottled with red-violet, longer than petioles; spathe lanceolate, green (B & K Yellow- green 6/10), 8-23 cm long, 1.5-4.5 cm wide, acuminate at apex, round to cordate at base; spa- dix stipitate to ca. 11 cm, violet-purple, 7-28 cm spiral, 5-12 flowers visible in the alternate spiral; tepals matte, punctate, minutely papillate, some- times with droplets, lateral tepals 1.8-2 mm wide, the inner margin broadly rounded; pistils emer- mens emerging in basal one third before alter- t , held in circle d pistil à I ;anther yellow or tinged purple, thecae moderately di- [2 ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 | varicate, pollen purple fading to white some- times yellow, with a pungent fruity aroma at an thesis. INFRUCTESCENCE pendent-spreading, ; 38 cm long, 3 cm diam.; berries ellipsoid to obo- | } void, 5-8 mm long, ca. 5 mm wide, orange to | red-orange (B & K Yellow-red 7/2.5); pericarp | thick; mesocarp juicy; seeds 2, oblong, greenish- white, flattened with transparent sticky append- ! ages at both ends, elongated at apex, rounded at | base. Figs. 22 and 23. The species is known from Costa Rica to Co- | lombia at elevations from near sea level to 1,00 | m. In Costa Rica the species occurs in premon- — tane wet and tropical wet forest. In Panama it ! also occurs rarely in premontane moist and trop: | ical moist forest. The species was confused with A. denudatum — Engl. from Colombia by Standley (1944) in a | Flora of Panama, but that species has a pedunc | much shorter than the petioles, as well as a short- ' er spadix. " 2 Anthurium brownii is a member of section lolonchium and is distinguished by its thick bad | with usuall dulat £ | that dry very prominent. The leaf tends e | silver or grayish on both surfaces. Also o } tive are the long, slender, violet-purple spa” — and the red-orange berries. Costa RICA. ALAJUELA: NNE of Bijagua, Bus Baker 9855 (F, MO). HEREDIA: vicinity Río La t | rande, Croat 35859 (MO). PUNTARENAS: ee =e de Osa, Koch 5065 (MO). san José: SW f | del General, Utley & Utley 4926 (DUKE). | ia 16 Anthurium burgeri Croat & Baker, e } (Supl. 1): 30. 1979. TYPE: Costa Alfon | José: ridges and steep slopes x Gener bra along road from San Isidro w ge18'N. | to Dominical (on Pacific consi) 1975. . 83*46W), ca. 1,000 m, 19 Novel D. polo ' Burger & Baker 10096 (F 17691 at MO} type: CR, K, MO, US, isotypes; Liv “a Epiphyte or terrestrial on steep e | short or long, 1-1.3 cm diam.; roots MU cm downward; cataphylls moderately thin, brow! long, blunt to acuminate at apex, E sub- (B & K Red-yellow 4/10), remaining ^ 4 + sequently deciduous. LEAVES Nr erect to spreading; petioles 15-20 em (0 9p | mm diam. midway, terete; geniculu des oP” | 2 cm long, weakly flattened adaxially; k (deyi? long-oblanceolate, moderately thic | CROAT KVALI 1983] somewhat reddish-brown or gray with purplish splotches), gradually acuminate at apex, narrow- y cuneate at base, 36-48 cm long, 6.8-8 cm wide, broadest at or just above the middle; upper surface glossy to semiglossy or matte, lower sur- f iglossy; midrib p i tly raised above and below, broad at the base, becoming acute + straight to the collective vein, flat and some- what obscure above, raised below; collective vein arising from near the base, more conspicuous than the primary lateral veins, 2-9 mm from margin, weakly sunken above, prominulous be- low. INFLORESCENCE spreading, as long as or longer than the leaves; peduncle 40-53 cm long, to 4mm diam., terete, longer than petioles; spathe moderately thin, light green sometimes tinged purple at margins (B & K Yellow-green 7/10) oblong-lanceolate to lanceolate, 6.5-9 cm long, L. -2.3 cm diam., broadest near the base, acu- mınate-cuspidate at apex, rounded to weakly cordate and slightly clasping at base, the margins sometimes joining petiole at slightly different levels; spadix sessile, greenish-brown to pur- plish-brown (B & K Yellow 6/7.5), 9-12.2 cm long, 5-7 mm diam. at base, to 3 mm diam. at apex; flowers rhombic to weakly 4-lobed, 2-3 mm wide in both directions, the sides straight to weakly sigmoid; 5-9 flowers visible in either spi- ral; tepals weakly glossy to matte, pale yellow- green with purplish inner edges, pale punctate, with a few scattered droplets, lateral tepals 1-1.3 mm wide, the inner margin broadly convex to m ost straight; pistils weakly emergent, green- Ri Vos ellipsoid-circular, ca. 5 mm ong, í a ike and exserted, dark violet-purple, with cere droplet appearing briefly, well be- then rst anthers emerge, dry and dark when an- a ume stamens emerging promptly from s € middle of the spadix or from the base, Pte so l mm above the tepals at anthesis, the M ca. 3 times longer than the anthers, soon E ing the anther to the level of the tepals; us rS yellow-brown, 0.4-0.5 mm long, 0.7-0.8 NM ag not contiguous; thecae weakly divar- Yellen gone elliptic; pollen orange (B & K appear; /2.5), drying tan to whitish, quickly dis- Spread; INFRUCTESCENCE pendent- bea spadix ca. 12 cm long, ca. 1.9 cm diam., Ee 8 berries in the basal half only; berries nar- Y ovate, 6-6.7 mm long, 3-3.4 mm wide, OF MEXICO AND MIDDLE AMERICA 261 acute to narrowly rounded at apex, red-orange (B & K Yellow-red 5/2.5); pericarp red-orange; mesocarp gelatinous, pale orange, with numer- ous raphide cells; seeds 1 or 2, elliptic, pale yel- low, sometimes weakly flattened, 3.2-3.5 m long, 2-2.1 mm wide, 1.2-2 mm thick, surround- ed by a sticky amber envelope. Fig. 28. The species is known only from the type lo- cality in Costa Rica in premontane rain forest. Anthurium burgeri can be distinguished by its oblong-obl late leaf blades with obscure pri- mary lateral veins, and more or less straight, weakly depressed collective veins arising from near the base. The midrib on the lower surface is sometimes tinged reddish and the cataphylls turn brown and remain intact. The inflorescence is equal to or longer than the leaves and the spadix is greenish to purplish-brown with dark violet-purple stigmas. It does not seem to be closely related to any other species in Central America. The sectional placement of A. burgeri is as yet uncertain but is probably best placed in i'alophylli lespite th tthatit may secti J p sometimes have short stems. CosrA RICA. SAN JOSÉ: above Alfombra along road from San Isidro del General to Dominical, Burger & Baker 10096 (CR, F, K, MO, US). Anthurium caperatum Croat & Baker, Brenesia 16 (Supl. 1): 32. 1979. Type: Costa Rica. Heredia: vicinity of Bajo la Hondura, along road between Paracito and the Río Claro, 1,100-1,400 m elev.; Croat 44518 (MO- 2598657-58, holotype; CR, F, US, isotypes; Live at MO). Epiphytic near the ground or terrestrial; roots (1)3-5 mm diam., directed outward and down- ward; stem 60-120 cm long, 2.5-5 cm diam.; leaf scars 2.5-4.5 cm wide, cataphylls coriaceous, green tinged with purple, 14-40 cm long, nar- rowly rounded at apex with a subapical apiculum ca. 3 mm long, drying pale brown, weathering into fine fibers. LEAVES erect-spreading; peti- oles 24-109 cm long, 6-15 mm diam., subterete, shallowly and bluntly sulcate; geniculum 1.5-5 cm long; blades ovate to narrowly ovate, sub- coriaceous, usually weakly bullate, abruptly to dually acuminate at apex, prominently lobed at base, 46-120 cm long, 30-69 cm wide, broad- est at point of petiole attachment; anterior lobe 26-90 cm long, the margins + straight; posterior lobes 12-30 cm long, flat or directed inward, 262 often overlapping, the sinus spathulate, obovate sunken near apex, convexly raised below, some- times weakly ribbed or striate; basal veins 4-9 pairs, (second) fourth to seventh coalesced (2)4.5— 9 cm, raised above and below, the posterior rib stout, curved or nearly straight, naked, rolled up along the outer margin; primary lateral veins 9— 18 per side, departing midrib at 55° angle, + straight to slightly ascending to a weak collective vein, sharply raised in weak valleys above, raised below; collective vein usually arising from fourth or fifth basal vein, sunken above and raised be- low, 1-4(10) mm from the margin, loop-con- necting the basal and primary lateral veins. IN- FLORESCENCE erect, much shorter than the leaves; peduncle 6—54 cm long, 8-13 mm diam., subterete or terete, as long as petioles; spathe thin, pale green (B & K Yellow-green 7/10), lan- ceolate-oblong, 10-28 cm long, 3.3-4.8 cm wide, broadest in lower third, sometimes erect, usually spreading to reflexed, often twisted and rolled up, inserted at 10-30? angle on peduncle; spadix sessile, pale green (B & K Green 8/2.5), (3)7- 18(28) cm long, 10-20 mm diam. at base, 6-12 mm diam. at apex; flowers 4-lobed, to 1.8 mm long, to 1.5 mm wide, the sides jaggedly sigmoid; 12-15 flowers visible in the principal spiral, 17— 19 flowers visible in the alternate spiral; tepals semiglossy, the lateral tepals to 1 mm wide, the inner margin + straight; pistils slightly emergent, very glossy, whitish green with darker green stig- matic area; the stigma elliptic, 0.2 mm long, glossy, exserted with glistening papillae; stamens emerging scattered throughout the spadix, some- times with the alternate stamens preceding the lateral ones; filaments ca. 1 mm long, anthers pale yellow, ca. 4 mm long and 5-6 mm wide; thecae oblong; pollen yellow, fading white. IN- FRUCTESCENCE usually pendent, spathe often absent; spadix 15-40 cm long, to 2.5 cm diam.; berries ovate to ovate-elliptic, acute to rounded and green at apex, greenish-white to yellowish- green at base (white fide label of A/len 24 19) to 9 mm long, ca. 5 mm wide; seeds 2, oblong and + flattened, dark brown, 3.5-5 mm long, 2.2- i mm wide, 1.5-1.8 mm thick. Figs. 24 and a: The species is known from Costa Rica and Panama and is to be expected in Colombia on ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo.. 70 the southwestern slopes of the Serrania de Pirre. It occurs on both the Atlantic and Pacific slopes. In Costa Rica, it ranges between 850 and 2,000 ; m but in Panama it has been collected as MF 400 m. Th est, premoutiie rain forest and lower monti rain forest. » A . - a4 itc laroe Anthirni n1 gi ovate leaves that have most of the veins of the , upper surface sunken, giving the leaf a slightly bullate appearance, and a collective vein which | arises from one of the lower pairs of basal veins. | Other characteristic features include the green, | brittle spathe, the pale green spadix, prominently j exserted stamens, exserted pistils, greenish ber- | The species. is similar to Anthurium pant- | : A. corrugatum Sodiro complex. nu were | placed in section Polyneurium by Engler. | VELA: hanit pei —— = PUNTARENAS & ALAJ Rio Aquilar on road to Lake pmi cisci Zapote, A. Smith 928 (F). € "Chil ntera, Standley 35575, 35663 (US near Car Fm dillera de Talamanca, William es uae, is 1433 ; o Tajam echin, Lent tabov Patillos, SE of Tapanti, pig 082 (CR, F); p | 03 (F); Tapanti Hy M Croat 36069, 36211, 36216 MO): 108 7 MO); TE to Quebrada Platanillo, Croat 3 pantí Watershed Preserve, SW of Pei e ue Baker 9475 (F, MO); vicinity Ri ara Blan ti Cariblanco o. Moore ore 6613 Cur o Las Vueltas, NE of Volcán Barba, 47518 (CR | P. vicinity of Bajo La Hondura, Croat l 7 1983] MO, US). HEREDIA & SAN JOSE: Cerro de Zurqui, Burger & Baker 9323 (F). SAN JOSE: Valley of Rio La Hondura, below La Palma, NE of San Jerónimo, Burger & Stolze 4921 (CR, F); Bajo La Hondura, Utley & Utley 4994 DUKE); forest above Rio La Hondura, Lent 3784 (F); between San Isidro del General and Dominical, Croat 35243, 35322 (MO). ~ Anthurium carnosum Croat & Baker, Brenesia 16 (Supl. 1): 36. 1979. Type: Costa Rica. San José and Cartago Provinces: along Inter- American Highway near Trinidad and Km 72 (about 20 km SE of El Empalme), cut over area, 2,600-2,800 m elevation; Burger & Baker 9545 (F-1831851, holotype, MO- 2385461, isotype). Epiphytic vine, rarely terrestrial; stems slen- der, canelike, erect or trailing, to 1 m or more long, 5-7 mm diam., with thin, grayish, smooth periderm; internodes 2-9 cm long, shorter near apex; roots slender, elongate, frequently branch- ing; cataphylls moderately thick, 2-6.5 cm long, with a subapical apiculum 2-3 mm long, encir- cling stem, persistent at upper nodes, drying thin, dark brown, deciduous except for a few pale, threadlike fibers. LEAVES with petioles 3-12 cm long, 2-3 mm diam., weakly sulcate; geniculum l-2 cm long; blades coriaceous, succulent, lan- ceolate-elliptic, narrowly elliptic or sometimes oblanceolate-elliptic, gradually long-acuminate apex, acute to attenuate at base, (7)9—15(17) cm long, (1.7)2.5-5(7) cm wide, broadest at or M the middle; both surfaces semiglossy, the Ower surface slightly paler; midrib convexly hes above, diminished and sunken before the es two thirds, raised throughout its length be- ; Primary lateral veins 9—13(20) per side, de- dns. midrib at 25°-50° angle, weakly sunken i Prominulous below, lesser veins incon- a - above, prominulous below; collective hod Ping from near the base, 2-7(12) mm esie e margin, obscure to weakly sunken above, oed raised below. INFLORESCENCE in ing, + equalling leaves; peduncle 7-13 cm Mt ster 2mm diam., 2^ to 2 times longer than lb es; spathe thin, pale green, broadly lanceo- " oblong, 2.5-4.3 cm long, 8-14 mm wide, i bi in the lower third, acuminate at apex, : © weakly lobed at base, inserted at ca. angle on peduncle; stipe 5-8 mm long in en 2-5 mm long in back, ca. 2 mm diam., nally Het green, becoming yellowish and fi- dia Tange in fruit, (2.5)3-5.5 cm long, 3-4 mm m. at base, 1.5-3 mm diam. at apex; flowers CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 263 rhombic to 4-lobed, to 4 mm long, to 1.5 mm wide, the sides sigmoid; 1-2 flowers visible in the principal spiral, ca. 2 flowers visible in the alternate spiral; tepals dark green, the lateral te- pals (1.5)2.5-3.5 mm wide, the inner margin convex; pistil emergent, green; stigma ellipsoid, ca. 0.3 mm long; stamens held just above tepals, the anthers yellowish-white, 0.3-0.5 mm long, 0.7-0.8 mm wide; thecae broadly ovate to sub- orbicular, weakly divaricate; pollen yellowish- white, drying white. INFRUCT ENCE pen- dent, the spathe persisting; the spadix 6-10 cm long, 8-10 mm diam.; berries ovoid, acute at apex, yellow becoming bright orange, 9-11 mm long, 6-8 mm diam.; mesocarp thick, gelatinous, pale orange; seeds usually 2, sometimes 3 or 4, broadly ovate, rounded on both ends, flattened, ca. 3 mm long, 2 mm wide, 0.8 mm thick. Figs. 29 and 30. Anthurium carnosum ranges from Costa Rica to Panama at elevations from (700)1,500 to 2,600(3,000) m. It inhabits premontane rain for- est, lower montane wet forest, lower montane, and montane rain forest life zone formations. In Costa Rica it is known from the Cordillera Cen- tral, Volcán Turrialba, as well as on the western end of the Cordillera de Talamanca in the vicin- ity of Tapantí and Cerro de la Muerte, but it is to be expected throughout the Talamanca range to Panama. The species is common at 2,400 to 2.600 m elevation on Cerro de la Muerte and shares this habitat with only one other aroid, the cordate-leaved Anthurium concinnatum Schott. In Panama, the species has been found more widely, being known from a number of localities in Chiriquí Province. Anth uriu i g d by its scan- dent habit, its thick, succulent leaves with strong- ly discolorous surfaces, and its usual montane habitat. It is closely related to A. flavescens Poepp. of Peru, but differs from that species in fewer lateral veins, a shorter peduncle, dark green te- pals (white for A. flavescens), and a longer, thin- ner spadix. Among Central American species, A. carnosum is most closely related to A. pittieri Engler and A. interruptum Sodiro. Anthurium carnosum differs from A. pittieri by its longer internodes, viney habit, non-decurrent spathe, and its thick, succulent leaves. No intermediates have been found and they do not appear to be variants of one another. Anthurium carnosum differs from A. interruptum in its higher altitu- Awtl 13111 FIGURES chamulense 30— ssp 30 oaxacanurr Suun carnosum C roat A Baker 1 Cre roat 48389 (Vy . Croat 48619.—31—32. Anthurium cerrobaulense Matuda, Croat 47640.— 33. Anthurium v9c [ Nadav) TVOINV.LOS INNOSSIN JHL JO STVNNV 0L 710A) 1983] CROAT dinal range, short-stipitate spadix, orange (vs. red) berry color, shorter internode length, and thicker leaves. The sectional placement of A. car- nosum is uncertain but it has been tentatively placed in section Xialophyllium. Costa Rica. No other location, O. Jiménez 29 (US). Probably CARTAGO or SAN JOSE: El Pára: TAGO & SAN JOSE: SE of El Empalme along the Inter- american X & E o3 3 ^ No ^4 Co 8 A e 3 oO ) = - a% 4 A D (US); near Finca La Cima, N of Copey, Standley 42635, 42682, 42726, 42732, 42747 (US); Lake Chonta, NE of Santa Maria de Dota, Standley 42161, 42307, 42313, 42364 (US); near Lake La Escuadra, NE of Copey, Standley 41956, 41990, 42060, 42061 (US); near Rio a” N of Santa María de Dota, Lent 3911 Anthurium cerrobaulense Matuda, Anales Inst. Biol. Univ. Nac. México 30: 112-113. 1960. TYPE: Mexico. Oaxaca: Cerro Bail (on the border of Oaxaca and Chiapas NW of Rizo de Oro located on Hwy. 190), November 15, 1957, MacDougall 276 (MEXU, holo- type). Epiphyte or on rocks, rarely terrestrial 17-40 m tall; stems 2.5-4 cm diam.; leaf scars 1-2 cm Wide; roots 4-5 mm diam., tan, descending; cata- i: ylis 2-5 cm long, green tinged with red-violet, t " ed at apex with a subapical apiculum ex- ending beyond the apex ca. 3 mm, drying me- lum brown and i gi us LEAVES ding reading; petiole 15-45 cm long, 3-6 mm s broadly and sharply sulcate adaxially, vi abaxially (sometimes tinged with red- a et); seniculum 1-1.7 cm long; blades ovate Re. at base, (10)23-37 cm long, 13-18.5 cm td €, broadest at point of attachment or just Ow (often broader at base when flattened); an- Ne Eius lobe 1425.5 cm long, the margins convex OF MEXICO AND MIDDLE AMERICA 265 semiglossy; midrib convexly raised to midway, nd below, the posterior rib naked, the outer interprimary veins sunken scarcely visible; collective vein arising from the first basal vein, sometimes loop-connected with the third or fourth basal vein, 0.5-1.5 cm from margin, sunken above, raised below. INFLO- RESCENCE erect-spreading; peduncle 22-40 m long, 4-5 mm diam., terete, pale green; spathe low; i ale green sometimes tinged with red at base of spadix; spadix green, heavily tinged red-violet along inner edges of tepals, 4—4.4 cm long, 7~8 mm diam. at base, 4-5 mm diam. at apex; the flowers 4-lobed, ca. 3 mm long, 3.5 mm wide, the sides sigmoid; 5-8 flowers visible per prin- cipal spiral, 3-6 flowers visible in the alternate spiral; tepals matte, with copious droplets when stamens are emerging; lateral tepals 1.2-1.5 mm wide, the inner margin turned up; pistil slightly emergent, purple; stigma a shallow depression; stamens developing from the base, the laterals followed by the alternates in rapid succession, with the leading spirals preceding the alternates by 1 or 2 spirals; stamens held above pistil on transparent filaments 0.4-0.6 mm long, 0.6-0.9 mm wide, then retracting to edge of tepals before next one appears; anthers pale yellow; thecae + oblong, somewhat divaricate; pollen cream, soon dispersing. INFRUCTESCENCE spreading-pen- dent; spathe persisting; spadix 9.5-12 cm long, 2.3 cm diam.; immature berries emergent ca. 4 mm, dull violet-purple; mature berries red, to red-orange, obovoid, illif pex, 1.5- 1.6 cm long, 1-1.7 cm wide, mesocarp thickly gelatinuous to mealy with numerous short raph- ide cells; seeds 1-2, obovoid, green to yellow- green, 9-10 mm long, 7.6-8 mm wide, flattened, with a flattened truncate gelatinous appendage at the base. Figs. 31, 32, and 34. 4 Anthurium cerrobaulense is restricted to Mex- ico in the SW corner of Chiapas in the Munici- ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 DF | X OAM » | elonenst FIGURES 34-37, 34. Anthurium cerrobaulense ` ium cerrop Matuda, Croat 48180. ense Matuda, Croat 47640.—35-37. Anthurium 1983] pios of Cintalapa and Arriaga at elevations of 830 to 1,600 m in “bosque pino-Encino.” The species is tentatively placed in section Be- lolonchium and is distinguished by its thick leaf blades with usually prominent, elongate, narrow posterior lobes that are directed usually upward at an angle of 45? to 90? to the midrib, and also by its relatively much paler lower blade surface with only obscurely visible tertiary veins. Anthurium cerrobaulense is most easily con- fused with 4. chamulense, which has leaves of similar coloration and texture, but that species is distinguished by having the blades merely sub- cordate or by having the posterior lobes merely rounded. MEXICO. CHIAPAS: north of Arriaga, Breedlove & Thorne 30579 (DS); vicinity Cerro Baúl, Breedlove 21695, 24716, 31356 (DS); Croat 47640 (MO); east of San Pedro Tapanatepec, Breedlove 9975 (DS). Anthuriu pel Matuda, Bol. Soc. Bot. México 24: 37. 1959. Type: Mexico. Oa- xaca: road between Cerro Pelón and Yetla, 2,600 m, MacDougall 405 (MEXU). Epiphyte or on rocks; to ca. 1 m tall; stems 4- 8 cm diam.; leaf scars 4-6 cm wide; roots thick, tan, descending; cataphylls coriaceous 9-13 cm long, acuminate at apex, subapical apiculum 1 cm long, drying dark brown, persisting intact at apex, splitting at base into coarse fibers. LEAVES àt apex (the acumen minutely apiculate), deeply lobed at base, 30-65 cm long, 20-50 cm wide, broadest near point of petiole attachment, the margins somewhat irregular and undulate; an- terior lobe 23-50 cm long, the margins convex; MES Y Inbes 7-20 cm long directed upward or at the sinus spathulate, acute at apex; both sur- aces matte to semiglossy; midrib convexly raised above and below, diminishing and sunken at apex above; basal veins 8 pairs, the first and second (sometimes third) free, up to 5 coalesced 1-2 cm, ised above and below; posterior ribs not mar- ee tissue turned up at margins; primary 3s veins 5-7 per side, departing midrib at angle, sunken above, raised below, arcuate- s ums on lower surface; collective vein ing from first basal vein or one of the up- CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 267 permost primary lateral veins, 10-18 mm from margin, sunken above, raised below. INFLO- RESCENCE spreading, shorter than leaves; pe- duncle 22-45 cm long, 8 mm diam., terete, '^— 24 as long as petioles; spathe subcoriaceous, me- dium green (B & K Yellow-green 6/7.5), tinged at margins with purple or very heavily tinged with purple inside, narrowly ovate, 9.5 cm long, 2.3-4 cm wide, broadest near base, caudate- acuminate at apex (the acumen ca. 1 cm long), obtuse to rounded at base and clasping, inserted at 30? angle on peduncle; spadix violet-purple, sometimes green turning violet-purple (B & K Red-purple 2/5), 5-11 cm long, 0.8-1.5 cm diam. at base, 5-7 mm diam. at apex; flowers square, 2.7-3 mm in both directions, the sides weakly sigmoid; 10-14 flowers visible in the principal spiral, 8-12 flowers visible in the alternate spiral; tepals matte, with dense, pale punctations, mi- nutely papillate; lateral tepals 1.8-2.2 mm wide, the inner margin convex; pistil not visible before anthesis; stamens developing from the base in a complete sequence, the leading lateral stamens 1-2 spirals ahead of the fourth stamen; anthers ca. 1.2 mm long, 0.5 mm wide; thecae ovate- oblong, scarcely divaricate. INFRUCTES- CENCE pendent; spadix to 15 cm long, 4 cm diam.; berries obovoid, mammilliform at apex, orange (?), 9-10 mm long, 7-9 mm diam.; me- socarp fleshy, thick; seeds usually 2, 7 mm long, 4.5-5.5 mm wide. Figs. 35, 36, and 37. Anthurium cerropelonense is endemic to the Sierra de Juárez in the Sierra Madre Oriental in Oaxaca at 2,400 to 2,800 m in cloud forests near the summit of the Atlantic slope. The forest type is “selva alta perennifolia.” All collections have been made along Highway 175. The species is characterized by its large, thick, ovate-cordate blades, narrow or closed sinus, posterior rib rather distant from the leaf margin, and by conspicuous tertiary veins, and also by its thick, dark violet-purple spadix. Unlike most Mexican species, which appear to be related to several other species, A. cerro- pelonense seems to be unrelated to or confused with any other species. It is tentatively placed in section Belolonchium. MEXICO. OAXACA: between Cerro Pelón & Yetla, MacDougall 405 (MEXU); vicinity of Ixtlán de J uárez, Almeda & Luteyn 1666 (DUKE); Graham & Frohlich 1105 (MICH); between Tuxtepec & Oaxaca, Croat 48180 (MO). 268 Anthurium chamulense Matuda, Anales Inst. Biol. Univ. Nac. México 27: 339. Fig. 2. a. Anthurium chamulense ssp. chamulense. TYPE: exico. Chiapas: San Juan Chamula (just of San Cristóbal de Las Casas) MacDougall 348 (MEXU, holotype—not found). Anthurium motozintlense Matuda, Anales Inst. Biol. Univ. Nac. México 43, Ser. Bot. (1): 55. Fig. 4. 1972 (Publ. 1974). Type: Mexico. Chiapas: Sierra Madre, El Rosario, Motozintla, 1,800 m elev. MacDougall 393A (MEXU, holotype). Anthurium monserratense Matuda, Anales Inst. Biol i éxic 2. Fig (MEXU, not seen). Epiphyte or on rocks; roots tan (dry), ca. 2 mm diam.; cataphylls sul i ,3.5-10 cm long, rounded at apex (subapical apiculum ca. 2 mm long), drying medium to dark brown (B & K Yellow-red 4/10), persisting as reticulate fibers. S usually erect; petioles D-shaped, 14— 32(47) cm long, 2-6 mm diam., sharply and broadly sulcate; geniculum 1.5-2 cm long; blades coriaceous, narrowly to broadly triangular, 13- 37 cm long, 5-26.5 cm wide, broadest 1-2 cm above base, acute to narrowly rounded and apic- ulate at apex (the apiculum 1—2 mm long), trun- cate to obtuse or broadly lobed at base; anterior lobe 13.5-30 cm long; posterior lobes 3.2-12 cm long; sinus arcuate to arcuate with decurrent pet- iole, rounded at apex; upper surface semiglossy, the lower surface matte to semiglossy; midrib bluntly raised at base, diminished and sunken at apex above, raised at base, flat at apex below; basal veins 3-5, the uppermost free (sometimes all free), third or fourth coalesced to 3.5 cm; posterior rib naked, turned up; primary lateral veins 7-19 per side, departing midrib at 45? an- gle, sunken but scarcely more conspicuous than lesser veins above, prominulous and darker than surface below; submarginal collective vein aris- ing from first or second basal veins; third basal vein loop-connected to collective vein, 2-15 mm from margin. INFLORESCENCES erect to spreading, shorter than or as long as the leaves; peduncle 27-48 cm long, 3 mm diam .; spathe moderately thin, pale green tinged pink at base and apex, ovate to ovate-lanceolate, 3.5-14 cm long, 1.3-6 cm wide, broadest just above base, abruptly to graduall i tapex, rounded to truncate at base, inserted at 10° angle on pe- ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 duncle, the area around stipe thickened, donut- shaped; stipe 3 mm long in front, 1.5 mm long in back, green; spadix green becoming slightly to heavily tinged with violet-purple (B & K Red- purple 2/5), 3-13 cm long, 5.5-10 mm diam. at base, 3-6 mm at apex; the flowers rhombic, 22- 2.3 mm long, 2.4-2.5 mm wide, the sides straight to broadly sigmoid; 5-6 flowers visible in the principal spiral, ca. 5 flowers visible in the al- ternate spiral; tepals matte, sparsely punctale, lateral tepals 1.6 mm wide, j quence beginning at base, sometimes with as — many as 3 emerging simultaneously, the leading stamens no more than 3 spirals ahead of com- pleted sequence; anthers cream-colored, me a tight cluster obscuring pistil, 0.8 mm long, Ù. mm wide, shrinking to 0.4 mm wide; thecae ovate, slightly divaricate; pollen white. IN- | FRUCTESCENCE pendent, 7-10 cm long, 1 i i vate cm diam.; spathe persistent, berries red, obo to round; seeds not seen. Fig. 38. Mex- ies i nly from southern The species is known only ue of two ico in Oaxaca and Chiapas and on ec subspecies. Subspecies oaxacanum E known only from the type MU n" m pi in “selva alta perennitolia, ™ Swan P; e : E 1 ead in Chiapas and cnecies Ex widespr is expected also in Guatemala. It ranges E 1,500 to 3,000 m and is perhaps rest regions of “bosque pino-encino. Mos u tions have been made from northern Chia the species has also been collected 1n the EA icipio of Motozintla in southeast e also in the Municipio Villa Corzo in sou Chiapas. One of the collections fro collecting area (MacDougall 376) 1s " aberrant in that it has a thicker blade wl conspicuous posterior lobes. This haps represents a distinct species incomplete to be certain. Anthurium chamulense is tentatively P byits section Belolonchium and is distinguls usuall coriaceous, ovate-triangular leaf blade U57 bes and broad sinus pE decidi ith the hibit a characteristic vein pattern with acodit | i isi e lower S | veins scarcely visible on th 20 e upt“ clearly visible and minutely etched 0 surface CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA series Geren SSp. oax- Croat 409 & FIGURES 38-41, 38. Anthurium Oe ETa. Croat 46426.—39. a VAM Croat 48389 (Type). —40—-41. Anthurium chiapasense Standl., 270 The species is probably closest to A. andicola, but that species can be distinguished by having usually conspicuous tertiary veins on the surface that are weakly elevated and darker than the sur- face. Anthurium chamulense is also very similar to A. cerrobaulense from southwest Chiapas be- cause they both have leaves with similar color- ation and texture, but that species has blades with usually conspicuous narrow, rather elongate, posterior lobes. The type of Anthurium monserratense Matuda was not seen but the description, illustration, and the location of the collection make it rather con- clusively synonymous with A. chamulense. Mexico. CHIAPAS: NE of Bochil, Breedlove 34643 (DS); Cerro Tres Picos, Breedlove 24972, 30125 (DS); MacDougall 376 (MEXU); El Rosario, MacDougall 393A (MEXU); Matuda 37603 (DS, MEXU); near Ocosingo, Breedlove 16021, 16033, 16034 (DS); Pueb- lo Nuevo Solistahuacán, Croat 46426, 47765 (MO); Luteyn & Almeda 3443 (DUKE); Mill 220, 595, 673 5 (DS; M unc. Zinacantán, Laughlin 536 (DS, DUKE, ENCB). b À nth ` h 1 p. oa ium Croat, ssp. nov. TYPE: Mexico. Oaxaca: along road from Teotitlán del Camino to Santa María Chilchotla, 4.4 mi beyond turnoff to Huatla de Jiménez, elev. 1,980 m, steep slopes with huge boulders, Croat 48389 (MO-2682519, holotype; F, K, MEXU, SEL, US, isotypes; Live at MO). Anthurium chamulense ssp. oaxacanum a ssp. cha- mulense laminis tenuioribus, nervis tertiaribus supra vix impressis subtus manifestis differt. Epiphyte; stems narrow, 1.5 cm diam.; leaf scars 1.2 cm wide; roots thick, brown; cataphylls subcoriaceous, 4-7 cm long, drying medium wn ( K Yellow 4/10), persisting as retic- ly sulcate, the margins sharper near base of blade; geniculum 1-1.3 cm long, sometimes tinged red- violet; blades subcoriaceous, triangular, 24-48 cm long, 8-13 cm wide, broadest 1-2 cm from base, gradually acuminate at apex, obtuse to truncate at base; anterior lobe 23.5-27 cm long, the margins straight; posterior lobes 4-5.5 cm long, directed upward; sinus arcuate to arcuate with petiole decurrent, rounded at apex: upper surface glossy, lower surface matte; midrib blunt- ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 ly raised at base, sunken to flat at apex above, raised below; basal veins 3-4 pairs, the second . to fourth coalesced 1—1.5 cm, prominulous above and below; posterior ribs naked, the outer mar- gins rolled up; primary lateral veins 3-6 per side, departing midrib at 30*-45? angle, sunken above, raised below, straight to collective vein; lesstr veins scarcely visible; collective vein arising from the first or second basal vein, running to the apes, 3-5 mm from margin, sunken above, raised be- low. INFLORESCENCE spreading, shorter or as long as leaves; peduncle terete, green, tinged with red-violet, 30-32 cm long, 3-4 mm diam., spathe green, subcoriaceous, + triangular, ca. 3 cm long, 1 cm wide; spadix violet-purple, tan at anthesis, 4.5-5 cm long, 5-6 mm diam. at base, 3-4 mm diam. at apex; flowers rhombic, 3 mm tepals matte, weakly punctate, ae mi late, lateral tepals ca. 1.5 mm wide, Pn sgh margin convex; pistils weakly emergent, $ linear; st emerging from base, laterals first, greenish white, obovate, depressed a at base, flattened, ca. 7 mm long, 9 F 3.1-3.4 mm thick, enveloped in a us : substance, slightly longer on ends than 2 of seeds. Figs. 33 and 39. Anthurium chamulense SSp. oaxacanii x known only from the type locality Oaxaca in "selva alta perennifolia 2,000 m. The taxon is recognized by its thinner, usually truncate, narrowly a leaves, sulcate petiole, green spathe an usually stipitate, violet-purple spa member of section Belolonchium. The subspecies oaxacanum is not Do: with any other taxon from that part © i It is separated from A. chamulense P mulense by having thinner leaf blades conspicuously impressed tertiary ve upper surface but with the tertiary vem lower surface clearly visible. In subspe" wo mulense the tertiary veins on the lower ” at n short, on the relatively :angulat dix. It i5 à | ——————— | in northem | on be es chê -— 1983] are scarcely visible. In addition the subspecies a alta perennifolia”) than ssp. chamulense (*bos- que pino-encino"). MEXICO. OAXACA: between Teotitlán and Santa Ma- ría Chilchotla, beyond turnoff to Huatla de Jiménez, Croat 48389 (F, K, MEXU, MO, SEL, U Anthurium chiapasense Standl., Publ. Field Mus. Nat. Hist., Bot. Ser. 22: 67. 1940 a. Anthurium chiapasense ssp. chiapasense. TYPE: Mexico. Chiapas: Monte Ovando (just NW of Escuintla) Matuda 1562 (MO, holotype). Anthurium cuspidatum Matuda, Madroño 10: 169. 1950, non Mast., 1875. Anthurium cruxiphyllum oc. Mex. Hist. Nat. 11: 93. 1950. porke near Colonia Jalapa, 32 km E of Escuintla, elev. 700 m, Matuda 18316 (MEXU, holotype; F, isotype). _ Epiphyte; stems 2-4 cm long, 2.5 cm diam.; internodes and leaf scars obscured by cataphylls; roots moderately thick, green, descending; cata- phylls subcoriaceous, 3-6 cm long, acute at apex (apiculate to 2 mm), drying brown (B & K Yellow 3/10), weathering to reticulate fibers and per- sisting. LEAVES erect-spreading; petioles 10-51 moderately thick, narrowly triangular, long acu- minate at apex, broadly lobed at base, 13-42 cm long, 5.5-25 c m wide, broadest at base; anterior lobe 11.5-36 cm long, the margins + straight; posterior lobes 2.5-10.5 cm long; sinus arcuate, round at apex; both surfaces matte, lower surface densely glandular-punctate; midrib acutely raised at base above, flat at middle, sunken at apex, acutely raised throughout below; basal veins 3- 4 pairs, the third and fourth coalesced 1.5-2 cm, flat to weakly sunken above, raised below; the Posterior rib naked, the outer margin turned up; Primary lateral veins 2-3 per side, departing midrib at 30° angle, weakly sunken, nearly ob- scure above, prominulous below, + straight to collective vein, lesser veins obscure; collective vein arising from the uppermost basal vein, 0.7— l.5 mm from margin, sunken above, aed be- low, INFLORESCENCE erect-spreading; as long 5-4 cm diam., terete; spathe green, tinged with Violet-purple, lanceolate; 4.5-7 cm long, 1-1.8 em wide, cuspidate-acuminate at apex, rounded at base, broadest j just above base, inserted at 50° angle on peduncle; spadix deep purple (B & K CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA aii Red-purple 2/2.5), 7.4-10.5 cm long, 5.5-8 mm diam. at base, 3—4 mm diam. at apex, the flowers square to sub-4-lobed, 2.3-2.5 mm in both di- rections, the sides straight to jaggedly sigmoid; = flowers visible in x the principal spiral, 778 TR z pm to weakly glossy, obscurely punctate, minutely papillate; lateral tepals 1.2-1.5 mm wide, the inner margins broadly rounded; pistils sa e not emergent until just before stamens emerge, green tinged with purple, weakly papillate; stig- ma linear, ca. 0.3 mm long, scarcely distinguish- able; stamens emerging from base, lateral sta- mens emerging to midway, before alternates emerge at base; anthers yellow, 0.5 mm long, 0.6 wide, held just above tepals; thecae oblong- iunii bee a divaricate; pollen yellow, fading UCTESCENCE not known. Figs. pi d i The species consists of two subspecies. An- thurium chiapasense ssp. chiapasense occurs in Mexico in southeastern Chiapas and southwest- ern Guatemala at elevations of 800 to 1,900 m. The type locality is probably tropical wet forest and is mapped as "bosque de oyamel" (Flores et al., 1971) on the Mexican vegetation map, but the species also occurs in “bosque cauducifolio" and “bosque pino-encino. acense (Matuda) Croat is endemic to southern Oaxaca and differs from the typical subspecies in having an olive green spadix at anthesis and blades with the primary lateral veins promi- nently raised on the lower surface. Anthurium chiapasense is in section Belolon- chium and is characterized by its thick blades that are much paler and conspicuously punctate on the lower surface with an oblong or narrowly triangular anterior lobe, 3-9 cm wide midway. The species is closely related to A. lucens and distinct from isl sid . Although both species share a rel- atively short, dark violet-purple spadix, subco- riaceous, glandular-punctate leaf blades, and bright red berries, A. chiapasense differs from A. lucens in having slender anterior lobes with con- cave margins. The margins of the anterior lobe of A. lucens are generally conspicuously convex. Anthurium chiapasense has a leaf shape more similar to that of A. verapazense, which also has subcoriaceous, glandular-punctate leaves. It dif- fers from A. verapazense in having a porpor- tionally much shorter spadix with the spathe more than half as long as the spadix at anthesis. In 272 addition, A. verapazense occurs from sea level to less than 1,000 meters, whereas A. chiapasense occurs from 800 to 1,800 m. Typical A. vera- pazense has a long, slender MEE well over twice as long as the spathe at anthes Anthurium chiapasense iden from both A. lucens and A. verapazense in usually oe blackened while the other t greenish to greenish-brown. It is Venatively iced in section Belolonchium. GUA LA. SAN MARCOS: Finca Armenia above San Rafael, pono 40900 (M T 9 MEXICO. uintla, Matuda 18315 (DS, AS, MEXU, NY); 9, 18316, 18363, 18491 (MEXU); Matuda 18573 (CAS, DS, MEXU); Matuda 18629 (MEXU, NY); Montebello, SE of Comitán, en s.n. (MEXU); Monte Ovando, Matuda 1562 (MEXU, MICH, MO); Matuda 17930 (MEXU); Mo- ioni Breedlove & Thorne 31026 (CAS, DS); Por- venir, Matuda 30186 (MEXU). MÉxico: San Angel, De la Cerda 81 (MEXU). b. Anthurium chiapasense ssp. tlaxiacense (Ma- México 32: 149, Fig. 6. 1961. TYPE: Mexico. Oaxaca: Cerro de la Virgin, Tlaxiaco (just E of Tlaxiaco); elev. 1,500 m. MacDougall s.n. (MEXU, holotype). Epiphytic or epipetric; stems to 6 cm long, 2 cm diam.; internodes and leaf scars obscured by cataphylls; roots moderately thick, green; cata- phylls subcoriaceous, 6.7-8 cm 3 IS ies “a - = G D Lad E © H 5 do T e o o o S Lon | Uu e ig 2 "0 & Fy sisting around stem. LEAVES erect- -spreading; petioles 37-64 cm long, 6-8 mm diam., + terete: geniculum 2.5—4 cm long, sometimes weakly sul- cate; blades subcoriaceous, narrowly triangular, 35-56 cm long, 13-30 c m wide, acuminate at apex (apiculate 5 mm), decals lobed at base, broadest at base; anterior lobe 28-45 cm long, the margins round to convex; posterior lobes 9— 12 cm long, directed upward: sinus hippocrepi- form, acute at apex; upper surface matte to semi- glossy, lower surface matte, glandular-punctate; midrib flat to convexly raised above, weakly sunken at apex, prominently raised below; basal veins 4—5 pairs, the third to fifth coalesced 1.5 cm, flat to weakly sunken above, raised below; posterior rib naked, turned up; primary lateral veins 5-6 per side, departing midrib at 209-25» angle, sunken above, raised below, straight to collective vein; lesser veins obscure above, ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 prominulous below; submarginal collective vein below. equalling leaves; peduncle 53-57 cm mm diam., terete; spathe subcoriaceous, yellow- green (B & K Yellow-green 7/5), lanceolate, 6- 7 cm long, 1.5-1.7 cm wide, margins thin, trans- parent, soon withering, inserted at 40? angle on peduncle; spadix dark yellow-green to gray-green, heavily tinged red-violet at outer margins of te- pals, 5.5-8.5 cm long, 6-9 mm diam. at base, 3- 5 mm diam. at apex; the flowers 4-lobed, 2.3-3 mm long, 2.4-2.8 mm wide, the sides sigmoid; 6-8 flowers visible in the principal spiral, 6-7 flowers visible in the alternate spiral; tepals matte, weakly punctate, densely papillate, lateral tepals 1.3-1.6 mm wide, the inner margin convex; pn raised, green, glossy; stigma elliptic, 0.4-0. 5m long, deep purple with large globose droplet ca. 1 week before stamens emerge; stamens emerg- ing from base, lateral stamens emerging through- out, alternates following within a few days. exserted well over pistil on fleshy, red- violet tinged filaments, retracting when anthers open, filaments ca. 5 mm long, ca. 6 mm wide, anthers pale yellow, ca. 0.6 mm long, 1.9 mm wi de; the- cae oblong, scarcely divaricate; pollen yellow fading c cream. INFRU eii ESCENTS arching- pendent; 10-12(20) cm long, 1.5-3. 5 cm i diei: (5-6 cm diam. with berries fully emerged); berries + we : pe Red 4/5); eked tratisiübent merous raphide cells; seeds 1-2, ye with purple, weakly flattened, 4.5-5 m ca. 3 mm wide, 1 mm thick, enveloped na parent, amber, gelatinous sack. Figs. 42, 43. an 6. Anthurium chiapasense ssp. tlaxiacense is en- demic to southern Oaxaca and has been collect at 1,400 to 1 500 m in seasonally dry forest. Anthurium p. tlaxiacense is dis tinguished from M WEE by its olive. green lateral veins prominently raised on E face. It is recognized by its coriaceous, ney triangular-hastate leaf blade, dark green and much paler and conspicuously &l punctate beneath. and MEx AXACA: between Pinotepa Nacio - Tasia 3 Croat 45927 (MO); S of Mise 126 P (MO: between a and Pochutla, Tlaxiaco, Mac Dongail s.n. (MEXU) andulàr- w — — -—A—— uie D CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA Fig : - CRUS 5 » 44. in thuriun URES 42-45. 42-43. A nthurium chiapasense ssp. tlaxiacense (Matuda) Croat, Croat 46126.—44. An m clarinervium Matuda, Selby 9-75-59.—45. Anthurium clarinervium Matuda, Croat 40646. > A Z > o e| 7 — len tm <= Z) na e! G 2 w le, n > Z O > E Q > x g m r- Fu Sones per 46. Anthurium c Was rendo ssp- eae ense (Matuda) Croat, Croat 46126.—47—48. Anthurium clavatum Croat & Baker, Croat stitial as Anthurium clavigerum roa 1983] CROAT— Anthurium clarinervium Matuda, Anales Inst. Biol. Univ. Nac. México 25: 203. 1954. TYPE: Mexico. Chiapas: NE of Ocozocoautla (col- lected by Ing. Ramirez August 5, 1951) Ma- tuda 25894 (MEXU, holotype). Terrestrial, deeply rooted among rocks, less frequently epipetric in thin soil on rocks; stems 1-2 cm diam.; leaf scars 1.7 cm wide; cataphylls subcoriaceous, 3-4 cm long, obtuse at apex with subapical apiculum, pale red-violet with pale green speckles, drying brown (B & K Yellow 3/2.5), splitting and persisting. LEAVES erect to spreading; petioles 7-16 cm long, 4-6 mm diam., terete, green or pale red-violet with green speck- les; geniculum 1—1.5 cm long; blades ovate, mod- erately coriaceous, abruptly to gradually acu- minate at apex, deeply lobed at base, 11—25.5 em long, 6.5-17 cm wide, broadest at middle or Just below, the margin flat, drying conspicuously undulate; anterior lobe 6.5—18 cm long, the mar- èns convex; posterior lobes 2.5—9 cm long; sinus spathulate to obovate, the lobes sometimes over- apping, acute at apex; upper surface matte, vel- vety, dark green, lower surface paler, matte; mid- rib convexly raised, diminishing at apex above, cols ascending to margin, loop-connecting; “ective vein arising from first basal vein; all veins clearly outlined by whi ish whi y white or greenish white area. INFLORESC vane violet-purple near margins (B & K is ‘hag 6/7.5), lanceolate, 3.7—6.5 cm long, misié i wide, broadest at base, narrowly acu- apex, obtuse at base, inserted at 40°- Ms cA on peduncle; stipe to 5 mm long in With viol mm long in back; spadix green, tinged (B&K oe sometimes appearing brown 15 cm lo so Breen 5/5 to Yellow 4/10), 4.7- Ong, 4-7 mm diam. at base, 3-3.5 mm 5 S flowers sub-4-lobed to square, 3.8- sigmoid: ng, 3.2-5 mm wide, the sides jaggedly : 3-4 flowers visible in the principal spi- a lateral tepals 2.8-3.1 mm wide, the inner Weakly concave; the pistils weakly emer- OF MEXICO AND MIDDLE AMERICA 275 gent usually as soon as spathe opens, green; stig- ma elliptic, 0.6-1 mm long, brushlike, exserted with large droplets 2-3 days before stamens emerge; stamens emerging rapidly from base on flattened, translucent filaments 1 mm long, 1.5 mm wide, lateral stamens emergent to midway before alternates appear, retracting and held just above tepals; anthers yellow, mm long, 0.7-1.5 mm wide, thecae elliptic, not divaricate, pollen cream fading to white. INFRUCTES- CENCE with bright orange, subglobose berries, ca. 1.5 mm long; ly, seeds not Figs. 44 and 45. E Anthurium clarinervium is probably endemic to Mexico, known only from areas with lime- stone outcrops between the Río Grijalva and Río de la Venta, north of Ocozocoautla de Espinosa, Chiapas, at 800 to 1,200 m elevation. Mr. Larry Bussle, a Tampa based horticulturist, claims to have seen Anthurium clarinervium along the In- teramerican highway in Guatemala, near the Mexican border, but no collections were made to substantiate this report. The species is a member of section Cardiolon- chium and is recognized by its broadly ovate, velvety leaf blades with all principal veins much paler than the surface and with a very narrow or closed sinus. The blades have basal veins usually free to the base or coalesced into a short posterior rib that is never marginal. The species has been confused with A. /euco- neurum, which also purportedly originates from southern Mexico but whose exact origin is un- known. The latter species is distinguished by having a generally larger, more narrowly ovate blade with an obovate or hippocrepiform sinus and a conspicuously well developed, naked, pos- terior rib (i.e., it is marginal to the sinus). Anth m clarinervium often occurs in mixed stands with A. pedatoradiatum, a species that flourishes in the same limest tcrop habitat in southern Mexico. The two species readily hy- bridize to form a somewhat intermediate plant, having the velvety texture of A. clarinervium but displaying a somewhat shallowly lobed margin representative of A. pedatoradiatum (see Croat 40647 In Chiapas the species is known as “Hoja de corazon.” It is now widespread in cultivation, at least in the United States. *hiiriii MEXICO. CHIAPAS: N of Berriozábal, Breedlove 20274, 24822 (DS); Breedlove & Dressler 29713 (CAS, DS); 276 vicinity Ocozocoautla de Espinosa, Breedlove 9990 (DS, XU); Breedlove 10365 (DS); Breedlove & Raven 13582 (CAS, DS); Breedlove 38172 (CAS, DS); Croat 40646 (MO); Matuda 25894 (MEXU). Anthurium clavatum Croat & Baker, Brenesia 16 (Supl. 1): 39. 1979. Type: Costa Rica. Car- tago: Tapanti Hydroelectric Reserve along Rio Grande Orosi, 4.5 km beyond small bridge which crosses the river inside the pre- serve, along road to diversion dam, elev. 1,500-1,700 m, Croat 36079 (MO-2381175, holotype; CR, F, NY, US, isotypes; Live at MO). Epiphytic or terrestrial; stems to 1 m long, 2— 3 cm diam.; internodes 1.5—2.5 cm long; roots 1 or 2 per node, ca. 5 mm thick, woolly-pubescent on drying; cataphylls moderately thick, 13-28 cm long, narrowly rounded at apex with a sub- apical apiculum, drying reddish-brown (B& K Yellow-red 4/10), subpersistent. LEAVES erect- spreading; petioles 90-140 cm long, 1-1.5 cm diam., subterete, narrowly and obtusely sulcate; geniculum 3-3.5 cm long; blades ovate, mod- erately thick, short-acuminate at apex (the acu- men minutely apiculate), deeply lobed at base, 35-70 cm long, 26-50 cm wide, broadest near point of petiole attachment; anterior lobe 25-48 cm long, the margins broadly convex; posterior lobes 12-22 cm long, directed downward; the sinus hippocrepiform, rounded to obtuse at apex; tds surfaces semiglossy; midrib convexly raised ve, diminished and sunken toward apex, Mid raised below; basal veins 5-7 pairs, up of them coalesced 2.5-6 cm, raised above midrib at ca. 55? angle, then spreading at 35% m m ESCENCE usu- pd slightly shorter than or about as long as the leaves; peduncle 43-110 cm long, to ca. 8 mm diam. midway, half as Fs as or equalling " lyc ous, pale gr (B & K Yellowgreen 8/ 7. 5), “ovate Iie 7- Ü cm long, 3.5-5 cm wide, broadest 1-2 cm above point of attachment, cuspidate-acuminate at apex (the acumen inrolled, 6-10 m m long), acute to obtuse at base, inserted on peduncle at ca. 80? angle; stipe 1-6 cm long in front, 2-8 mm long in back, ca. 6 mm diam.; spadix clavate, green (B & K Yellow-green 8/ 10) to orange-tan, 4,5. ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 7.5 cm long, 1.2-1.5 cm diam. at base, 1.5-2.2 cm diam. at apex; flowers 4-lobed, (1.4)2.3-2.7 mm (dry) in both directions, the sides sigmoid; 20 or more flowers visible in the principal spiral, 13-16 flowers visible in the alternate spiral; te- pals semiglossy, the lateral tepals 1.2-1.3 mm wide (dry), the inner margin broadly rounded; pistil emergent but not raised, green; stigma ob- long-elliptic, slitlike, ca. 0.5 mm long; stamens emerging, scattered, beginning near the middle, the lateral stamens exserted throughout the spa- dix before the alternates emerge; anthers white, held in a moderately close circle, contiguous or nearly so at upper edge of tepals, ca. 0.4 mm long and 0.5 mm wide; thecae ovate, moderately di- varicate; pollen white. Immature INFRU ES- CENCE to 10 cm long and 3.3 cm wide; berries not seen. Figs. 47 and 48. The species is known only from central Costa Rica on the Atlantic slope of the Cordillera de Talamanca in the region of the type locality. It inhabits lower montane rain forest from 1, to 1,800 m elevation. Anthurium clavatum belongs in section Cal omystrium and is perhaps most closely related to A. fraternum Schott, A. hoffmanii, and 4. nymphiifolium C. Koch & Bouché, which all share thick, persistent cataphylls. Anthurium clavatum differs from. all of these by its clavate sad Another feature is the broa with a long cusp. TA RICA. CARTAGO: no other m rande Tapanti, pcne et al. 6334 (F); Lent 940 0v 2 998 ( of Orosi, above Finca La Concor " linger & o 1508 (CR, F, US); road from Pes to Taus- Tausito, beyond bridge over Río Gran Ha- US), 47050 (MO); road between Trinidad am cienda El Volcan Turrialba, Luteyn 3259 (DUKE). dl., Anthurium clavigerum Poepp. in Poepp. d Nov. Gen. & Sp. 3: 84. 1845. TYPE: Pampayacu, Poeppig 1423 (W, not seem © photo) t. M panduratum Mart. ex Schott, One enbl. 5: 273. 1855. Type: Brazil. Rio Ma: us 3115 (M, not seen; US, photo). | Wo- Araha | holtonianum Schott, Oesterr. Bot- n Anthurium repandum Sc 7: 317. 1857. Type: Columba. die s.n. (K, not seen; SEL, photo). | —— M Ó€ 1983] Anthurium pandi Schott, Oesterr. Bot. Z. 8: 182. Bon Rica. San Miguel, Wendland Brun Anthurium karen N. E. Brown, Gard. Chron. pt. iod 881. E Colombia. Antioquia, Kal- brey n. (K, not seen). p ul E usum. T. WAS. Field Mus. Nat. Hist., Bot. Ser. 11 — .Panduratum var. dc EA Engl., V 23 B: 279. 1905. Type: Brazil. Para, Burchell ' ME (K, not seen). Anthurium obovatum Gleason, Bull. Torrey Bot. Club a Anthurium monsteroides Steyerm., Fieldiana, Bot. 28: 93 . Type: Venezuela. Méri ío Onia near nam N. of Mesa Bolívar, Steyermark 56727 Epiphyte; scandent; stems 2 m long, 3-4 cm diam.; roots dense, 3-10 mm diam., brownish- green; cataphylls coriaceous, to 12 cm long, weathering to medium brown fibers, ultimately deciduous. LEA VES erect- spreading; petiole 65— 150 cm long, terete or weakly sulcate; geniculum 1.5-2 cm long; blades digitately discoid to re- niform in outline, to 2 m wide, the leaflets 25- es cm or more long, 4—12 cm wide, sinuate or obate, acuminate at apex, acute at base, the up- a surface matte, the lower surface semiglossy; the midrib of the leaflets convexly raised above e below, diminished and sunken at apex above; ue lateral veins 3-4 per side, raised in is cuba above, raised below, departing midrib at he asi arcuate-ascending to form collective n apical portion of leaflet. INFLORES- Mis pee ne ndent; to 90 cm long, duas orter than petioles; spathe coriaceous, Ps Me De lanceolate, 30-65 cm long, 3-11 62,5) vip TPN lavender-purple (B & K Purple Mee m long, to 2 cm diam. at base, to ae 5 apex; flowers rhombic, ca. 3 mm in both dire sigmoid: z e the sides straight to slightly pg develo Pendent; berries obovate, purple, oblon ping in basal portion only, seeds 1-2, white, ng. Figs. 49 and 50. An an clavigerum is found from Nicara- ied, * Guianas, Brazil, and Bolivia. In Costa Occurs in premontane wet forest and trop- CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA PA ical wet forest life zones at sea level to 700 m. In Panama, the species also occurs in premon- tane moist forest and tropical moist forest. It has also been collected at a reportedly 1,000 m ele- vation in what is apparently lower montane wet forest or lower montane rain forest. This species is not confused with any other in Central America and can be recognized by its pedatisect leaves with sinuate to lobed margins and by its huge pendent inflorescence. The species may have leaf blades to about 2 m broad making it the largest species of Anthurium in Central merica. It is a member of section Dactylo- phyllium and is most closely related to A. sinu- atum Benth. ex Schott of South America. CosrA RICA. ALAJUELA: vicinity Artezalea, NE of Villa Quesada, Molino et al. 17843 (F, NY); between Cafias and Upala, NNE of Bijagua, Croat 36291 (MO); Burger & Baker 9963 ). LIMÓN: NE of Bribrí, aj als ied (MO); Hacienda Tapezco & Hacienda La (DU , MO, NY). P Burger oan Mas 4770 (F, NY); Playa Blanca Road near Rincón de Osa, Liesner 2153 (MO); SW of Rincón de Osa, Raven asd pde NY). AN: near Cano Chontaleno, NE of El Castillo, Neill 3378 (MO). ZELAYA: along Rio p N of Bluefields, Harmon & Fuentes 5083 (MO); vicinity Siuna, at Caño Majagua, Neill 1818 ecd Stevens dnd (MO); trail from Cerro Saslaya to San José del Hormiguero, Stevens 7067 (MO); S of et at test Agrícola Heel on road to Colonia Manantiales, Stevens "et (M to Colonia Agrícola Yolaina & Colonia La UE (MO); along Río Sucio, Sd 8028 (MO); m Puerto Cabezas to penes S Neill 4433 pe hia Rosita and Puerto Cabezas, Stevens O); NE of Siuna, along Cañ Madriguera, wa 3727 (MO). Anthurium clidemioides Standl., Publ. Field Mus. Nat. Hist., Bot. Ser. 22: 3. 1940. Type: Costa Rica. Alajuela: Villa Quesada, San Carlos, altitude 800 m, Valerio 1726 (F-933729, ho- lotype). Epiphyte; stem scandent; internodes elongate, 2-15 cm long, 23 mus thick, covered with corky, lenticel nt. LEAVES radiot petioles 2-5.5 cm long, sheathed half their length or less, the base encircling the stem, narrowly flattened | with a medial ridge once. tlv mnlti faintly Iuiti-115 J: 5 ANNALS OF THE MISSOURI BOTANICAL GARDEN E €————— a — aeg P alienam rz eS ewe N Y 53 2 ndl. gerum Poepp., Croat 4654.—51. Anthurium clidemioides 9! Figures 50-53. 50. Anthurium clavi e ~ , : tuos. Schott, Croat 26493. — 53. Anthurium cordatotriangulum w Daniels s.n.—52. Anthurium concinnatum Matuda 38276 (Type). 1983] CROAT 5 mm long, weakly tuberculate; blades ovate- cordate, bullate, chartaceous, acuminate at apex, lobed at base, 7-14 cm long, 4-9.5 cm wide, broadest 1-2 cm above point of petiole attach- ment; the anterior lobe 6-11 cm long; the pos- terior lobes rounded, 1-2 cm long; the upper surface matte; basal veins 4—5 pairs, sunken above; the lower surface paler; collective vein arising from the first basal vein. L S- CENCE spreading; peduncle obscured by base of spathe or rarely to 5 cm long; spathe green, lan- ceolate, ca. 5 cm long, ca. 1 cm wide; spadix purplish, 4.2-6 cm long, 5-7 mm diam. at base, 3-4 mm diam. at apex: the flowers rhombic to 4-lobed, 5.3 mm long, 3.5 mm wide; 3-5 flowers visible in the principal spiral, 4—5 flowers visible in the alternate spiral; tepals semiglossy; pistil weakly raised; stigma punctiform, ca. 0.6 mm diam. INFRUCTESCENCE intendi ing-pende berries orange, ns: ellipsoid, eei mm long, 4-6 mm w rying dark maroon; seeds (1)2—6, E ean. flattened, ets or black, 7 mm long, 3.5 mm wide. Fig. Ü This species is known from the Atlantic and e. lowlands of Costa Rica in tropical wet Orest from sea level to 800 m elevation. Anthurium clidemioides is placed in section «x num (Croat & Baker, 1978) and is easily du the most distinctive species of Anthurium * to its bullate leaves and near absence of a e uncle. In its sterile state, it could be mistaken Species of Piper, Dioscorea, Pilea, or even some Melastomataceae. MO). pun en? Davidson & Donahue 8735, 8824 2922 (M B Corcovado National Park, Liesner nthurium h — concinnatum Schott, Prod. Aroid. Eee An Schott var. 2 rii (Schott) Engl, Monogr. Phan. Ta 1879. Type: Costa Rica. Cartago: urrialba, Oersted s.n. (Photo of Schott a Drawing #709; NYBG Neg. #N:S. D epiphytic or sometimes prow 9 l m or more long, 2.5-4 cm diam.; odes obsc Scured by weathering ctaphylli S coriaceous, 7-15 cm long, drying internod “ataphyl] OF MEXICO AND MIDDLE AMERICA 279 brown, splitting at base, remaining intact at apex. LEAVES erect to spreading; petioles subterete, weakly and bluntly sulcate, 35-75 cm long, 0.6- 1.2 cm diam.; geniculum 1.5-3 cm long; blades subcoriaceous, narrowly ovate, 37-58 cm long, 21-34 cm wide, abruptly acuminate at apex, deeply lobed at base; the sinus spathulate, the lobes often overlapping; both surfaces glossy; the midrib and primary lateral veins raised above and below, lesser veins prominulous; basal veins to 7 pairs, all basal veins coalesced to 1.5 cm; the collective vein arising from the first basal vein, loop-connecting the primary lateral veins and running to apex. INFLORESCENCE spreading-pendent; peduncle 35-65 cm long, shorter than petioles; spathe oblong-lanceolate, pale green turning purplish, 8.5-27 cm long, 1.5- 5 cm wide, abruptly cuspidate at apex; stipe to 2.5 cm long, sometimes arching; spadix green turning purplish, 5-26 cm long, sometimes arched at base, 8-16 mm diam. at base, 3-8 mm diam at apex; flowers rhombic; 8-9 flowers visible in the principal spiral, 12 or more flowers visible in the alternate spiral; pistils raised; stamens well exserted, held at the edge of tepals and + above pistil; anthers white; thecae brown, persistent, giving spadix spiny appearance. INFRUCTES- CENCE pendent; spadix to 30 cm long; berries orange; mesocarp pulpy, orange with numerous, linear raphide cells; seeds 1-2 obovoid, 4.3 mm long, 3.5 mm wide, pale yellow, indented at base, darker area at apex. Figs. 52 and 54. The species is found in Costa Rica and possibly also in Colombia and Ecuador at elevations of 2,000 to 3,000 m in lower montane rain forest. The species grows at higher elevations than any other Anthurium in Costa Rica or Panama. Anthurium concinnatum is a member of sec- tion Belolonchium and is recognized by its large, subcoriaceous, cordate blade and its curved stipe, its persistent, exserted stamens and high altitu- dinal range. The species usually has long stems that, although sometimes rooted in the soil, are usually leaning on trees for support. CosrA RICA. ALAJUELA: Volcan Poás, Lems s. CARTAGO: Cerro Carpintera, Standley MAT6 (US): Cer- ro de La Muerte, Croat 35378, 35407 (MO); re El Cañón, S of Cartago, Williams et al. t (N Pee vicinity El Empalme, Alm da & 3807 (M Almeda 3823 bate aeg al. 16151. bees a Lems s.n. (NY Lo ); N x e e & Nakai 3788 (MO); La Estrella, Stanaiey National Park, Davidson 8231 dg en 4401 (BR); -— — d ho - OS NHP 1DOOA, ) MORNA W Hoss 1874qQ0102 ta rnianijup/ ^ LS — TS [ 490fpgp NOOS UNUMGOSUOD WNIUNYIUV ^99—SS —' 9C£ 9C IDOAD 'YOuoS UINIDUUIOUOS tunianujuy ‘YS "LG6-vs SAAN Z Ww a [^a < g — < E Z < — Q em x - O u e z E T sa L o N -— < Z Z < 1983] CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 281 vicinity Poás, Tonduz 10681 (BR). CARTAGO & SAN JOSE BORDER: along Interam Highway, Burger & Lies- ner 6479 (F, MO); 7018 (F, NY); Burger & Baker 9537 (F, DUKE); Luteyn 2975 (DUKE); Raven 20945 (MO); Tablazo, Brade 2537 (BR). HEREDIA: Cerro Caricias, N of San Isidro, Standley & Valerio 52301 (US); between e et al. 238 (F, MO); Volcán Barba, Pittier & Durand 1345 (BR); Utley 4571, 4572, 4573 (DUKE). PUNTA- Almeda 3672 (CAS); western as t o de la Muerte, Croat 32854, 32892 (MO); near Finca La Cim pey, Standley 42059, 426 R Lagun- illa, Pittier 10447 (BR, nta María Anthurium consobrinum Schott, Oesterr. Bot. Wochenbl. 5: 66. 1855. Type: Nicaragua, Dept. of Rio San Juan, Rio San Juan, Fried- richsthal s.n. (Photo of Schott Aroid Draw- ing #353; NYBG Neg. #N.S. 3874). Rosulate epiphyte; stems short, 1.5-2 cm diam.; roots very dense, short, ascending, nar- rowly tapered, white to greenish white; cataphylls subcoriaceous, 5.5-15 cm long, caudate-acumi- hate at apex, drying rusty brown (B & K Yellow 6/2.5), persisting intact, ultimately deciduous. LEAVES erect; petioles broadly sulcate, rounded abaxially, 2.5-9 cm long, 0.6-1 cm diam.; ge- Ré: aa 0.5-2 cm long; blades oblanceolate, sub- k inta 19-64 cm long, 3.5-20 cm wide, E ptly acuminate at apex, acute to narrowly unded at base; both surfaces matte to semi- -— the midrib flat at base above, acutely raised Ward apex, convexly raised below; primary lat- bi veins 12-18 per side, raised above and be- em S Msi Shorter than leaves; the peduncle 20—43 das x 5-6 mm diam., several times longer a loles, terete; spathe linear-lanceolate, em Wide ous green, 4.5-7.5 cm long, 0.8-2.5 thea abruptly acuminate at apex, round to pinkish. - base; spadix greenish-white, turning tm diam a strongly curved, 6-9 cm long, 1-1 5 L^ the flowers 4-lobed, 2.7-3.4 mm in moid: nos the sides weakly to jaggedly sig- ral, vida i flowers visible in the principal spi- tepals m Owers visible in the alternate spiral; to 22 atte; the pistil pyramidal and exserted ‘*“ mm; the stigma ellipsoid, 0.5-0.7 mm long, becoming brushlike; stamens emerging slowly from base of spadix, exserted on fleshy, colorless filaments ca. 1 mm long, then retracting to surface of tepals; anthers ovate-triangular, white, 0.8-1 mm long, 0.8-1 mm wide; thecae narrowly ovate; pollen white. INFRUCTES- CENCE pendent; spadix to 13 cm long; berries translucent, greenish-white with apex reddish to purplish, narrowly acute to prominently beaked, ca. 1 cm long, 5 mm wide; mesocarp juicy; seeds 2, greenish-white, oblong-elliptic, 3-3.4 mm long, ca. 2 mm wide, 1.5 mm thick, enveloped in a gelatinous, transparent envelope that extends ca. 1 mm beyond seed on both ends, broader at apex. Figs. 55 and 56. The species ranges from Nicaragua to Costa Rica from sea level to 700 m elevation. It is a common species from tropical wet forest and premontane wet basal belt transition life zones on the Caribbean slope. It is to be expected in Panama in Bocas del Toro Province. The species is distinguished by its rosulate habit, oblanceolate blades, short sulcate petioles, massive whitish root system, spreading inflores- cence with upturned, nontapered spadix, mark- Anthurium consobrinum is probably most closely related to A. fatoense of Costa Rica and Panama and differs from that species by its abax- ially rounded petioles, by its nearly oblong spa- dix, and by berry coloration. Anthurium fatoense has a quadrangular petiole, a cylindroid-tapered spadix, and pale yellow or orange berries. An- thurium consobrinum was placed in section Ep- iseiostenium by Schott but is probably a member of section Pachyneurium. A Rica. no other location, Proctor 32168 (LL); Taylor 11593 (NY). ALAJUELA: vicinity of Bijagua, Al- meda & Nakai 4043 (MO); Burger & Baker 9857 US); between Naranjo & Aguas Zarcas, NE of Quesada, Croat 46949 (MO); Santa Maria National Park, Lies- Ensayos, NW of Zarcero, Croat 43615 (MO). HEREDIA: near Puerto Viejo, Croat 35671, 35695 (MO); Taylor A r Ti MO). LIMON: road between Cahuita &L & Burger 151 (F, MO); Finca Castilla, Dodge & Goer- ger s.n. (MO); Hacienda Tapezco & Hacienda La Suerte, 282 W of Tortuguero, Davidson & Donahue 6723, 6768, 8413, 8461, 8733 (MO). NICAR . RIO SA AN: near Cano Chontaleno, NE of El Castillo, Neill 3332, 3376, 3420, 3421 (MO); Neill & Vincelli 3622 (MO). zELAyA: Bluefields, N of Base Camp, Proctor et al. 26913, 27010 (LL, MICH, NY, US); Caño Costa Riquita, SW of Colonia Naciones Unida, Stevens 4966, 5058, 6411, 6412 (MO); Caño Zamora on Río Rama, Stevens 88 O); S of bridge at Colonia Agrícola Yolaina, Stevens 4819, 4826, 6308 (MO); road to Colonia Agrícola Yolaina & Colonia La Esperanza, Vincelli 166, 168 (MO); Salto La Oropén- dola, Stevens 8954 (MO). Anthurium cordatotriangulum Matuda, Anales Inst. Biol. Univ. Nac. México 36: 109. 1966. Type: Mexico. Chiapas: Motozintla; Sierr. Madre de Chiapas, El Rosario (ca. 8 mi S of Motozintla), 1,800 m elev., MacDougall 393 (MEXU, holotype). 3 cm long; blades broadly ovate-triangular, sub- coriaceous, gradually acuminate at apex, broadly lobed at base, 37-64 cm long, 26-46 cm diam., broadest at point of petiole attachment or just below; the anterior lobe 30-44 cm long; the pos- terior lobes 11—23 cm long, directed outward at ca. 45? angle from petiole; the sinus parabolic to hippocrepiform, round at apex; the midrib raised above and below; basal veins 5 or 6 pairs, the third to sixth coalesced 3-6 cm, raised above and below; primary lateral veins 5 or 6 per side, de- parting midrib at 60° angle, + Straight, usually prominently loop-connected into collective vein in apical half of blade; the collective vein arising from the first basal vein or sometimes from one of the lower primary lateral veins. INFLORES- CENCE usually longer than leaves; peduncle subterete, to 57 cm long, ca. 5 mm diam.; spathe moderately thick, green, lanceolate, 14-18 cm long, 2.5-4.5 cm wide, long-acuminate at apex, rounded and + clasping at base; Stipe to 15 mm long in front, 6 mm long in back; spadix purplish (dried), 10-13 cm long, 8-10 mm diam. at base, 4-5 mm diam. at apex; the flowers rhombic, 2.7— 3.1 mm long, 1.5-2.2 mm wide; 7-9 flowers vis- ible in the principal spiral; tepals weakly punc- tate; pistils not emergent; stigma 3-3.5 mm long, broadly elliptic; stamens emerging from base and held just above tepals in a tight cluster; thecae oblong-elliptic, scarcely divaricate. INFRUC- TESCENCE not seen. Figs. 53 and 58. ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 Anthurium cordatotriangulum is known for | certain only from the southeastern corner of l Chiapas near the Guatemalan border in Moto- zintla and Unión Juárez in the Sierra Madre de Chiapas at 1,800 to 1,900 m in oak-pine forest. Two other closely related species, A. montanum and A. chamulense, have been collected at the type locality of A. cordatotriangulum. A collec- | tion from San Carlos Yautepec in the Sierra de Lacheguia in South Central Oaxaca matches the f type of A. cordatotriangulum very well but itis» difficult to explain the otherwise apparent ende- mism of this species. (This specimen bears the unpublished herbarium name of A. oaxacari- picolum Matuda but since the collection differs | in no appreciable way from A. cordatotriangi- | lum it will not be recognized here.) The ree around San Carlos Yautepec is very arid and nearly devoid of habits for Anthurium. I suspect that the label data referring to San Carlos Ya | tepec was intended to refer to the municipio of the same name which is very large and no doubt — includes good Anthurium habitats in the south- | ern part. The species is placed in section Belolonchium | and is distinguished by its almost deltoid blades. — its coarsely elevated basal and primary la veins (on drying), broad ovate-lanceolate spathe, | and long-tapered, violet-purple spadix. Altho Matuda compared the species with A. macdou- | galii (A. andicola), it is perhaps most easily fused with 4. montanum, which occurs 1n | | spicuously long-tapered. » f MEXIco. CHIAPAS: El Rosario, MacDOUg a4 (MEXU); Matuda 38628 (MEXU); Volea © Matuda 38626 (MEXU). oaxaca: San Carlo pec, Matuda 38276 (MEXU). n z = —— [4] bs. | an ko] £e e 2 Ey e 2 a em un " e Qa -y »* e Id = ce S kd —— NN, — rr. Anthurium cotobrusii Croat & Baker, pre 16 (Supl. 1): 43. 1979. Type: Panama. ^. | riquí: above San Felix, along mining 18-27 miles off Pan-American High (above Chamé on turnoff to Po elev. 1,200-1,500 m, Croat 33058 p 2381190-91, holotype; CR, F, NY, US! types; Live at MO). Terrestrial or sometimes epipetric 0 ic; stem 1-3 cm diam. usually less than long; roots greenish or grayish-brown, thick sella: r epiphy 15 cm - 1983] Fic 0 Mi diary Croat CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 283 onto De Er. OU h on a m averia E Mosaic, hotosintla,en Doi “nire de Chiapas. ype).— 61. . ES 58-61. 58. Anthurium cordatotriangulum Matuda, Mac ee: id ( aon — 59-60. Anthurium & Baker, Croat 33058 (T Anthurium abonis Engl., Neill 2 284 smooth, to ca. 6 mm diam.; cataphylls 9-11 cm , sometimes tinged purplish, turning , the ac sharply, ie uh broadly lobed at base, are cm long, 9.5-—25 cm wide, widest at the base; the anterior lobe (118-307) cm long, the margin + straight to convex; the pos- terior lobes (5)8-11(15) cm long, turning up; the sinus much broader than deep when flattened, obtuse to acute at apex; both surfaces semiglossy; midrib obtusely raised on both surfaces at base, more sharply raised toward apex on upper sur- face, finally diminishing at apex; primary lateral veins 5-8 per side, weakly to sharply raised (es- pecially toward base) above, weakly raised be- low, arcuate-ascending; basal veins 3-5 pairs, 3 or 4 of them coalesced 4-6 cm; posterior rib naked throughout its length, gradually curved to almost straight, turning under; lesser veins less conspicuous; the collective vein arising from the pri from the first basal vein. INFLORESCENCE erect-spreading, shorter to longer than the leaves; peduncles ter- ete, (4.5)20—46 cm long; spathe green, pale green, or green with purplish veins or heavily suffused with purple on inner surface, sometimes tinged purple only at base medially, ovate to narrowly ovate, (2.5)5-9.5(12) cm long, (1.7)2.7-4.5 cm wide, weakly boat-shaped to flat, acuminate at apex, acute to rounded or shallowly cordate at the base, inserted at 60°-70° angle on peduncle; stipe green or green tinged with purple, 10-20 mm long in front, 2-6 mm long in back; spadix green to dark purple-violet at anthesis, (2.5)5— 25 cm long, 7-15 mm diam. at base, 3-7 mm diam. at apex, with a pungent, sweet scent when droplets are fresh; flowers sub-4-lobed, 3.5-5 mm long, 3.5-4 mm wide, sometimes ca. 3.5 mm in both directions; 3-6 flowers visible in the prin- cipal spiral, 7-10 flowers visible in the alternate spiral; tepals dark violet-purple or yellow-green with purple margins, sparsely punctate, more densely so near margins, semiglossy to matte, the inner margins broadly convex; pistil green; stig- ma linear, purple-violet or green, ca. 0.6 mm long, forming conspicuous droplets for 1-3 weeks before emergence of stamens, droplets persisting when the first and second stamens are opening; ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 —€— drying and covered by theose when third Ne pe ae inacl to yellowish, held pe above the tepals, 0. 6-03 mm long, 1 mm wide; filament short, retracting; pollen pale orange, fading white. INFRUCTES- CENCE erect-spreading; berries pale orange (B & K Yellow-red 8/7.5), obovoid, 12-15 mm long, 8-11 mm diam., rounded at apex, the apex dark- er, depressed with 4 short, slightly elevated ridges; mesocarp juicy, orange; seeds usually 2, oblong- elliptic, 6.5-7.5 mm long, 2.5-4 mm wide, en- cased in a flattened, translucent envelope that projects up to 3 mm beyond the apex of the seed and up to 1 mm beyond the lateral margins. Figs. 57, 59, and 60. Anthurium cotobrusii is found in Costa Rica and Panama in premontane wet forest at ele- vations of 1,000 to 1,500 m. This species is most easily distinguished by its subcoriaceous leaves which have more OF less obscure reticulate veins when fresh but have con- spicuous reticulate veins when dry, and by its prominently tapered, violet-purple spadix with a tight cluster of stamens. In a dried state the flowers have their tepals turned somewhat up- wards, which, in conjunction with the ee stam s, es. the spadix a rough. appe ens, giv p i po d emergent berries that are quadr prominent ridges radiating from the center corners. The species is now placed i in chium. It was erroneously placed previously " the section Cardiolonchium by Croat and Bak * 1980. It is most easily confused with 4. "T : choanum, which has similar leaves, but the latte has a short cylindroid, scarcely tapered sp? dis that is usually nodding and hooded by the SP? when in flower (the spathe remaining € n section Belolon- teflexed spathe. In addition, the bert! i ranchoanum are smoothly rounded at e along coffee fincas Costa RICA, PUNTARENAS: above Panama Rio Coto Brus, near Cotan, N of La Unión (on ie Croat 26674 (MO). 25: Anthurium cubense Engl., Bot. Jahrb. Syst- ait 364. 1898. Type: Cuba. Ad Monte Toro tothe — — » M——— —— — 1983] 300 m, Eggers 5402 (K, holotype; B, iso- type) Epiphytic or terrestrial; to 40 cm tall; stem short, to 3 cm diam.; roots green, pubescent at base, to 8 mm diam., ascending at upper nodes; and persisting be- ges erect; petioles 5-10 cm long, 5-11 mm diam., subterete or ob- scurely quadrangular, broadly sulcate; geniculum 0.5-1 cm long, sometimes tinged purplish; blades oblanceolate-elliptic, subcoriaceous, graduall acuminate at apex, obtuse to narrowly rounded at base, 26-83 cm long, 9-25 cm wide; the upper surface semiglossy; the midrib bluntly raised above and below; primary lateral veins 6—7 per side, raised above and below; lesser veins less suffused with violet; spathe subcoriaceous, sub- ovate, olive green, tinged with violet-purple at margins, 2.5-16 cm long, 0.6-5 cm wide; spadix weakly ellipsoidal, 5.5—10 cm long, 0.8-1 cm diam. midway, white or lavender (B & K Purple 6/7.5); flowers rhombic or weakly 4-lobed, 2.1- 2.5 mm long, 2.4-2.6 mm wide, the sides weakly sigmoid; 9-12 flowers visible in the principal spiral, 14-16 flowers visible in the alternate spi- ral; tepals matte, lateral tepals 1.2-1.4 mm wide; pistil weakly raised, greenish; stigma linear to 0.4 mm long: ^mm emerging in scattered pattern hs sae the erus stamens emerging almost lo the ape held just d. tepals agninat pistil, obscuring pistil When all stamens develop; anthers white tinged Ted-violet: thecae Hie ollen white. IN- CTESCEN i NCE spreading; berries red, ob- Ong-elliptic, 9-12 mm ds 4 mm wide; seeds 1 or 2, 5-5.5 mm long, 3 mm wide. Figs. 61 and 62. à horn cubense is found in Cuba, Yucatan, WA T parts of eastern Guatemala on the At- Pacifi slope; also Nicaragua to Panama along the — CS Slope. The species is also known from from s : Colombia and Venezuela. It is known moist ^de dry forest and drier parts of tropical Orest at low elevations. The species is ap- a a pure dee. in Nicaragua where it is lo- Anthurium cubense is placed in section Pachy- Short i CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA Prin and i is recognized by its short petioles, UIAU tiiv 1v 285 the peduncle as short as or only slightly longer than the spadix), its cylindroid, pale violet-pur- ple spadix usually weakly tapered toward both ends, and its bright red berries. Engler (1905) also recognized material of this species under the name A. recussatum Schott from Cuba. That name, however, has been misapplied and actually refers to A. hookeri Kunth from the Lesser Antilles and northeastern South America. Anthurium hookeri is not related to A. cubense, although it has a similar rosulate habit. It differs chiefly in having punctate leaves with unusual scaliform secondary veins and by having super- volute vernation. A RICA. GUANACASTE: S of Peñas Blancas, Lies- ner 4652 (MO). PUNTARENAS: vicinity of tel n Nature Reserve, Burger & Liesner 6627 (F, TATA ALTA VERAPAZ: Cubilhuitz, a heim 8606 (U NICARAGUA. C pan N of fue Sievers i 610 86, 87 (MO). GRANA e of Volcán Mombacho, Atwood & Neill AN82 Pere "Neill 2730 (MO). Ma- ES: S of vui. Nichols 1725 7 (MO); Vincelli Ican Masearas near Santa Cruz, Stevens 6626 (MO). 2: ZELAYA: along Cano Majagua, Stevens 6894 (MO). Anthurium cuneatissimum (Engl.) Croat, comb. nov. Anthurium consobrinum var. cuneatis- simum Engl., Pflanzenr. IV 23B: 176. 1905. Tyre: Costa Rica. Limón: Llanuras de Santa Clara, 250 m, Donnell Smith 6811 (B, ho- lotype; US, isotype). Epiphyte; stems very short; roots dense, de- scending, fuzzy (dried), ca. 3 mm thick; cata- phylls coriaceous, 6-9 cm long, acuminate at apex, drying tan (B & K Red 9/10), weathering into longitudinal fibers. LEAVES erect-spreading; petioles usually quadrangular with sharp edges, sometimes bluntly angular abaxially, often with a sharp medial rib abaxially, 6-32(40) cm long, m wide; geniculum 0.8-2.5 cm long; pater obovate-elliptic, subcoriaceous; (20)25-60 cm long, 8.5-21 cm wide, long-acuminate at Eig acute to cuneate at base; upper surface semi- glossy to glossy, lower surface semiglossy; the midrib broad, scarcely raised above, sometimes edial rib, narrowing and sunken at apex, prominently raised below, sometimes with 1-2 sharp ribs; primary lateral veins (5)10— 15 per side, departing midrib at 30*-40* angle, + straight to collective vein or weakly arcuate- ascending, raised in shallow valleys above, raised “SEIPE IDOAD “SEW t'unippidsno uanianuiy ^S9—7 SOI LZ 1DOA,) “WEY tunipp -1idsno WiNLNYyIUP “p9—'KOS#YE IDOAD “VEOID (CI8ugp) wniuissimpouno wunianijuy ^€9—'70£ ZC INAN “Pug osuoqno wUiniunylUp “TI 7$9-C9 SAAN e ul a) x < O ES! < e e < = A 4 m e - le, "7| o = Lu I E LL o u — « Z, rA < 1983] CROAT below; lesser veins flat above, flat and conspic- uous below; collective vein arising from one of the primary lateral veins, 1-3 mm from the mar- gin. INFLORESCENCE erect-spreading, usually shorter than leaves; peduncle 21-58 cm long, sharply 1-ribbed abaxially, 2—4-ribbed from base to just above middle; spathe ovate, cucullate, dark green, coriaceous, 2.5—6(9) cm long, 2.1-4 cm wide, abruptly to gradually acuminate at apex, rounded to truncate at base, inserted at 50°-70° angle on peduncle; spadix sessile, clavate, pale yellow-green or lemon-yellow, 1.9—5 cm long, ca. 8 mm diam. midway, weakly tapered at base, round at apex; flowers 4-lobed, 1.6-2.2 mm long, 2-2.5 mm wide, the sides jaggedly sigmoid; 16— 20 flowers visible in the principal spiral, 13-16 flowers visible in the alternate spiral; tepals matte, My papillate, lateral tepals 0.6—0.9 mm inner margin straight to convex and turned s un pistil; the pistil emergent to 0.5 mm, white to pale greenish-white; stigma brushlike, ca. 0.2 mm long, with small droplets ca. 2 weeks before drying, 2-3 days before first stamens emerge; lateral stamens emerging si- multaneously on all flowers, soon followed by third and fourth stamens, exserted on long, whit- ish filaments, 0.3-1 mm long, 0.2-0.3 mm wide, Which slowly retract holding anthers in a tight circle around and above pistil; anthers pale yel- ow, 0.3-0.4 mm long, 0.5-0.6 mm wide; thecae ellipsoid; pollen pale yellow fading to white. IN- FRUCTESCE E with spadix 5-11 cm long; berries obovoid, greenish-yellow to white, 4—4.3 mm long, 2.5 mm thick, sharply pointed at apex; seeds 2. Fig. 63. A Anthurium cuneatissimum is known from the a Slopes in Costa Rica and Panama in dent wet forest life zones from 20 to 450 m e; 10n in Costa Rica and at much higher el- ations of 800 to | ,000 m in Panama. an Pecies is in section Pachyneurium and is , inguished by its long, usually quadrangular, » Coriaceous, ovate spathe, short vede s -yellow spadix, and white obovoid berries. um ed only with A. —— which ith à similar short cl ometimes yellow- SPadix. That species diis however, in hav- i ng much narrower leaves with 20-30 promi- nently su OF MEXICO AND MIDDLE AMERICA 287 lanceolate to broadly lanceolate and about twice as long as the spadix (versus ovate and about as long as the spadix for A. cuneatissimum). This species was not distinguished in the treat- ment of the genus Anthurium for Costa Rica (Croat & Baker, 1979) even though we initially had material separated along these lines before finding that the 3-sided versus 4-sided petiole character breaks down. Special thanks go to Mi- chael H. Grayum (University of Massachusetts) for pointing out additional characters that made their separation possible. OSTA RICA. ALAJUELA: NNE of Bijagua, Burger & Baker 9820 (MO); Croat 36275; 36296; 36454; 36480A . HEREDIA: near Puerto Viejo along road near Río Sucio, Croat 35751 (MO). goi que cuspidatum Mast., Gard. Chron. 3: 428. 1875. Type: Colombia? Hort. Vietch., January 1875 (K). h l., Bot. Jahrb. Syst. 25: 386 1898. Type: Costa Rica. Limon: Talama pita 100 m, Tonduz 9233 (B, Palacin po, designated Croat & Baker, 1979). ae donnelsmiti Engl. LB Jahrb. Syst. 25: 389. E: Cos a. Limón: Río Jimé- nez, rid id ide pus "280 m, April 1894, J Donnell | Smith 4979 (B, holotype; US, isotype). Anthurium orteganum Engl., Bot. Jahrb. Syst. 25: 387. Type: Colombia. Cauca: near Po n, L que dies m, Lehmann 5989 (B, holotype; F, K, i Anthurium ier Engl., Pflanzenr. IV 23B: 119. 1905. : Costa Rica. Limón: Suerre, Llanos de tis Sn 300 m, J Donnell Smith 6812 (B). Epiphyte or terrestrial; stems often to 50 cm long, ca. 3 cm diam.; internodes short; cataphylls 7-15 cm long, subcoriaceous, drying pale red- dish-brown, weathering into fibers. LEAVES erect-spreading; petioles terete, 30-118 cm long, 2 mm diam.; geniculum 2-3 cm long; blades usually ovate-oblong, sometimes narrowly ovate, nently lobed at base, 25-62 cm long, 11-40 cm wide; the sinus (when pane is paved) pasé iform to sp ] and below; primary lateral veins | p 15 per side, sunken above, raised below, often reddish be- neath; lesser veins sunken above, raised below; collective vein arising iig one of the lowermost basal veins, raised below, 1-2 mm from margin. INFLORESCENCE erect-spread- ing, shorter than or equal to the leaves; peduncle 28—46 cm long, terete; spathe membranaceous, 288 oblong-lanceolate, pale green, sometimes tinged purple, 7.5-18 cm long, 1.8-2.5 cm wide, grad- ually acuminate at apex, round at base; spadix green to usually dark violet-purple, 6.5-18 cm long, ca. 5.5 mm diam. at base, 4.5 mm diam. at apex; flowers rhombic, 3.5-3.8 mm long, 2.8- 3 mm wide, the sides straight to weakly sigmoid; 4-5 flowers visible in the principal spiral, 5—6 flowers visible in the al piral; tepals glossy; pistils weakly emergent, violet-purple; stigma linear, ca. 0.5 mm long; stamens emerging in scattered manner from base, the basal one quar- ter of spadix with 4 stamens before those in the apical one half emerge, held in tight contiguous cluster obscuring pistil; anthers violet-purple; thecae oblong, scarcely divaricate; pollen pale purple fading to white. INFRUCTESCENCE pendent; spadix to 27 cm long; berries purple- violet to reddish-violet, ovoid-ellipsoid, 6-8 mm long, ca. 5.3 mm wide. Figs. 64 and 65. The species is found from Nicaragua to Ec- uador at elevations from 400 to 1,700 m or rarely lower in very wet areas, in premontane wet, trop- ical wet, premontane rain, and lower montane rain forest life zones. It is one of the more com- mon species of Anthurium. Anthurium cuspidatum is distinguished by its thin, usually oblong-ovate blades with many veins Anthurium cuspidatum was placed in section Polyneurium by Engler but has since been shown to be an atypical member of section Pachyneu- rium because of its involute vernation. It is su- perficially close to A. williamsii of Costa Rica, Panama, and northwestern South America, with which it shares a nearly identical inflorescence and similar thin, veiny leaves. Anthurium wil- liamsii, in section Polyneurium, differs in having supervolute vernation and a more ovate blade with more basal veins that are free or nearly free to the base. Anthurium cuspidatum is quite vari- able in leaf shape with blades sometimes ovate, similar to A. williamsii, but when this is the case, ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 A. cuspidatum can be distinquished by having a definite posterior rib (coalesced basal veins). j pidat number ofsibling species in the coastal and sub-Andean portions of Colombia and Ecuador. Many of these species may prove to be synonymous but others have radically different inflorescences while sharing similar leaf blades with numerous primary lat- eral veins. Included in the complex are A. mon- ticolum Engl., A. longicaudatum Engl., A. oreo- philum Sodiro, and others. Costa RICA. ALAJUELA: NNE of Bijagua, Croat 36258, Hule, Bak- , MO); Finca Los Ensayos, N cero, Croat 43590 (MO); E rim of Lake Hule, Luteyn 3 ); between Naranjo and Aguas Zarcas, Cr 46958 (MO); N of Rio Naranjo along road to Lows Utley & Utley 5343 (DUKE); San Carlos, Smith 165 (MO); between Volcán Pods and Volcán Barba, ie 35466 (MO). cARTAGO: road between Juan Vinas an Turrialba, Croat 36830 (MO); road between ee to Quebrada Platanillo, Croat 36653, 36673 (MO) i bridge and 1a 5 35607 (MO). LIMÓN: Hacienda Tapezco-H Suerte, W of Tortuguero, Davidson 6746, 69 between Siguerres and Limón, Baker 185 (F. pe Burger et al. 10522 (F, MO); SW of sae E & Utley 5445 (DUKE); Talamanca, Pittier & 5 vodt 9233 (BR, US); between Turrialba and Limon, ion 43307 (MO). SAN José: slopes above Aloma road from San Isidro del General to Dominical u MO); between Río Cascajal Do- 3785 (MO); between San Isidro del General and minical, Croat 35321 (MO). n 1815 NICARAGUA. MATAGALPA: Cerro Musün, ge : "n (MO). Rio SAN JUAN: near Caño Chontaleno, Castillo, Neill 3426 (MO). Anthurium davidsoniae Standl., Publ. Field A Nat. Hist., Bot. Ser. 22: 4. 1940. TYPE: Dis- ama. Chiriquí: Bajo Chorro, Boquete ^ trict, 1,800 m, 17 January 1938, M. E- US. vidson 134 (F 915728, holotype; MO. isotype). Usually terrestrial, rarely epiphytic; elongate, less than 1 cm diam.; internodes at apex, longer in lower portion of San phylls moderately thick, to 12 cm long, shal- reddish-brown. LEAVES + erect; petioles g lowly sulcate with sharp marginal ribs aaar stems short cata- 1983] 13-27 cm long, 2.5-3.5 mm wide; blades ovate side, sunken, straight or weakly arched to col- lective vein; collective vein arising from the first or second basal vein, 6-8 mm from margin. IN- FLORESCENCE erect, shorter than leaves; pe- duncle 9-19 cm long; spathe lanceolate, pale green, 3-5 cm long, 5-10 mm wide; spadix pale yellow-green, 2-4 cm long, 5-7 mm diam. at base, 2-3 mm diam. at apex; flowers rhombic to 4-lobed, 2.3-2.7 mm long, 2.7-3 mm wide; 3-5 flowers visible in the principal spiral, 4—7 flowers visihle in th 14 pi kt p 1 ++ , lateral tepals 1-1.2 mm wide; pistil emergent, green; stigmas exserted, brushlike; stamens emerging but may retract (not evident on dry specimens). INFRUCT ESCENCE with spadix 3.5—6 cm long; berries greenish-yellow; mesocarp syrupy; seeds obovoid, yellow-green, darker at apex with nu- eg raphide cells visible on the surface. Fig. Anthurium davidsoniae is known from Costa Rica and western Panama, from 1,300 to 2,000 m elevation in premontane rain and lower mon- lane rain forest. The Species is related to both A. pallens and A. micr ospadix, although more closely to the lat- ler. All of these species share a similar habit with Pe stems and a similar inflorescence. An- urium davidsoniae differs from both others by aving a much broader, more ovate, more con- ‘picuously bullate leaf blade. All three of the above species occur in similar Ses forest habitats and evaluation of herbar- ud ooh alone is unsatisfactory in separa- Fiela "nsidering the variability of Anthurium. that th I inced however, Ee dy are all good species. Anthurium david- Dada and A. microspadix are both semi-erect "d are usually terrestrial but require esos to lean on for support. Anthurium davi E ou a pendent epiphyte. Anthurium in an Vos is placed in section Xialophyllium aidaa oc I refer to as the A. microspadix ftom cs his group, which differs substantially Such as a. members of section Xialophyllium - Caucanum Engl., may ultimately have cl CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 289 to be considered a separate section. All species in the group share thin, prominently veined, often bullate leaves, elongate slender stems, and com- monly green inflorescences and fruits. ; ; RENAS & ALAJUELA: serve, Burger & Baker 9734 (F, MO). SAN JOSE & HERE- DIA: Cerro Zurquí, Almeda 3687 (CAS). SAN jos: NW of Cascajal, Taylor 11303 (NY). Anthurium durandii Engl., Bot. Jahrb. Syst. 25: 401. 1898. Type: Costa Rica. Inter la Di- vision et l'Alto del Palmital, January 1891, Pittier 3868 (B, hololectotype; BR, CR, iso- lectotypes, designated Croat & Baker, 1979). Anthurium littorale Engl., Bot. Jahrb. Syst. 25: 405. 1 . TYPE: Costa Rica. Marais de Sierpe, March 1892, Pittier 6837 (BR, holotype; CR, isotype). Epiphytic or rarely terrestrial; stems short; in- ternodes sometimes to 3 cm long; roots short, sometimes purplish; cataphylls moderately thin, 5.4—6 cm long acuminate at apex, drying pale brown, weathering to reticulate fibers, ultimately deciduous. LEAVES erect-spreading; petioles 9— 25 cm long, 4-5 mm diam sulcate adaxially, rounded abaxially, sparsely, dark glandular-punctate; geniculum 1.5-3 cm long; blades elliptic, coriaceous, short-acuminate at apex, obtuse at base, 10-31 cm long, 4-15 cm wide, broadest near middle; both surfaces matte, dark glandular-punctate, the lower surface paler, the upper surface drying with an alveolate pat- tern; midrib ly raised at base above, acute at middle, weakly sunken toward apex, convexly raised below, narrowing toward apex; primary lateral veins 8-15 per side, departing midrib at 45° angle, straight to collective vein, flat to weak- ly sunken above, raised below, lesser veins ob- shorter than or equalling leaves; peduncle 12-40 cm long, ca. 3 mm diam.; spathe moderately thin, pale green to pinkish green (B & K Yellow- green 7/7.5), oblong-lanceolate, 4-7 cm long, 1— 1.5 cm wide, broadest near the base, abruptly acuminate at apex, acute at base, inserted at ca. > Zz ^ > (ed n © es - ie» m - un un O C & w Q 4 > a e » En: Q > x J m rA Fioures 66-69. 66. Anthurium durandii Engl., Fennell s.n.—67. Anthurium eximium Engl., Orchid Jungle 4/80. —68. Anthurium eximium Engl., as ruces Ror, Gard 69 Anthurium flexile Schou., Croat 40285 1983] CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 291 72 - ai andl.. Croat 48592.—71-72. Anthurium fatoense K. Krause, 4 nthiirnim o . 70. Anthurium davidsoniae Standl. n ?94« 3^ ^ "a " c 7 4949.—73. Anthurium flexile Schott, Croat 36467. FIGURES 70-73 292 65° angle on peduncle; spadix pale green turning reddish brown, narrowly tapered toward apex, 10-16.5 cm long, 4-9 mm diam. at base, 2-4 mm diam. at apex; flowers 4-lobed, 3-4 mm long, 2.5-3 mm wide, the sides straight in the direction of the principal spiral, sigmoid perpen- dicular to the spiral; ca. 3 flowers visible in the principal spiral, ca. 5 flowers visible in the al- ternate spiral; tepals semiglossy, lateral tepals 1.8— 2.2 mm wide, the inner margin convex; pistils emergent but not raised, square, green; stigmas scarcely visible, ca. 0.2 mm long, drying black; stamens emerging from the base, laterals first, scarcely exposed, quickly retracting before alter- nates emerge; pollen white. INFRUCTES- CENCE with spadix to 22 cm long; berries ovoid, probably orange. Fig. 66. Anthurium durandii is known only from Costa Rica on the Pacific slope in San José and Pun- tarenas Province at elevations from near sea level to 1,600 m in tropical wet and premontane rain forest. It was initially confused with A. paludosum by Standley, which was reported from the Atlantic slope of Costa Rica. Anthurium durandii resem- bles A. paludosum superficially but differs in having the cataphyll weathering into a dense fi- brous network and in having longer petioles, one third to equally as long as the blade and in having a more or less elliptic leaf blade. It is a member of section Porphyrochitonium. Engler cited two specimens in his original de- scription of A. durandii. The second of these, Tonduz 8422 (BR) is A. gracile (Rudge) Lindl. Costa Rica. No other location, Pittier 3868, 6837 (BR). PUNTARENAS: Finca Loma Linda, SW of Cañas Gordas, Liesner 645 (NY, US); 23] (MO, NY, US); along Fila Las Cruces, Luteyn 3855 (DUKE): Osa Pen. insula, W of Rincon de Osa, Burger & Liesner 7322 Vito de Java, Raven 21758, 2184] (DS). sAN José: below Santa Maria de Doto, Croat 55203 (MO). Anthurium eximium Engl., Bot. Jahrb. Syst. 25: 412. 1898. Type: Costa Rica. Puntarenas: in forest near Punta Mala, Tonduz 6768 (BR, hololectotype (two sheets); CR, isolectotype; d 979). esignated Croat & Baker, | Epiphyte; stem elongate, ca. | cm diam.; roots thick; cataphylls persisting intact i LE etioles broadly sulcate, rounded to 3-ribbed abaxially, 7-12 cm long, 6-7 mm diam.; blades ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 narrowly oblanceolate, subcoriaceous, 40-80 cm long, 5-11.5 cm wide, narrowly attenuate at base, long-acuminate at apex; primary lateral veins 10- 12 per side, raised above and below (dry), de- parting midrib at 40?-45? angle, weakly ar vein; g the base or one of the lowermost primary lateral veins, 1-4 mm from margin. INFLORES- CENCE shorter than leaves; peduncle 29-32 cm long; spathe lanceolate-ovate, green, 5-7.7 cm long, 2-2.6 cm wide, clasping at base, spadix creamy white, oblong to clavate, 4.5-6 cm long, 0.8-1 cm diam. at base, 9 mm diam. at apex 10-12 flowers visible in the principal spiral; te- pals matte, longer than wide; stamens exserted, emerging in a somewhat irregular fashion in a complete sequence, the leading stamens 1-2 spi- rals ahead of the third and fourth; pollen yellow, diam.; berries bright red, obovoid, ca. 4.6-7 mm long, 4-5 mm wide; seeds 2, 2.7-2.9 mm long. 1.5-1.7 mm wide. Figs. 67 and 68. The species is found in Costa Rica in tropical wet forest formations from sea level to 800 m elevation. Until recently it was known only from the Osa Peninsula and the adjacent Golfo Dulce region, but now it has been collected in San Jose Province on the road to Purrita. Anthurium eximium can be distinguished by the oblanceolate, epunctate leaves, the persisting i he cataphylls, the clasping, lanceolate e t the persisting cataphylls. E In his monograph, Engler (1905) cited e imium as a synonym of A. consobrinum. e shapes in both species are very similar. P iy Schott's illustration of A. consobrinum € je: shows a long, slightly tapering spadix with Pm and berries greatly exceeding the tepals. n de ium consobrinum is a common species ° ae Atlantic lowlands and numerous collections consistent with Schott's drawing. The bet A. consobrinum are white with purple, SP orange apices as opposed to the solid brigh berries of A. eximium. In addition, the A. consobrinum are early long-exserte thesis. held at about the level of the stamens at an ti Anthurium eximium is a member of puc Pachyneurium despite certain features it thick with section Calomystrium, including the i —M m 1983] CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 293 persistent, intact cataphylls and the presence of linear raphide cells on both blade surfaces. Costa Rica. No other location, Pittier 6768 (BR). PUNTARENAS: Osa Peninsula, Corcovado National Park, Liesner 2846, 2996 (MO); above Palmar Norte, Croat 35121 (MO); Punta Mala, Tonduz 6768 (BR); Rincón e Osa, Burger & Liesner 7286 (MO, NY); Liesner 1934 (MO); Santo Domingo of Golfo Dulce, Tonduz 7260 (BR, CR, US). san José: below Santa Maria de Doto, Croat 55204 (MO). Anthurium fatoense K. Krause, Notizbl. Bot. Gart. Berlin-Dahlem 11: 611. 1932. TYPE: Panama. Colón: Dos Bocas, Río Fato Val- ley, 40-80 m elev., Pittier 4227 (US). Rosulate epiphyte; stems to 60 cm long, usu- ally much shorter, ca. 1 cm diam.; dense mass of greenish roots, acute and ascending at tips; cataphylls subcoriaceous, sometimes tinged with red, 5-10 cm long, drying medium brown, per- sisting + intact. LEAVES erect-spreading; peti- oles quadrangular, sulcate, 2.5—5(9.5) cm long, 5-7 mm wide; geniculum 7-10(20) mm long; blades oblong-oblanceolate, gradually acuminate ps apex, acute at base, 24-83 cm long, 6-26 cm Wide, broadest at middle or above; both surfaces semiglossy: midrib acutely raised above, square and prominently raised below; primary lateral veins 8-16 per side, departing midrib at 45? an- gle, arcuate-ascending, acutely raised above, raised and sometimes reddish below; submar- kis ive vein arising in the apical quarter CENCE 2-9 mm from margin. INFLORES- Mus go reading to arching-pendent, ai an or equalling leaves; peduncle 14—48 hg, subterete, faintly ribbed; spathe pale Miss Sometimes tinged purplish, narrowly lan- ties P 3.5-9 cm long, 0.7-1.3 cm wide, nar- ou inate to abruptly acuminate at i os tuse to subcordate at base; spadix pale or green tinged reddish, 5-15 cm long, mm diam. at base, 2.5-4 mm diam. at apex; rs rhombic to weakly 4-lobed, 1.8-2.8 mm lo : ng, 2-3 mm wide, the sides straight parallel to *Xsert se in a + complete sequence, well s ed and held against sides of exserted pistils; ers white, 0.5-0.6 mm long, 0.5-0.8 mm wide; thecae broadly ovate; pollen white. IN- FRUCTESCENCE with spadix to 26 cm long; berries cream to greenish-yellow or orange, ob- long-ovoid, 4-7 mm long, acute at apex. Figs. 71 and 72. The species is known from southern Costa Rica and Panama, occurring principally on the Carib- bean slope from sea level to 1,000 m in wetter parts of tropical moist forest as well as in tropical wet, premontane wet and lower montane wet forest. It also occurs on the Pacific slope in the Fortuna area of Chiriquí Province. Anthurium fatoense is a member of section Pachyneurium and is considerably variable in the size of the leaves and inflorescences as well as in berry color. It is most easily confused with A. consobrinum in Costa Rica and is distin- has the stamens emerged in a close cluster with the pistil not at all apparent. In the flowers of A. consobrinum, the pistil g ll ahead ofthe stamens and dwarfs the stamens in size, extend- ing several millimeters above the stamens. De- spite the close vegetative resemblance of the two species, they are probably not closely related. Costa RICA. ALAJUELA: between Naranjo & Aguas Zarcas, NE of Villa Quesada, Croat 46976 (MO). LIMON: S of Punta Cahuita, Croat 43197 (MO); between Rio Bananito & Cahuita, Burger et al. 10485 (MO). Anthurium flexile Schott, Oesterr. Bot. Z. 8: 180. 1858. a. Anthurium flexile ssp. flexile. Type: Costa Rica. Heredia: Pedregel, Wendland 932 (GOET, holotype). Anthurium myosuroides (H.B.K.) Endl. var. angusti- folia Engl., Bot. Jahrb. Syst. 25: 382. 1898. TYPE: Costa Rica. Limón: Matina, Pittier 9797 (Pittier 9767?) (BR, CR, US). Scandent epiphyte; the stem elongate; inter- nodes 2-6 cm long, ca. 3 mm thick; roots slender, —2 cm long, long the int des; cata phylls to 4 cm long. LEAVES spreading; petioles 4-11 cm long, flattened to sulcate adaxially; blade elliptic or narrowly elliptic, (10)13-26.5 cm long, (2.5)3-9(10.5) cm wide, acuminate to long-acu- minate at apex, narrowly subcordate, pliveined at base, the upper surface semiglossy; midrib sunken; collective vein arising from the base or sed with the midrib for 0.2-2.5 cm from base. INFLORESCENCE usually arching-pendent; penduncle 4-23 cm long, terete; spathe green, * — = 294 lanceolate, 3.5-9.5 cm long, 0.7-1.8 cm wide; stipe absent or to 2 mm long; spadix violet-pur- ple at anthesis, 4.5-13 cm long, 0.3—0.5 cm diam. at base, usually less than 1 cm diam. at apex. INFRUCTESCENCE with spadix to 2 cm diam.; berries oblong-ellipsoid to narrowly ellipsoid or narrowly ovoid (with four longitudinal ridges on drying), orange to red, 0.5-1.2 cm long, with numerous round to slightly elongated raphide cells visible on drying; mesocarp orange, pulpy; seeds (1)2(3), black, suborbicular, flattened, ca. 6-9 mm long. Figs. 69, 73, and 75 Anthurium flexileis a member of section Poly- phyllium and consists of two subspecies. The ssp. flexile ranges from Mexico to Panama at 0 to 1,000 m elevation in tropical wet forest life zones. It is distinquished from ssp. muelleri by the char- acters listed in the Key. For a discussion of these differences see the discussion of ssp. muelleri. BELIZE, CAYO DISTRICT: Maya eta Cocks- comb, Schipp S-192 (MICH, MO). NN CREEK DISTRICT: Sittee River, Kendal, Schipp 192 (F, NY). TOLEDO DISTRICT: Colombia Forest Station, Proctor 36087 (MO); Maya Mountains, Gabriel Camp, Boutin & Schlosser 5153 (MO); near Monkey River, Gentle 2 (MO, NY); Rio Blanco branch, Stevenson 110 Cost TA RICA. ALAJUELA: vicinity of Bijagua, Burger & — 987 1 (F), Croat 36467 a: Llanura de San Carlos, Molina et al. 17658 . GUANACASTE: viet of » Gentry 2 on ies Standley & Va- lerio 44253, 4 290, 45020, 45997, 4611 DIA: Pedrega ARENAS: SE of the Rio Claro, Burger & Matta 4803 (F). SAN JosE: Tucurrique, Tonduz 12750 C ATEMALA. IZABAL: Río Chacón Machaca, Johnson 7 (UCLA): Cerro San Gil, along Rio Bonito, Stey- ermark 41693 (F), 41886 (F, NY). SAN MARCOS: vicin- ity Finca Armenia, above Africa, Croat 40954 (MO). SUCHITEPÉQUEZ: no other rae. Lind 73 (WIS). HONDURAS. ATLANTIDA: near Point Triunfo, Wilson 439 (NY); Mg a Ceiba, Yuncker et al. 8202 (F, MICH, MO, NY, US); Lancetilla Valley, Chickering 206 (MICH): co 42654 (MO); Molina 10454 (EAP, F); Standley 53105, rise US); 56544 (F); Yuncker 4615 (F, MO). GRACIAS A DIOS: Mosquitia Region, Rio Platano, Gentry et al. 751 18 anes XICO. CHIAP cinity ? Meuse d Ma- tuda 30187 (M EXU); vicinity p Mun Breedlove 24248, 265 Hoover 135 (MO) MEXU); N of Pi- chucalco, Miranda 6628 (ME XU); between Ranchería a Ma San Joaquin, Gilly & Her- ndez 169 XU). TABASCO: icinity Cerro de opis Gilly & Hernandez 277, 289 (MEXU); E of eapa, Conrad et al. 2881 (MO); Croat 40129 (MO); ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 Zunáü, vicinity Tapijulapa, Magaña 3127 (MO). VER- ACRUZ: no other location, Rosas 1055 D Be ical Station at Tuxtla, Calderon 2186 (F, XAL); méz-Pompa 4565 (F), Rosas & Yilalpando Hh zada 577 (F, : 2529 (XAL, 'MEXU); Goméz-Pompa & Nevling 518 je Vazquez 854 (XAL); NW of La Palma, Toledo EXU); Río Solostichil, Vazquez 1625 (XAL). cola eleki 7089 (MO); S of bridge a ia Agríco Kolini sion 6466 (MO); Vincelli 212 (MO); Wof M 3h 50 (MO); ne Río Okanwás, E of Rosita, fu «m aor vicinity "Wani including Rio Uli, Stevens 72 (MO). b. Anthurium flexile ssp. muelleri (Macbr. ) Croat & XAL); vicinity Hidalgotitlán, Dorantes (B). MA- | Puerto Cabezas, Stevens | Baker, Selbyana 2: 236. 1978. An | muelleri Macbr., Candollea 5: 348. 19 5 Anthurium mexicanum Engl., Monog hoe 2: 105. 1879, non Liebm., 1849. TYPE: | Mexico. Veracruz: Orizaba, Mueller 993 (NY, eS BR, LE, isolectotypes. here designated) Anthurium flexile ssp. muelleri differs from the typical subspecies in that the lea | ves are lance" — date |. ate, broadest below the middle, weakly cor at the base and the first pair of basal y Th not fused with the midrib or very rarely 50. 4. It spadix may be stipitate 1-1.5 € m. Fig. tions occurs in Mexico and oM at eleva of 1,000 to 1,700 m and is replaced by the yp subspecies in southern Central America. IA Mexico. No other location, Linden 136 Orbs) ve pas: Pueblo Nuevo Soliana ei p ij (CAS. Calzada 2 UEZALTENANGO: between Finca Pirineos Standley dii — 86990 (F); n Jerónimo, Finc Y Y vicinity Sa sh 3450 (P) Standley 83 Oy A: Vol Steyermark 37212, 37240 (F). SOLAL sam | 07 (P. | 1983] 295 CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA SA E e mes Li D> EM pote * TEID SITE ae KL 74 FIGURE 74. Anthurium flexile ssp. muelleri (Macbr.) Croat & Baker, Molina et al. 17658. N3QIVD 'IVOINV.LOS I300SSIN JHL AO STVNNV formosum Schott, Croat 35775.—77. Anthurium gracile (Rudge) Ya y ee 75-78. 75. rarer flexile —. RS rom 36467. —76. Ant hp M vns »at 381491 78 inthur n hacumense wl ultivated Climat n Mic 1983] Anthurium formosum Schott, Oesterr. Bot. Z. 8: 181. 1858. Type: Costa Rica. Cartago?: near Naranjo, Wendland 583 (GOET) Anthurium oe Engl., Bot. Jahrb. Syst. 6: 277. 885. Type: Colombia, Lehmann 2582 (B). Epiphytic or often terrestrial; stems often to 1 ted do subapical apiculum, drying tan, persisting intact. AVES erect-spreading; petioles 40-150 cm long, 0.6-1 cm diam., subterete, weakly flattened or weakly sulcate adaxially; geniculum 2-2.2 cm long, dues pinkish; blades ovate, 30-80 cm long, wide, acuminate and turning down at apex, diy lobed at base; anterior lobe 31-54 cm long; the posterior lobes 13-20 cm long; sinus broadly hippocrepiform to spathulate; both sur- faces glossy; midrib acutely raised above and be- low; primary lateral veins 7-9 per side, sunken in valleys above, raised below; interprimary veins sunken above, raised below; basal veins 7-8 pairs, the fourth to sixth coalesced 2.5-3 cm; the pos- terior rib naked, bins up; collective vein usu- yansi fthe lowermost basal veins, frequently from the seas or third basal vein, mm from the margin. INFLORESCENC erect-spreading; peduncle 21—47 cm long, stout Spathe subcoriaceous, ovate or plora ADR. —24 cm Né ui on Meo spadix stipitate 12-15 kc i: e 2-3 mm ack, white to violet- y e, = oblong to Pisis to both ends, 5.5- em long, 1.2-1.5 cm diam. at middle, 7-9 d us spiral, 10-20 flowers visible in the al- yos E tepals glossy, the lateral tepals 1— gulate: , the inner margins rounded to an- ple, ui pistils early emergent; stigmas deep pur- ultimately brushlike; stamens emerging from a of ex, the lateral stamens emerging ughout ens emerge, held in Shad Re E over pistil; filaments pale lav- som 5 anthers pale lavender to deep red-violet, etimes white tinged with red-violet at mar- oe 32 cm long; berries dark reddish-vi- nara Purple, often reddish-orange at ya wide Y acute at apex, 10-12 mm long, 4-5 m e seeds 2, 7-8.4 mm long, 3-3.5 mm vids: '88. 76 and 79. CROAT —ANTHURIUM OF MEXICO AND MIDDLE AMERICA 297 The species is found in Costa Rica, Panama, Colombia, and perhaps also Ecuador [see Gentry 9621 (MO)] at elevations of 500 to 1,500 m. In Costa Rica and Panama the species occurs in premontane wet forest and more commonly in premontane rain forest and lower montane rain forest. It is often one of the most common species at middle elevations, where it is generally found along shady road banks and sometimes forms large stands. Anthurium formosum is variable both in size of leaves and inflorescences and in the color of the spathe and spadix, both of which vary from white to purplish-violet at anthesis. The spathe and spadix, if white at anthesis, tend to become more colored with age. The collective vein is also quite variable, sometimes arising from one of the lowermost basal veins or sometimes from the second or third basal vein. The species is a member of section Calomys- trium and is distinguished by its usually large cordate blade with a prominent posterior rib, by its large ovate or oblong-elliptic, subcoriaceous spathe, glossy tepals, sharp, early-emergent pis- tils and reddish-violet to violet-purple berries. Leaf blades usually dry olive green with char- acteristic reddish-brown veins and with purplish and/or brownish blotches. Costa RICA. CARTAGO: Finca Navarro, Maxon 647 nillo, Croat 36680 (MO); SW of P. ; SE of Platanillo, Croat 3671 6, vicinity Platanillo, pone FH 251, Fes (US); val- 6 (BR); near Narajo, ET). ae eus Cariblanco, Croat 35775, 35834 (MO); Tirimbina, Proctor 32286 (LL); ra Blanca, Croat 35567, 35634 (MO); S of Siguirres, S rar 43308 (MO). PUNTARENAS: Las Cruces Botanical Garden, Croat 32967 (MO). SAN JOsÉ: vicin- ity Bao Han ndura, Croat 44498, 44519 (MO), Lu- teyn 3317 (DUKE, MO); Utley & Utley, 4995, 4996 Anthurium friedrichsthalii Schott, Oesterr. Bot. Wochenbl. 5: 65. 1855. Anthurium gracile Sin d Lindl. var. friedrichsthalii (Schott) En t. Jahrb. Syst. 25: 370. 1898. TYPE: Pes nae Insula Cativo, Friedrichsthal s.n. (W, Type destroyed; Photo of Schott Aroid Drawing #332, NYBG Neg. #N.S. 3819). Anthurium linearifolium Engl., Bot. "-— Syst. 25: 370. 1898. Type: Panama. Canal Zone: Río Chagres, Lehmann 4538 (B, K). ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 FIGURES 79-82 Croat 14137.—81. dut 32903 halii sa (m friedrichsthalii Schott, Croat 16557.—82. Anthurium hacumense Engl. e Anthurium formosum Schott, Croat 35775.—80. Anthurium friedrichst 1983] Epiphyte, usually pendent, often loosely at- tached by only a few roots; stem usually less than 15 cm long, 1-1.5 cm diam.; roots thin; cata- phyll thering t li fibers. LEAVES pendent; petioles 1.5-14 cm long, mostly less than 10 cm long, ca. 2-3 mm diam., terete or flattened adaxially; geniculum 5-8 mm long; blades linear, 12-56 cm long, 0.8-4 cm wide, acuminate at apex, subacute at base; upper sur- face matte to semiglossy, the lower surface con- spicuously glandular-punctate; the midrib acute- y raised, becoming sunken at apex above, convexly raised at base, flattened at apex below; primary lateral veins 8 or more per side, flat above, obscure below; collective vein arising from the base, 3-5 mm from margin. INFLORES- CENCE pendent; peduncles 19-30 cm long, ca. 2.5 mm diam., terete; spathe linear, pale green, l-5 cm long, 3-5 mm wide, rounded at base, gradually acuminate at apex, withering and soon deciduous; spadix pale green, sometimes tinged purplish, turning brownish, 3-15 cm long, 4-5 mm diam.; the flowers 4-lobed, 2.8-3.8 mm in both directions, the sides straight to weakly sig- moid; 3-4 flowers visible in the principal spiral, 2-3 flowers visible in the alternate spiral; tepals matte, minutely papillate, lateral tepals 1.6-2 mm wide; pistil not exserted; the stigma linear, 3 mm long; Stamens emerging + slowly from base, held at edge of pistil, soon completely retracting; an- thers cream, ca. 0.3 mm long, 0.9 mm wide; thecae ellipsoid; pollen white. INFRUCTES- CENCE pendent: spathe absent; spadix to 25 cm ? cm diam.; berries pale yellow-orange, ir- Ly Shaped broader than long, ca. 5 mm ong, 5-8 mm wide, narrowing to apex, truncate with a central depression; seeds 3-4, + ovate, M or yellowish, ca. 2.7 mm long. Figs. 80 long, un doces is found from Guatemala to Co- at elevations from sea level to usually m. In Costa Rica and Panama it is common MEL moist forest and is also known from lane wet and tropical wet forest. int Anthurium friedrichsthalii is recognized by its Pendent habit and the t mo i L B * lin Te typical members of that section only in its far leaves and pendent habit. CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 299 COSTA RICA. ALAJUELA: N of San Miguel, Croat 35650 (MO). cARTAGO: vicinity Santa Teresa, Luteyn 3249 Tapezco-Hacienda La Suerte, Davidson 7102, 7104 (LAM). PUNTARENAS: vicinity Monteverde, Luteyn 3398 (DUKE); Primack et al. 428 (DUKE); Wilbur et DUKE); Rincón de Osa, Liesner 1809, 1930 (MO). SAN JOSÉ: La Palma, Luteyn 3308 (DUKE). Anthurium gracile (Rudge) Lindl., Edward's Bot. eg. 19: t. 1635. 1833. Pothos gracilis Rudge, PI. Guian. Rar. 23, T. 32. 1805. Type: French Guiana, Martin s.n. (Herb. Rudge, BM). Anthurium scolopendrinum (W. Ham.) Kunth, Enum. l. 3: 68. 1841. Pothos scolopendrinus W. Ham. Type: French Guiana, Desvaux s.n. (P). Anthurium longipes Matuda, Bol. Soc. Bot. México 14: 23. 1952, non N. E. Brown, 1882. Type: Belize, Gentle 26244 (MICH). Epiphyte, usually erect; roots white; stems 1.5— 30 cm long, 7-10 mm diam.; internodes short; cataphylls membranaceous, 2-9.5 cm long, drying reddish or brown, persisting intact, ultimately deciduous. LEAVES + erect; petioles 1-20 cm long, ca. 3 mm diam., subterete flattened adax- ially; the geniculum ca. 5 mm long; blades mod- erately thin, oblanceolate, 11—32 cm long, 3-8.5 cm wide, acuminate at apex, gradually tapered to long-cuneate at base; the upper surface matte to semiglossy; the midrib prominently raised al- m p bove, raised below; primary lateral veins numerous, nearly obscure; collective vein arising from base, 4-7 mm from margin. IN- FLORESCENCE + erect; peduncles 13-40 cm long, 1-3 mm diam., longer than petioles; spathe membranaceous, red-violet, lanceolate to ob- long-elliptic or ovate-elliptic, 1.3-2.5 cm long, 3-7 mm wide, abruptly acuminate at apex, rounded at base; spadix sessile, purplish-brown, 0.6—6 cm long, 2-4 mm diam. midway; flowers 4-lobed, 5-5.6 mm long, 3.8-4.3 mm wide, the sides jaggedly sigmoid; 2-3 flowers visible in the principal spiral, 3-4 flowers visible in alternate spiral; tepals matte to semiglossy, punctate and minutely papillate; pistils weakly emergent, the apex + flat; stigma exserted, buttonlike, 0.8-0.9 mm wide; stamens emerging in scattered man- ner; anthers cream, ca. 0.7 mm long, inclined inward over pistil; thecae ovoid, slightly divar- icate; pollen white. INFRUCTESCENCE pen- dent; peduncle to 60 cm long; spadix to 10 cm long, 1.5 cm diam.; berries bright red, + globose, slightly longer than broad, 6-8 mm long; me- 300 socarp gelatinous, + translucent, lacking raphide cells; seeds 2, cylindroid, rounded on both ends, ca. 4mm long, 1.6-1.8 mm diam., with an oblique depression at one end and the other end greenish, sticky gelatinous appendages at both ends. Fig. 71. The species is found from Guatemala and Be- species is known only from topical moist forest. Anthurium gracile has a large root system that may sometimes become infested with black, stinging ants. It is distinguished by its thin, ob- lanceolate, epunctate leaves, persistent, intact cataphylls, the relatively few-flowered spadix, the whitish roots, and the bright red, globose berries. Anthurium gracile has been confused with A. friedrichsthalii, which differs in having coria- ceous, broadly linear leaves, conspicuously punctate on the lower surface, cataphylls that soon weather to persistent fibers, flowers more numerous on the spadix, and berries that are pale yellow-orange, broader than long, and truncate at the apex. Anthurium gracile is also confused with A. bakeri Hook. f., which has punctate leaves with a collective vein that is sunken and more conspicuous than the primary lateral veins. Anthurium gracileis the only described species of section Leptanthurium and is either an ex- tremely variable species or comprised of some additional undescribed sibling species in the An- des of South America (Croat, 1976). COSTA RICA. ALAJUELA: vicinity of Las Chiles, Holm & Iltis 653 (P). CARTAGO: vicinity of Turrialba, Godfrey 66241 (MO); Tonduz 8422 (BR). LIMÓN: vicinity of Siguerres, Burger & Liesner 6921 (MO); N of Puerto Viejo, Baker & Burger 91 (DUKE): Shirores, Tonduz 9231 (BR). GUATEMALA. IZABAL: vicinity Lechuga, Steyermark 39615 (MO). PETEN: vicinity El Ceibal, Steyermark 46166 (MICH). HONDURAS. ATLANTIDA: vicinity La Ceiba, Yuncker et al. 8666 (MICH, NY, RS). vono: village of Las Flores, Yuncker et al. 8187 (MO). CARAGUA. BOACO: N of Río Las Cañas, Stevens & Krukoff 5846 (MO). cHoNTALEs: E of Santo Tomas Neill 549, 7396 (MO). MATAGALPA: SW of Rio Tuma ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 bridge, Stevens 6020 (MO). Rio SAN JUAN: no other location, Atwood 5169 (MO). ZELAYA: no other loca- tion, Nichols 584 (MO); Svenson 5100 (MO); NW of Alamikamba, Stevens 8172 (MO); S of Cerro Bakan, Vincelli 319A (MO); N of El Empalme, ye s d vicinity Quebrada El Toro, along Rí 375 (MO); SW of Río Kukalaya, Stevens 8505 (MO); E of Rosita, Neill 4444 (MO); Road from Siuna to Matagalpa, Stevens 7522 (MO); vicinity Siuna, Neill 3807, 4151 (MO); W of Siuna, Stevens 6899 (MO). Anthurium hacumense Engl., Bot. Jahrb. Syst. 25: 363. 1898. Type: Costa Rica. Puntare- nas: along banks of Río Hacum (Platanar) near Buenos-Aires, 250 m elev., Tonduz 6536 (B, holotype; BR, CR, isotypes). Rosulate, epiphyte; stem short; roots few, di- rected upward; cataphylls coriaceous, ca. 9 cm long, drying dark brown, weathering to coarse linear fibers. LEAVES + erect; petioles sulcate, rounded to weakly angular abaxially, 2-12 - long, 1-1.5 cm diam.; geniculum 2-2.5 cm long. blades narrowly to broadly oblanceolate, coria- ceous, 25-81 cm long, 7.5-29 cm wide, acumi- nate at apex, attenuate at base; upper surface glossy, lower surface semiglossy, conspicuously punctate; midrib broad and flat at base S: wing a lateral veins 7-15 per side, departin. mary vel p etn below, n arising Ee spathe subcoriaceous, yellow-green ting be violet-purple, oblong-elliptic, 5.5- 2.2-3.5 cm wide, withering and early stipe becoming elongate to 29 cm at yr 5) spadix purple-violet (B & K Red-purple Hec ) 10-20 cm long, 1-2.5 cm diam. at base. k mm diam. at apex; flowers sub-rhombic to lobed, 2.4-3.6 mm long, 2-3.5 mm wide, sides + straight to jaggedly sigmoid; visible in the principal spiral, the altern@ irregular with 9—12 flowers visible; te with numerous tiny droplets; pistils flat to emergent, pinkish to dark violet-purple. fly-shaped; the stigma oblong-elliptic, a ps putter in a moderately prompt sequence s pre of spadix to both ends, the lateral stamen | weakly | | | | : | 1983] ceding alternates by several spirals, the stamens complete at apex before basal flowers complete, emitti gast g, sweet scent at anthesis; anthers dark violet-purple; exserted on fleshy, pale vi- olet-purple filaments, soon retracting to level of tepals, held in a cluster covering pistils; thecae oblong, slightly divaricate; pollen white or cream turning lavender. INFRUCTESCENCE with spadix to 23 cm long; berries bright red (B & K Red 6/5), narrowly obovoid, acute to mammil- liform at apex, ca. 1 cm long, 4-5 mm wide; seeds 2, white, weakly flattened, 2.5-2.7 mm long, 1.7- 2 mm wide, 1-1.5 mm thick. Figs. 78 and 82. The species is fc 1 from Costa Rica, Panama, and probably in Colombia in the Department of hocó, ranging from sea level to 1,560 m and occurring principally in tropical wet forest, but also in premontane wet forest. In Costa Rica the Species occurs in San José Province (El General Valley) and Puntarenas Province (Osa Peninsula and the adjacent valley of the Rio Diguis). Al- though the species is restricted to the Pacific slope in Costa Rica, it is known in Panama only from the Atlantic slope and along the Continental Di- vide on the Pacific slope. Ri cous hacumense is strikingly different ks all other species and can easily be distin- &uished in flower or fruit by its exceptionally long igh ione on the lower leaf surface and the picuous collective vein, (0.8)1—2.4 cm re- mote from the margin. rapes was placed by Engler in section lahceoí rium apparently owing to its large ob- Olate leaves and rosulate habit, but it should Placed Instead in section Porphyrochitonium of its glandular dotted lower surface and nvolute leaf vernation. i A hacumense is perhaps most closely ronpi : curvilaminum Croat (ined.), which MEL 88h petioles, smaller but similarly co- relatively s: es with a prominent collective vein Hh emote from the margin, and a spadix use lack of j A Tela al ri Or oft : len exceeding the length of the spadix. Cosr PDA CARTAGO: near Selva, Taylor 11463 (NY). Nas: Corcovado National Park, Liesner 2849, CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 301 2851 (MO); N of Palmar Norte, Croat 35194 (MO); Rincón de Osa, Liesner 1698 (MO); Utley 1190 (DUKE); Santo Domingo, Engler 10094 (BR); Playa Blanca, Valerio 323 (BR); vicinity San Vito de Java, Croat 32903 (MO). SAN JOSÉ: between San Isidro del General & Dominical, Croat 35340 (MO). Anthurium halmoorei Croat, sp. nov. P ico. Nayarit: along Hwy. 28 between Tepic and Jalcocotán at 15.5, elev. ca 1,000 m, Croat 45337 (MO-2690192-193, holotype; CAS, K, MEXU, MICH, SEL, iso- types; Live at MO). Planta epiphytica aut rupestris; caudex brevis, cras- sus, radicibus densis crassisque; petiolus 15-19 cm lon- gus, 12-18 cm latus, laterib pl tis. adaxile late sulcatus, abaxiale obtuse angulatus aut costatus; la- mina oblanceolata aut oblanceolata-elliptica, subcori- acea, 52-108 cm longa, 15-43 cm lata, marginibus late undulatis, nervis primariis lateralibus ad margi- nem liberibus; inflorescentia effusa, foliis breviora; pe- dunculus 32-38 cm longus; spatha ovata aut elliptica, viridis, 13-16 cm longa; spadix viridis aut purpurata, 9-21 cm longa; baccae flavovirentes. Rosulate, epiphyte or on rocks on very steep slopes; stems thick; internodes and leaf scars ob- scured by root mass; roots dense, descending, ca. 1 cm diam.; cataphylls subcoriaceous, 20-23 cm long, apiculate at apex, drying tan to brown, per- sisting intact at apex, dilacerating at base. LEAVES erect-spreading; petioles 15-19 cm long, 12-18 mm wide, flattened laterally, broadly sulcate adaxially, obtusely angular or 1-ribbed abaxially; geniculum 1-1.5 cm long; blades ob- lanceolate to oblanceolate-elliptic, subcoria- ceous, short-acuminate at apex, attenuate at base, 52-108 cm long, 15-43 cm wide, broadest at middle or just above, the margins broadly un- dulate; the upper surface semiglossy, the lower surface matte and much paler; midrib broad and flat at base, convexly raised at middle, narrowed and flat at apex above, acutely raised below, the abaxial rib of the petiole extending onto midrib; primary lateral veins 12-14 per side, departing the midrib at a 407-55? angle, prominently raised above, raised below; straight then curving toward apex near margin; interprimary veins less con- spicuously raised above, visible and flat below; secondary veins prominulous above, flat below; collective vein arising from a primary lateral vein near the apex, 6-11 mm from margin, flat above, prominulous below. INFLORESCENCE spread- ing, shorter than leaves; peduncle 32-38 cm long, 9-10 mm diam., subterete, obscurely angled, 2- 2.5 times longer than petioles; spathe subcoria- 302 ceous, green (B & K Yellow-green 6/10), some- times tinged purple, ovate to broadly ovate, 13- 16 cm long, 6.2-9.7 cm wide, broadest near base, acuminate at apex, acute at base, the margins undulate, inserted at a 70° angle on peduncle; spadix green, purple, or green heavily tinged pur- ple (B & K Red-purple 2/5), 9-21 cm long, 11- 27 mm diam. at base, 5-7 mm diam. at apex; flowers 4-lobed, 2.4-2.8 mm long, 2.7-3.2 mm wide, the sides sigmoid; 8-14 flowers visible in the principal spiral, 12—20 flowers visible in the alternate spiral; tepals matte, weakly punctate, minutely papillate, lateral tepals 1.5-1.8 mm long, the inner margin convex; pistil slightly raised, green tinged with purple in area of stigma; stigma linear, 0.4-0.6 mm long, a minute droplet ap- pearing 4—5 days before first stamens emerge, dry with white exserted papillae when anthers open; wide, retracting to hold anthers held at edge of pistil, mounded with pollen; anthers orange; the- cae ellipsoid, scarcely divaricate; pollen orange B & K Yellow 7/2.5), drying white. IN TESCENCE pendent; spathe usually i brown square around style, 12-15 mm long, 7- 9 mm diam.; mesocarp fleshy with numerous raphide cells; seeds 2, pale tan, subglobose, weak- ly flattened, 5-6.5 mm long, 3.5-4.1 mm wide, ca. 2.5 mm thick, with a sticky append ge at base and apex. Figs. 83 and 87 Anthurium halmoorei is restricted to western Mexico in the states of Nayarit, Jalisco, Colima, and Michoacán in seasonally ve ry dry forests from 450 to 1,450 m elevation. It occurs usually on rocks on very steep slopes in forested areas. The vds is in section Pachyneurium and is characterized by its generally oblanceolate or ob- eint. blades with usually free-end- ing primary lateral veins, and by its petiole which is acute Me ub but especially by its ovate to ovate-elliptic spathe, green to purplish spadix and greenish-yellow mature berries The species is probably most ro related to A. schlechtendalii ssp. Jimenezii, which differs in having the petiole subquadrangular, usually flat abaxially and in having a lanceolate spathe and red berries at maturity. That KREE occurs only in Guerrero and southern Oaxa XICO. JALISCO: vicinity Autlán de Navarr eves 10224 (MICH); Wilbur 2289 (MICH); be. ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 tween El Tuito & Puerto Vallarta, Delgado 359 (MICH); Talpa "de Allende, McVaugh 23408 (MICH). MICHOACÁN: vicinity Aguililla, McVaugh 24741 (MICH); Coalcomán de Matamoros, Rzedowski 16707 (ENCB, MICH). NAvaRiT vicinity Jalcocotán, Gentry & Gilly 10734 (MEXU, US); McVaugh 12140 (MICH) Tepic, Lewis s.n. (MICH); Ferris 6816 A W of Te- pic, Dressler 1024 (UC); between Tepic and Jalcocotán, Croat 45337 (CAS, K, MEXU, MICH, MO, SEL). Anthurium hoffmannii Schott, Oesterr. Bot. Z 8: 181. 1858. Type: Costa Rica. San José: Chica Uruca, Hoffmann 571 (B, destroyed; Photo of Schott Aroid Drawing #662, NYBG Neg. #N.S. 3897). Anthurium saad Engl., Bot. pu^ d 25: - 98. LEcroTYvPk: Costa Rica. San cisco de E 1,100 m, Tox 7 7176 (US. hololectotype; F, M, NY, isolectotype; designated Croat & Baker, 1979). Epiphyte or terrestrial; stems to 60 cm or lon- ger; internodes 1—2 cm long; leaf scars obscured by cataphylls; roots few, green, 3-8 mm a cataphylls coriaceous, 7-15 cm long, drying dish brown, persisting intact. ma S with pet- iole erect-spreading, 20-90 cm long 4-8 1: diam., terete or weakly sulcate; ges um "o cm long; blades ovate, moderately thick, 1 cm long, 9-38 cm wide, acuminate at apex, deep- ly lobed at base; anterior lobes 11-44 cm o: posterior lobes 4-12 cm long; sinus parabo d spathulate; the upper surface semiglossy, the e surface glossy; the midrib convexly raised à 23 acutely raised below; primary lateral veins per side, weakly sunken above, raised ene al veins 4—7 pairs, 3—4 coalesced 1-4 cm ( free to base); posterior rib naked, weakly t upward; collective vein arising from first 3-10 vein, sunken above, acutely raised ce eret rf mm from margin. IN RESCENCE € slightly arching; peduncle 16-50 cm long. ! mm diam., terete; spathe, pale yellow-£r€€ coriaceous, oblong-ovate to lanceolate, long, 1. Min m wide, inserted at 70 an j : . flowers wide, the sides straight to sigmoid, 5-8 CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 5337 (Type). — 84. Anthurium hoffmannii Schott, FIGURES 83-86. 83. Anthurium halmoorei Croat, Croat 4 Ant ann n Anthurium interruptum Sodiro, Selby 8 we as Anthurium huixtlense Matuda, Croat 40583.—86. 5-32. Han 58-1 GURES 87-90. Fi huixtlense Matuda € Wi. roat 5 pope appa QU Cro thurium 199 ' at, Croat — 37 TA ME "iy k unthii Poep ‘reat Anthurium hoffmannii Schott, Burger & Baker 10092.—89. Anthurium N3Q3VD 'TVOINV.LOS8 THNOSSIW JHL AO STVNNV POE 0L 10A] 1983] visible in the principal spiral; 8-13 flowers vis- ible per alternate spiral; tepals glossy, obscurely punctate, lateral tepals 1.4 mm wide, the inner margin + straight; pistils weakly and acutely raised, white, minutely papillate; stigma linear, 0.3 mm long, a minute inconspicuous slit with minute droplets briefly apparent ca. 1 week be- fore stamens emerge; stamens emerging rapidly, scattered throughout spadix, the lateral stamens first; anthers white, held in + contiguous square at edge of pistil, 0.4—0.7 mm long; thecae ovoid, weakly divaricate; pollen white. INFRUCTES- CENCE arching-pendent; spadix to 15 cm long, to 2 cm diam.; berries red, ovoid, ca. 8 mm long, ca. 5 mm wide; pericarp with elongate raphide cells; mesocarp clear, gelatinous; seeds 2, brown, a. 4 mm long, ca. 2 mm wide. Figs. 84 and 88. Anthurium hoffmannii is found in Costa Rica and Panama at elevations mostly between 700 and 1,700 m, principally in tropical wet, pre- montane rain and lower montane rain forest but also in premontane wet forest. This species is in section Calomystrium and is closely related to (and perhaps not separate from) A. huixtlense, which ranges from Mexico (Chiapas) to northern Nicaragua, and to 4. monteverdense from Costa Rica. It is also re- ated to a meer of unnamed taxa from Pan- ama. All these species share a similar thick, ovate, Prominently basally lobed blade, and thick, per- sistent cataphylls that dry reddish-brown and re- Main intact in age. Leaves generally have nu- ae Short, linear, raphide cells on the upper ant "e which can be seen both before and after mith H991 (NY). anillo, py 36725 (MO). AS: ríe poe 4655 (BR); Buenos Aires, BR); Las Cruces Botanical Garden, a » Burg : I DS Webster et al. 12403 (US); S of San Vito, Raven 3 (DX N JOSÉ other location, Luteyn 3282 42054 (us ler 455 (BR, UC, US); Standley 42015 Gua S; Alajuelita, Pittier 8839 (BR); vicinity upe, Pittier 8764, 9858 (BR); Tonduz 71 & t at Hips rae vicinity San Isidro del General, Burger oat oat 35266 A (MO); Burger & Visconti 10679 (MO); PS 66 (MO); Jiménez 3836 (F); Molina et al. (CAS); Skutch 3076 (GH, MO, NY). A peres. huixtlense Matuda, Revista Soc. Mex Ist. Nat. 11: 91, Fig. 1. 1950. TYPE: Mex- ico. Chiapas: near Cons Piedra de Huixtla, CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 305 20 km NE of Huixtla, 900 m elev. Matuda 18615 (MEXU, holotype). Au bochilense Matuda, Bol. Soc. Bot. México 1956. Type. Mexico. Chiapas: Bochil, near San Med de Las Casas, MacDougall 363 (MEXU, holotype). Terrestrial or epiphytic; stems to 14 cm long, 3 cm diam.; leaf scars 2-3 cm wide, obscured by cataphylls at upper nodes; roots few, ca. 5 mm diam., descending; cataphylls coriaceous, 5-23 cm long, long-acuminate at apex (the acumen apiculate ca. 1-2 mm), ing dark tan (B & K Yellow 5/2.5), persisting intact. LEAVES erect- spreading; petioles 15-70 cm long, 6-9 mm diam terete; geniculum 1-5 cm long; blades ar riaceous, ovate to broadly ovate, 17-70 cm long, 10.5-36 cm wide, broadest just below point of petiole attachment, short-acuminate at apex, deeply lobed at base; anterior lobe 14-50 cm long, the margins rounded; posterior lobes 5-20 cm long, flat to downturned, sometimes over- lapping; the sinus triangular to hippocrepiform, sometimes "s rhombic i in arger leaves when flat- tened; th sparse, Caoa raphide cells, the lower surface a basal veins 5-6 pairs, the third to fifth Ead 2-2.5 cm, raised; the posterior rib naked, turned up on the outer margin; primary lateral veins 3- 5 per side, prominulous above and below, de- parting midrib at 30-40? angle, straight or arch- ing near collective vein; lesser veins less con- spicuous; collective vein arising from one of the lowermost primary lateral veins or from the first basal vein, flat to weakly raised, 4-10 mm from margin. FLORESCENCE erect-spreading, shorter than leaves; peduncle 21-39 cm long, 4- 7 mm diam., terete; spathe coriaceous, pale whit- ish-green (B & K Yellow-green 9/10), oblong- lanceolate, 4.5-15 cm long, 2.7-4.5 cm wide, broadest at middle, abruptly long-acuminate at apex, obtuse to rounded at base; the stipe 12 mm long in front, 2.5 mm long in back, pale green; spadix reddish-violet to lavender (B & K Red 8/ 5. red-purple 5/2.5 to purple 6/5), rarely white, 4.2—20 cm long, 6-10 mm diam. at base, 5-6 mm diam. at apex; the flowers 4- lobed, 2.3-3.1 mm long, 2.5-3.5 mm wide, the sides jaggedly sigmoid; 5-8 flowers visible in the principal spi- ral, 5-9 flowers visible in the alternate spiral; tepals semiglossy, with white inconspicuous punctations, smooth to very minutely papillate, 306 lateral tepals 1.5-2.2 mm wide, the inner margin straight; pistil emergent, white, glossy; stigma 0.3— 0.6 mm long, linear; stamens emerging moder- ately rapidly from the base, lateral stamens emerging to apex before alternates emerge at base, held in tight cluster obscuring pistil; anthers white to pale yellow, 0.6-0.7 mm De 0.6-0.8 mm or not at all de; ber- ries orange, globose; seeds green, up to 3 per berry. Figs. 85 and 89 Anthurium huixtlense ranges from southern Mexico (Chiapas) to Nicaragua. The species is a typical member of section Cal- omystrium and can be recognized by its thick, persistent cataphylls that turn reddish-brown and remain intact, and by its relatively long, usually reddish-violet to lavender spadix with early emergent pistils and orange berries. In Flora of Guatemala (Standley & Steyermark, 1958) this species was called A. concinnatum, an unrelated species from high elevations in Costa Rica and Panama. The species is closely related to and is perhaps only subspecifically distinct from A. hoffmannii, which ranges from Costa Rica to Panama. An- thurium hoffmannii has a moderately thick green spathe and a white spadix. In addition, the tepals of A. hoffmannii are generally matte. In A. huix- tlense the tepals are glossy and the pistil i 1s rather usually soon hid- den by the "tight cluster of sminnn giving the spadix a coarse, almost “prickly” look. The spa- dix of A. hoffmannii has the pistils scarcely exserted at or near anthesis so that the spadix does not have a “prickly” look. Spathe and spadix color are quite variable for A. huixtlense. Spathe color may be green to white and spadix color ranges from creamy white to lavender or reddish-violet. For some popula- tions the color remains consistent but since con- siderable color variation may be found within a population the spadix color is not a useful taxo- nomic character in A. huixtlense The plants of A. huixtlense with lavender spadices are similar to those of 4. armeniense, which has somewhat pinkish spadices and a white spathe. However, that species is distinguished by having the tepals matte rather than glossy and in having an early emergent pistil that is as long or longer than the stamens when they emerge. ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 Another close relative of both A. hoffmannii and A. huixtlense is A. erythrostachyum Croat (ined.) from Panama. The latter species differs from both of the former species by its caudate- acuminate spathe with the acumen up to 3 cm long. It differs also from A. hoffmannii in having the spadix colored somewhat lavender. Anthurium huixtlense has been confused with A. fraternum Schott since the time of Engler. 8 mian and A. fraternum share a similar leaf shape (as do most members of the section Cal- omystrium) and a similar spathe and spadix shape, but A. fraternum differs in having a dense array of discrete, punctiform, glandlike dots on the upper surface of the blade on drying an nd has aS astuiiinsi sida That n occurs in the Sierra Nevada de Santà | Maria of northeastern Colombia. Anthurium a dense array of minute, reddish-brown blotches — on the lower surface. Anthurium fraternum is not known from Central America. Pm 4 BELIZE. CAYO DISTRICT: vicinity La Flor, Croat 23 78 Oy. Bernhardt $ EL SALVADOR. SAN SALVADOR: La Palma, s.n. cdd did ALA. ALTA VERAPAZ: vicinity Cobán, C 41470, "41479 (MO); Tuerckheim 3795 (GH); F ke Cubilhuitz, Tuerckheim 8611 (GH); Río Ez Johnson 723 (U Luteyn & Almeda 3524 Mod W of d t MO). ESCUINTLA: Volcan Oo ia ge N of Highway M Harmon d eA. 1926 (MO). QUEZALTENANGO: v pei lín, Standley 86887 (F). SAN MARCO; s: Fin pue Dwyer 14459 MOX SOLALÁ: Vols Atitlán, mark 47997 (F, GH). SUCHITEPEQUEZ: Santa ed iv 12 (WIS). zacAPA: Volcán de Monos, Steyi ne ba, Yune ncker A URAs. ATLANTIDA:vicinity La Ceiba ring et al. 8741 F. GH, MO); Lancetilla Valley, Chicke P: 204 (MICH); Croat 42673 (MO): doni P cH : 7m d h- cinity Peñas Blancas, Hernandez idi Lisa Santa Cruz de Yojoa, rohs EPEQUE: vicinity Nueva Ocotoped Molina 30950 (MO). SANTA BARBARA CH, MO oa, Croat 42751 (MO); Yuncker 4859 pers 7968 cinity San Fernando, Tenejapa, Breedlove 7560 (CAS); Tuxtla Pcie ; 25976 e 1983] MacDougall 375 (MEXU). veracruz: Hidalgotitlan, b is UR ee (XAL); Vazquez 51 (XAL ICARAGUA. No other location, Williams et al. 24733 m NUEVA SEGOVIA: S of Jalapa, Neill 1666 (MO). Anthurium interruptum Sodiro, Anales Univ. entr. Ecuador 15: 301. 1902. Type: Ec- uador. Western slopes of the Cordillera de Angamarca, Nov. 1899. Sodiro s.n. (B, ho- lotype) Anthurium volti Standl. & Steyerm., Publ. Field Mus t. Hist., Bot. Ser. 23: 210. 1947. Type: Guatem ÉD Alta’ Verapaz: along Rio Icbolay, 300- be is. 6 March 1942, Steyermark 44776 (F, ho- ot Scandent epiphyte; stems elongate with 1 or more short internodes alternating with a long internode, 8-33 cm long, appearing to have sev- eral leaves at each node, often with a short, leafy, branch with several short internodes at thes nodes; rooting profusely at each node; cataphylls ng, drying brown, dilacerating and ultimately deciduous. LEAVES spreading; peti- oles * terete, 1.8-13 cm long; blades lanceolate to elliptic, oblong-elliptic or eres sub- coriaceous, 5-21 cm long, 2.2-9.5 cm wide, acu- minate at apex, acute to rounded at vene 5-21 cm long; the midrib raised above, flat to prom- inulous below; primary lateral veins 9-12 per side, departing midrib at 45°-55° angle, + ob- scure above, usually obscure below; collective vein arising from the base, prominulous above, : t below, 4-6 mm from the margin. INFLO- io CE erect-spreading; peduncle 4.5-14 cm la & longer than petioles; spathe green, oblong- neeolate, 3.7-5 cm long, 1-1.5 cm wide, acu- Minate at apex, acute at base; spadix green, be- aan, Cung 2.5-4 cm long; flowers quu ca. 3 mm in both directions, the sides hei ca. 3 flowers visible in either spiral; h ent, green, mammilliform; stamens ld above PUn longo as iod l n 1 mm long; pollen pale yellow. IN- CTESCENCE spreading-pendent; spadix to ^m long; berries red, broadly obovoid, ca. 7 m : m long, flattened at apex with central depres- 9n. Figs, 86 and 95 mie 1$ known from Guatemala to Pan- the Pacif or, and surely will be found also on Panama j € slope of Colombia. In Costa Rica and ical “ot It is known from premontane wet, trop- rain for Premontane rain, and lower montane est is elevations from 150 to 1,000 m. In Osta Rica the species is known on the Atlantic CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 307 slope from northwestern Alajuela to southeast- ern Cartago Province. In Panama it is known from both the Atlantic and Pacific slopes. Anthurium interruptum is in section Xialo- phyllium and can be recognized by its scandent stem with greatly, elongated internodes alter- nating with two or more successive abbreviated internodes giving the appearance of several leaves being clustered at a node. In addition, no other scandent, entire-leaved species is known to have bright red berries. Anthurium i. tieri and A. carnosum. Steyermark, in Flora of Guatemala (Standley & Steyermark, 1958), com- bined A. radicosum and A. pittieri, apparently due to their vegetative resemblance. However, A. pittieri can be differentiated by the long-de- current spathe and shorter internodes. They also differ in altitudinal and geographical distribu- tion. Anthurium pittieri is known only from Cos- ta Rica and Panama while A. interruptum ranges from Guatemala to Ecuador. The former species also has a range of 50 to 2,100 m altitude but mostly above 1,000 m. In Ecuador the species may range to about 1,300 m. Anthurium inter- t resembles A pit- 100 and 1,000 m. Anthurium carnosum differs from A. interruptum in its nearly uniform inter- node length, thicker leaves and altitudinal range of 1,500 to 2,600 m. BELIZE. TOLEDO DISTRICT: Pueblo Viejo, Schipp S-678 S). CosTA RICA. ALAJUELA: pies Bijagua, along road to Pies Paci & Baker 9861 (MO); ves t 36432 (MO). c s Mons de Ouch rada AGO: roa tween Pla RAM qur 36631, 36655 (MO); SE of Platanillo, Croat 36715 (MO); between Río Taus & Quebrada Burger et al. 10029 (DUKE, MO, NY); Turria Río Reventazón, Lent 693 (M Puerto Viejo de Sarapiquí, Jiménez 101 (MO). LIMÓN: S of Siguirres, pis 43323 (MO). NICARAGUA. ZELAYA: road to Colonia Agrícola Yo- — & Colonia La Espera ranza, Stevens 631 E Bra vens 5 7521 (M e^ Anthurium kunthii Poepp. in Poepp. & Endl., Nov. Gen. et Spec. 3: 84. 1845. Type: Peru. Loreto: MS Poeppig s.n. (W, destroyed; isotypes ? Epiphyte, scandent, stems usually to 1 m or longer, 1-1.5 cm diam., internodes 4-20 cm long; leaf scars 8-1 x mm wide; roots numerous at "o node, 8.5 cm long, drying yellow to tan (B & K eie 308 red 9/10), persisting as linear fibers at each node. LEAVES spreading; petioles 20-60 cm long, 2- 3 mm diam., terete, shallowly sulcate; geniculum 0.3-0.9 cm long, more deeply sulcate than pet- iole; blades ovate, 22-60 cm long, 20-40 cm wide, 5-, 7-, or 9-pedatisect, p leaflets moder- ately thin, elliptic to ovate-elliptic, 7-27 cm long, 3-8 mm wide, long-acuminate at apex, acute to attenuate at base; petiolules to 5.5 cm long, 1-2 mm diam., broadly and sharply sulcate; upper surface semiglossy, lower surface semiglossy to matte; midrib prominulous above, prominently pars below; primary lateral veins 5-18 per side, sunken jd raised below, departing midrib at 459-55? a Oop-connecting, weakly sunken mm from margin. INFLORESCENCE spreading to arching, longer or shorter than the leaves; pe- duncle 20-55 cm long, terete; spathe green or sometimes slightly tinged purple, oblong-lanceo- late, 5.5-15.5 cm lon .8-2.5 cm wide, acu- minate at apex, acute to obtuse at base, inserted at 50°-70° angle on peduncle; stipe 5-9 mm long in front, 0.5-3.5 mm long in back; spadix pale green fading to tan at anthesis, 7-21 cm long, 5- 10 mm diam. at base, 1-4 mm diam. at apex; the flowers rhombic to 4-lobed, 3.5-4 mm long, 2.5-3 mm wide, the sides straight to sigmoid; 5— 9 flowers visible in the principal spiral, 5-7 flow- ers visible in the alternate spiral; tepals matte, minutely papillate, lateral tepals 1-1.5 mm long, the inner margin convex and upturned; pistils m wide; thecae ellipsoid, widely divaricate; pollen white. INFRUCTES- CENCE pendent; spadix to 25 cm long; berries purple to violet-purple. Figs. 90 and 91. Anthurium kunthii ranges from Costa Rica to Peru. In Costa Rica it is known only from trop- ical moist forest in Limón Province. In Panama it occurs in tropical moist and premontane wet forest on both slopes at elevations below 500 m. Because of its palmately compound leaves, A. kunthii is confused only with A. pentaphyllum var. bombacifolium. Anthurium kunthii differs by having long, slender peduncles, nearly equal- ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 ling the petioles in length, and by having a slen- der, elongate, usually greenish spadix. It also dif- fers from A. pentaphyllum var. bombacifolium in having leaflets distinctly petiolulate and more or less equilateral at the base. Anthurium pen- taphyllum var. bombacifolium usually has at least the outer leaflets inequilateral at the base, a short peduncle, and a pale violet-purple spadix. The species is a member of section Dactylophyllium. Costa RICA. LIMON: Talamanca, Ocampo 1904 (MO). NICARAGUA. ZELAYA: Cerro Baka, Coperna; Pipoly 4920 (MO). Anthurium lancetillense Croat, sp. nov. TYPE Honduras. Atlantida: Lancetilla Valley, cà 10 mi SE of Tela; in forest preserve along Rio Lancetilla, on trail to water reservoir, 10-150 melev., Croat 42672 (MO-2583478- 79, holotype: EAP, K, MEXU, SEL, TEFH, US, isotypes; Live at MO). Planta terrestris; caudex ad 15 cm longus, Pest 17-82 cm longus, teres; lamina anguste ovata, rescentia effusa-erecta, foliis breviora; pe 36 cm longus; sped lees interdum intra Wer lanceolata, 8.5 onga, 2-3.2 cm lata; et atropurpureus, x a ac 9, I pollen auran cum; baccae aurantiacae, ovoidea Terrestrial; stems to 15 cm long, 2.5-3 cm ; leaf scars conspicuous, 2.3-3 cm wide: roots din. 6—8 mm diam., descending; cataphyl subcoriaceous, 9-13 cm long, acute at ape. nutely subapiculate, drying tan, dilacerating base, remaining intact at es persisting aroun base of leaves. LEAVES wi ee 82 cm long, 5-10 mm pope terete; genic 1.5-3 cm long; blades narrowly ovate, ately thick, gradually long-acuminate lobed at base, 38-54 cm lo e, broadest midway between base an ie petiole attachment, margins weakl an anterior lobe 24-37 cm long, the margins mee ed to concave; posterior lobes 12-19 cm en à the sinus spathulate : hippocrepiform, à acu surfaces semiglos n e, re acutely pi spicuously upward; primary lateral veins eee side, departing midrib at 45°-50° angle. T2! raised ve midrib, sunken near collective vein a above below; lesser veins less conspicuous, MEM SS ASS aa a 1983] and below; collective vein arising from the first basal vein, sunken above, raised below RES E o pa green, sometimes heavily tinged with violet-pur- ple on inner surface, lanceolate, 8.5-16 cm long, 2-3.2 cm wide, gradually long-acuminate at apex (the acumen inrolled ca. 3.5 cm) rounded at base, inserted at 45? angle on peduncle; spadix dark violet-purple (B & K Blue-purple 2/10), 11—20.5 cm long, 7-12 mm diam. at base, 3-5 mm diam. at apex; flowers rhombic to weakly 4-lobed, 3.3- 5 mm long, 2.8-4.5 mm wide, the sides straight to weakly sigmoid; 5-7 flowers visible in the principal spiral, 4—12 visible in the alternate spi- ral; tepals matte, obscurely punctate, minutely papillate, lateral tepals 2.3-4 mm wide, the inner margin concave; pistils violet-purple around sigma, green below, weakly emergent; stigma ellipsoid, 1 mm long, opening ca. 2 weeks before first stamens emerge, large milky droplets ap- parent for several days, drying before first sta- mens emerge; stamens emerging in a relatively rapid sequence from the base, the lateral stamens emerging in basal one-quarter before alternates emerge, exserted on translucent filaments, 0.7 mm long, 0.5 mm wide; anthers pale orange, 0.6— 0.7 mm long, ca. 0.8 mm wide, held in a contigu- one- = long, 8-11 mm diam., acute at apex; me- 6.2 mm wide. Figs. 92 and 96. Anthurium lancetillense is known only from Onduras in the Lancetilla Valley at elevation f fom 10 to 600 m. It occurs in wet forest. Pres Pecies isa member of section Belolon- an and is characterized by its narrowly ovate, anette thick leaf blade, its green, subcoria- wi parte that is sometimes tinged violet-pur- res n surface, its deep violet-purple opar pes nge pollen, and ovoid orange berries. e saes lancetillense is most clearly related in u— rosum, which Occurs only in Mexico Bitis i It can be distinguished by its longer, €ct petioles and shorter more erect inflo- ves. Anthurium lancetillense has larger, CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 309 ovoid berries while A. umbrosum has smaller, globose berries with 4-radial ridges at apex. An- thurium umbrosum is also found at much higher elevations ranging from 1,200 to 1,800 m. HONDURAS. ATLÁNTIDA: Lancetila Valley, Croat 42672(EAP, K, MEXU, MO, SEL, TEFH, US); Stana- ley 53556 (F, US); Webster et al. 12706 (DAV, US); vicinity of San Alejo, Standley 7594 (F). Anthurium lancifolium Schott, Oesterr. Bot. Z. 8: 180. 1858. Type: Costa Rica. Cartago: Volcán Turrialba, Wendland 518 (GOET). Anthurium allenii Standl., Publ. Field Mus. Nat. Hist., Bot. Ser. 22: 66. 1940. Type: Panama. Coclé: north rim of El Valle, elev. 800-1,000 m, Allen 1854 (MO). Anthurium firmum Engl., Bot. Jahrb. Syst. 25: 460. 898. Type: Costa Rica. Puntarenas: Cañas Gor- das, 1,100 m, Pittier 11130 (B, holotype; BR, CR, isotypes). Anthurium densinervium Engl., Bot. Jahrb. Syst. 25: 86. 1898. Type: Costa Rica. Along Río Hondo near Madre de Dios, elev. 200 m, Pittier 10346 (CR, holotype; B, isotype). Epiphyte or terrestrial, 30-90 cm tall; stems 1-1.5 cm diam.; the internodes very short; roots short, 3.5 mm diam.; leaf scars obscured by per- sisting cataphylls; cataphylls 4—7 cm long, drying reddish-brown weathering into coarse fibers. LEAVES spreading; petioles 15-45 cm long, 3- 5 mm diam., obscurely and narrowly flattened adaxially sometimes with a faint medial rib; ge- niculum ca. 1.5 cm long, tinged purplish; blades subcoriaceous, lanceolate to ovate-lanceolate or oblong-elliptic, 20-30(50) cm long, 4-12 cm wide long-acuminate at apex, cuneate at base; both surfaces semiglossy with black glandular punc- tations, more conspicuously so on lower surface; midrib narrowly and convexly raised above, be- coming more acute at apex, prominently raised below; the primary lateral veins 15-25 per side, departing the midrib at 35° angle, sunken in val- levs above, raised below; lesser veins obscure; collective vein arising from near the base, rela- tively straight, extending to the apex, 5-15 mm from the margin. INFLORESCENCE - erect, shorter than leaves; peduncle 26-60 cm long; spathe green, oblong-lanceolate, 4.3-8.5 cm long, 0.8-1.9 cm wide, broadest at base, abruptly cus- pidate at apex; spadix gray- or greenish-white, 4.5—9.5.cm long, 5-7 mm diam. at base, 4-5 mm diam. at apex; flowers rhombic to 4-lobed, 3-3.3 mm long, 3.3-3.6 mm wide, the sides straight to 310 broadly sigmoid; ca. 5 flowers visible in the prin- cipal spiral, ca. 8 flowers visible in the alternate spiral; tepals glossy, smooth, the lateral tepals ca. 2mm pendent; spadix to 15 cm long; ber- ries violet-purple, ovoid to globose, ca. 3-5 mm wide; mesocarp pulpy; seeds white, broadly el- lipsoid, to 4 mm long, 3 mm wide, 2 mm thick, flattened. Fig. 93. Anthurium lancifolium ranges from Costa Rica to Colombia (Antioquia and Chocó) at eleva- tions of 100 to 1,600 m (especially 800 to 1,200 m) in regions of premontane wet, tropical wet, premontane rain, and lower montane rain forest life zones. It is the most common and widespread species in section Porphyrochitonium and is also one of the most variable. Blades vary in sha from linear-lanceolate to oblong-lanceolate to narrowly ovate with their dried coloration vary- ing from green to olive-green to brown The species is recognized by its short stems with persistent, weathered cataphylls, frequently terrestrial habit, generally elongate, gradually acuminate, glandular-punctate blades that more or less equal the petiole in length, by the green, lanceolate spathe, and especially by the grayish- white, moderately short, weakly tapered spadix that has stamens persisting after anthesis, unlike most species in the section, that promptly retract their stamens. The species is known to have two varieties and more varieties or perhaps even dis- tinct species are expected to be segregated from this variable, wide ranging species as living ma- terial ofthe group is studied in detail. The typical variety has dark violet-purple berries from the beginning of berry grins: whereas the va- riet hi tc before maturity becoming faintly Sole Videt at the base at maturity. Leaves of variety albifructum also differ from those of the typical variety found on Cerro Colorado where they occur together. Those of variety albifructum are narrowly ovate, rather than lanceolate, and dry yellowish- -green on the lower surface. Most blades of the typical lancifolium dry rather dark. Anthurium lancifolium is similar to A. aureum Engl. of South America, which has leaves drying golden green beneath. ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo.. 70 A noteworthy collection (D’Arcy 11259) from | Panama, in Colón Province at Río Iquanita (390 | m elevation) may ultimately prove distinct. It | has unusually small leaves, measuring only 21 4 cm | eee 1 h t } 11 1 2 4 th ered to fibers as is the usual case in A. /ancifo- } lium | Costa RICA. ALAJUELA: near Cariblanco, Burger & — 11104 (MO). cARTAGO: between Morav via and | tes Platanillo, Croat 36629, 36693 (MO); Vol- | Wendland 518 (GOET). HEREDIA: Sof | Cari iblanco, Croat 35793 (MO). sei vi of Guapiles, Standley 37512 (US). PUNTARENAS: Cañas > Gordas, Pittier & Tonduz 11130 (BR). SAN ue Finca | ava LAYA: vicinity Quebrada El Toro, along | Río pon "Vincelli 375A (MO). | Anthurium lentii Croat & Baker, Brenesia l6 (Supl. 1): 56. 1979.Type: Panama. Chiriquí: } vicinity of Cerro Colorado Mine Develop- or | ment, 28 miles above San Felix, 1, 200-1,500 m, Croat 33204 (MO-2381193, holotype: | Live at MO). Epiphyte or terrestrial, 56-120 cm tall; con a coriaceous, (5)8-19 cm long, turning brown, MU- cronate at apex, remaining intact, event e lacerating. LEAVES with petioles erect to aa spreading, 10-70 cm long, (5)8-9 m sal | broadly sulcate _— Meum s more de er | geniculum 2 lish; bla a -erect cate near ape ially, sometimes tinged purplish; cm long, sulcate, usually pu ovate to narrowly ovate, coriaceous, semi-€ ‘il or spreading from petioles, rounded to em apiculate at apex, truncate to subcordate OF ide, date at base, 15-50 cm long, 12.5-37 em b | broadest at point of petiole arae the the | gin pale green, often slightly tu anterior lobe to 46 cm long cael ma and often irregular where basal veins mee ri margin; the posterior lobes to 14 cm ed rected conspicuously upward; the em at | row to broad (when present), = lower sif brown" n ish glandular-punctate; the aver | raised above, diminished oid apes, veins | raised below and usually -— en many } usually 3 or 4 pairs, sometim with as them 6 pairs on cordate blades, up e 4 om nea | | alesced 0.5-3 cm, sunken or raised in 1983] the midrib on upper surface, sharply raised and s + straight and connecting pairs xi basal veins, + flat and nearly as prominent as the primary lateral veins above, slightly raised feiss. collec- tive vein arising from uppermost basal vein, 2- 5 cm from the margin midway on blade, raised at base above, sunken at apex, tinged purple be- low. INFLORESCENCE erect to spreading, usually much longer than leaves; peduncle 30- 88 cm long, 5-10 mm diam., terete, frequently mottled with purple, 0.66 to 3 times as long as petioles; spathe coriaceous, green tinged wit purple (B & K Yellow-green 6/7.5), narrowly ovate-oblong to oblong-lanceolate, 8—14 cm long, (2)2.8-3.9 cm wide, broadest just above point of attachment, abruptly acuminate at apex, obtuse lo acute at base, the margins often tinged purple, inserted at 459-80» angle on peduncle; stipe 5- 25 mm long in front, 2-6 mm long in back, ca. 8 mm wide, green, often mottled with purple; Spadix violet- -purple (B & K Red-purple 3/10), 12-25 cm long, 8-13 mm diam. at base, 5-8 mm diam. at apex, sometimes weakly arched in basal fourth; flowers rhombic, 2.8-4.5 mm long, 2.3- 35 mm wide, the sides straight to weakly sig- moid; 7-13(15) flowers visible in the principal spiral, 5-8(10) flowers visible in the alternate Spiral; tepals glossy, green-punctate, with small ete at anthesis, lateral tepals 2-2.3 mm wide, * Inner margins + straight, not overlapping; m green, exposed at base as soon as the spathe heu ; stigma ca. 1 mm long, green, usually square Pe shaped, brushlike with a small it appearing briefly, 1 or 2 days before the be OM emerge, the papillae exserted above n I stamens emerging moderately rapidly piat € base, held in a close circle at edge of 1; anthers white, held over pistil, opening as vide; rund emerge, ca. 0.5 mm long, 1.5 mm d Xeon ellipsoid, not divaricate; pollen Ns FRUCTESCENCE spreading; spathe din: Facts the spadix 29-38 cm long, 1.5-2.5 cm "i rries Mec square in wrote lu. tapered to + square at apex, ca. 10 m -—" green, white at iita exserted ie mis -5 mm before maturity, often not ma- dis 8 in apical one-half to one-fifth of the spa- ds not seen. Figs. 94 and 97. 2 ti CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 311 The species ranges from Costa Rica to Colom- bia (mountains of northwest Chocó adjacent to Panama) at elevations of 350 to 1,800 m. It oc- curs only along the central cordillera. In Costa Rica the species is known only from premontane rain and lower montane rain forest in the region around Tapantí but it possibly extends through- out the Talamanca ouman page: In Panama, + the species hash tropical wet, and lower montane wet forest ium Cerro Colorado in Chiriquí Province to the Cerro Jefe region in Panamá Province, then appears again only near the Colombian border in Darién. Anthurium lentii is in section Digitinervium and is perhaps most closely related to A. ovati- folium Engl., a species from Ecuador and Peru, which differs in having red berries. Anthurium lentii can be recognized by its thick, distinctively bicolorous, heavily punctate, truncate to cordate leaf blades with several, heavy, prominently as- cending basal veins. Other distinctive features include the prominently exserted pistils and the white, oblong, more or less tetragonal mature berries. TA RICA. CARTAGO: road between Moravia & Oacbradá Platanillo, Croat 36591 (MO); beside Río Taus, SW of Taus, Lent 2989 (F); N of Tapanti, Lent 2219 (F, US); near — N by ENE of Tapanti, Baker et al. 217 (CR, MO, US); Rio Grande de Orosi, W of Tapanti, 1 907, 908 (CR, F). Anthurium leuconeurum Lem., Ill. Hort. 9, Pl. 1862. Type: The description is based on a ‘plant from southern Mexico collected by Ghiesbrecht and cultivated in Ghent, Belgium, by Ambr. Vershchaffelt; the illus- tration in Ill. Wort (9: 5.314. 1862) serves as the type. Terrestrial; stems and internodes very short; — moderately coriaceous, acute at apex, LEAVES erect-spreading; petioles 30— 70 cm seem 4-7 mm diam., terete or weakly sulcate adaxially, the blades held parallel to pet- iole; om l. 5-3. 5 cm long; blades aga eep- » lobed at base, (12)20-38 cm long, I cm wide, broadest " ee of petiole attachment; a. 2.9613 cm 1 long, directed. toward base; the si si- the iem mutes: velvety, dark green, the lower surface much paler; midrib and basal veins flat to weakly sunken above, much paler than sur- 312 face; the area on either side of the veins paler than rest of blade; basal veins 5—6(7) pairs, at least the third and fourth coalesced 1—4.5 cm, arcuate-ascending, the second and sometimes the third joining the margin well above the middle of the blade; the posterior ribs straight to grad- ually curved, naked (at least on the larger blades); primary lateral veins 3—4 per side, departing the midrib at an acute angle, then gradually curved toward the margin, loop-connecting with collec- tive vein; collective vein arising from the first basal vein or sometimes one of the lowermost primary lateral veins, 0.7-1.4 cm from margin. INFLORESCENCE erect, at least as long as leaves; peduncle 40-68 cm long, 4-6 mm diam., terete; spathe narrowly ovate to ovate-lanceo- late, green, 4.5-10 cm long, 1-2.5(3.5) cm wide, broadest just above base, narrowly acuminate at apex, weakly cordate and clasping at base, in- serted at ca. 20° angle on peduncle; spadix sessile or nearly so, green, (4)6-13 cm long, 6-8 mm diam. at base, 4-5 mm diam. at apex; the flowers rhombic, 3.8-4 mm long, 3-3.5 mm wide, the sides straight parallel to spirals, smoothly sig- moid perpendicular to spirals; 6-7 flowers visible in the principal spiral, 4—5 flowers visible in the alternate spiral; tepals sparsely punctate, lateral tepals 2.1 mm wide, the margin broadly rounded thin and turned up against pistil; pistils elevated but not emergent until after anthesis of stamens; the stigma 0.5 mm long, elliptic, slitlike; stamens emerging in a prompt sequence from the base, the stamens in the basal one-third complete be- fore the alternates emerge in the apical one-third; anthers held above tepals and over pistil but not obscuring stigma, 0.9 mm long, 0.8 mm wide, evenly spaced; thecae oblong-elliptic, scarcely di- varicate. INFRUCTESCENSE pendent; berries broadly obovoid, rounded at apex, bright orange, 9-10 mm long, 8-10 mm diam.; pericarp mod- erately thick, with dense raphide cells; seeds 2, narrowly obovoid to obovoid, brown(?), slightly flattened, 6-7 mm long, 5-6.5 mm wide, 3.6-4 mm thick. Fig. 98. * Anthurium leuconeurum was collected in southern Mexico in about 1860 and apparently has not been recollected in the wild. All known herbarium collections are from living collections, presumably derived from the original living col- lection in Belgium. If the species stil] exists in the wild it is surely a narrow endemic similar to A. clarinervium. I have strong suspicions that 4. ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 leuconeurum represents a plant of hybrid origin, probably offspring of A. berriozabalense and A. clarinervium. This cross has been made by Banta — [see Aroideana 6(1): 26-27. 1983] and the re- sulting hybrid looks much like the type drawing of A. leuconeurum. The species is a member of section Cardiolon- — chium and is characterized by having a velvety | leaf blade with usually paler major veins similar to A. clarinervium, to which it is apparently re lated. It is distinguished from the latter species by its generally larger and more narrowly ovate leaves that have the basal veins coalesced into a naked posterior rib, 1—4.5 cm long. Anthurium | clarinervium generally has the basal veins freeor | if they are coalesced the posterior rib is never — naked. ! Engler (1905) cites Ghiesbreght (187 7) (note | difference in spelling). It is not certain whether this is his collecting number or not but it implies that he saw an actual specimen. Ifso, it is unlikely that the specimen still exists. Anthurium leuconeurum was the subject ol much hybridization after its introduction into — Europe in the early 1860s. Perhaps the most well known of these hybrids is A. macrolobum Hort. (A. dentatum André is a synonym), which i5? cross between A. leuconeurum and A. pedalo- radiatum. j Although the species was originally described by Lemaire as acaulescent and epiphytic, 1t rs no doubt a terrestrial plant with a definite (a though short) stem. Since all known collections have been 3 from living collections in cultivation, no kj mens are complete with the stem. In addii have seen no bonafide living collection of F. species. Because of these shortcomings the scription is less complete than most. made Mzxico. Cultivated at California Botanical bec 54.1349 (UC); cultivated at Fairchild Tropica Frantz 3288 (FTG). Anthurium lezamae Matuda, Bol. Soc. Bot. i xico 19: 19. 1956. Type: Mexico. y, Mun. Bochil, Bochil (cultivated in San EA tóbal de Las Casas), MacDougall M inally collected by W. Lezama) (MEXU. lotype). 1.5 cm diam. - Terrestrial; the stems short, ternodes very close; cataphylls thi — -— | — i 1983] drying reddish-brown, soon weathering at least at base into thin, mostly longitudinally oriented fibers. LEAVES with petioles 17—41 cm long, 2- 4 mm diam., terete; geniculum 1-1.5 cm long; blades ovate to ovate-triangular, gradually acu- minate at apex, deeply lobed at base, 13-28 cm long, 9-18.5 cm wide, broadest at the middle or at the base, th gin slightly hyaline (on drying) weakly revolute; anterior lobe 12.5—20.5 cm long, the margin almost straight to broadly rounded, (sometimes somewhat bulged at base of lobe), the margins sometimes somewhat variable with one margin slightly convex, the other slightly concave; the posterior lobes directed downward or slightly outward, 6-11.5 cm long, 3.5-6.5 cm wide; the sinus usually obovate, rarely parabolic, rounded or obtuse at apex; the upper surface with the epidermis drying with a characteristically raised, alveolate reticulation (papillate when resh), the lower surface drying more or less smooth; the midrib raised on both surfaces; basal veins 4-5(6) pairs raised on both surfaces, the third and fourth (fifth) coalesced 4.5-6 cm, all V butthe uppermost arcuate-ascending, joining the margin, the second basal vein joining the margin well above the sinus and sometimes well above the midrib at 70°-80° angle, sometimes curving laterally at almost a 90° angle to the midrib, Prominulous above and below, scarcely or not atall more prominent than the interprimary veins or tertiary veins; the tertiary veins weakly raised above and below; the collective vein arising from the uppermost basal vein, 5-13 mm from the pee weakly raised on both surfaces. INFLO- E mm long in front, 1-3 mm long in back; spadix Hb 4—6.5 cm long, 5-6 mm diam. at base, 74 mm diam. at apex, held at 180? angle from Meus flowers rhombic, 3-4 mm long, (dry) Ni a -8 mm wide, (dry) the sides straight, parallel 2i ide gradually sigmoid perpendicular to the Piral; 4-6 flowers visible in the principal spiral, Ei flowers visible in the alternate spiral; tepals Ty glossy, minutely and densely punctate, CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 313 lateral tepals 1.8-2 mm wide, the inner margin very broadly rounded, the alternate pair with inner margins straight to concave; pistil greenish, not emergent; stigma ca. 1 mm long, linear; sta- mens emerging promptly, from the base, the lat- eral stamens preceding the third and fourth sta- mens by only a few spirals, held just above the tepals in a close circle around the stigma; anthers 0.5 mm long, 0.7 mm wide; thecae semicircular, flat, not at all divaricate. INFRUCTESCENCE not seen. Figs. 99 and 100. Anthurium lezamae is endemic to Mexico, in northern Chiapas at 900 to 1,500 m elevation. It is known from the municipios of Bochil, El Bosque, and Ocosingo. The plants grow on usu- ally steep, rocky slopes in pine-oak forest in sea- sonally dry areas. A specimen from adjacent Oa- xaca (Schultes & Reko 867, ECON) is possibly also this species but it has a broader spathe and lacks the typically alveolate epidermal pattern on the upper blade surface. The species is a member of section Cardiolon- chium and is recognized by its narrowly ovate- cordate leaf blades with the poorly developed primary lateral veins scarcely more prominent than the i pri y veins and the tertiary veins Also characteristic is the unusual reticulation created by the epidermal cells on drying. The species is perhaps most closely related to (and possibly inseparable from) A. leuconeurum, which differs in having the midrib and basal veins paler than the rest of the blade surface. In ad- dition, the blade surface of A. /euconeurum dries obviously papillate with the individual papillae distinguishable, whereas A. /ezamae dries with the epidermal cells drying minutely alveolate (honey-combed). Unfortunately, A. lezamae has not been studied alive so I do not understand the significance of these differences. MacDougall 302 and two unnumbered MacDougall collections from an area believed to be southwestern Oaxaca (between La Gloria and Rio Grande and between Buena Vista and La Gloria) differ from A. /ezamae only slightly in leaf shape but lack the alveolate epidermal pat- tern on the leaves typical of that species. It is possible that these represent a distinct species but too little is known about them to be certain. MEXICO. CHIAPAS: vicinity of Jitotol, Breedlove & Dressler 29832 (DS); Breedlove & Smith 32558 (DS); FIGURES p lium Sen ott -94. reat 2705 9a. 25 A m 4 urium kunthii Poepp., Croat 38121.—92. inthurium lentii Croat & Baker, Croat oe lancetillense Croat, Croat 42672 (Type).—93. 3722 Anthurium lancifo- VIE N3G?3IVO TVOINV.LOd P3IOOSSIN AHL HO STVNNV 0L 10A] CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA . —96. Anthurium lancetillense Croat, Qu 42 2 rd. —98. Anthurium leuconeurum Lem., 7467 r, Las Cruces Bot. Ga FIGURES te 95. Anthurium Hip genis Sodiro, Selby 58- 1975-32 i Croat & B (Cul Anthurium lentii tivated a at pesas s, Ayr, Aust ralia). i 3 Vor. 70 316 ANNALS OF THE MISSOURI BOTANICAL GARDEN [ NA A MEXICAN: FLORA MEXICANA FLOR ^ MEE nA MATUDA "ENBARIUM pamu! Col, thomas aaLeugal) S62 me w Antburium Leracea Maluda ap sov, en bosque humede en 5eghil, 1tivedo Por frof.i = eg die $e Centre zorgi der ináilganista,B.C Anthurium s 99-102. 99-100. Anthurium lezamae Matu MacDougall 362 (Type).— 101. gipe Phone Mimi. Croat 48326.—102. Anthurium XH Dye Matuda, Croat 3975: lon- 1983] CROAT vicinity Ocosingo, Breedlove 27925 (DS); San Cristó- bal de Las Casas, MacDougall 362 (MEXU). OAXACA: between La Gloria & Rio Grande, MacDougall s.n. 302 (MEXU). Anthurium longipeltatum Matuda, Anales Inst. Biol. Univ. Nac. México 37: 75, Fig. 1. 1967. Type: Mexico. Oaxaca: between Vista Her- mosa and Puerto Eligio, Comaltepec Ixtlán de Juárez, elev. 1,200 m in humid forest, October 5, 1963, MacDougall 563 (MEXU, holotype). Epiphyte or terrestrial; stems green, ca. 18 cm long, 2.5-3 cm diam.; roots few, dark greenish- brown, descending; leaf scars 2-3 cm wide; cata- phylls coriaceous, 6.5-15 cm long, long-acumi- nate at apex, drying dark brown, splitting at base, subpersistent. LEAVES with petioles erect to spreading, 29-56 cm long, 5-7 mm diam., terete; &niculum 1—4.5 cm long; blades oblanceolate to oblong-ovate, moderately thick, long-acuminate at apex, lobed at base, 33-56 cm long, 14-28 cm wide, broadest just below middle; anterior lobe 4-40.5 cm long, the margins convex; posterior lobes 4.5210 cm long; the sinus parabolic to tri- angular, sometimes the lobes overlapping, acute al apex; upper surface semiglossy to glossy, lower surface semiglossy, obscurely to conspicuously 5 pairs, usually free, sunken or prominulous in valleys above, raised below; primary lateral veins 8-12 per side, departing midrib at 357-50" angle, Sunken above, raised below, arcuate-ascending to collective vein; lesser veins flat to weakly es above, flat below; collective vein arising rom the first basal vein, sunken above, raised E low, 5-10 mm from margin. INFLORES- E erect-spreading, longer than leaves; pe- € 30-80 cm long, 3-8 mm diam., terete; SPathe subcoriaceous, green sometimes faintly LM Purplish (B & K Yellow-green 8/10), ob- «né lanceolate, 9-14.5 cm long, 1.5-2.2 cm wide, atrowly acuminate at apex, obtuse at base, in- *rted on peduncle at 30°-50° angle; spadix olive- AR (B & K Yellow-green 6/5) to dark purple a ^s Blue-purple 2/10), 9-35 cm long, 7-9 eh am. at base, 2-4 mm diam. at apex; flow- thombic to 4-lobed, 2.5-4 mm long, 2.3-3 mm wide, the sides + straight parallel to spiral, T sigmoid perpendicular to spiral; 6-7 vidi. Visible in the principal spiral, 7-9 flowers * in the alternate spiral; tepals glossy, mi- OF MEXICO AND MIDDLE AMERICA 317 nutely papillate, sometimes with droplets as sta- ens emerge, lateral tepals 0.8—1 mm wide, the in b mergent oblong-linear, 0.5-0.6 mm long, purplish with droplets apparent a few days before stamens emerge; stamens ging from the base, the lat- eral stamens preceding alternates by 1—2 spirals, exserted briefly on pinkish flattened filaments, which retract, holding stamens at edge of tepals against pistil; anthers pale yellow, held + con- tiguous, 0.3-0.4 mm long, 0.5-0.6 mm wide; the- cae ellipsoid; pollen yellow to white, fading to ream. INFRUCTESCENCE pendent, spathe withered; berries subglobose to obovoid, round- ed at apex, red to red-orange (B & K Red 6/7.5), 8-8.5 mm long, 5.5-6.5 mm wide; mesocarp fleshy, + transparent with raphide cells; seeds 1— 2, ovoid, cream or greenish-white, 4 mm long 2.5 mm wide, 2 mm thick with sticky appendages at both ends. Figs. 101, 102, and 104. The species occurs in northern Oaxaca, Sierra de Juárez, Veracruz near Orizaba, and is ex- pected in nearby Puebla in Mexico in tropical wet forest from 600 to 1,500 m elevation. The species is a member of section Belolon- chium and is distinguished by its moderately thick, deeply lobed, oblong-ovate blades, which are dark green above and paler below, by its long slender spadix, by having basal veins frequently free with little or no posterior rib, and red to red- orange, subglobose berries. e collections of A. /ongipeltatum from northern Oaxaca are somewhat obscurely punc- tate whereas those collected near Valle Nacional are more conspicuously punctate. The spadix color varies from olive-green to violet-purple. PN A 4 :41 3 VV WIMI A. verapazense, a species ranging from Belize to Honduras, but that species has a consistently vi- olet-purple spadix and usually has a well devel- oped posterior rib. Matuda considered A. /ongipeltatum to be similar to A. titanium (A. xanthosomifolium) but it is in no way similar or related except that the two species have posterior lobes. Anthurium ti- tanium differs principally in lacking punctate dots on the lower blade surface. The etymology of the name is curious. Al- though Matuda named the plant “‘longipeltata”’ there is no part peltate. Perhaps this is a typo- graphical error from the intended “‘longipetiola- tum.” 318 ANNALS OF THE MISSOURI BOTANICAL GARDEN 103 FIGURE 103. Anthurium louisii Croat & Baker, Field Museum Drawing 82395. [Vor. 70 O w O > T a z z Cc e — S x i, Los] z tr ia Q o > Z i») z J g p tm » < tr E eo > Figures 104-107. 104. a ser ltatum Matuda, Croat 48326.—105. depo ob lucens Standl. ex Yuncker, Croat 41541.—106. An- thurium michelii Guillaumin, Croat : 2.— 107. Anthurium microspadix Schott, Croat 4096 MEXICO. OAXACA: Comaltepec, MacDougall 563 (MEXU); road from Teotitlán del Camino to Santa ilc t 3 MO); King 2143 (MICH); Moore & Bunting 8917 (BH); summit of Cerro de Cuhuatepete Cerr vicinity Tehuacán, Vera Santos 3340 (MICH). Anthurium louisii Croat & Baker, Brenesia 16 (Supl. 1): 59. 1979. Type: Costa Rica. Car- tago: about 2.5 km north by east northeast of Tapantí, 9?45'N, 83?47'W, 1,500-1,600 m elevation; forest remnants on steep slopes, Baker, Utley & Utley 220 (F-1758200, ho- lotype; MO-2385464, isotype. Epiphyte or terrestrial, to 30 cm d stems erect, to 1 cm diam.; internodes 1-4 cm long; leaf scars obscured by cataphyll fibers; anit 4—7 cm long, gradually acuminate at apex, drying brown, quickly weathering into coarse persistent fibers. LEAVES dispersed along stem at each node; petioles (15)20—30 cm long, ca. 3 mm diam., terete, narrowly sulcate; geniculum ca. 1 long; blades ovate to broadly ovate, subcoria- ceous, gradually to abruptly long-acuminate at apex, obtuse or truncate to rounded at base, 14— 17 cm long, 8-13.5 cm wide, broadest slightly below the middle; both surfaces conspicuously glandular-punctate; midrib raised above, dimin- ished toward apex, prominently raised below; midrib at 45° angle, + straight, loop-connected by the collective vein, raised above and below; lesser veins inconspicuous; collective vein aris- ing from the base, 4-12 mm from the margin. NFLORESCENCE, erect-spreading; peduncle (15)25-38(45) cm long, considerably longer than the leaves; spathe somewhat thin, green becom- ing reddish, lanceolate-linear, 4—6.5 cm long, ca. 1 cm wide, broadest at base, gradually acuminate at apex, slightly clasping at base, inserted at ca. 20° angle on peduncle; stipe very short or absent; spadix green to yellow-green, usually coiled at maturity, (4)6-18 cm long, 3-4 mm diam. at ase, 2-3 mm diam. at apex (dry); flowers rhom- bic, 2-3.5 mm long, 2-2.5 mm wide; the sides + straight; 2-3 flowers visible in either spiral; lateral tepals 1.6-2.1 mm wide, the inner margin roadly rounded; pistil green; stigma ca. 0.5 mm long, ellipsoid; stamens probably weakly exsert- ed above the tepals, well spaced; filaments ca. 0.2 mm long; anthers ovate, 0.4-0.5 mm lon : 0.5-0.6 mm wide; thecae ellipsoid, slightly di- varicate. INFRUCTESCENCE with the spathe ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 at least sometimes deciduous; spadix 8-15 cm long, 3-4 mm diam., usually coiled in age; berries ovoid to subglobose, green becoming orange-red, ca. 5 mm diem and 5 mm wide, maturing si- roughout the spadix, drying with a buttonlike bak Fig. 103. Anthurium louisii is known only from Costa Rica in premontane or lower montane rain for- est, apparently restricted to middle elevations on the Atlantic slopes of the Cordillera Central and the Cordillera de Talamanca, 1,300 to 1,800 m. The species is in section Porphyrochitonium and can be recognized by its broadly ovate leaves with glandular punctations on both surfaces and by its usually coiled spadix (in fruit). The type collection was found on a steep bank and formed the dominant ground cover. Anthurium louisii is related to A. angusturense Engl. from Colombia. The latter species differs chiefly by its lanceolate leaf blades. Anthurium louisii also resembles A. circinatum Croat from Panama in Chiriquí Province. anti, Baker, Utley ay CARTAGO: NNE of Tap NY, SEL US; cal station, Utley & Utley 4332 (DUKE; road from Tapantí to Taus and Tau NE of R Orosi at Tapantí, Utley " po^ 5035 (DUK P) ve Copey, Cordillera de Talamanca, William 16506 (EAP, F); region beyond Muñeco, 3 m ton 9 (P. heros church, SW of Orosi, Utley & U Utley 1 N José: La Palma, Coronado, Werklé s.n. (CR). pe Anthurium lucens Standl. ex Yuncker, e Nat. Hist., Bot. Ser. 17: 317 l. 1 Ah Type: Honduras. Comayagua: near i : ue, ote, hills above the plains of aon & Var elev. 1,350 m, Yuncker, T 5844 (F, holotype; GH, MIC isotypes). Nat. Anthurium seamayense Standl., Publ. Field Git His t "PL 5. 1940. TYPE: a. Alta Verapaz: Seraxcaj. Finca Mer RA e pss tih 960 m, C. L. Wilson ( h age ex jd Field va 7, Pi. 6. 1938. is ar El Achiote; h above the plains of umen. elev. à. ir Yuncker, Dawson & Vouse 5935 (F 85768 o Anthurium yunckeri Nat. Hist. lotype). r, Biol ae paca Matuda, Anales TM M exico ac. México 27: 339. 1957. be elev. WT | CROAT 1983] m, MacDougall 337 (MEXU 109434, holotype; CAS, isotype). Epiphyte or terrestrial; stems to 20 cm long, ca. 4.5 cm diam.; internodes and leaf scars ob- scured by roots and cataphylls; roots dark brown, medium thick, descending; cataphylls subcoria- ceous, 6.5-7.5 cm long, heavily tinged red-violet, the apex acuminate with subapical apiculum 2- 3mm long, drying brown (B & K Yellow 4/7.5), weathering to persistent reticulate fibers, the apex remaining intact. LEAVES with petioles erect to spreading, (12)26—72 cm long, 0.6-1 cm diam., subterete; geniculum 2.5-3.5 cm long, some- times faintly tinged red-violet; blades narrowly ovate to oblong-ovate or ovate-triangular, mod- erately thick, gradually acuminate at apex, deep- ly lobed at base, 30-46 cm long, 18-24 cm wide, broadest at the base or just below point of petiole attachment; the anterior lobe 22-32 cm long, the margins usually convex, sometimes + straight; below; basa] veins 5—6 pairs, the first and second free, the remaining coalesced 2-2.5 cm; posterior m naked one half to one third its length, usually With inner margin turned up; primary lateral veins 7-7 per side, departing midrib at 557-60" angle, Straight or weakly curving to collective vein, in above, raised below; lesser veins scarcely heec above, visible and + flat below; collective bii eth from the first basal vein, in larger k a rom fourth or sometimes fifth basal veins, i en above, raised below, 6-15 mm from the eoe INFLORESCENCE usually held erect, ors ng or pendent in larger plants, shorter than 5-8 qualling petioles; peduncle 33-75 cm long, rues diam., terete, sometimes tinged red- Mis spathe moderately thin, green heavily kir With red-violet, lanceolate, 4—8.5 cm long, M Sa wide, narrowly acuminate at apex, des ren base, inserted at 45? angle on pedun- -l "s ix deep violet-purple, 10.5-13 cm long, mm diam. at base, 3~3.5 mm diam. at apex; ebd visible in the principal spiral, 8-9 & ple in the alternate spiral; tepals matte Dillate Y glossy, weakly punctate, minutely pa- > With scattered droplets, lateral tepals 1.5- flowe i OF MEXICO AND MIDDLE AMERICA 321 1.6 mm wide, the inner margin broadly rounded; pistils weakly raised, green, sometimes purplish; stigma linear, ca. 0.6 mm long, opening as sta- mens emerge; stamens emerging from the base in a moderately rapid sequence, the laterals first, soon followed by alternates, the leading stamens several spirals ahead of third or fourth stamens; anthers bright yellow, held at the edge of tepals, ca. 0.6 mm long, ca. 0.3 mm wide; thecae ovate, widely divaricate; pollen pale yellow. INFRUC- TESCENCE pendent; the spathe usually decid- uous; berries bright red, obovoid to oblong-el- lipsoid, usually developing only in basal one quarter to one third, round to truncate at apex, 8-18 mm long, 6.5-18.5 mm diam.; mesocarp juicy, clear, with raphide cells; seeds 1—2, green- ish white, oblong-elliptic, slightly flattened, 4—5 mm long, 2.5-4 mm wide, 1.8-2.1 mm thick, obliquely concave at apex. Figs. 105 and 108. Anthurium lucens ranges from Mexico (State of Chiapas) to Honduras usually from 350 to 1,500 m (rarely to as low as 150 m in Guatemala to as high as 2,000 m in Chiapas). In Mexico the species occurs only in **bosque pino-encino" and **selva alta perennifolia" in northern Chiapas. In Guatemala it has been collected only in the de- partments of Huehuetenango, Alta Verapaz and Izabal. Leaf shape is extremely variable in A. /ucens but all populations share in common coriaceous, punctate leaf blades, an inflorescence shorter than or equal in length to the petiole, a relatively short violet-purple spadix with a spathe usually more than half as long as the spadix at anthesis, and red, obovoid, 1-2-seeded berries. Generally the anterior lobe of the blade is convex along the mareins y g o ave and wre similar to A. chiapasense except that they dry greenish to greenish brown and not black- ened as is the case with A. chiapasense. Honduran collections of A. /ucens are some- what unusual in that they are geographically iso- lated, the species not usually collected in south- ern Guatemala. Other collections worthy of mention are those of MacDougall 337 (type of A. mapastepecense) and Matuda 19647. While both of these collec- tions are unusual in being south of the range of re lieved to be within the range of variability for A. lucens considering that the species is so wide- spread and variable. 322 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 l um macht Anthuriu URES 108-111. 108. M mee Stan Rune Mas uda, Ramirez 144 ( montanum Hemsl., Croat 4716 x Yuncker, Croat 41541.—109. were yia TRE Schott, Croat 40906.— 1983] Anthurium lucens is in section Belolonchium and by having the spathe aiii less than half as long as the spadix at an Berry and seed size are usto variable in A. lucens. For example while Croat 41541 has ber- ries ca. 7 mm long and seeds ca. 4 mm long, AMETS 9709 and T have berries 11 to 14 m long and seeds 7 to 8.5 mm long respec- ds Breedlove 9709 is irata noteworthy seeds appear to be qualitatively similar and are all believed to represent the same species. TEMALA. ALTA VERAPAZ: vicinity Telemán, Croat 41541 (MO); Cerro eon yy Steyermark 45570 (F, MO); 45624 (F, LL); 45649 (F); vicinity Finca Cubil- huitz, Steyermar rk 44740 © Finca Sipacate, Cook & Griggs 434 HUEHUETENANGO: near Barillas, Steyermark 49736 (F, US). 1zaBaL: El Estor, Harmon & Dwyer 4340 (MEXU, ; NW of Lake Izabal, Jones & Facey 3389 (NY); along Rio Bonito, Steyermark 41723 (F, = DUR ntonio & El " a 22588 (F, NY); vicinity poco is s & Molina 18083 (F), above the plains of UU E Yuncker et al. 5844 (F, GH, MICH, MO, US); 5935 EROS CHIAPAS: SE of Comitán, Carlson 1913, 2334 ides ); 2294 (MICH); Munc. Tenejapa, Breedlove Ta (DS); Ton 1574 (CAS, DS, MEXU); E of Leere ; Breedlove 9709 (CAS, DS, ENCB, D 14971 (DS, LL, MEXU); Bre edd Raven 11308 (DS, CAS); ne 21226, 275814 (DS); Park, 4 (DS): Montebello National ragga & fec 29535 (DS); E 38754 DS. M unc. Ocosingo, Dressler 1452 (GH, MICH, NY, Tm of Ocozocoaut tla de Espinosa, Breedlove 275 00 ); da qu ied. Mi 19647 (MEXU); be hrs Palenque & B of San un Bodas 33385 (CAS, DS). Anthurium D Matuda, Anales Inst. Sion): Chinantla, 1,200 m, Ramirez 144 (MEXU, holoty pe). woe short, internodes short; cataphylls thin, ai cm long, drying brown, weathering to lin- bers, probably deciduous. LEAVES with esse 6-7 cm long, 4-5 mm diam., broadly "cate adaxially, rounded abaxially; geniculum ks mm long: blades oblanceolate, narrowly acu- ad "t at apex, narrowly cuneate at base, 45-48 ong, 6-8 cm wide, broadest above middle, CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 323 matte on both surfaces (dry); the midrib prom- inently and narrowly raised above and below (dry); primary lateral veins 10-15 per side, de- parting midrib at 20°-30° angle, curving down and joining collective vein; reticulate venation prominulous in dry specimens; collective vein arising from one of the primary lateral veins near the base, 3-4 mm from margin. INFLORES- CENCE equalling or longer than leaves; pe- duncle 40-48 cm long, to 5 mm diam., terete; spathe narrowly lanceolate, inserted at 45° angle on peduncle; spadix green, 20-24 cm long, 1.3 cm diam. at base, 6 mm diam. at apex; the flow- ers rhombic, 2-2.4 mm long, 2.4-2.6 mm wide; 13-14 flowers visible in the principal spiral, 9- 10 flowers visible in the alternate spiral; Tees non emergent; stigma agri vis 0.5 m INFRUCTESCENCE not seen. n. Fig. 109. Anthurium machetioides is endemic to central Mexico on the Atlantic slope. The type was from the region of Chinantla in northeastern Oaxaca (a region comprising part of 4 districts south of Valle Nacional) at 1,200 m and a second col- lection was made relatively near at 200 m in the District of Tuxtepec (probably San José Tuxte- pec between Valle Nacional and Tuxtepec). The species is characterized by its oblong-lin- ear leaves with a collective vein from near the base and primary lateral veins that are scarcely In describing the species Matuda indicated its relationship as section Leptanthurium; however, despite the fact that it has a collective vein from the base, its affinities are most likely with An- thurium schlechtendalii in section Pachyneu- rium, which often has a collective vein from near the base in juvenile leaves. It differs from A. schlechtendalii in its proportionally narrower leaves, the collective vein and by its proportion- ately much longer spadix. ico. OAXACA: near Chinantla, Ramirez 144 pres Tuxtepec, Martinez-Calderon 844 (MICH). Anthurium michelii Guillaumin, Bull. Mus. Hist. Nat. (Paris) 31: 263. 1925. Type: Panama. Bocas del Toro: hills beyond Fish Creek La- goon, Wedel 2276 (P, holotype). — amethystinum Croat & soapy I - sin 1x 20. 1979. Abia Mets u road between Moravia (east of T uis, SE of Turrialba) and Quebrada S Plátanille, 3-5 km from 324 Finca Racine in Moravia; elev. 1,200-1,300 m; disturbed primary forest, Croat 36601 (MO 2395446, holotype; CR, F, K, NY, SEL, US, iso- types; Live plant at MO). Epiphytic or terrestrial; stem usually very short, 6-25 cm long, 1.5-3 cm diam.; roots dense, 1.5— 5 mm thick, descending; cataphylls subcoria- ceous, 8-23 cm long, cordate-acuminate at apex, drying brown (B & K Red 9/2.5), persisting, weathering into fibers. LEAVES held nearly erect; petioles 9-25 (40) cm long, 5-9 mm diam., flat- tened laterally, sulcate with sharp margins, acne to rounded abaxially; geniculum 1—1.6 cm long; blades elliptic to oblong-elliptic or m saa ually acuminate at apex (the acumen down turned), acute to obtuse at base, "ep cm uen 5.5-22 cm wide, broadest at the middle or slight- ly below the middle, the margin + flat, contin- uous with margins of n petiole; both surfaces matte to semiglossy; m ly raised and diminished toward apex above. acotely raised below; basal veins consisting only of the lowest pair of primary lateral veins; primary lateral veins 6-15 per side, sunken above, prominulous be- low; interprimary veins weakly sunken above, prominulous below; collective vein moderately straight, prominent, continuous, arising from the second pair of primary lateral veins, sunken above, prominently raised below, 5-9 mm from the margin. INFLORESCENCE pendent or arching-pendent, usually shorter than the leaves; peduncle 30-58 cm long, 3-5 mm diam., much longer than petioles, prominently 4- or 5- Ab especially near the base, irregularly striate toward m Yellow-green 7/10), oblong-lan- ceolate to oblong, 3-15 cm long, 1-2.5 cm wide, broadest in lower third, inserted at 45° angle on peduncle; stipe 0.8-3.5 cm long; spadix yellow- green (B & K Yellow-green 6/7.5), (4.5) 8.5-20 cm long, 6-10 mm diam. at base, 4-6 mm diam. at apex; flowers rhombic, 3-4.2 mm long, 2.8— 4.3 mm wide, the sides straight to jaggedly sig- moid; (4)7-11 flowers visible in the principal spiral, 5-7 flowers visible in the alternate spiral; tepals green, semiglossy, sometimes with scat- tered nectar droplets; lateral tepals 2-2.7 mm wide, the inner margin + straight, somewhat turned up against pistil; pistils exserted 0.5 mm, pointed before maturity, the exserted portion broadly ovoid; stigma linear, 0.4-0.6 mm long, exserted, brushlike, with conspicuous droplets for about 10 days, becoming dry and brown, about ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 1 week before first stamens emerge, dry and exserted when stamens emerge; the lateral sta- mens emerging rapidly from the base, open in basal three-quarters before the first alternates be- gin to emerge at the base, sometimes emerging in a more scattered fashion, held in a close circle around pistil; anthers white, 0.5-0.6(0.8) mm FRUCT 2s green; the spadix 5-20 cm long, 2-3 cm | berries obovoid to ovoid or pyramidal, aos at apex, glossy, dark purple to violet- | purple, 7-12 mm long when dry, paler violet- | iem in lower third; mesocarp gelatinous; "i 2, flattened, ca. 5 mm long, 4 mm wide, 3 m m| thick. Fig. 106. . m.; The species is known from Costa Rica and | Panama on both slopes at altitudes from 40010. | 1,400 m, mostly in premontane wet and tropical | wet forest. Anthurium michelii is distinctive for its pe green or whitish underleaf surface and bright | purple berries. Also characteristic are the oblong- | elliptic leaves, broadly and sharply sulcate e iole, and the frequently conspicuously stipitale | spadix. | The species is a somewhat atypical member of section Pachyneurium, and a ieee to be lated to any other species in the se The type specimen of A. michel until recently, is aberrant and unusua ly n for the species and, despite the fact that s b little resemblance to the type of A. amet : l num, the two are now known to represen same species. d studied j STA RICA. ALAJU Balsa de San PP | Utley” & Utley 1850 (P) Hacienda la ans 20147 Rafael, canton de Aguas Zarcas, Williams et at a! (F); E of Rio San Rafael & S of the hot aie = La Marina Burger & Stolze got (F); La & Stolze 4994 (CR, F, MO, NY); Austin Smith NY1258 (F, NY); road from 4 “tte to Lake Hule, W of Hwy. 9, Baker, Utley (F); NW of Cariblanco, Luteyn 3 3348 ( of Bijagua along the new road to Upala, Burs! fe (F, MO} Pu Tu cARTAGO: Moravia camp CR); along Camino f SE. of Plat ee dass 36743 (MO); o, Cro 3660! tween Moravia and Quebrada ron of (CR, F, K, MO, NY, SEL, US). H Bess] Colonia Virgen del Socorro, E of Car Ms 2827 e finca of Sr. Carolo Molina, Utley & Utley jg SAN José: La Hondura, Standley 36313 (US); Las | 1983] Leon 1148 (CR); along road from San Isidro del Gene- tal to Dominical, Burger & Baker 10106 (F); Croat 35259(MO); 1 mile beyond divide between San Isidro del General and Dominical, Croat 35323 (MO). Anthurium microspadix Schott, Oesterr. Bot. Z. 8: 180. 1858. Type: Costa Rica, near Naran- jo, Oersted s.n. (Type not found; Photo of Schott Aroid Drawing #322, NYBG Neg. #N.S. 3812) Anthurium porrectum Schott, Oesterr. Bot. Z. 8: 180. 1858. Type: Costa Rica, near Desengafio, Wend- land 841 (GOET). Anthurium fapinostachyum Schott, aig, Bot. Z. 8: 80. 1858. : Costa Ric Aroid Drawing #331, NYBG Neg. #N.S. 3818). Anthurium colombianum Engl., Bot. Jahrb. Syst. 25: 37 8. TYPE: Colombia . Putomayo: valley of rosa near Santiago, E of Pasto, elev. 2,000- 2,5 Anthurium trom Engl., Bot. Jahrb. Syst. 25: 380. 1898. Type: Guatemala. Alta Verapaz: Pan samalá, 1 om m, Tuerckheim 864 (B, holotype: US, isotype Anthurium ei Sodiro, Anales Univ. Centr. ss Mir 15: 393. 1902. Type: Ecuador. Pichin- negal, Sodiro s.n. (B). ‘nari lepturam Sodiro, Anales Univ. Centr. Ec- 5: 458. 1902. Type: Ecuador. Pichincha: sid Hopes of Volcán Pichincha, Sodiro s.n. Anthurium nanegalense Sodiro, Anales Univ. Centr ae p 460. Aa Type: Ecuador. Pichin- j As ha: Gualea, Sodiro s.n. (B). nthurium Satius Sodiro, Anales Univ. Centr. Ec- 5. 1907. Type: Ecuador. Tungurahua, iro s.n. Anthurium barbanum K. Krause, Notizbl. Bot. Gart rlin- Beni 5 9: 270. 1925. Type: Costa R AES ar Barba, 4. C. Brade 2510 (B). urium e Standl., Publ. Field Mus. Nat. r. 22:68. 194 0. Type: Panama. Chi- riquí: Río Chiriquí Viejo Valley, 1,300-2,000 m, White 160 (F, holotype; MO, isotypes). AE nt epiphyte or semierect, terrestrial herb; él S elongate, frequently to 1 m or more long, the j cem. , sparsely long-rooted at the nodes; thi internodes (0.6)1-8(11) cm long; cataphylls «» 4.5-7.5(11) em long, acute and apiculate at ier Promptly weathering, soon deciduous, EAVES with a few pale fibers persisting. petiole ersisting at the upper 6-10 or so nodes; Shallo : erect- -spreading, subterete, narrowly and diii: wly sulcate, 5-16(24) cm long, 2-2.3 mm + obl &eniculum 5-10 mm long; blades usually kagak sometimes lanceolate, narrowly ob- ul iptic or oblong-ovate, abruptly or grad- Y acuminate at apex, usually truncate to CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 325 rounded at base, sometimes acute or subcordate, 9-25 cm long, 2.5-11 cm wide, moderately thin; the upper surface semiglossy; midrib raised in a valley above, diminished and sunken toward the apex, prominently raised below; primary lateral veins 4-11 per side, departing midrib at ca. 45? angle, sunken above, raised below, straight or slightly curving to collective vein; interprimary veins usually present; tertiary veins usually prominent on lower surface; basal veins 1—3 pairs, usually free at base, all but the uppermost soon joining the margin; collective vein d num the first basal vein, 2-5 mm from the INFLORESCENCE erect, shorter man the E peduncle terete, 5-15 cm long; lanceolate- elliptic, olive-green or pale green, sometimes tinged purplish, 3-4(5) cm long, to 1.7 cm wide, abruptly acuminate at apex, weakly cordate at base, inserted at 45?—50? angle on peduncle; stipe usually 2-5 mm long, sometimes absent, or to 10 mm long; spadix pale yellow to yellow-green or green, sometimes tinged with purple, 1.5—4 cm long at anthesis, usually 4-5 mm diam. near base, 3-3.5 mm diam. near the apex; flowers square to weakly 4-lobed, 2.5-3 mm wide in both directions, the sides straight to jaggedly sigmoid; 4—6 eee in -the principal. spiral, 6-7 1 flowers uu tepals semi- ossy, piney papillate, large nectar droplets scattered over the surface when anthers are emerging, lateral tepals 1.5-2 mm wide, the inner margin broadly convex; pistil weakly emergent, medium green, slightly darker than tepals, semi- glossy; stigma colorless, raised, brushlike, ca. 0.5 mm long, with stigmatic droplets 1—3 days before anthers emerge; stamens emerging slowly from the base, the laterals preceding the third and fourth stamens by only a few ro anthers cream, inclined over pistil, ca. 0.5 mm long, ca. m wide; thecae ovate-elliptic, slightly di- FRUCTES- diam.; berries yellowish-green, subglobose, ca. 5 mm diam. Figs. 107 and The species ranges from Mexico to Ecuador at elevations from 800 to 2,300 m in premontane rain, lower montane wet, and lower montane rain forest. It is one of the most variable species of Anthurium in North America. In Panama it is most easily confused with A. pallens (see that species for a discussion of the differences). In South America it is apparently closely re- lated to A. amoenum Kunth in Peru and Bolivia, 326 A. E sehe and A. brachypodum Sodiro from Ecua Sodiro acted several species based on the sisse Sr absence s a ADS P" pei pb lengths nd the shape and sepe size of the spathe i in udi to the length of the spadix. These characters have been Observed to be too variable even within a single population to warrant their use in the separation of species. Anthurium microspadix has been placed in section Xialophyllium. Mention should be made of an error in the translation of Lehmann's handwritten label by Engler. The type locality should read *'Sibun- doy" rather than “Dibundog BELIZE. Camp 32, Guatemala Survey, Schipp S-679 Cos Rica. No other location, Bogner s.n. (MO); ier 66. 3863 (BR); Tonduz 1888 (B R). ALAJUELA, N : d : Almeda & Nakai 3848 (MO); N of San Ramón, Utley 4590 (DUKE, MO); Volcán Barba, Tissine 1600 (WIS); S of Zapote, Utley 4655 (DUKE); vicinity Zarcero, D 43564 (MO); Smith 115, 404, 485 (MO); 638, 826, 112 v CARTAGO: trail to Meterological Sta- tion, Utley 4 KE); S of Mufieco, Utley & 1 (MO); vicinity Palo Verde, Luteyn 3329 RE uteyn & Wilbur 4381 (DU KE); vicinity Pla- tanillo, Croat 36819 (MO); between Quebrada Honda nd the Río o La Hondura, Croat 44503 riblanco, nsi 35847 (MO); Río Las Vueltas, Taylor 17705 (NY); Tarraz zu, Tonduz 7865 (BR); N of Vara Blanca, Croat 35622, 2 (MO); Vara Blanca de Sarapiquí, Skutch 350 MICH : Re Baker 9643 (M Canas Gordas, Pittier 11128 (BR); sa Cotón, along Río Coto S Croat 26643, 266504 (MO ); W uteyn 3847 A) (DUKE). s N JOSE & HEREDIA: slopes of Cerro o Zurqui, pre 3688 (CAS); Río Para Blanca, Burger et al. ANNALS OF THE MISSOURI BOTANICAL GARDEN "| [Vot. 70 10268A (MO). SAN José: NW of Canaan, Burger & Liesner 7086 (MO); NW of Cascajal, Taylor & TE 41 (NY); Cordillera Central above La Palma, Li teyn 3307 (DUKE, MO); Copey, Stork 155 0 (MICH; La Palma, Brenes 15055 (NY); Tonduz 9698 (BR); between alm ndura, Luteyn 425) ! m . eral, Croat 35265, 35300 (MO); N of San Jeronimo, Utley & Utley 2894 (MO); Santa Maria de Dota, Pittier 2339 GR) GUATEMALA. ALTA VERAPAZ: Coban, Tuerckheim 1297 GR NY, US); Cubilhuitz, Tuerckheim 8605 (NY, | U Moca, Johnson 143 (US); AS 864 (P); SE of Tactic, William et al. 40601 9 Finca El Porvenir, Steyermark 37975 (MO); La Trin- idad, Croat 40906 (MO); vicinity xs Rafael, he] | et al. 26202 (NY). zacapa: Mt. Virgen, Steyer 42870 (F) ONDURAS. C (MO); vicinity Sigistenedias hir. 22847207 | Cordillera ven, RAZÁN: C r » 13885 d 3184 (U S). s SANTA BARBARA:.E of Lake Yojoa, & Hazlett 3882 (M oat | MEXICO. CHIAPAS: between xtapa & Pi por OH 47798 (MO); vicinity ent "Breedlove ^ $77 eo (DS); vicinity Lake cepted Miranda 2689 (MEXU); M tebello National Park, pares ae 5 120 pi Munc. Ocosingo, DAE : 517 (MEXU, uacán, pe ven & Br er Tuxtepec & Oaxaca, Vaf) s ih osa, Moore & Bunting 89 ICARAGUA. JINOTEGA: NW of ar 9173 (MO); near Cerro c (MO); Neill N163 (MO); Matagalp: a: Hu DON M Jinotega to Matagalpa, Croat 4306 grs ii Ma of siga Croat 4311 . (MO); a & Molina de Ostuma, Molina 2038 phe Molina LIS 30538 (MO); Neill 1575 (M lia 27724 (MO); El Porvenir, pode 67 l 3 vicinity Cerro Saslaya, Neill 3812, Stevens Anthurium montanum Hemsl., Diagn ji v gott EI Mexic. 36. 1878. TYPE: mre de Fuego, above Calderas at 2,700 m. (K). 2. £i -15 cm long, Usually epiphytic, stems 5 roots nume | ous, thick, descending; cataphylls 2 12 cm long, rounded at apex with su pi 1983] ulum, ca. 3 mm long, drying reddish-brown (B & K Yellow-red 5/7.5) usually persisting as fibers around stem. LEAVES with petioles spreading to erect, (6.5)22-75(99) cm long, 5-15 mm diam., usually D-shaped, sometimes subterete, broadly sulcate, the margins acute; geniculum 1.5-3 cm long (sometimes tinged with red in larger plants); blades triangular to ovate, subcoriaceous, grad- ually to abruptly acuminate at apex, broadly to deeply lobed at base, 12-55 cm long, 7-30 cm wide, broadest at base; the anterior lobe 11-40 cm long; posterior lobes 5.5—18.5 cm long, usu- ally directed upward; the sinus rarely arcuate, usually hippocrepiform or spathulate, rounded at apex; the upper surface semiglossy to glossy, the lower surface semiglossy; the midrib broadly raised at base above, acutely raised to middle, sunken at apex, convexly raised below; basal veins 3-7 pairs, the fourth to seventh coalesced 1-5.5 cm, raised above and below; the posterior ribs naked, the outer margin rolled up; primary lat- eral veins 6-11 per side, departing the midrib at 45°-60° angle; weakly sunken above, raised be- low, straight to the collective vein; lesser veins sunken above, raised below; collective vein aris- ing from the first basal vein, } bove, raised below, 5-14 mm from margin. INFL - CENCE spreading, shorter than or equalling leaves; peduncle 17—49(65) cm long, 3-8 mm diam., terete to weakly ribbed, shorter than pet- loles; spathe subcoriaceous Qw-green 6/7.5 to 5/10), narrowly ovate to eee 6-8.5 cm long, 1.5-2.7 cm wide, Toadest just above point of attachment, cau- us pistils weakly emergent, green; stigma lin- as Sell mm long; stamens developing deve from the base, the full complement cedi ped promptly, the leading stamens pre- ing alternates only by 2 or 3 spirals; anthers m to tan, 0.2 mm long, 0.5 mm wide, exsert- CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 327 ed on short, transparent filaments, quickly re- tracting; thecae oblong-ellipsoid, slightly divar- icate; pollen yellow to cream, fading to white. TESCENCE pendent; spadix to 20 cm long, to 2 cm diam.; berries oblong to obtusely triangular, truncate and usually broadest at apex, medium orange (B & K Yellow-red 7/2.5), 10- 13 mm in both directions; mesocarp firm, thick with few raphide cells; seeds 1-2, pale greenish- white, weakly flattened, + reniform, 5-7.5 mm long, 4-6 mm wide, 3-4 mm thick, minutely appendaged at both ends. Figs. 111, 112, and 15. Anthurium montanum is known only from southeast Chiapas and southwest Guatemala at elevations of 1,200 to 2,900 m in wet mountain forests and cloud forests (not designated as part of the Holdridge Life Zone system). In Chiapas the species is known principally from “bosque de oyamel" but probably also “bosque pino- encino." The type locality of A. montanum, namely Volcán de Fuego, is located in southwest Guatemala along the frontiers of the departments of Escuintla, Chimaltenengo, and Sacatepéquez. According to annotations by N. E. Brown on the type collection, it consists of a mixed collection of two species, namely A. andicola and A. mon- tanum. Despite my reluctance to question the generally excellent observations of Brown, I can see no differences in the two parts of the collec- tion and certainly one part does not appear to be A. andicola. The species is in section Belolonchium and is recognized by its narrowly ovate blades, its usu- ally D-shaped, weakly glaucous petioles, long- pedunculate inflorescences with a matte, green to violet-purple spadix, and thin dilacerating cataphylls. Anthurium montanum is confused with A. huixtlense, a more wide ranging species usually occurring from about 300 to 1,500 m (to 1,800 m fide Flora of Guatemala). The latter species is distinguished from A. montanum by having coriaceous cataphylls that turn reddish-brown and weather to slender fibers while t p and either green or violet-purple. Anthurium montanum is also similar to A. umbrosum, a species from northern Oaxaca with similar narrowly ovate leaf blades and dilacer- ating cataphylls, but that species differs in having > fs y » m Nn © "TH - a” m z [7] 22] Q a m w > e. > a e > r- Q > nw gJ m y i : f Fissures 112-115. 112. Anthurium montanum Hemsl., Croat 41015.—113. Anthurium montanum Hemsl., Croat 47235.— 114—115. Anthurium nakamurae Matuda, Croat 47407 CROAT 1983] a terete petiole and a leaf blade that is relatively straight along the margin. Anthurium montanum is perhaps most closely related to A. titanium, a species from southeast Chiapas and adjacent parts of Guatemala and sympatric with A. montanum in the lower parts of its elevational range (i.e., below 1,800 m). The two species share a D-shaped, weakly glaucous petiole and leaf blades of similar color and tex- ture, but those of A. titanium attain generally much larger size and are almost round to broadly ovate in outline, while those of 4. montanum are generally narrowly ovate. other species with which 4. montanum can be confused is A. chamulense, with which it is Sympatric in eastern Chiapas. They share a sim- ilar blade shape but the latter species is distin- guished by having blades more coriaceous. Anthurium montanum is apparently closely related to 4. cordatotriangulum, which differs in having a more deltoid leaf blade with conspic- vously elevated veins, a generally longer and ka. south of Motozintla, but the only species wo now occurring in the area is 4. mon- Gua $ Standley ¢ Php p CHIMALTENANGO: no other location, F) above Las Cal di 80173, 80174, 80190, s Calderas, Stan "d = Libertad, Standley 51172 (F). QUEZALTENAN- o F Pa e: Bn ar (UMO). SAN MARCOS: vicinity meda 3460 (DUKE y cinit coy Mar- mi Croat 41015 (M MO); San Pedr Sh ei ii. TT EY olcá n Tajumulco, Steyermark 37063 (F). lara SU 467 : UEZ Voices Santa C 46664, i : E M tenango, Tejada 308, 310 rts $7235 eo petals vicinity El Rosario, Croat 40751, (MEXU, M : Er ins Monte Ovando, Matuda 2571 CH); Matuda 4212 (DS, MEXU, MICH, NY, y oU Mari 4219, 19663 (MEXU); Matuda Gran ; t. Pash- aa Pal nal, Matuda 17805 (MEXU); E side 9). A mem monteverdense Croat & Baker, Bre- esia 16 (Supi. 1); 62. 1979. Type: Costa OF MEXICO AND MIDDLE AMERICA 329 Rica. Puntarenas: Monte Verde Biological Preserve (operated by the Tropical Science Center), Kennedy 582 (F, holotype). piphyte, to 1.5 m tall; stems thick, moder- ately short, the internodes very short; gym S apex, pasen lobed at base, 29-50 cm long, 18.5~-34 wide; the anterior lobe 22-36 cm ong, dis unn broadly convex; the posterior obes 7-12 cm long; sinus spathulate to hippo- crepiform, longer than broad, rounded at apex; upper surface y with sparse, + obscure, raphide cells, lower r surface sparsely and obscure- ly dark dotted; the midrib raised above and be- low; primary lateral veins 3—5 per side, departing midrib at 40?—50* angle, E straight until nesr the margin, then p y arcuate joining the margin; basal veins 5-6 pairs, 2-4 of them coalesced 1.5-4 cm, arcuate and joining he margin, + loop-connecting the secondary veins; posterior rib prominently curved, naked for most of its length, turned up along its outer margin; sees vein arising from the upper- most | or 2 primary lateral veins. INFLORES- CENCE shorter ia ave peduncle 35—40 (100) cm long, 6-8 mm diam.; spathe coriaceous, white, broadly ovate, 8-10(17.5) cm long, 5-6 (8.5) cm wide, acuminate-cuspidate at apex, rounded at base, broadest at or slightly below middle, held erect, inserted at ca. 45° angle on peduncle; spa- dix green to creamy yellow or yellowish-tan, 5.5— 14.5 cm long, 1-2.2 cm diam.; flowers rhombic to weakly 4-lobed, 2-2.2 mm long, 2.2-2.5 mm wide (dry); 10-12 flowers visible in either spiral; the lateral tepals 1.2-1.3 mm long, the inner mar- gin broadly convex; pistil and stamens not stud- ied. INFRUCTESCENCE not seen. The species is apparently endemic to the mountains in west central Costa Rica at eleva- tions of 1,400 to 1,600 m, in the Provinces of Alajuela and Puntarenas in lower montane moist and premontane wet forest. Anthurium monteverdense belongs in the sec- tion Calomystrium and is closely related to A. hoffmannii, but that species differs in having an ovate-lanceolate spathe. Costa RICA. ALAJUELA: W of Monteverde Reserve, Dryer 1529 (F); Zapote, Austin Smith 2893 (F); Zar- cero, Austin Smith H564 (F). PUNTARENAS & ALAJUELA: 330 Monteverde Reserve, Burger & Baker 9710 (MO); Kennedy 582 (MO). Anthurium nakamurae Matuda, Revista Soc. Mex. Hist. Nat. 11: 93, Fig. 2. 1950. TYPE: Mexico. Chiapas: Cascada, Siltepec, elev. E m, June 2, 1949. Matuda 18668 (MEXU-85863, holotype). N e o Epiphyte; stems very short; roots thick, green, descending; cataphylls subcoriaceous, 4-6 cm in red-violet, acuminate at apex with short subapical apiculum, drying medium brown (B & K Yellow 4/2.5), weathering into reticulate fibers. LEAVES erect to spreading; petioles 10.5— 40 cm long, 4-5 mm diam., usually terete, some- WCaKLy ,B m ma 2 cm long; blades lanceolate to narrowly trian- gular, acuminate at apex, truncate to weakly lobed at base, 14.5-45 cm long, 4.5-12 cm wide, broadest at middle or just above; the anterior lobe 15-42 cm long, the margins straight, some- times convex in apical one half of blade; poste- rior lobes 2-5 cm long; sinus arcuate with petiole decurrent to parabolic; both surfaces semiglossy; midrib flat at base above, raised near middle, sunken at apex, raised below; basal veins 2-3 vein usually arising from the first or second basal vein, sometimes from one of the primary lateral veins near the base, flat above, prominulous be- low, 3-5 mm from the margin. INFLORES- CENCE spreading, equal to or longer than leaves; peduncle 30-47 cm long, 3-5 mm diam.; spathe thin, lanceolate to narrowly ovate, green heavily tinged red-violet, 4.5-6 cm long, 1.9-2.5 cm wide, long-acuminate at apex, weakly cordate at base, inserted at 90? angle on peduncle; spadix sessile or stipitate, the stipe sometimes to 1 cm (rarely to 4 cm long), olive-green or purple, 3.5-8.2 cm long, 6-7 mm diam. at base, 3-5 mm diam. at apex; the flowers 4-lobed, 3.2-3.5 mm long, 2.6- 3.3 mm wide, the sides sigmoid; 5-6 flowers vis- ible in the principal spiral, 8-10 flowers visible in the alternate spiral; tepals matte, densel pillate, lateral tepals 1-1.6 mm wide, the inner margin convex, turned up, the outer margins tinged red-violet; pistils emergent, red-violet: stigma linear, ca. 0.2 mm long, scarcely open; stamens developing rapidly from the base the laterals first, followed quickly by the alternates, ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 the leading stamens emerging only 1-2 spirals ahead of the third and fourth stamens; anthers pale yellow, 0.6 mm in both directions, held over pistil, retracting to edge of tepals but not contig- uous; thecae ellipsoid, scarcely divaricate; pollen bright yellow, fading to white. INFRUCTES- CENCE pendent; spathe persistent; spadix to 17 cm long; berries red, obovoid, round at apex, ca. 10 mm long, 7 mm wide; seeds 2, pale green, flattened, 5-6 mm long, ca. 3 mm wide. Figs. 114 and 115. Anthurium nakamurae is apparently endemic to eastern Chiapas in the vicinity of Siltepec and La Grandeza at elevations of 1,200 to 2,100 m. My collection was made in a cloud forest area that appeared to be either tropical wet forest or premontane rain forest. The species is a member of section Belolon- chium and is characterized by its usually subte- rete, sometimes weakly sulcate petiole, its mod- erately thick, oblong to oblong-elliptic leaf blade. which is usually truncate (rarely subcordate) at the base and usually broadest above the . Blades are often weakly constricted somewhat above the base. Anthurium nakamurae is most similar to A. parvispathum, which has leaf blades oblong t0 oblong-elliptic with the base rounded to truncate, but that species differs in having much thicker blades that are generally much paler and matte on the lower surface. Moreover, typical f ar L JC ME ET ; ic A. nakamurae is reported principally epiphy" and, where I have seen it, grows high up 1n that appear to persist intact rath to reticulate fibers as in typical A. nakam : uix- MEXICO. CHIAPAS: vicinity Siltepec on road to - tl 413 (DS); vicinity mend CN Siltepec, Croat 47407, 47419, 47461 OE ja cada, Siltepec, a 18668, 19660 (M , Grandeza, Matuda 5576 (F, MEXU). urae. Anthurium nelsonii Croat, sp. nov. TYPE p ico. Oaxaca: Reyes (Reyes Etla, NNW ° Jsof of Oaxaca) 2,500-3,466 m; E. W. Ne 1758 (US-358640, holotype). Planta terrestris, ca. 1 m alta; caudex ca. 4 Cm pe petiolus 29-48 cm longus, sulcatus late acuteque, 2- na ovata aut triangularis-ovata, 30-57 cm long, iore 56 cm lata, basi cordata profunde, in pagina = pe- albida hebetataque; inflorescentia effusa- ; j 1983] dunculus 30-43 cm longus; spatha viridis, 10-13 cm longa, 1.5-3 cm lata; spadix purpureus, 6.5-15 cm lon- gus; baccae obovoideae. Terrestrial, less than 1 m tall; stems ca. 4 cm diam.; roots 3 mm diam.; tan, smooth to woolly pubescent (dry); cataphylls ca. 8 cm long, drying tan to reddish-brown, weathering to moderately thin fibers. LEAVES with petioles 29—48 cm long, -8 mm diam., sharply and broadly sulcate adaxially; the geniculum 1-1.5 cm long; blades broadly ovate to ovate-triangular, 30-57 cm long, 22-56 cm wide, broadest at base, acute to nar- TOWIy ded and apiculate at apex, deeply lobed at base, the margin sinuate and undulate; ante- rior lobe 25-45 cm long; posterior lobes 11.5- 25 cm long from apex of sinus to the outermost point; the sinus arcuate or arcuate with petiole decurrent; upper surface semiglossy, lower sur- face matte, often whitish with glistening trans- lucent circular cells scattered in the epidermis (visible at high magnification on dried leaves); midrib convexly raised above; the basal veins 5- 9, all except the first and second (third) coalesced 5-11.5 cm, raised on both surfaces, drying darker than the surface; posterior ribs broadly arcuate, naked most their length; primary lateral veins 5- per side, almost straight or gradually curved upward, arcuate-ascending near margin and join- ing Margin, raised on both surfaces; collective vem arising from one of the lowermost primary lateral veins (rarely as low as the first basal vein), -7 mm from margin, weakly raised on both Ps oe INFLORESCENCE spreading-pendent (); peduncle 30-43 cm long, 3-5 mm diam., hi shorter to longer than petioles; spathe sub- d green, lanceolate to ovate-lanceolate, ibid ne long, 1.5-3 cm wide, broadest just ed is ase, narrowly acuminate at apex, round- e E iate at base, inserted at ca. 45* angle T ^ uncle; the stipe 3-25 mm long in back, Pe viel. long in front, 3-4 mm diam.; the spadix "Purple, 6.5-15 cm long, 6-9 mm diam. at the as mm diam. at apex, sometimes curved; mcd ng rhombic, 3.6-4.7 mm long, 2.6-3.9 Rig (dry), the sides scarcely to jaggedly ral, 46 6-8 flowers visible in the principal spi- ms flowers visible in the alternate spiral; Pals 2 ig de^ minutely papillate, lateral te- roun ig mm wide, the inner margin broadly mm lo » Pistils not emergent; stigma oblong, mode ng; stamens emerging from the base ina i ng slow progression; the anthers held in Es cluster completely obscuring the pistil, " mm long, ca. 1.4 mm wide; thecae narrowly — CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 331 ovate, scarcely divaricate; pollen not seen. IN- FRUCTESCENCE pendent, berries often de- veloping only in the basal one half or one third; ong, 3 cm diam.; berries obo rounded and scurfy, 10-13 mm long, 9-10 mm i icarp moderately thick, leathery at apex, ense in lower three quarters of berry; seed rown, considerably flattened, raphide cells sparse and minute, 9-13 mm long, 7-9 mm wide, 3-4 mm thick, weakly depressed at apex, with a sticky, Figs. 117 and 118. Anthurium nelsonii is endemic to Mexico, oc- curring in the State of Oaxaca in the arid region north northeast of the city of Oaxaca between latitudes 17°10'N and 18°N at elevations of 1,000 to 2,500 m. The principal vegetation types of the region are “selva baja caducifolia" and **bosque pino-encino.” The species is named in honor of the first col- lector, E. W. Nelson, who made the first collec- tions in 1894. The collections have in part been confused with A. andicola, a species occurring in wetter parts of northern Oaxaca. Anthurium nelsonii is a member of section Be- lolonchium and is distinguished by its broadly ovate-triangular blade with well developed pos- erior lobes and a broad sinus and by its sharply and broadly sulcate petiole. The most charac- teristic feat d the principal feature by which it is distinguished from A. andicola is its matte, often whitish lower blade surface. The latter species is semiglossy on the lower surface. MEXICO. OAXACA: no other location, Conzatti 1724 (MEXU); Tomellín Canon, Rose & Rose 11351 (NY, US). Anthurium nizandense Matuda, Bol. Soc. Bot. México 24: 35, Fig. 1. 1959. TvPE: Mexico. Oaxaca: Nizanda near the Isthmus of Te- huantepec (16?40'N, 95°2'W) dry, rocky hill, MacDougall s.n. (January 6, 1959) (MEXU, holotype). lotype). 332 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 nother Vas veteoail Croat " C termi vis 19. 116. Anthurium nakamurae Matuda, Croat 47407.—117-118. Antht A ATES A Pa Sa RI T nizandense Matuda, Croat 45756. Y I rium nelson! i 1983] Usually terrestrial or epipetric; stems thick; leaf scars 0.7-1.2 cm long; roots numerous, thick, T em obovate to obovate-elliptic, subcoriaceous, 25- 36.5 cm long, 10.7-18 cm wide, broadest just above middle, acute at apex and base (the tip glandular); the upper surface matte to semiglos- mary lateral veins 6-8 per side, departing midrib at 50*-60* angle, prominent above and below; straight to weakly arcuate-ascending to collective vein; lesser veins scarcely visible and flat on both surfaces; collective vein arising from one of the primary lateral veins near middle of blade, 3—9 ?-11 mm diam. at base, 4 mm diam. at apex; the flowers rhombic, 1.8-2.5 mm long, 2.8-3 =m wide, the sides sigmoid; 7-9 flowers visible in the principal spiral, 5-6 flowers visible in the alternate spiral; lateral tepals 1.5 mm wide, broadly rounded to weakly concave; the pistils ee stigmas broadly ellipsoid; stamens just I! in à prompt, complete sequence, held m : ve tepals in a contiguous circle around Ns, anthers ca. 0.4 mm long, 0.6 mm wide; TES cL liPSoid, weakly divaricate. INFRUC- dix 6 3 E pendent; the spathe withered; spa- em long, 1.5 mm wi e; berries round, et nt White, 8 mm long; mesocarp transpar- obo ^on Pulpy; seeds 1 (2?) per berry, broadly With Pale yellow (B & K Yellow 9/2.5), tinged long, 4 dn at base and on one side, 5.2-5.4 mm 374.7 mm wide, 3.4-4 mm thick, enve- Nerei à gelatinous sack; appendage minute, With sien t gnarled at apex on side, tinged Town. Figs. 119 and 120. CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 333 Anthurium nizandense is endemic to Mexico in southern Guerrero and southern Oaxaca, where it occurs at less than 850 m in seasonally very dry habitats on steep, rocky cliffs. It is known only from the vicinity of Tierra Colorada in Guerrero and at Nizanda in Oaxaca. The species is a member of section Pachyneu- rium and is recognized by its greenish-white ber- ries and by its thick, narrowly to broadly elliptic or oblanceolate-elliptic, relatively long-petiolate leaves that are matte on the lower surface and ti hibit a d y of minute, linear raphide cells on drying. It has at least the lower primary lateral veins extending to the margin. Matuda reported the petiole to be subquadran- gular, but this was possibly due to a misinter- pretation of the dried specimen since this char- acter is not apparent on the live material. The epidermal cells on both surfaces of the blade are convexly raised. In the area of Mexico where it occurs, A. ni- zandense could be confused only with A. schlech- tendalii ssp. jimenezii, which is more generally abundant and occurs in similar habitats. An- thurium nizandense is distinguished by its pro- rtionately longer petioles and its more typi- cally elliptic blade, which is distinctly matte on the lower surface. In A. nizandense the blades are 1.3-2.5 times longer than the petioles. In A. schlechtendalii ssp. jimenezii, the leaf blades are usually semiglossy beneath and usually 6-10 times longer (rarely as little as 3 times longer) than the petiole. MEXICO. GUERRERO: between bridge over Río Omi- tlán & confluence of Río Papagayo, Croat 45756 (MO); Tierra Colorada, Kruse 1595 (MEXU). oaxaca: Ni- zanda near Isthmus of Tehuantepec, MacDougall s.n. EXU). Anthurium obtusilobum Schott, Oesterr. Bot. Z. 8: 181. 1858. Type: Costa Rica, near San Miguel, Wendland 777 (GOET, holotype; W, isotype). Epiphyte or terrestrial; stems ca. 1.5 cm thick, to 30 cm long; roots numerous, moderately thin; cataphylls 10-15 cm long, drying tan, persisting intact. LEAVES with petioles + erect, subterete, obscurely sulcate, 19-60 cm long, 6—9 mm diam.; geniculum ca. 2 cm long; blades ovate, moder- ately thick, 15-46 cm long, 9.5-28 cm wide, gradually to abruptly acuminate at apex, deeply | lobed at base; anterior lobe 21-33 cm long, the > Z Z > - n o "Tj - le» m E [77 ^l Ó C a w Q " » ys O > - Q > a J m 2 FIGURES a? 20—123. 2 Anthurium nizandense Matuda, C roar ma 56.—121. Anthurium obtusilobum Schott, Croat 49997.—122. Anthurium ochranthum eina yeh ‘roat 4569 -123 Anthurium ochranthum Croat 33965. CROAT LAVAL 1983] margins convex; posterior lobes 9-16 cm long; sunken above, raised below; basal veins ca. 8 pairs, 3-4 coalesced 2.5—4 cm; posterior ribs na- ked, turned up along the outer margin; collective vein arising from one of the lowermost basal veins, extending to apex, 3-7 mm from margin. INFLORESCENCE spreading, shorter than leaves; peduncle (7)15—43 cm long, shorter than white, 7-11.5 cm long, 2.3-2.7 cm acuminate at apex, obtuse at base; spadix cream to pale yellow-green, turning dark lavender to 6 at apex; the flowers 4-lobed, ca. 3 mm long, ca. 4 mm wide, the sides jaggedly sigmoid; 5-10 pec. visible i in the principal spia 8—9 flowers isi lateral Nai ca. 1.8 mm wide. the inner margin up- turned; pistils raised, white to pale green; stigma brushlike, 0.4-1 mm long, glistening papillae exserted 4-5 days before first stamens emerge, turning brown; stamens emerging slowly from base, exserted on short white filaments, held over 0.3-0.9 mm wide; thecae ellipsoid, not di- Varicate; pollen white. INFRUCTESCENCE with Spadix to 15 cm long; berries red-violet, ovate, acute at apex; seeds not seen. Fig. 121. Anthurium obtusilobum ranges from Costa Rica to Colombia (Choc6) at elevations of 600 to 1,550 m in premontane wet, tropical wet, premontane rain, and lower montane wet forest. d The species is in section Calomystrium and is Osest to A. hoffmannii, sharing with that species thick, persistent, reddish-brown, entire cata- Costa D Near San Miguel, Wendland 777 ALAJ ariblanco, Luteyn 3222 AXE. 3341 (DUKE, MO); W of Fort Arenal, Taylor & Taylor 11585 (MO); between UA Aguas Zarcas, Croat 46953 (MO); between vicini E pote, Croat 46920 (MO); ity San Almeda & Nakai 3849 (CAS); OF MEXICO AND MIDDLE AMERICA rtuna, NW of 335 (DUKE, MO); vicinity of Vara Blanca, ep 3906 (DUKE); San Luis de Zarcero, Smith 1 i , Smith 4 : Án Wd 47081 (MO); Muneco Standley & Torres 5171 2 (US); near Selva above Río Taus, Edad & Taylor 11436 (NY); near Tapanti, e 27 (DAV, T Lent 840 ( to Tau pe & buc s 593 ERE urrialba, Mae 4335 6, 43358 (MO). HEREDIA: N ye Vara Blanca, Vicinity Río La ie Grande, Croat 36024 (MO). LIMÓN: Toro Maris: Tessene 1422 (WIS). PUNTARENAS: Las Cruces Botanical Garden, Croat 32966 (MO). SAN JOSÉ: Laguna la Chonta, Standley 42358 (US); La Palma, Luteyn 2999 (DUKE Anthurium ochranthum C. Koch, Ind. Sem. Hort. erol Q6. : Wendland s.n. (P, ATE RET N B BP LE, W, isolectotypes; designated Croat & Baker, 1979) Ee EREA Schott, Oest henb 9. 1857. TvPE: ata ee a huri j | Enel.. Bot Jahrb Syst scale LI 1898. TYPE: Costa Rica. Limón: Tal in forest near Shirores, 100. m eig " Pittier & Tonduz 9225 (B Anthurium isakianu m Engl., Bot. Jahrb. Syst. 25: 423. 1898. Type: Costa Rica, Talamanca, in forest near Taki, 200 m elev., Tonduz 9510 (B, holotype; CR pe). Anthurium baileyi Standl., act red Mig Nat. Di t. Ser. 22: 66. 194 40. T anal bun Ead, Island, der & 1 Bailey 196 (F. holotype Terrestrial, rarely more than 1.5 m tall; stem to 45 cm long, 2-4 cm diam.; cataphylls 11.5- 13 cm long, drying brown, weathering to coarse linear fibers. LEAVES erect-spreading; petioles terete or obscurely sulcate, 30-70(90) cm long, 5-8 mm diam., blades ovate-triangular, mod- erately thin, 39-75 cm long, 18-48 cm wide, acuminate at apex, deeply lobed at base; the an- terior lobe 23-52 cm long; posterior lobes 11- 22 cm from sinus to outermost point; sinus nar- rowly triangular; the upper surface matte to semi- glossy, the lower surface matte; the midrib raised above and below; primary lateral veins 4—6, raised in valleys above, raised below; collective vein arising from the uppermost basal vein. INFLO- RESCENCE erect, usually shorter than or some- times longer than leaves; peduncles 24—63 cm long; spathe, green or pa ale green sometimes tinged with red-violet, lanceolate, 10-15(32) cm long, 3-5 cm wide; spadix green turning bright yellow, 7-22 cm long, 5-8 mm diam. at base, 4-6 mm diam. at apex; flowers 4-lobed, ca. 2.6 mm long 336 (dry), ca. 2.3 mm wide (dry), the sides jaggedly sigmoid; ca. 6 flowers visible in the principal spiral, ca. 9 flowers visible in the alternate spiral; tepals glossy, lateral tepals 1.7-2 mm wide, the inner margin convex; pistil weakly emergent, green; stigma brushlike; stamens emerging scat- tered throughout length of spadix, usually the lateral stamens first, the alternates quickly fol- lowing; anthers held over pistil; thecae yellow- orange; pollen orange. INFRUCTESCENCE with spadix 11-24 cm long; berries violet-purple at apex, white at base, obovoid, 7-9 mm long. Figs. 122 and 123. Anthurium ochranthum is known from Costa Rica to Panama from sea level to 1,350 m, most commonly in tropical wet forest but also in wet- ter parts of tropical moist forest and in premon- tane wet and premontane rain forest. The species is placed in section Belolonchium and is recognized by its terrestrial habit, bright yellow, long-tapered spadix, violet-purple and white berries, and the leaf blades that have the anterior lobe sometimes concave along the mar- gins. It is most closely related to A. pluricosta- tum. See the discussion following that species for differences. COSTA RICA. ALAJUELA: NNE of Bijagua, Croat 36285, 3647 1 (MO); between Naranjo & Aguas Zarcas, Croat 46951 (MO); near Rio Segundo, ca 36857 (MO). CARTAGO: SE of Platanillo, Croat 36809, 36733 (MO). eal Burger & Gentry 9100 (MO). L : i es & Rio Pacuare, Burger & Liesner 6880 (MO, NY . Anthurium oerstedianum Schott, Oesterr. Bot. Z. 8:180. 1858. Ie Costa Rica, near Naran- jo, Oersted s.n. (W, destroyed? Photo of Schott Aroid Davi #228, NYBG Neg. #N.S. 3817). Anthurium cuspidifolium Schott, Oesterr. Bot. Z. 8: 180. 1858. Type: Costa Rica, Oersted s.n. (W, pein by Photo of Type at W, Field Mus. #29812 (MO); Photo of M Aroid Drawing #314, NVDG Neg. #N.S. 3816). Terrestrial; stems short, to 2 cm diam.: ; roots umerous, short, tan; cataphylls 7-12 cm long, moderately thin, apiculate at apex, drying brown, persisting intact, ultimately deciduous. LEAVES usually + erect; petioles + quadrangular, (12)30- 65 cm long, ca. 0.6 cm wide; the geniculum 1.5— 2 cm long, remote from base of blade 10-20 cm; blades ovate to narrowly ovate or lanceolate- elliptic, moderately thin, 20-66 cm long, 8.5-30 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 cm wide, caudate-acuminate at apex, obtuse to truncate, sometimes rounded to subcordate at base, abruptly: Deum on petiole bom surfaces | semiglossy; y e, raised. | and obscurely ribbed below; primary lateral veins / 11-20 as side, raised in aie above, n: below; less collective vein arising from one of the lowermost nate at apex, rounded at base; spadix pale green, 5-12 cm long, ca. 9 mm diam. at base, ca. 6 mm diam. at apex; flowers sub-rhombic to 4-lobed, 2.2-2.5 mm long, 2.2-2.7 mm wide, the sides parallel to spirals usually straight, those perp white, 0. 7-1 mm long, 1-1. 2r mm wide; - ellipsoid ith abundant, A pollen. INFRUCTESCENCE not seen. Fig. 12 pS | Anthurium oerstedianum is endemic to a | Rica at elevations of 900 to 1,300 m, principally forest. Anthurium oerstedianum is distinguished by Hd its strictly terrestrial habit as an understory forc ] | herb and by its geniculum which is located 10- 20 cm below the base of the blade. Other dn guishing features are its 4-sided petiole and i dig spadix. T he species isa meme a: donde petioles, its remote geniculum, an of a densely rosulate ha Anthurium iui are which is sns mous with A. oerstedianum, was e MacBride (1936) from Peru. However, the a imens cited by him [MacBride 5622 (P: Schun m, 572 (F)] are referable to A. flavescens PoePP- — Costa RICA. SAN JOSE: between San Isidro del Ge a ne | ral dud Dominical, Croat 35267 (MO); about e | i Ci xn O > - | E = M = oy a [2 S x oO "T S m i O O > Fa J z J J es m » S m e ie! » FIGUR T^ 12 7. 1983] beyond divide between San Isidro del General and Dominical, Croat 35293 (MO). Anthurium ovandense Matuda, Anales Inst. Biol. Univ. Nac. México 30: 113-115, Fig. 7. 1960. Tyre: Mexico. Chiapas: Municipio Escuintla, Cerro Ovando (north of town of Escuintla) 1,400 m elev., Matuda 32769 (MEXU, holotype). Note: Type description lists Matuda 32768 as the type but the only specimen found at MEXU is numbered 32769. Anthurium ocotepecense Matuda, Anales Inst. Biol. iv. Nac. México 27: 345, Fig. 4. 1957. TYPE: Mexico. Oaxaca: Cerro de Larion, ; Ocotepec, MacDougall 32] (MEXU, holotype (?) not seen). Epipetric or terrestrial to 1 m tall; stems 10- 30 cm long, 2-3 cm diam.; leaf scars 2-2.7 cm Wide; roots numerous, 5-6 mm diam., scurfy, brown; cataphylls green, 8—11 cm long, narrowly rounded at apex (acumen with subapical apicu- lum), drying tan, weathering at base into coarse bers, remaining intact at apex. LEAVES erect- Spreading; petioles terete, 29-75 cm long, 5-8 mm diam., weakly glaucous; geniculum 1.8-2 em long, sometimes narrowly sulcate; blades nar- Towly-triangular, chartaceous, narrowly acumi- hes at: sunken at apex above; basal veins 4— Keri the fourth and fifth coalesced 4—5.5 cm, i Eo and below; primary lateral veins 5- EE side, sunken above, raised below; inter- tive ed veins sunken above, flat below; collec- s *in arising from the uppermost basal vein, mm from margin. INFLORESCENCE a Pendent; peduncle 32-72 cm long, 4-5 lanceo am., terete; spathe green, subcoriaceous, pi Es 5-7.5 cm long, 1.5-2.4 cm wide, cus- a ate at apex, truncate to weakly 2 sm at base; stipe to 2.5 cm long in front, to 9 Ong in back; spadix green, 5-11 cm long, : 4-7 flowers visible in the principal spi- Tal, p : : 2-4 flowers visible in the alternate spiral; CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 339 tepals glossy with abundant sticky droplets; pis- tils weakly raised; stigma linear, 0.5 mm long; mens emerging rapidly from base, held in tight usually developing only in lower one quarter of spadix. Fig. 129. Anthurium ovandense is endemic to Mexico in southern Oaxaca and Chiapas in moist forest at elevations of 800 to 1,600 m. The species is a member of section Belolon- chium and is characterized by its usually epi- petric habit, its thin narrowly ovate leaf blades with conspicuous posterior lobes, naked, gradu- ally curved posterior rib, and an anterior lobe that is concave and generally undulate along the margins. The species is similar to A. umbrosum from northern Oaxaca in generally much wetter areas and also to A. lezamae mostly from central Chia- pas but both of these species are characterized by having their anterior lobe distinctly convex. Matuda 37655, a sterile collection made at the type locality of A. andresloviense Matuda, was identified as that species by Matuda, but it is probably best placed with A. ovandense. It differs from typical A. ovandense in that the blade dries relatively more glossy on the upper surface and in having a distinct constriction at the base of the anterior lobe. Although the type of A. ocotepecense was not seen, the illustration of the type appears to in- dicate that it is a synonym of A. ovandense. The type locality was not determined. Although I in- vestigated part of the Municipio of Ocotepec and found it extremely arid, perhaps there may be other parts of the region that are more mesic. I was unable to locate a “Cerro de Larion" at Santo Tomás where the type was collected. The region where the type of A. ocotepecense was collected is substantially to the north of other collections. MEXICO. CHIAPAS: Escuintla, Monte Ovando, Ma- tuda 32729 (MEXU). oaxaca: Cerro Concordia, Mor- ton & Makrinius 2727 (F, US); vicinity La Galera, Croat 46164 (MO); Pluma Hidalgo, MacDougall 715 MEXU). Anthurium pallens Schott, Oesterr. Bot. Z. 8: 180. 1858. Type: Costa Rica, near Desengafo, 2,600 m, Wendland 840 (GOET). Epiphyte, scandent, internodes 5-9 mm long; 340 cataphylls 4.5—5 cm long, drying reddish-brown, weathering into linear fibers and persisting. LEAVES erect-spreading; petioles subterete, 7— 25 cm long, 2.5—6.5 cm diam., narrowly sulcate; blades lanceolate-linear to narrowly oblong-el- liptic, 18-34.5 cm long, 2.5-6.5 cm wide, long- acuminate at apex, acute to obtuse or rounded at base; midrib raised above and below; primary lateral veins 5-9 per side, sunken above, raised below; lesser veins i i ; collective vein arising from near the base. INFLORES- CENCE + erect, shorter than leaves; peduncle 10-16 cm long; spathe green, lanceolate, 2.5-3.2 cm long, 0.8-1 cm wide; spadix green turning yellow (sometimes orange), 3-4.5 cm long, 3-6 mm diam. (dry) at base, weakly tapered at apex; the flowers rhombic, ca. 3.1 mm long, 3 mm wide (dry); ca. 3 flowers visible in the principal spiral, ca. 5 flowers visible in the alternate spiral. IN- FRUCTESCENCE with spadix to 6 cm long; berries green to yellow-green (?), + globose. Figs. 124 and 125. = Anthurium pallens is known from Costa Rica and Panama at 1,000 to 2,000 m elevation in premontane rain forest and lower montane rain forest life zones. Anthurium pallens is easily confused with A. microspadix with which it shares similar habi- tats. It differs from A. microspadix by having blades acute to rounded at the base and clustered at the ends of relatively short t hes with short internodes. It also has longer petioles and blades that dry characteristically gray-green and weakly glossy on the lower surface. Anthurium micro- spadix has long internodes with the leaves more widely and regularly spaced and blades that are rounded to subcordate at the base and generally matte on the lower surface upon drying. While both species may be epiphytic, A. microspadix generally is rooted in the soil and is an erect or semierect, vine-like plant, reclining or using trees for support. Anthurium pallens is usually a short- stemmed epiphyte that is frequently pendent. Although mature berries have not been seen, they are no doubt yellowish-green and subglo- bose like those of 4. microspadix. The species is a member of section Xialophyllium. MO); Wilbur et al. 15982 (DUKE); pass between Vol. ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 cán Irazü and Volcán Barba, Davidse & Pohl 17024 (MO). Anthurium paludosum Engl., Bot. Jahrb. Syst. 25: 401. 1898. TYPE: Colombia. Valle: near Buenaventura, Lehmann 3811 (B, holo- type). Epiphyte; stems to 30 cm long; roots numer- ous, grayish-green, smooth, descending; cata- phylls coriaceous, 5-6 cm long, persisting + m- tact, drying brown, ultimately deciduous. LEAVES spreading-pendent; petioles 3-14 cm long, 3-5 mm diam., D-shaped to subterele, broadly and obtusely sulcate; geniculum 1-1.5 cm lon inti oblanceolate. subcoriaceous, 10-38 cm long, 3-13.5 cm wide, short-acuminate at apex, acute to attenuate (rare- ly obtuse to subrounded) at base; both surfaces semiglossy, sparsely punctate; midrib convexly raised above and below, sunken at apex above, primary lateral veins 12-14 per side, departing midrib at 40° angle, sometimes weakly sunken, usually + obscure on upper surface, prominulous on lower surface; collective vein arising from the base and extending + straight to apex, 3-12 mm from margin, flat and nearly obscure on upper surface, prominulous on lower surface. INFLO- RESCENCE erect-spreading, shorter than leaves; peduncle 7.5-15.5 cm long, 3-5 mm diam., ur ete, tinged red-violet; spathe moderately pe whitish to green or purplish, withering s00n afte anthesis, oblong-lanceolate, 5.5-7 cm long, 4 mm wide, acuminate at apex, rounded at bast Rhi x 1 11 + + : y elliptict inserted at ca. 60° angle on peduncle; $ ip tinged orange, becoming violet-purple, E i ral, 7-9 flowers visible per alternate ore matte to semiglossy, tinged violet-purple at ¢ emergent; stigma linear; stamens emergin jv the base, the lateral stamens emerging with 0? y á apex of the anther visible, retracting before d ternate stamens emerge, leaving only pa white pollen. INFRUCTESCE 6-25 cm long; berries orange, red-orange, tinged with orange, + globose to ovoid, ; diam.; seeds 3-4, ellipsoid-oblong; qe long, yellowish, except green on one ene 126. mim Fig tS TEL E Rd nondum uns 1983] Anthurium paludosum is known from Costa Rica to Colombia at elevations from sea level to 400 m. It ranges along the Atlantic slope only west of the Isthmus of Panama. To the east of the Isthmus of Panama it occurs on both slopes until reaching South America, where it is re- stricted to the Pacific slope. The species occurs in wetter parts of tropical moist forest and pre- montane wet and tropical wet forest life zones. Anthurium paludosum is a member of section Porphyrochitonium and is recognized by its co- riaceous, almost oblong leaf blades that are more or less the same color on both sides with a pale green midrib on the upper surface, and by its thick, persistent, nondilacerating cataphylls and reddish-orange berries. No other species in this section i is known to have cataphylls that remain intact. In the revision of the Costa Rican Anthurium (Croat & Baker, 1979), A. eggersii Engl. was in- cluded as a synonym of this species, but exam- ination of the type now precludes that possibility use that species has prominently dilacerat- ing cataphylls. In other respects the two species appear to be markedly similar. = Rica. No other location, Pittier 6837 (BR). (MO). ibd between Bribrí & Bratsi, Burger et al. 10470 Wi 175 (MO). een Limón & Puerto Viejo, Baker & Burger Anthurium panduriforme Schott, Prod. Aroid. 536. 1860. Type: Costa Rica. San Miguel, Wendland 776 (GOET). PC Dandurarum: Schott, Oesterr. Bot. Z. 8: 182. hi 58 Mart. ex Schott, 1855. Type: Costa e San Miguel, Wendland 776 (GOET, holo- Pase phyte or occasionally terrestrial; stems usu- bei dà than 30 cm long, occasionally to 1 m Phy Oots numerous, scattered aoe stem; cata- devin tan (B & K Yellow- red 9/10) usually PATEAT x a few thin fi- in sometimes persisting. LEAVES moderately scu 25-73 cm long, 3-7 mm aN 1 ULLI ropan z pu blades "Certe cabra erect or ding from the erect petioles, 20—65 cm long, ata ii m wide, gradually to abruptly acuminate Hise. » Prominently lobed at base; anterior lobe lobes ong, constricted near base; posterior nearly orbicular (or rarely oblong), some- CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 341 times directed upward from point of petiole attachment: sinus s usually spathulate y veins less conspicuous: basal veins 5—9 pairs, 3- 4 coalesced 1—4.5 cm; posterior rib naked, weak- ly turned up along outer margin; collective vein arising from one of the lowermost basal veins and extending to apex, 3-7 mm from the margin. CENCE erect, shorter than leaves; peduncle (5)11—20(32) cm long, 4-6 mm diam., terete; spathe moderately thin, pale green, some- times tinged reddish, oblong-lanceolate, 6.3-19 m long, 2.7-3.5 cm wide, abruptly acuminate : apex, rounded at base, inserted at 60? angle on peduncle; spadix golden yellow (B & K Yel- low-red 8/10), 7-18 cm long, 7-9 mm diam. midway, tapered to 4-6 mm at apex; the flowers sub-4-lobed, 2.5-3 mm long, 2.9-3.3 mm wide, the sides + straight to jaggedly sigmoid; 6-9 flowers visible in the principal spiral, 6—10 flo ers visible in the alternate spiral; tepals yellow, glossy, lateral tepals 1-1.5 mm wide, the inner margin convex; pistils weakly emergent, pale green, translucent; stigma linear, ca. 0.3 mm long, brushlike and glistening as anthers emerge; sta- mens emerging in scattered manner slowly from base to apex, exserted to ca. 1.3 mm on white or translucent, flat filaments, the exserted portion of the filaments ca. 0.3 mm long, ca. 0.5 mm wide; anthers yellow, 0.5-0.6 mm long, 0.7-0.9 mm wide; thecae divaricate; pollen golden yel- low. INFRUCTESCENCE arching-pendent; the spadix to 23 cm long; berries pale green, ovoid, 5-7 mm long; seeds 2, greenish-white, 2.6-3 mm long, 2.3-2.5 mm wide, 0.8-1 mm thick. Figs. 127 and 130. - The species is known from Costa Rica to Co- lombia and probably Ecuador at 400 to 1,560 m elevation in tropical wet, premontane rain, and lower montane rain forest. In Costa Rica, the species ranges from Bijagua in the north to Ta- pantí in the south on the Atlantic slope. In Pan- ama the species occurs on both slopes. Anthurium panduriforme is distinguished by its almost orbicular posterior lobes (sometimes more elongate in Panama, particularly collec- tions made at Cerro Tute, near Santa Fe, Vera- 342 guas), its characteristically yellow-green, shiny lower blade surface, and its golden yellow spadix and is no doubt related to other South American species such as A. argyrostachyum Sodiro of Ec- uador that have numerous closely parallel pri- mary lateral veins. Costa Rica. San Miguel, Wendland 776 (GOET). ALAJUELA: NNE of Bij ‘esi ' Croat 36286 (MO); N of Coliblanco, Utley & Utley 4687 : Sy San Ramón, Utley & gedi 4583 (DUKE); valley of o Sarapiquí, between Vara Blanca & Cariblanco, us 6614 (BH). HEREDIA: vicinity Bajo La Hondura, Croat 44515 (MO); N of Vara Blanca, Croat 35587 (MO). LIMOn: vicinity Guapiles, Standley 37255 (US). Anthurium parvispathum Hemsl., Biol. Cent. Amer., Bot. 3: 432. 1885. Type: Guatemala. Baja Verapaz: Santa Rosa, Salvin & God- man 408 (K, holotype). Anthurium hypoleucum Standl., J. Wash. Acad. Sci. RR 11: 610. 1932. Type: Guatemala. Santa Rosa, 1,600 m, Tuerckheim II.2214 (US. 932125, holotype; B, isoty pe). Epipetric; stem short; roots thick, tan, some- times directed upward; cataphylls ca. 7 cm long, drying brown, persistent, weathering to coarse linear fibers. LEAVES spreading; petioles 16-50 ong, 6-9 mm diam., broadly and sharply sulcate, weakly glaucous; geniculum 1.3-1.5 cm long; blades oblong to lanceolate, subcoriaceous cous, the lower surface only slightly paler; ce convexly raised above, diminished and sunken at apex, more acutely raised below; basal veins one pair, sunken above, raised below; primary lateral veins 6-10 per side, departing midrib at 40° angle, sunken above, raised below; interpri- veins weakly sunken above, flat below; lesser veins obscure; collective vein UN from the basal vein, 8-10 mm from ma sunken above, raised below, INFLORESCENCE spreading, shorter to as long as leaves; peduncle 31-52 cm long, subterete, 1-2 faint ribs; spathe subcoriaceous, green (B & K Green 5/7.5), lan- ceolate, 3.5-7 cm long, 1-2 cm wide, long-acu- minate at apex, clasping at base, inserted at 180° angle on peduncle; spadix sessile, green (B & K ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 RASA 7/7.5) becoming pale yellow, 3.5- 9 cm long, 7-10 mm diam. at base, 3-5 mm diam apex; flowers 4-lobed, ca. 2.3 mm in both raitis: the sides jaggedly sigmoid; 5-7 flowers visible in the principal spiral, 7-9 flowers visible in the alternate spiral; tepals matte, lateral tepals ca. | mm wide, the inner margin convex pistils green, weakly emergent; stigma slit-like, stamens emerging from the base, laterals emerg- ing first, followed quickly by alternates, the lead- ing stamen preceding fourth stamen by 1-3 spi- rals. INFRUCTESCENCE pendent, mature berries not seen. Fig. 131. Anthurium parvispathum is endemic to Gua- temala and is known principally from Baja Vera- paz at 1,500 to 1,800 m elevation, but it is also known from single collections from the Sierra de la Minas in Zacapa Department and at Mazaté- nango in the Dept. of Suchitepéquez on the Pa- cific Coast. The species occurs principally in open or exposed areas on rocks in areas of dry forest. The species is a member of section Belolon- chium and is characterized by having a Sess spadix and oblong-lanceolate leaves that are a least initially glaucous (often drying whitish) on the lower surface and by having the primary lat- eral veins equally as prominent as the basal vein. The species is also characterized by occurring 0n rocks. Anthurium parvispathum is similar to A. pi kamurae from southeast Chiapas but differs i having thicker leaves, which are generally muc paler and weakly glaucous on the lower eu In addition, it is always found growing 0n roc xtent. shape and venation, but that species has m vated broader leaves that have conspicuously os pe au fused with some leaf formal of A. 5 but that species differs in having the pasal v p much more conspicuous than the primary " veins. ATEMALA. BAJA Yanan Pantin, Williams et al. p 94 Williams et al. 43354 (F); Santa RUM 214 91242, 91258, 91271 (F); Tuerekheim 2596 (US sueo ee , T6 $e ZACAPA: Río Nuevo Sitio, betwee dee UNE Suvermant 42212 (NY). 1983] CROAT Anthurium pedatoradiatum Schott, Bonplandia 7: 337. 1859. a. Anthurium pedatoradiatum ssp. pedatoradia- tum. TvPE: Mexico, cultivated in Vienna, illustrated by Schott Aroid Drawing #781, 782, NYBG Neg. #3920, 3921; Photo MO. Anthurium pedatifidum Regel & Linden, Gartenflora 15: 66. 1866. Type: Plate 501 accompanying the description (see discussion). Anthurium mirandae Matuda, Anales Inst. Biol. Univ. ac. México 22: 375. 1951. Type: Mexico. Chia- pas: El Ocote, NW of Ocozocoautla, elev. 575 m, Miranda 6312 (MEXU). Terrestrial, usually more than 1 m tall; stem 1.7-3 cm diam.; leaf scars 2 cm wide; roots mod- erately thick, directed downward; cataphylls sub coriaceous, 3.5-4 cm long, rounded at apex, drying medium brown (B & K Yellow 5/5), per- sisting intact at apex, splitting at base. LEAVES erect; petioles 35-87 cm long, 5-6 mm diam., terete; geniculum 1-1.2 cm long; blades broadly ovate, 22-33 cm long, 38-60 cm wide, broadest at middle or just below, to 13-pedatisect, the lobes confluent 4-7 cm from base, the base De lobed, the apices of the lobes narrowly minate, the margin of the lobes entire, the median lobe 18-33 cm long, the outermost lobes 18-27 cm long, inequilateral and auriculate, fal- My e a xiu the median lobe sharply raised above fifth cg Ls basal veins 3-5 pairs, the third to lize esced 1.5-1 8 cm, the fourth to fifth co- Dei to 5.5 cm, raised above and below; pri- idea ateral veins of the median lobe 4-8 per shee “parting the midrib at ca. 40° angle, sunken vi A Taised below, loop-connecting; collective tsi the median lobe arising from the first or Primary lateral vein. INFLORESCENCE ia than leaves; peduncle 5-23 cm long, 1/7) ERS spathe green (B & K Yellow-green SO late, 7-9.2 cm long, 1.5 cm wide, long i he 50 angle on peduncle; stipe 18 mm violet Ck; spadix greenish tinged with purple- to dark purple-violet (B & K Red-purple cm long, 5-7 mm diam. at base, 3- am. at apex; flowers rhombic to sub-4- $i Pig 8 mm long, 4.2-4.4 mm wide, the tini Le. 6-7 flowers visible in the principal legale flowers visible in the alternate spiral; "yu ; cde minutely papillate, the lateral tepals top Ei wide, the inner margins broadly $ > Pistil raised but not emergent; stigmas T6; stamens emerging in a slow progression lobed OF MEXICO AND MIDDLE AMERICA 343 from base, the alternate stamens following soon ft b d pistil; filaments whitish, translucent 0.5 mm long, 0.1 mm wide, flattened; anthers white to pale yellow, 0.5—0.6 mm long, 0.5-0.6 mm wide; the- cae narrowly ovate, slightly divaricate; pollen pale yellow to cream, fading to white. INFRUCTES- CENCE unknown. Figs. 134 and 141. Anthurium pedatoradiatum consists of two subspecies, both endemic to Mexico. The species occurs in southern Veracruz, Chiapas, and Ta- basco ranging in elevation from near sea level to 800 m. Most collections have been made above 500 m in a region of “selva alta perennifolia” north of Ocozocoautla, a region with extensive outcrops of limestone. The species is in section Schizoplacium and is distinguished by the pedatisect leaf blades with entire lobes, erect inflorescences that are shorter than the leaves, and the green, lanceolate spathe. In Mexico it is possible to confuse the species with some forms of A. podophyllum, which may only in northern Oax guished by having much thicker leaf blades. The subspecies pedatoradiatum is distin- guished by its terrestrial habit and differs from subspecies Aelleborifolium by having leaf seg- ments rather uniformly united at the base with no segments divided to the base of the leaflet. In subspecies Aelleborifolium, one or more of the segments is free or nearly free to the base. The plate accompanying the type description of A. pedatifidum shows a plant with one leaf blade divided in the manner characteristic of subspecies pedatoradiatum and another seem- ingly divided to the base, but nothing is men- tioned in the description about this range of vari- ation and it can only be assumed that the second leaf represents an error by the artist. The taxon that is here considered a synonym was supposed to differ from A. pedatoradiatum by a supposedly different geniculum and fewer number of blade segments ta, Mi M . TABASCO: Munc. Hui- manguillo, Gilly & Hernandez 103 (MICH). b. Anthurium pedatoradiatum ssp. helleborifoli- um (Schott) Croat, comb. nov. Anthurium > ys = > ws n o 7 - ie» tm = [77 vel OQ C a w e! aa » ^ A > E Q > ^x Jg m pA Fioures 132-135, 132-133. Anthurium pedatoradiatum ssp. helleborifolium (Schott) Croat, Croat 4008 1. —134. Anthurium pedatoradiatum Schott, Kew 5O04-69-04135.— V5, Anthurium pen "hyllum v bombacifolium (Schou) N ison, Croat 32910 à URES 136-139. 136. Anthurium pittieri Engl., Kamemoto 260.—137. Anthurium pittieri Engl., fi. Cullen nii iC roat, Croat 47134.—139. Anthurium pittieri var. fogdenii Croat, Croat 467 Croat 35981.—138. Anthurium pittieri var. VOINAWNYV TYIGGIN GNV OOIXJN JO WATYNHINY—LVOU) [£861 SpE ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo.. 70 FiGu JRES 140-143. 140. Anthurium paige 141. Anthurium pedatoradiatum Schott, (Hort. Par bombato (Schott) Madison, Folsom s.n.— 143. um ssp. helleborifolium (Schott) Croat, Croat ‘roat Neg. $500). —142. Anthurium pentapy Anthurium pluricostatum Croat & Baker, Croat 081. bes var. 3206. ) 1983] helliborifolium Schott, Bonplandia 10: 148. 1862. TvPE: Mexico. No specimen cited in original description (F, hololectotype, FM 29818; Photo of cultivated plant prepared by Schott here designated). Terrestrial; ca. 1 m or less tall; stems thick, 9— 25cmlong, 2-3 cm diam.; internodes very short; leaf scars 2 cm wide, usually obscured by cata- phylls; roots medium thick, descending; cata- phylls subcoriaceous, 2.5-5.3 cm long, some- times tinged reddish, drying medium brown (B & K Yellow 5/5), splitting at base, remaining intact at apex, persisting around stem. LEAVES With petioles erect, (27)48-56(63) cm long, 8-9 mm diam., terete; geniculum 1-1.7 cm long; blades, 7-11 pedatisect, moderately thick, broadly ovate in outline, deeply lobed at base, 27-33.5 em long, 18-48 cm wide, broadest at middle or Just below; the lobes free or nearly so to base, the outermost lobes confluent to a greater degree, sometimes falcate, acuminate at apex, the me- dian lobe 21-27 cm long, 5.7-7.9 cm wide, broadest just above middle, outermost lobes 12- 18 cm long, markedly inequilateral and usually auriculate; sinus narrowly to broadly spathulate to hippocrepiform, acute to rounded at apex; the upper surface glossy, lower surface semiglossy; the midrib of the median lobe acutely raised above, diminished and sunken at apex, raise below; primary lateral veins 3-4, departing mid- rib at 409.509 angle, sunken and nearly obscure above, raised below; basal veins 5-6 pairs, the rst and second free to base, the remaining co- alesced 1,5—4 cm, raised above and below; the Posterior ribs weakly curved, naked ca. one half their length: collective vein arising near the base lobe, or from one of the primary lateral veins hear middle, 1-5 mm from margin. INFLO- ( NCE erect, shorter than leaves; peduncle )20-30 cm long, 3-5 mm diam., terete; spathe ü Ny thin, green, lanceolate, 5.5-11 cm dx -6-1.5 cm wide, gradually acuminate at d X, rounded at base, inserted at ca. 45° angle s pr ne spadix green (B & K Yellow-green 3 ) 4-8 cm long, 4-6 mm diam. at base, 2- T at apex; the flowers rhombic to sub- Fog 2.5-3 mm long, 2.8-3.2 mm wide, the the Smoothly sigmoid; 3-5 flowers visible in Principal spiral, 4-6 flowers visible in the ione Spiral; tepals matte, densely papillate, aoc. mm wide, the inner margin > Pistils emergent, only slightly raised, CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 347 green, densely papillate; stigma linear, 4-5 mm long, dark slit with red-violet punctations; sta- mens emerging from the base, lateral stamens preceding alternates by 1-2 spirals, exserted on long, translucent filaments, 0.5—0.8 mm long, 0.9— 1 mm wide, retracting to hold stamens at edge of pistil; anthers yellow to tan, sometimes deep purple; thecae ellipsoid, somewhat divariate; pollen cream to yellow. INFRUCTESCENCE not seen. Figs. 132, 133, and 140. Anthurium pedatoradiatum ssp. helleborifoli- um is endemic to southern Mexico in Chiapas and Tabasco where it is common in “‘selva alta perennifolia” in the vicinity of Teapa, Tabasco. All collections seen were made at elevations of less than 300 m. Plants are terrestrial, commonly well rooted in the soil but generally on well drained soil, usually over porous limestone, commonly on rocky slopes. Anthurium pedatoradiatum ssp. helleborifoli- um differs from ssp. pedatoradiatum in having 1-3 of the medial blade segments free or nearly free to the base. In addition, those collections of ssp. helleborifolium grown side by side in the eenhouses with ssp. pedatoradiatum have con- sistently darker green, thicker leaves. Madison (1978) treated ssp. helleborifolium as synon- ymous with A. pedatoradiatum but I feel they are separable at least at the subspecific level. All of the material observed thus far has been easily separable on the key characters although some specimens with a poor state of preservation ap- peared to have free segments because of torn segments. A collection of Bunting and Davies (117) collected near Teapa was deemed inter- mediate by Bunting (1965) who reported that it differed in “having the outline of the former but with the central lobe free to the very base." I have found that the outline of the leaf blades of the species are very generally the same and the division of the leaf blade of the Bunting and Davies specimen is well within the range of vari- ation for the species. Since no type specimen was cited I have se- lected as a lectotype a collection prepared and annotated by H. C. Schott from a cultivated plant in Vienna. Although the specimen itself has been destroyed, the photographs of the specimen are Sa aay “1 T1 = 34 Y i lata #90010 WIUCLY Vio MEXICO. CHIAPAS: between Ixtapa and Teapa, Croat 47899 (MO), vicinity Pichucalco, Croat 40081 (MO). 348 TABASCO: N of Cerro de Azufre, Gilly & Hernandez 296 (MEXU, MICH); base of Cerro de Madrugal, Croat 47906 (MO); Tacotalpa, Cowan 2062 (MO); E of Te- apa, Croat 40130 (MO); Conrad & Gallegos 2869 (MO). Anthurium pentaphyllum (Aubl.) G. Don in Sweet, Hortus Brit. ed. 3: 633. 1839. Anthurium pentaphyllum var. bombacifolium (Schott) Madison, Selbyana 2: 278. 1978 Anthurium bombacifolium Schott, Oesterr. Bot. Z. 8: 182. 1858. Type: Costa Rica. Near Atenas, Hoffmann 779 (B, destroyed; Photo of Schott Aroid Drawing #829, NYBG Neg #N.S. 3934 Anthurni ] CAL onplandia 7: 165. 1859. bsc Mexico. Colipa, ln 15799 (C?, not in karwinskyi Schott, Oesterr. Bot. Z. 9: 101. 859. Type: Mexico. Veracruz: Colipa, Karwin- Sky ). Epiphyte; stems creeping up tree trunks; in- ternodes 2-9 cm long; cataphylls caducous or persisting as coarse fibers. LEAVES erect-spread- me petioles 19-44 cm long, eu 5 mm diam., I T lets, moderately t thin, leafs oblanceolate to nar- rowly elliptic, 13-30 cm long, 2.5-9 cm wide, the medial leaflet pas ya the lateral leaflets inequilateral, abruptly to cuspidate-acuminate at apex, cuneate-attenuate at base; midrib raised above and below; primary lateral veins 6-8 per side in each leaflet, sunken above, raised below lesser veins obscure above, raised below; collec- tive vein arising from base, 6-14 mm from mar- gins, + straight to apex of leaflet. INFLORES- CENCE erect; peduncle 1—6 cm long; spathe pale green to purple, lanceolate to ovate, moderately thick, 6-10 cm long, 1—2.8 cm wide, early de- ciduous; spadix pale purple-violet, 2. 2-10(15) cm long, 8-20 mm diam. at base, tapering at apex. INFRUCTESCENCE erect-spreading; berries red to dark red-violet, + globose, 5-8 m m diam.; seeds ovoid to oblong, ca. 5.3 mm GR and 4 mm wide. Figs. 135 and 142. Anthurium pentaphyllum var. bombacifolium is known from Mexico to Panama at elevations from near sea level to 800 m, occurring princi- pally on the Atlantic slope in Mexico and Middle America "M on both slopes i in Costa Rica and Panama. Th tropical moist, , premontane wet, and tropical w wel forest life zones, making it one of the most eco- logically variable species of Anthurium in Cen- tral America. eneri ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 The species is a member of section Dactylo- phyllium and is most easily confused with A. kunthii, which has similar leaves, but A. kunthii differs in having a longer peduncle, three-quar- ters the length of the petiole, and in having a long, slender, greenish spadix and lateral leaflets which are more or less equilateral at the base. BELIZE. BELIZE: Gracy Rock, Sibun River, Gentle 1741, ; Gentle 1491 (MICH, MO). cayo Camp 6, Gentle 2417 (MICH); Caves se of Mountain Cow, Whitefoord jose Cuevas, tius John's Hunting Camp, Dwyer 11560 (MO); Macaw River, road to Millionario, e ré Liesner 12343 (MO); Mountain Cow Ridge, Gentle 3553, 3526 (MICH); Vaca, Gentle 2551 (MICH); Val- MICH). STANN ke S key River, Gentle EE 4010 (MIC H). 65, Costa RICA. ALAJUELA: vicinity Bijagua, Cro at 362 36424 (MO); Fortuna, n 11615 (NY); beyond San Ramón, Luteyn 3384A (DU AE a of Zarcero, 43585 (MO). HEREDIA: Puerto V 6948 ). PUNTARENAS: Corcovado NE al Park, Liesner 2831 (MO); above Palmar No de 35124 (MO); between Palmar Sur and Piedras B iT " dras Croat 32910 (MO); S of Puerto Armuelles, Liesner O). pove 2i GUATEMALA. ALTA VERAPAZ: Finca Argentina ee Papalha, Croat 41565 (MO); road to mee cinity of Puerto Ba i H ; Cerro Cel no other location, Lundell 2990 (MIC ) Cer cal cam, Steyermark 461 hs "ET 81 tmd, Lundell 3054 (MICH); Tikal Bartlett a ; Uaxactün, Bartlett 12466 (MICH): p (MICH, is tz CHE ONDURAS. ATLÁNTIDA: nity ed et al. 8597 (MO); riis d "Chickering Tan Croat 42637 (MO); d 32 (MICH). MO. of N n& prier 6421 (U love 26341 (DS); N of El T (DUKE); NE of Escuintla, Croat 43824 Pe 980 (MEXU, MO); Matuda 16566 det UCLA); Mi 40235, 40268, 40339 (MO); vicinity of des Croat 40072, 40148 (MO); N of Tapachu?? OI = = rr i — e — M 1983] ee (MO); NW of Turquiz, Croat 47476 (MO); E re and Río del Corte, ERAT aeo s.n. (MEXU); Temas- cal, Sousa 1826 (M Valle Nacional, Croa t re & B XU); So m 9 (MO); NE im Valle deere Moo ing 16 (BH). SAN LUIS POTOSI: vicinit etnias c 39274 (MO); Rzedowski 10992 (CAS, MICH). zacoalcos, ede 8721 (MICH); vicinity of Martí- nez de La e and Misantla, Gomez-Pompa & Riba 201 (MEXU).- Moore & Bunting 8948 (BH); TUS Sarukhán & Martinez s.n. (M ENID PUEBLA: La Gra ja, panei & Martinez s.n. (MEXU). AGUA. ZELAYA: Caño Co ps Riquita, Stevens 5057 ? (MO; S of rias: at Colonia Agricola Yolaina, Stevens 4820, 6353 (MO); W of Rio Wawa Ferry, Ste vens 8646 (MO); near Rio adem Neill 3801 (MO); SW of Waspán, Atwood 3716 (M (MO). Anthurium pittieri Engl., Bot. Jahrb. Syst. 25: 372. 1898 a. Anthurium pittieri. var. pittieri. TYPE: Costa Rica. Heredia: Rancho Flores, Volcán Bar- ba, 2,505 m, Pittier & Durand 865 (BR, hol- olectotype, designated Croat & Baker, 1979). Epiphyte, sometimes somewhat scandent; Stems less than 30 cm long; internodes short, 1— 1.5 em diam., leaf scars conspicuous; roots few, descending; cataphylls subcoriaceous, (3)8- 12.5(14) cm long, apiculate at apex, drying dark làn, persisting intact, ultimately deciduous. AVES spreading; petioles 3-15 cm long, 4-7 mm diam., subterete, usually bluntly sulcate to bes fattened; geniculum 2-3 cm long; blades /ly ‘la 1 on elliptic ¢ or sklone, moderately thick, (8. 5)10 L43 em long, 1.5-9 cm wide, acuminate at apex, acute lo obtuse or rarely rounded at base; upper surface uid to semiglossy, lower surface semiglossy; ow i broad and convexly raised above and be- E iminished and sunken at apex above; pri- bud lateral veins 17-20 per side, + obscure i ; Prominulous to flat and + obscure below, lecti ing midrib at ca. 50° angle, straight to col- Ve vein; ceps vein arising near the base, au rm low, gin. INFLORESCENCE + Prod in ually shorter than leaves; peduncle (5)1 1- lan ng; spathe green, subcoriaceous, oblong- Pe me (3)7-15 cm long, 1.2-2.5 cm wide, p tpuminate at apex, decurrent onto pe- vus 1.5 cm, ipe to attenuate at base, in- &K Yel dd n peduncle; spadix green (B 3)7-8 m ovareen 7. 5), (2.2) 1214.5 cm long, m diam. at base, (2)4-5 mm diam. at CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 349 apex; flowers rhombic, 6—6.3 mm in both direc- tions, the sides straight to weakly sigmoid; 2-3 flowers visible in the principal spiral, 3-5 flowers visible in the alternate spiral; tepals matte, mi- nutely papillate, larger plants with tepals usually mounded giving spadix a knobby appearance, lateral tepals ca. ‘ mm wide, the inner margin straight; ^ae usually exserted, green, matte; stigma ellip ca. 0.6 mm long, brushlike, exserted in prc with a small droplet for 3 or 4 days ca. 2-3 weeks before anthers emerge, retract faece anthers at sides of pistil; anthers cream-colored ca. 0.5 mm long, ca. 1 mm wide; thecae tan, ellipsoid, not divaricate; pollen pale yellow fading to white, soon dispersing. IN- TESCENCE pendent; spadix 5-15 cm long; “ar yellow to orange in Panama or re- portedly white in Costa Rica, obovoid, acute at apex, 1-1.5 cm long, 7 mm wide; seeds 2, green, flattened. Figs. 136 and 137. The species is known from Costa Rica and Panama at 50 to 2,100 m elevation in tropical wet and premontane rain forest life zones. A specimen recently collected by Ernesto Ancuash (1230) under the auspices of Brent Berlin’s an- thropological studies in Peru (Dept. of Amazo- nas, Rio Cenepa) is apparently also this species. It would be the first collection of the species from South America. Anthurium pittieri is a member of section Oxy- carpium and is distinguished by its coriaceous, epunctate leaf blades, premature emergent ber- ries, and especially by the long-decurrent spathe. The species is variable in leaf blade shape and in the degree of emergence of the pistils. Smaller plants tend to have the tepals flat with the pistils not noticeably emergent while larger plants have the pistils somewhat protrudent with the tepals turned somewhat upward. Specimens with un- usual leaf shape include Antonio 2139 from Pan- ama with oblong leaves, 6.5 times longer than wide and Croat 35295 from Dominical (Puntare- nas, Costa Rica) with leaves almost as wide as long. It is most easily confused with A. interruptum and A. carnosum, both of which differ in usually having more elongate internodes, in lacking a — decurrent spathe, end in having either green is reportedly Libet to A. decurrens Poepp. from 350 Amazonian Peru but differs from that species in having a much shorter stipe. Costa Rica. No other location, Bogner 917 (M). dies CARTAGO: no other location, Madison 6968 (MO); La Carpintera, Allen 512 (MO); Brade 2514 (BR); Rodrigues 496 (UC); El adig Standley & Torres 51671 (US); between Moravia & Quebrada Platanillo, Croat 36646 (MO); vicinity lea Standley 39903 (US); S of Tapanti, Burger & Liesner 6724A pov Lent 904 (CAS); SE of €— Holm & Iltis 194 (NY). GUANACASTE: H Santa María, Dodge 1 al. 6366 (MO), Monteverde, Primack et al. 436 DUKE). HEREDIA AND SAN JOSE: Slopes of Cerro Zur- qui, Luteyn & Wilbur 4468 (DUKE, MO). HEREDIA: Las Lajas, vicinity San Isidro, Pittier 14057 (NY); N Oo jo 3694 (MICH, NY, US); 3704 (MICH); Utley & Utley 4978 (DUKE); slopes of Volcán Barba, A/meda 519 (DUKE); Burger & Stolze 6068 (NY). PUNTARENAS: Monteverde, Utley 150 (DUKE); S of San Vito de Java, Raven 21927 (CAS). SAN José: slopes of Cerro eire Utley 404 | (DUKE); la Palma, Luteyn 3303 Dominical, s roat 35925 (MO); San Josecito, Primack & L 352 (DUKE); Tablazo, Pittier 7990 (BR); Utley & Utley 3876 (DUKE). b. Anthurium pittieri var. fogdenii Croat, var. nov. Type: Costa Rica. Guanacaste: along the road between Santa Elena and Monte- verde, ca. 2 mi from Santa Elena-Monte- verde junction, elev. ca. 1,500 m, 10°20’ N, 84°47'W, Croat 47134 (MO-2769759, ho- lotype; CR, K, SEL, US, isotypes; Live at MO). Planta epiphytica, parva; caudex 3-6 cm ee gra- cilis; lamina ovata-elliptica aut ovata, 4.5-8 cm longa, 2.3-3.5 cm lata, subcoriacea, epunctata, nervis supra l uris; inflorescentia effusa- erecta, foliis voies a * ne 5 cm lo ongus; spatha viridis, 2-3 cm longa, 6-9 mm lata; spadix viri- dis, 1.5-2.2 cm tik Panda kaea baccae ignotae. ev viter impressis Epiphyte; stems 3-6 cm long, 0.7-1 cm diam.; roots green, pareen thin; cataphylls siibosid: aceous, 1.5-2 cm long, rounded at apex, drying dark tan, thin at base, persisting around stem. S spreading; petioles 1.3-7.3 cm long, 2- 3mm diam., terete, shallowly and sharply sulcate at base of blade; geniculum not apparent; blades ovate to ovate-elliptic, moderately thick, short- ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 acuminate at apex, round to obtuse at base, 4.5- cm long, 2.3—4.7 cm wide, broadest at middle; upper surface glossy, lower surface semiglossy; midrib raised at base above, sunken in apical one quarter of blade, convexly raised below; pri- mary lateral veins 5-10 per side, departing mid- rib at 50? angle, straight to collective vein, etched above, obscure below; lesser veins scarcely vis- ible; collective vein arising near the base or from a primary lateral vein in the basal one half of blade, 2-4 mm from the margin. INFLORES- CENCE erect-spreading, shorter than or as long as leaves; peduncle 3.5-1 1.5 cm long; spathegreen (B & K Yellow-green 7/10), subcoriaceous, €l- liptic to obovate, 2-5.2 cm long, 6-15 m m wide, decurrent on petiole 0.8—1.5 cm, obtuse at base, short acuminate at apex, the apex tightly in- rolled, the apiculum hooked toward spadix; stipe ca. 1 cm long; spadix green, 1.5-4 cm long, 3-4 mm diam. at base, 2-3 mm diam. at apex, flow- ers 4-lobed, ca. 2.5 mm long, 2.3 mm wide, the sides weakly to jaggedly sigmoid; ca. 3 flowers visible in the principal spiral, ca. 4 flowers visible in the alternate spiral; tepals matte, dene margin straight to convex; pistils weakly emer- gent; stigma minute, elliptic; stamens emerging from the base, lateral ‘stamens emen to mió- wav b ge at base; an- thers white, ca. 0.4 mm in both pub. held over pistil in tight cluster; thecae ellipsoi scarcely divaricate; pollen white. INFRU CENCE not seen. Figs. 138 and 139. Anthurium pittieri var. fogdenii is endemic T the vicinity of Monteverde (a biological y^ serve operated by the Tropical Science On San José), where it occurs with the typical be of A. pittieri. The variety fogdenii is known property near the Monteverde Preserve. distin- Anthurium pittieri var. fog denii is guished by its small size, green spadix wi - anthers, slender stems, and ovate to ovi tic, subcoriaceous, epunctate leaf blades E etched veins on the upper surface and wi scure veins beneath. The taxon differs from var. pittieri by i: the veins prominently etched on the upper t face. This character disappears on drying $° se in the dried condition it is not possible b 1983] CROAT arate the two taxa readily. Although A. pittieri is placed in section Oxycarpium, var. fogdenii looks much like members of section Porphyrochitoni- um, differing only in the absence of glandular punctations on the leaves. CosrA RICA. GUANACASTE: between Santa Elena and Monteverde, Croat 47134 (CR, K, MO, SEL, US); vicinity Monte Verde Preserve, Croat 46771 (MO); PUNTARENA LAJUELA: in and around Monte- verde Preserve, Burger & Baker 9722 (F). Anthurium pluricostatum Croat & Baker, Bre- nesia 16 (Supl. 1): 68. 1979. Type: Costa Rica. Limón: ca. 1 mile NE of Bribri, ca. 40 miles SW of Limón (on coastal road); steep ravine above Río Catarata, elev. 50-100 m, Croat 43243 (MO 2584480-81, holotype; R, F, K, US, isotypes; Live at MO). Terrestrial; stem 2.5-3 cm diam., to 23 cm long; leaf scars conspicuous, ca. 2.5 cm diam.; Toots greenish, ca. 4 mm diam., moderately nu- merous and long, descending; cataphylls mod- erately thin, 6-12 cm long, acute to acuminate at apex, soon turning brown and weathering into s dense mass of persistent, coarse fibers. LEAVES with petioles erect, 26-68 cm long, 6-8 mm diam., subterete, narrowly flattened to weakly and ob- solely sulcate adaxially, almost smooth to char- acteristically several-ribbed or -striate; genicu- lum 1.2-1.7 em long; blades narrowly ovate, moderately thin, acuminate at apex, deeply lobed at base, 40-68 cm long, 20-38 cm wide, broadest near the point of petiole attachment, the margins broadly convex: anterior lobe 32-52 cm long, Posterior lobes 10-19 cm long, directed down- ward or inward; the sinus spathulate, obtuse to rounded at apex: upper and lower surfaces semi- glossy; midrib obtusely raised at base above, be- coming acutely raised and then diminishing to- e apex, prominently raised below; basal veins Pairs, 3 or 4 of them coalesced 3-4 cm, cominently raised above and below; the pos- boc rib thick, naked, almost straight, the outer margin flat or weakly rolled upward; primary Mis raised below; interprimary veins usually he Scarcely raised above, flat below; secondary ks : Weakly sunken above, darker than surface uS at below; collective vein arising from the Posee basal vein, 3-5 mm from margin, d Y sunken above, slightly raised and darker Ow. INFLORESCENCE erect-spreading, OF MEXICO AND MIDDLE AMERICA 351 equalling or much shorter than the leaves; pe- duncle 17—43 cm long, 4-5 mm diam., terete to obscurely angular, sometimes heavily suffused with dark purple, 0.33—0.75 times as long as pet- -purple 7/7.5), oblong-lanceolate, 13-18.5 cm long, 2-3.7 cm wide, broadest in lower quar- ter, gradually long-acuminate at apex, rounded to obtuse at base, stiffly spreading, sometimes twisting, inserted at 50-70? angle on peduncle; stipe ca. 9 mm long in front and 3 mm long in back, ca. 6 mm diam.; spadix pale lavender or green tinged with pale lavender (B & K Red- purple 8/10), yellowish-green near the tip or sometimes pale reddish-purple throughout (B & K Red-purple 3/5), 8.5-14.5 cm long, 7-9 mm diam. at base, 3-4 mm diam. at apex; flowers rhombic to weakly 4-lobed, 2.6~3.5 mm long, 2.4-3.6 mm wide, the sides straight to smoothly sigmoid, rarely jaggedly sigmoid; 7-9 flowers vis- ible in the principal spiral, (8)10-15 flowers vis- ible in alternate spiral; tepals semiglossy, papil- ate, with a few droplets wl th pening, the lateral tepals 1.3-1.8 mm wide, 3—5-sided, the inner margin straight or nearly so; pistils pale green becoming salmon-lavender (as tepals) with purplish dots, slightly mounded and weakly exserted above the tepals; stigma linear, 0.6—0.8 mm long, a pale, darker-edged slit, exuding drop- lets about 1 week before first stamens emerge, the nectar persisting even after all stamens are open; stamens emerging rapidly from the base, the lateral stamens emerging throughout before the alternate stamens begin to emerge at base, held just above the tepals in a circular cluster ca. 1.3 mm diam. (or s ti alf emergent with the exposed part ephemeral); anthers white, 0.5-0.6 mm long, 0.9-1 mm wide, peeping out and opening while still small or the thecae stand- ing on end, opening vertically and disintegrating when touched; thecae obovoid, slightly divari- cate; pollen white. INFRUCTESCENCE with purplish berries, acute at apex. Figs. 143 and 152. — ] on yh The species is known from Costa Rica, and in Panama in Chiriqui Province on the Burica Pen- insula. It is certainly to be expected in Bocas del Toro Province as well. It occurs in a wide variety of life zones in Costa Rica, ranging from pre- montane wet transition forest to premontane wet and tropical wet forest at elevations from near ANNALS OF THE MISSOURI BOTANICAL GARDEN hur- gy Na (Cham. & Schlecht. rap Hiwan 263.— 145. A & ium podophyllum (Cham. & Schl Kunth (Cultivated at Antosh’ s).— Anthurium prolatum Croa Baker, Croat 47087.—147. An rola in Croat & Baker, Croat 46 S 144-147. 144. erigi -) f { 1983] 150 FIGURES 148-151. roa & Fen : ^s ` at 35566. — 150. Anthurium purpureospathum Croat, Croat 3 33984. ex , K. Krause, ( roat 3: CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 353 oat 47113.—149. Anthurium protensum Schott, rotensum Schott, Cr h Meise 8094 (Type).— 151. Anthurium ramonense Engl. > z Z > P an O E - ie» m = un o le C & w A o! "1 » A 0 > a Q > zx g m z 154 Fissures 152-155. 152. ——— cgo hector po iur & Baker, Croat 43205.—153. idin purpureospathum Croat, Croat 38094. — 154. E* at 433404.—Y55. Anthurium ranchoanum En . Croat 48519A. Anthurium ramonense Engi. ex 1983] sea level to 1,000 m on both the Pacific and Atlantic slopes. Fes ed "disks pei hd isa member of sec- ] by its ter- NA habit, subterete petiole with 7-10 low ridges, and spadix that is lavender to reddish- purple or green tinged with lavender, usually with a discolorous yellowish-green portion near the apex. The species is also unusual in having flow- ers with stigmatic droplets persisting while sta- mens emerge. Anthurium pluricostatum is related to A. ochranthum. The latter species, although occur- à consistently terrestrial habit that is a diagnostic pia for species occurring at medium to low acter being of much less value in phi forest regionó). Staminal dehiscence is somewhat variable. Usually the anthers are fully exserted although the filaments are not exposed. Sometimes (even in the same population) anthers are only about half emergent and the exposed portion of the thecae loses its form when the pollen matures. The species displays an unusual pollination Possibility since the stigmas produce copious droplets EE before the stamens emerge, the us o open RE Speci leci 1 sis only we d might be tds St be considered peas The behavior Presents a clear possibility for Md fertilization. along Río Mu E of Bratsi between Limón & Shi- Bamb MO); near Río Sixaola, SW o u, NE of Bratsi, Croat 43265 (MO); NE of Bribri, F, pos món, above Ri MO, US); eid Río Catara n o 2 & m 4 ES] i A wy N e "8 aS w SE o [o E. 9 2 o 5 Land -g e ; le Domi n San Isidro del General and Mical, Croat 35269, 35324 (MO). Anthurium podophyllum (Cham. & Schlect) Kunth, Enum. PI. 3: 80. 1841. Pothos podo- CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 355 phyllum Cham. & Schlecht., Linnaea 6: 22. 1831. TYPE: Mexico. or Hacienda de la Laguna, Schiede & D B; not seen, illustistedd: by Schott dM Vend 2791, W). ns & Galeotti, Bull. Acad. y. Sci. Bruxelles 1o 122. 1843. Type: Mex- , ico. i Ver eracruz > Zac cualpan, semi 6055 pa dia 7: 337. 1859 Type: Mexico. (Photo ay on Aroid | Drawing #805, NYBG Neg. #N.S. 3924; Photo Anthurium i ge awe ea nell Schott, pees d: 338. YPE: neater ee in Vienna signe of Schott ‘roi Drawing #811, NYBG Neg. £N.S. Pothos Nepos M. Mart 3926; Anthurium T Linden ex Schott, Prod. Aroid. 1860. TYPE: none listed. Anthurium membranuliferum puse & Ohlend. ex , Monogr. Phan. 2: 196. 1879. (Known only in a ag ) Usually terrestrial or epipetric, often more than 1 m tall; stems 2.5-3.5 cm diam.; roots descend- ing; cataphylls, coriaceous, 7.5-1 1 cm long, grad- ually acuminate at apex, drying medium brown, persisting as linear fibers. LEAVES with petioles erect-spreading, 39-69 cm long, ca. 9 mm diam., D-shaped; geniculum 5 cm long; blades subcoria- ceous, broadly ovate, deeply to broadly lobed at base, (3)5-10 pedatisect, the lobes usually 5-12 pinnatifid, fused 8—10 cm at base, 16—47 cm long; the outermost lobes 7.5-27 cm long, auriculate; sinus hippocrepiform to nearly arcuate, rounded at apex; upper surface glossy, lower surface semi- glossy; midrib of median lobe raised, rounded, diminished and flat at apex above, raised below; primary lateral veins 2—3 per side, ne midrib at 40° angle, sunken above, raised below basal veins 3—5 pairs, second and third toulesced to 2 cm, the third and fourth coalesced 3.2 cm; the posterior ribs straight; collective vein arising — the first — lateral vein, 8-9 mm from margin, sun above, raised below. IN- FLORESCENCE sreading-pendent eer than leaves; penduncle 38-100 cm long, to 6 mm diam., terete, maie to or longer than petioles; spathe subcoriaceous, yellow-green (B & K Yel- ole s 5), lanceolate, 4. 5-9 cm long, 1.6- 2.2 cm t acuminate at apex, ded at 45° angle on peduncle; nue 9 mm long in front, 5 mm long in back, 4 m diam.; spadix green, sometimes tinged with ig ple (B & K Yellow-green 6/5), 4—14 cm long, 7- 11 mm diam. at base, 3-5 mm diam. at apex; flowers rhombic, 3.2-4.6 mm long, 2.5-4.6 mm wide, the sides straight to weakly sigmoid; 6-8 356 flowers visible in the principal spiral, 7—9 flowers visible in the alternate spiral; tepals matte, mi- nutely papillate, lateral tepals 1.2-2.2 mm wide, the inner margins broadly rounded, turned up against pistils; pistil emergent, green; stigma lin- ear, 0.6 mm long, dry, brown, crusty before sta- mens emerge; stamens emerging in a moderately slow sequence from base, exserted above pistil, then withdrawing, the leading stamens in any phase only a few spirals ahead of succeeding sta- mens; filament green; anthers white, held over pistil in tight cluster, ca. 0.6 mm in both direc- tions; thecae narrowly ovoid, dark brown, slight- ly divaricate; pollen cream-colored. INFRUC- SCENCE pendent, spathe persisting; spadix to 4.5 cm diam.; berries dull orange (B & K Yel- low-red 7/5), obovoid to obovoid-ellipsoid, rounded at apex, 13-19 mm long, 7-15 mm diam.; mesocarp thick, juicy, orange with small raphide cells; seeds 1 or 2, pale green, flattened, oblong to oblong-obovoid, 9-10 mm long, 5-6 mm wide, 3.3-4.3 mm thick (sometimes with large al t subterete seed) with appendages at both ends, the lower appendage flattened, about as broad as the seed. Figs. 144 and 145. Anthurium podophyllum is endemic to the At- lantic coast of Mexico in Veracruz and Oaxaca atelevations of 750 to 2,000 m in *selva mediana subperennifolia." The Oaxaca populations oc- cur at higher elevations (up to ca. 2,000 m), and have leaf lobes mostly entire to sinuate, rarely lobed as in pre-adult plants of A. podo- phyllum from Veracruz. Veracruz populations usually occur below 1,000 m and have more fine- ly dissected leaves with the lobes pinnatifid. Anthurium podophyllum is in section Schizo- placium and is characterized by its subcoria- ceous, pedatisect leaf blades with the lobes usu- ally sinuate to pinnatifid, and by its long inflorescence usually equalling or exceeding the length ofthe leaves. The Oaxaca populations have leaves similar in shape to A. pedatoradiatum, but the latter is recognized by its much thinner leaf blades, lobes with uniform margins, and gener- ally shorter peduncles. The species is rare in the wild, having been collected in only a few relatively adjacent local- ities since the type collection was made in Za- cualpan. Mexico. No other location, Karwinski s.n. (F); Mar- tens 137 (BR). OAXACA: between Teotitlán del Camino & Santa María Chilchotla, Croat 48367 (MO). vera- CRUZ: Barranca de Santa María, Moore & Bunting 8852 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 (BH); Jalapa, Sousa & Ramos 4744 (MEXU); Mirador, Liebmann s.n. (P); Municipio Totutla, Ventura 3660 (ENCB); Zacualpan, Galeotti 6013 (BR); Purpus 6364 (MO, NY, UC, US); Purpus 11130 (F, MO). Anthurium prolatum Croat & Baker, Brenesia 16 (Supl. 1): 72. 1979. Type: Costa Rica. Here- dia: Istaru Farm, Tirimbina, Sarapiqui; pri- mary forest; 220 m elev., Sept. 1971, Roy Lent 2140 (2 sheets) (F-1724125, holotype). Epiphyte; stems 2-2.5 cm diam.; cataphylls thick, 30-36 cm long, subpersistent, acuminate at apex, drying tan, weathering into fibers at the base. LEAVES pendent; petioles 30-35 cm long, —8 mm diam., terete, sulcate; geniculum 1-1.2 cm long; blades strap-shaped, gradually long- acuminate at apex, narrowly acute at base, 108- 142 cm long, 8-10 cm wide, broadest near the middle; p i ly raised above to v€ near the apex, raised below; primary lateral veins numerous, departing midrib at ca. 55° angle, scarcely or not at all distinguishable from the interprimary veins on upper surface, only slightly more prominent on lower surface; interpriman veins usually numerous; collective vein arising from near the base, 3-4 mm from margin throughout. INFLORESCENCE pendent, duncle to 39 cm long, ca. 4 mm diam., slightly longer than petioles; spathe medium thick, pale . : lo green, spreading, lanceolate-linear, to 29 cm ; " uminaté ON A. ae J5 peduncle; spadix brown (fide label), to ' long, 7-8 mm diam. at base, ca. 4 mm di 6 flowers visible in the principal s , ers visible in the alternate spiral; tepals $ the lateral tepals 2-2.5 mm wide, the inner maf- m long, least partly retracted on drying). INFR CENCE pendent; berries probably oran 146 and 147. ge. Figs: i from Anthurium prolatum is known only > tree) ah sae men mi es diris sii ica d j and H | northern Costa Rica in Alajuela F 00m Provinces in tropical wet forest at 200 elevation. The species is a member of the secti neurium and can be distinguished by e very long and narrow leaves, cataphylls, e unusual: spadi*. on Pachi CROAT KVALI 1983] and spathe. It is most easily confused with A. protensum, however, A. prolatum differs from that species by having its spadix more than twice as long as any observed for even the largest A. protensum, a spathe considerably more narrow and lacking the purple markings characteristic of A. protensum, and the flowers twice as long (0.5— 2.5 mm long in A. protensum, 3.4-4.5 mm long in A. prolatum). In addition, the common A. protensum has not been collected below 1,000 m and is most common at 1,800 m elevation. Costa Rica. eu along Hwy. 15, between Na- ranjo and Aguas Zarcas, NE of Villa Quesada, Croat 46979(MO). aides E ofCachí, Croat 47087 (MO). m Istaru Farm, Tirimbina, Sarapiquí, Lent 2140 Anthurium protensum Schott, Oesterr. Bot. Z. 8: 81. 1858 Anthurium protensum. ssp. protensum. TYPE: Costa Rica. Heredia: Volcan Barba, 2,500- 2,700 m, Oersted 1992 (type not found; Photo of Schott Aroid Drawing #427, NYBG Neg. #N.S. 3848). Ecco Cre Engl., Pflanzenr. IV 23B: 68. 1905. Typ ta Rica. Cartago: Las Vueltas, Tucur- rive, 635 m, Tonduz 12818 (B, holotype; P, US, Sotypes Anthurium costaricense Engl., Bot. Jahrb. Syst. 25: 398. 1898. Type: Costa Rica. Above Agua Gee cite 1,800 m qe nn s.n. (Drawing of Engler Ara 274 at P). Epi TM often pendent; stems short; cata- Ris 12-22 cm long, drying tan, weathering into fibers at base and persisting. LEAVES reading pendent petioles 11—35 cm long, 5-7 m diam., + terete, sometimes flat or weakly a adaxially: geniculum 1.5—2 cm long; blades : P-shaped 32-80 cm long, long-acuminate at has. acute to rounded at base; upper surface á * to semiglossy, lower surface semiglossy; e midrib acutely raised above, convexly raised ne-Pendent, shorter than leaves; petioles 15-45 late ong, 6-7 mm diam.; ; spathe oblong-lanceo- :Yellow-green to ois. usually at least tinged all Mah 10-15 cm long, 2.5—4 cm wide, usu- 2 pes + erect; spadix pale violet-purple, 8- diam, ng, 6-9 mm diam. at base, 2-4 mm 153 T apex; the flowers rhombic to 4-lobed, mm long, 2-2.2 mm wide; 4-9 flowers OF MEXICO AND MIDDLE AMERICA 357 visible in the principal spiral, 6-12 flowers vis- ible in the alternate spiral; tepals matte; pistils green, emergent; stigmas with droplets 2-4 days before stamens emerge; stamens emerging from weakly flattened, ca. 6 mm long, 4.5 mm wide, 3 mm thick. Figs. 148 and 149. The species is found in Costa Rica and Panama from 635 to 2,700 m, principally in premontane rain and lower montane rain forest. Anthurium protensum is a member of section Pachyneurium and is distinguished by its elon- gate, narrow leaf blades with many primary lat- eral veins, by its more or less terete petiole, and by its long, ovate to lanceolate spathe, pale vi- — to aaa or greenish spadix and orange berri The species is closely related to A. prolatum, A. brenesii, and A. seibertii, all species with sim- dur elongate leaves. See the commentary follow- ing these species for distinguishing characteris- tics. Anthurium protensum consist of two subspe- cies. The subspecies protensum occurs only in Costa Rica, while ssp. arcuatum Croat (ined.) occurs in Panama, ranging from Chiriquí to Ve- raguas at 1,300 to 1,750 m in premontane rain forest. The latter differs in being an erect plant with smaller leaves with fewer, more prominent sunken veins, an acute lower midrib, an arching inflorescence with shorter, usually ovate spathe, and a smaller spadix. COSTA RICA. ALAJUELA & ee Monteverde Nature Preserve, rte et al. 10792 (F). ALAJUELA: 8 (NY); vicinity Zarcero, Croat MO). CARTAGO: Car- pintera, Brade 25 greja, Wil- liams et a 182 ( E of Pacayas, Luteyn 3250 (DUKE); SE ra Platanillo, Croat 36714 (MO). GUAN- a, Greenman & Greenman 5465 (MO); (NY); betw Santa Elena ACAS E Pa rs 15 an Max 3706 (MICH, US). PUNTARENAS: vicinity Jalaca Farm Golfo Dulce Area, Allen 5211 (US); Monteverde, Ken- 358 nedy 605 (US); meee "ee (DUKE); SE of Santa Elena, oe et al. 203 Aserri arbaca, Standley 2 (US); Bajo La donde, Lent 1446 (WIS) ae 36334 (US); La Palma, Luteyn 3301 ( Piedra, Luteyn (DUKE); vicinity Vito San Cristóbal Sur, Utley & Utley 3059 (MO). Anthurium purpureospathum Croat, Selbyana 5 . 1981. TYPE: Panama. Bocas del Toro: Station Milla 7.5 on Changuinola-Al- mirante Railroad, less than 100 m above sea level, along slender ridge to WW II com- munications facility, ca. 2 km NW of ruins of U.S. Army military barracks, Croat 38094 (MO-2388335, holotype; CAS, F, K, MO, PMA, SEL, US; isotypes; Live at MO). E Epiphyt te or terrestrial at higher hidden by roots; roots dense, 3-4 mm likens: descending; cataphylls coriaceous, 7-14 cm lonk te a reticulate fibers. LEAVES erect to _ pet- ioles 12-31 cm long, 4-10 mm diam., + quad- rangular, flat to broadly and sharply sedeate: usu- ally prominently 3-ribbed abaxially; geniculum 1-2.5 cm long; blades oblong-oblanceolate to ob- long-elliptic, coriaceous, acuminate at apex, ob- tuse to acute or attenuate at base, E cm long, 7.5—-40 cm wide, broadest at or al middle; upper surface matte to semiglossy, swe surface glossy to semiglossy; midrib acutely raised above and below, diminishing at apex above; primary lateral veins (6)1 1-14 per side, departing midrib at 45° angle, raised in valleys above, raised be- low, giana loop-connected in upper half; interprimary veins flat or prominulous above, uiaiia below; secondary veins flat above and below; collective vein arising from one of the primary lateral veins near the middle, 4-8(13) mm from the margin, sunken above, raised below. INFLORESCENCE erect; pedun- cles 3-38 cm long, 0.8-2 cm diam., purplish or green, subterete, equal to or as long as petioles; spathe medium thick, purple to purplish with green mottling especially near base at center or green mottled with purple (B & K Red-purple 2/ 5) at anthesis, oblong to narrowly ovate, 4.5- 17.5 cm long, 1.5-4 cm wide, broadest į Just above the base, cuspidate at apex, obtuse to rounded at base; spadix violet-purple (B & K Purple 2/ 7.5), (4)5-22 cm long, 1-2.5 cm diam. at base, 4-8 mm diam. at apex; flowers rhombic to weak- ly 4-lobed, 2-2.3 mm long, 2.5-3 mm wide, the sides gradually to jaggedly sigmoid; ca. 14 flowers ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 visible in the principal spiral, ca. 7 flowers visible in the alternate spiral; tepals matte, with small scattered droplets, lateral tepals 1.4-1.6 mm wide, the inner margin flat against pistil; pistils slightly raised, purple becoming green and finally orange; stigma linear, ca. 5 mm long, with droplets for 4—10 days, drying with exserted papillae 4-6 days before first stamen emerges; stamens exserted ca. | 1 mm, emerging in a slow progression from the | base, the lowermost flowers with their full com- plement before the middle flowers have any sta- mens exserted, held in a tight, contiguous circle; filaments 0.2-1 mm long, flattened, transparent, tapered to anther, retracting to hold an nther at edge of tepals; anthers ca. 0.7 mm in both di- rections; thecae ellipsoid, usually held kpe o ae d 5), drying white. INFRUCTESCENCE erect to spreading; the spathe drying, usually persist- ing; spadix 9-19 cm long, to 4 cm diam.; ries narrowly ovoid, acute at apex with a round, mi- nute, mammilliform style, orange, red-orange to red (B & K Red 6/7.5), to 10 mm long; the me- socarp pasty, transparent; seeds | or 2, ca. 4mm long and 2 mm wide, pale greenish-yellow. Figs. 150 and 153. The species is endemic to Panama and occurs principally from near sea level to 200 m in E del Toro Province in wetter parts of tropical oe forest and premontane wet forest. It is to be è premontane rain forest at 1,200 to 1,500 m ap pears to differ in no way from the pop" vs at lower elevations except that it has qi lea and consistently terrestrial habit. The terres " habit is not unusual for populations at TR elevations where a greater accumulation p bris exists and where the steeper slopes eq better drainage. The species should be ex at medium elevations as well. Anthurium purpureospathum is pe s by its rosulate habit, more or less 0b leaf blades, almost coalesced 1.5—2.1 cm; posterior ribs nak midrib mary lateral veins 2—4 per side, air ective at 50° angle, raised above and below bi vein arising from first basal vein, 8-10 m 5 margin. INFLORESCENCE longer than — peduncle terete, 27-61 cm long: spa lanceolate, 2.5-8.5 cm long, 5-12 ™ broadest below middle, acuminate at es ing at base; spadix green, 2-5 cm long, fio diam. at base, 2 mm diam. at oe aw sides | rhombic, 4 mm long, 2.8 mm wide, ; spiral straight; 4 flowers visible in the principa e, 5 flowers visible in the alternate SP j | | | | l | 1983] Fici ‘ "lone grense > lum rione ae ns CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA —X rue ars age Íe thee ta M Ims Noe mapie, ep le "PW Ae LI 163 £e 160 ebria deeds tr gne Matuda, Mac Dougall ie Shp te Shee 163 tuda, isi Dougall 4 Type).— Anthurium rzedowskii Croat, Croat P^ se Matuda, Mac velis 459 ye Anthurium inth 4- 366 ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 punctate, lateral tepals 1.5 mm wide, inner mar- gin convex, turned up; pistil emergent, green; stigma slitlike, 7 mm long; stamens held over pistil; anthers cream, ovoid, 3 mm long, 4 mm wide; thecae ellipsoid, scarcely divaricate. IN- FRUCTESCENCE not seen. Figs. 161 and 163. Anthurium ri. 2 is known only from the Rio Negro and Rio Mono Blanco in the vicinity of the Oaxaca and Chiapas frontiers. Neither lo- cation has been located so the elevation is un- certain, but most elevations along the western Chiapas border range from 200 to 1,000 m, and only Cerro Baul on the Oaxaca border, southwest of Cintalapa, is very high, extending up to about 1,800 m The species is a member of the section Belo- lonchium and is recognized by its thin, broadly ovate leaves with short, linear raphide cells vis- ible on the upper surface on drying and by its slender tapered, greenish spadix with the tepals thin and clear along the inner margins and oth- erwise distinctly punctate with raphide cells. Anthurium rionegrense is perhaps most easily confused with two other thin leaved, basically ovate-cordate species, A. ovandense and A. le- zamae. Anthurium ovandense differs in having the anterior lobe usually concave along the mar- gins whereas A. lezamae differs in being more narrowly ovate and in having an alveolate epi- dermal pattern on drying. MEXICO. OAXACA: Rio Negro, between Chiapas and Oaxaca, MacDougall 459 (MEXU); Río Mono Blanco, MacDougall 7940 (MEXU). Anthurium rzedowskii Croat, sp. nov. Type: ico. G Croat 45520 (MO-2690028-31, holotype; Live at MO). Planta epiphytica aut rekes caudex ad 30 cm longus, 2-3 cm diam.; lamina ovata aut ovata-trian- Mesi 24-64 cm longa, 15-46. cm lata, basi cordata reacea + ; PUPA M , Saepe nerv collectivo secondario; nlrescenia effusa-erecta, fo- liis breviora; spatha viridis, 3.5-13 cm longa, 1.6-6 cm lata; spadix viridis, 3. 8-13 cm longus; baccae im- maturae purpureae Epiphytic or plebem, stems green, less than 30 cm long, 2-3 cm diam.; leaf scars 1.5 cm wide; roots 5-6 mm diam., M green to tan, descend- ing; cataphylls subcoriaceous, 6-10 cm long, pale green, acute at apex with slender subapical apicu- lum to 3 mm long, drying dark brown, the apex remaining intact, splitting at base and persisting. LEAVES with petioles erect to spreading, subte- rete or flattened adaxially, rarely with single abaxial rib; geniculum 2-2.5 cm long; blades ovate to ovate-triangular, acuminate at apex shallowly to deeply lobed at base, 24-64 cm long, 15-46 cm wide, broadest below middle or near point of petiole attachment; anter lobe 16-49 cm long, the margins convex; posterior lobes 1- 20 cm long; sinus arcuate to arabai pes with decurrent petiole; upper surface semiglossy, lower surface matte to semiglossy; the midrib flat to weakly, convexly raised above, narrowing and sunken at apex, prominently raised below; basal veins 3-7 pairs, the first free, the remainder co- alesced 1-3.5 cm, the fourth to sixth sometimes coalesced up to 6 cm, raised above, prominulous elow; primary lateral veins 7-17 per side, de- parting midrib at 352-65? angle, weakly raised in valleys or sunken above, raised below, straight to weakly curving to collective vein; interpri- rom the first basal vein, frequently with a sec ondary collective vein 0.4-1 mm m from the me gin, arising from the second basal vein and €x tending irregularly to apex or running to poc in apical quarter of blade, sunken above, ral below. INFLORESCENCE erect- spresti shorter than leaves; peduncle 19-56 cm long: : mm diam., terete; spathe moderately thic at base, inserted at 80°-90° angle on s pedunde stipe 9 mm long in front, 5 mm long i 38 spadix green (B & K eu 2 7/1.5). 13 cm long, 8-10 mm diam. at base, diam. at apex; the flowers rhombic to sulb-4-lobed ca. 2.3 mm long, ca. 2.9 mm wide, the "owes obscure, straight to weakly sigmoid; 6-7 visible in the principal spiral, 8-9 flowe in the alternate spiral; tepals matte, de ms minutely papillate, small droplets 0n te stamens emerge, lateral tepals ca. 2. | mm the inner margin broadly rounded; pistils emergent before stamens emerge, ciate : on nce emerging Bou the base in a rapid seque owed lateral stamens emerging first, quic as by alternates; the anthers creamy whi 7 | ) à j CROAT RAKVA I1 1983] not opening until the third or fourth stamens emerge, exserted briefly on translucent filament, held inward over pistil, ca. 1.1 mm long, ca. 1.2 mm wide; thecae oblong-ovoid, slightly divari- cate; pollen creamy white, fading to white. IN- FRUCTESCENCE pendent; spathe persisting; immature berries purplish, round at apex, ma- ture berries not seen. Figs. 162 and 164. Anthurium rzedowskii is known only from Mexico in the states of Guerrero, Oaxaca, and Chiapas at elevations of 1,400 to 2,900 m in “bosque pino-encino.” It is named in honor of Dr. J. Rzedowski of the Instituto Polytechnica in Mexico City, one of the first collectors of the species The peties is a member oF section Belolon- chium ovate blades with the collective cree usually arising from the first basal vein or sometimes arising from one of the primary lateral veins and fre- quently with a secondary collective vein extend- ing irregularly ni the margin to the apex or almost to the a tis E. font similar to A. subcordatum in having coriaceous leaf blades with the collec- tive vein arising from the first basal vein or one of the primary lateral veins, and in frequently having a secondary collective vein arising from the second basal vein and extending along the Margin to the apex or almost to the apex. An- thurium subcordatum differs in having truncate lo subcordate leaf blades, a violet-purple spadix; iorum in Guatemala, El Salvador, and Hon- uras, á Psi ICO. CHIAPAS: Municipio of Siltepec, Breedlove ON D 31867 (DS); Croat 47367 (MO); Matuda yi (DS, MEXU). GUERRERO: between Milpillas and Oyac de ‘Alvarez, Croat 45520 (MO); Municipio of MICH); Omil- 46056 AXACA: between Oaxaca and Pochutla, Croat 056, 46188 (MO). Anthurium salvadorense Croat, Selbyana 5 (3- 4): 333. 1981. Type: El Salvador. Ahuacha- Pan: 1.5 miles SW of Tacuba on road to San pe Menéndez; riverbank, Rió Cho- 700 m elev. Croat 42169 (MO- 2583750, holotype; CAS, F, ITIC, K, SEL, US, isotypes; Live at MO). e Terrestrial or rupicolous; stems 2-3 cm diam., Sually well rooted: roots to 1 cm thick, descend- OF MEXICO AND MIDDLE AMERICA 367 ing; cataphylls moderately thick, 5-8 cm long, mm wide, broadly am, to slightly 3- ribbed abaxially, broadly and sha medial rib; geniculum 1-1.5 cm long; blades elliptic to ob- lanceolate, medium thick, gradually acuminate at apex, turned downward (the acumen ca. 1.5 cm long), obtuse to rounded and auriculate at base, 25-50 cm long, 8-26 cm wide, broadest at or above the middle; upper surface matte to semiglossy, lower surface matte; midrib obtusely raised above, diminishing and sunken at apex, convexly raised below, ribbed near base; primary lateral veins 8-14 per side, departing midrib at ca. 45? angle, raised at midrib, diminishing and sunken near margin above, prominently raised rply sulcate, with a m striate, 1-ribbed adaxially and sometimes abax- ially, longer than petioles; spathe ovate-lanceo- late, medium thick, pale green (B & K Yellow- green 8/10), held erect, enshrining and wrapped around base of spadix, 4.5—6.5 cm long, 2.4-2.7 cm wide, broadest just above base, narrowly acute, inserted at 10-25? angle on peduncle; spa- dix pale green (B & K Green 8/5), 2.5-6.5 cm long, 8-12 mm diam. at base, 4-7 mm diam. at apex, held at a slight angle from the pedun- cle; flowers + irregularly 4-lobed, 2-3.3 mm long, 2.5-3.6 mm wide, the sides jaggedly sig- moid; ca. 10 flowers visible in the principal spi- ral, ca. 9 flowers visible in the alternate spiral; tepals matte, densely papillate, e tepals 1.4— 1.7 mm wide, turned up against emerging pistil; pistils weakly emergent, green; de stigma ellip- tic, ca. 0.4 mm long, erect and brushlike, with a small droplet apparent 2—4 days before first sta- mens emerge; stamens emerging rapidly in a complete progression from the base, lateral sta- mens emerging at apex when the third and fourth stamens are emerging at the base; anthers pale green, 0.8-1.2 mm long, 0.7-1.3 mm wide, held in a close, contiguous circle around and over the stigma, held at about the same height as the style; thecae weakly divaricate; pollen pale yellow, soon fading white. INFRUCTESCENCE spreading; spathe persisting and erect; the spadix to 9 cm 368 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo. 70 Cro oat. nsa | C. 45128 as : e deg A zedowskii Croat, Croat 47367.—165. Anthurium salvado roat ype).— Inthurium salviniae Hemsl. 7 i 167. Ant uriun viniae Hemsl., Croat 36480. (CUVA st Scl Serani T 1983] long, ca. 2 cm diam.; berries broadly obovoid, 5.5-9 mm long, 4-9 mm wide, rounded to flat or weakly indented at apex, orange-red (B & K Yellow-red 5/2.5 to Red 4/10); pericarp moder- ately thick, with dense punctiform raphide cells, especially in apical half; mesocarp pulpy or mealy with numerous raphide cells; seeds 1 or 2, ob- ovoid to broadly obovoid, yellowish-white to tan, -6.5 mm long and 5-5.8 mm wide, 3 mm thick, densely covered with punctiform raphide cells. Figs. 165 and 168. Anthurium salvadorense is known only from El Salvador and Guatemala at 500 to 800 m elevation. It was first collected by Sisto Alberto Padilla in 1922 in the Department of Ahuacapán in El Salvador. The species was later collected by Paul C. Standley in the adjacent Department of Jutiapa in Guatemala. The species is a typical member of section Pachyneurium and the earlier collections were confused with 4. schlechtendalii, to which it is related. Anthurium lua, Pes Bde ey g NC by being usually a well rooted plant with fewer, v large fleshy roots and a usually almost el- liptic blade, but especially by its inflorescence, bois has a pale green, ovate-lanceolate spathe that is weakly convolute at the base and erect, enshrining the usually short, pale green spadix. Anthurium schlechtendalii differs in having usu- ally oblanceolate blades, a coriaceous, more lan- colate, prominently reflexed, usually purplish SPathe, and a more elongate, purplish spadix. SALVADOR. AHUACHAPAN: no other locality, Padi- néndez fed El Imposible, vicinity San Francisco Me- (Cag s. Pernhardt s.n. (MO); Croat 42092 (MO), 42169 $ F, ITIC, K, MO, SEL, US). ALA. JUTIAPA: vicinity Jutiapa, Standley (OM US); SW of San Cristóbal, Dunn et al. 23222 (F, uw ^ road to Finca Agua Fría, Croat 41879 Anthurium salviniae Hemsl., Diagn. Pl. Nov. Mexic. 36. 1878. Type: Guatemala. Volcán de Fuego, Salvin s.n. (K). Anthurium giganteum Matuda, Madrofio 10: 169. 1950, Don Engl. 1898. Anthurium enormispadix Ma- r^v Revista Soc. Mex. Hist. Nat. 11: 94. 1950. D Mexico. Chiapas: Escuintla, Salto de Agua, atuda 18043 (MEXU, holotype; UC, isotype). Robust rosulate piphyte, rarely on rocks; stems Ort with mass of thick roots; cataphylls thick, eos cucullate, rounded at apex, drying brown, ining intact, ultimately becoming fibrous at CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 369 base. LEAVES erect; petioles flattened or sulcate adaxially, rounded abaxially, 5-23 cm long, 1- 2 cm diam.; geniculum 1.6-2 cm long; blades obovate to oblanceolate, moderately thick, acu- minate at apex, attenuate, sometimes cuneate (rarely rounded or subcordate) at base, (32) 40- 180 cm long, (9)13-50 cm wide; both surfaces t sometimes tinged purple, 4—45(50) cm long, 1— wide; gradually cuspidate-acuminate at apex, rounded to truncate at base; spadix dull green soon tinged throughout with lavender (B & K Purple 6/5), 6-34(47) cm long, 0.6-1.5 cm diam. at base, 2-7 mm diam. at apex; flowers rhombic, 2.4—2.5 mm in both directions, the sides straight to sigmoid; 6-7 flowers visible in the principal spiral, 8—12 flowers visible in the al- ternate spiral; tepals matte, lateral tepals 1.2-1.3 mm wide, the inner margins broadly rounded; pistils green, emergent but not raised; stigma lin- ear, ca. 0.9 mm long; st ging in a slow progression from the base; anthers pale yellow, held at edge of tepals and laying over pistil; the- cae ellipsoid; pollen pale yellow-orange, fading to white. INFRUCTESCENCE with spadix to 60 cm long; berries red, oblong-elliptic, rounded at apex, 1-1.5 cm long; seeds 2, oblong, flattened, 3.3 mm long. Figs. 166 and 167. The species occurs in western Mexico (Chia- pas) along the Pacific slope of Guatemala and along both the Atlantic and Pacific slopes from Nicaragua to Panama. There is also a single col- lection from Honduras (Atlantic slope) that ap- pears to be this species. The species also occurs in Colombia and has been collected in Chigorodó and Dabeiba in Antioquia. Anthurium salviniae occurs from near sea level to 1,400 m, in wetter parts of tropical moist forest, in premontane wet, tropical wet, and also, rarely, in premontane rain forest. Anthurium salviniae has long been called A. tetragonum but that name is a synonym of A. schlechtendalii, a species ranging from Mexico to Honduras or Nicaragua on the Atlantic slope. Although leaves of the two species are very sim- ilar, A. salviniae is a member of section Pachy- neurium and can be distinguished by its long- tapered, pale lavender spadix, long-tapered > Z z » E n o "T 4 x m - vA O e = z OQ - » zZ O > = Q > ^x g m A 3 " vi M € EN ! 3 E " T 1 T Fiaures 168-171 168. Amram salvadorense rent. Croat 42092.— 169. Anthurium ee (Aubl.) Engl., Croat XX-6.—170. Anthurium iby 25828.—171 sthaeriunm scherzerianum Schott, Croat 3 scandens ssp. pusillum Sheffer CROAT KUALI 1983] spreading spathe, and fist-shaped, cucullate cata- phylls. In addition the leaves of A. salviniae usu- ally dry darker and thinner. Anthurium schlech- tendalii usually has a shorter, blunter spadix and a short, thick spathe that is usually reflexed and often twisted. The spadix is darker violet-purple and is not at all glaucous at anthesis. A collection in cultivation at Missouri Botan- ical Garden from the Rio Maje region of Panama in Panamá Province is unusual in having dark violet-purple rather than red berries. CosrA ES ALAJUELA: N of Bijagua, Croat 36480 (MO); NW of Zarcero, Croat 43640 (MO). CARTAGO: SE of oen Croat 36808 (MO); NE of Tapanti, Lent 1034 ( GUANACASTE: vicinity of Tilarán, Bin A Valerio 44445 (US). lH ge! vicinity San Vito de Java, Croat 32898 (M. GUATEMALA, ESCUINTLA: DARE Escuintla a ta Lucía Cotz, Standley 63469 (F). RET ofAjaxá, Standley 88227 (F); W of Retalhule, Standley 87499, 8839] nidad, Croa nd San- P m “SucuereovEz E of Mazatenango, Croat 32786 E ONDURAS. ATLANTIDA: Lancetilla Valley near Tela, Standley 53228 (US). EXICO. CHIAPAS: eran Croat 43845, 43877 eg: Matuda s.n., 18630 (MEXU); Matuda 16377 Piera MO); Matuda 16776 (F, MEXU); Matuda i (MEXU, U UCLA); Matuda 18381 (DS, MEXU); atuda 18519 (DS, MEXU, UCLA); between Finca 17573 mit of Monte Ov: EO; vao Miranda 1 E Moin U). G RÍO SAN an n del Norte E tow), $m ith 87 (MO). ZELAYA: s. ve of Cerro Docente, Stevens 6793 (MO). — pum (Aubl.) Engl., Mart. Fl. Bras. 2: 78. à. Anthurium scandens ssp. scandens. Dracon- Hunt scandens Aubl., Hist. Pl. Guiana Franc. 6. 1775. TYPE: Frendi Guiana. (Draw- ing by Plumier in Amer. Nascentium Icon. tab. 74. 1703. BM, P Pothos violacea Me Prodr. zs 1788. Type: Ja- Dr maica, Swartz isting see: acontium repens Deso ., Fl. "Antill. 7: t. 499. 1829. Án Hur had (Plate 499 serves as the type.) , Melet. 22. 1832. An Y artz s.n. (not trim violaceum forma iit Kunth, Enum. Pl. 68. 184 Tyre: (not se Anthurj Pm VIO. “AE 1 th. Enum. ng hott, Oes Bot . . TYPE: ta "Schiede m Seen.) (Photos of Schott Aroid Drawing #279, OF MEXICO AND MIDDLE AMERICA 371 — In #N.S. 3804 and Schott Aroid Draw- in 2, NYBG Neg. £N.S. 3805. yo ds dolosum Schott, Oesterr. Bot. Z. 8: 179. 8. Type: Guatemala. Near San Pedro and San sie e zd id. hololectotype; K, solectotype; here designated). ANIME m Schon: Oeste Bot. Z. 8: 180. 1858. Type: Costa Rica. San José, Hoffmann 508 (Photo of don Aroid Races #270, BG Neg. #N.S. 3807). Anthurium virgosum Schott, Oesterr. Bot. Z. 9: 100. 59. Type: Brazil. Río de Janeiro, Riedel (K). Epiphytic creeper; stems usually less than ] m long; roots numerous along stem, ca. 3 mm diam., gray-green; cataphylls 3-6 cm long, drying brown, weathering to reticulate fibers and persisting around stem. LEAVES spreading; petioles sharp- ly sulcate, rounded abaxially, 2-8.5 cm long, 2- 3 mm diam.; geniculum 2-3 mm long, nearly obscure; blades ovate-elliptic to lanceolate-ellip- tic, moderately thick, (3.5)6-13 cm long, (1.5)2- 5 cm wide, short-acuminate at apex, acute to rounded at base; upper surface semiglossy, lower surface conspicuously brown punctate; the mid- rib acutely raised above, diminished and sunken in apical half of blade, convexly raised below; primary lateral veins 3-10 per side (sometimes more), departing midrib at 35°-40° angle, weakly sunken above, + obscure above and below; col- lective vein arising from the base, extending straight to apex, 3-5 mm from the margin. IN- FLORESCENCE erect to pendent, shorter than leaves; peduncle 1.5-6.5 cm long, 1- diam., equalling or longer than petioles; spathe 6 mm ly acuminate at apex; spadix yellow-green, some- times becoming lavender, 1.3-2 cm long, 2-3 mm diam. midway; flowers rhombic, 2-4 mm in both directions, the sides straight to sigmoid; 2-3 flowers visible in either spiral; tepals semi- glossy, lateral tepals ca. 1.2 mm apa the inner margin weakly concave to + straight; pistil greenish-white, scarcely emergent; stigma brush- like, exserted, ca. 0.2 mm long; stamens emerg- ing in a prompt and complete sequence from the base, held at edge of tepals and inclined over pistil; anthers white; thecae ovate, somewhat di- varicate; pollen white. INFRUCTESCENCE poe ig the spathe perststitg. spadix 2.5-3.5 m long, to 7 mm diam ; berries usually pale irai to almost white, sometimes purple, obo- void, rounded at apex, 5-8 mm wide; seeds 1 or 2 per locule, pale yellow, ovoid, ca. 2 mm long, 1 mm wide. Fig. 169. 372 Anthurium scandens is known from southern Hee the most ecologically diverse as well, occurring in tropical moist, premontane wet, tropical wet, premontane rain, and lower mon- tane rain forest life zones. The species is a member of section Tetrasper- mium and is distinguished by its more or less scandent habit, elongate internodes with usually persistent cataphylls, usually oblong-elliptic, glandular-punctate blades, reflexed green, more or less ovate spathe, cylindroid, usually pale green spadix, and pale lavender to almost white ber- ries. According to Sheffer et al. (1980), the great variability in the species is due at least in part to the fact that it represents a polyploid complex containing diploids, tetraploids, and hepta- ploids. The cytology of the group is insufficiently known, so it is not certain whether the different ploidy levels are correlated with morphological variation. The species is closely related to A. trinerve, with which it has long been confused. The dif- ferences separating the two species are outlined in the key. Anthurium scandens is most easily distinguished by its spathe, which is reflexed at anthesis, whereas the spathe of A. trinerve is per- sistently erect even after anthesis. BELIZE. CAYO DISTRICT: no other location, Lundell 6255 (MICH); vicinity of Cuevas, S of Milli lumbia Forest Station, Croat 24309 (MO, DAV): Dyer 9946 (MO) 2506 (BR); E of Cachí, Croat 47082, 47083 (MO): Cart jerih nari vote ne SE of Cartago, Harmon & Fuentes 6315 (M ; W of Turrialba, Croat 36846 Volcán Irazü, Moore 6 66 (MO); La P n 5446 (MO). HEREDIA: S of San Miguel, Lent 36 (DAV); Vara Blanca, ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 Skutch 3665 (MICH, — iin Barba, Brade 2509 rte, W of Liesner 2906 (MO); Rincón de Osa, Liesner 1701, 1804, 1997 (MO); N of San Vito, Davidson 2 d: ae Ste- vens 232 (MO); san José: without furthe: y, Brade 2501 (BR); Pittier 2445 (BR); at Casa tis 5237 (DUKE); Lent 2090 Sii N of at 43386 y between Río Paqu Mes Rio Viejo "Pam Goerger 9826 (MO); N of San Isidro del General, Har- mon & Fuentes 6229 (MO, UMO); between San Isidro del General and Do minical, Croat 3 5333 (MO); vicin- i itti. BR); D 1255 (BR); & CARTA a Mata, Utley 4743 (DUKE) EL SALVADOR. SANTA ANA: Finca La Cumbre, e talvo 3252 (MO); Weberling & Legen 1086 (GOE UATEMALA. ALTA VERAPAZ: n, Tuerckheim 1355 (MO), 1923 (BR, MO); vicinity Cubilhuitz, wi ermark 44357 (MO); road to El Estor, Croat 4] j (MO); near Turuct SUR 70725 (MO). BAJA We APAZ: along hwy. c i to Cobán, about 1 mi turnoff to Re pM 41146 (MO). HUEHUETE Bus 1128 (DAV); Harmon & Fue SACATEPEQUEZ: SW of Atene ue ai 42 2 p» SAN MARCOs: Finca El Porvenir, Steye ias 1 (MO). ania ae near Mss nango, Bé pis 388 (BR). ZAC ere "Rillito from Volcan de os, Steyermark 42342 Yuncker TLANTID Smith 8 5 (MO); n near El Ac hiote, : (MOI COPAN: W of Copan, Croat 42519 (MO): poet & Herna AE PARAÍsO: trail, Danlí to Finca La Emilia Moli (MICH). MoRAZÁN: Hoya Grande, Williams : 10037 (MO); between Tegucigalpa and El Zam oo plage Munic. of Berriozábal, B -— 20286 (MICH, MO); Miranda 6723 3 MEX MO o & Brossard 6327 (BH); Cerro Bail, C SW of Comitan, Carlson 1820 (MICH); ) vicini ie on (MEXU, MICH); Hs 17911 (MEXU); pae coautla de Espinosa, Carlson 2103 (MICH R ^ie 40565 614 (MEXU); S of Palenque, Thorne & Lathr ^ ji (RSA); Munic. Pichucalco, Gilly & Her Ken- (MICH); ears Pre Pueblo Nuevo Solin Mati nedy et al. 3 (M Oy; San Juan 18175 QUT Siltepec, Matuda 18301 MEX) NO: tween Tapachula and Unión Juárez, Croa f i : ruins of Bonampak, Hoover 192, 208 ( | 1983] CROAT Autlan, McVaugh 10219 (MEXU, MICH); Wilbur & Wilbur 1738 (MICH). OAXACA: between cers del Camino and Santa Maria Chilchotla, Croat 48323, 48385 (MO); between Tuxtepec and Oaxaca, Croat (ME aes ra 226 (UH); Teapa, Birdsey 235 (UC); Con- 3 (MO); Gilly & Hernandez 213 (MEXU, tr VERACRUZ: Schnee V (MEXU); Rancho la La- guna, Chazaro 771 (JAL); near Pedre al Rumbo, Lot et al. 1815 (MEXU); queries Dorantes 1800 (MICH, MO); Ventura 10382 (MEXU, UMO); Zola & Baker 98 (XAL); Atoyac, Matuda 1411 (ME XU) MICR MO); Lake Catemaco, Gonzales 856 (MEX ; Sousa 19 (MEXU); E of Córdoba, Lundell r Lenten 12532 (MICH); Matuda S- 150 (MICH); Matuda 838 (MEXU, MICH); between Córdoba & Fortin, Miranda 620, 4836 MICH); Kerber 294 (BR); Lot 537 XAL); Hidalgotitlán, Dorantes 3002 (XAL); Dorantes 3088 (M 2535, d us 113, 225, 439, 952, 1048, Za, ci Moro 703 (XAL); Mid Ventura 7153 oS nuin, Gomez- Pompa & Riba 161 oani Gor méz- 605 EXU); vicinity Zacualpan, Galeotti did Pur dee us 5764 ( MO AeKa Santos N CH); Valdivia 2141 (X. (XA aan GRANADA: Finca aera Atwood AN64 can Mombacho above Finca Las Delicias, N56 (MO); SE flank of Mombacho, Neill N30, 126 (MO); Stevens 4363 (MO); Croat 39116 (MO). : Cerro Zamaria, Atwood AN118 (MO); NW e Abanas, Cron 42989 (MO); eg El Stevens 3500 MO); between MO). MATAGALPA: vicini ity Tucero, Ps „Stevens 3300 (MO). z vA; Caño Zam la 26 (MO); trail from Cero o El Inocente to Cerro Sas- Mon 6696 (MO); vicinity Want, Stevens 7341 b. Anthurium scandens ssp. pusillum Sheffer, Aroideana 3(3): 86. 1980. Type: Costa Rica. rtago: Southeast of Tapantí, Río Grande de Orosi, 1,400 m, 16 April 1967, R. W. Lent 830 (F, holotype). OF MEXICO AND MIDDLE AMERICA 373 Anthurium scandens ssp. pusillum occurs from Honduras to Colombia and Venezuela at ele- vations of 600 to 2,500 m in lower montane moist, premontane wet, lower montane rain, and premontane rain forest life zones. The subspecies pusillum differs from the typ- ical subspecies in usually having a much smaller spadix (usually with fewer than 13 flowers), usu- ally 2210 mm long. Sheffer (1974) believes that the subspecies may represent one of the diploids upon which the polyploid complex has been built. Fig. 170 CosTA RICA. CARTAGO: E of Cachí, Croat 47062 (MO); Carpintera, Stork 1143 (MICH); S of Muñeco, Luteyn 3880 (DUKE, MO); Tapantí, Croat 47048 (MO). SAN JOSE: along CR 220, Luteyn 3593 oo San Jose- p EK E iani od UE ARAGUA TEGA y Ar ranjuez, Stevens 5606, 5936 (MO). ZELAYA: aes "idm Cerro El Inocente to Cerro Saslaya, Stevens 6695 (MO). Anthurium Mbermdinam Schott, Oesterr. Bot. Wochenbl. 3. 1857. Type: Guatemala, — n. (B, holotype (2 sheets); BP, iso- ). Epiphytic or terrestrial; stems very short, roots numerous, thin; cataphylls thick, to 1.5 cm long, caudate at apex, drying brown, persisting as fi- bers. LEAVES spreading; petioles terete, 4-20 cm long, 1-3 mm diam.; geniculum 4-7 mm long; blades moderately thick, linear to elliptic or lanceolate, 5-26 cm long, 1.5—6.5 cm wide, narrowly acuminate at apex, obtuse to cuneate at base; upper surface semiglossy, the lower sur- prominulous below; primary lateral veins 8-11 per side, departing midrib at 45° angle, nearly obscure above and below, loop-connecting to collective vein, lesser veins obscure; collective vein arising from near the base or one of the lowermost primary lateral veins, 1-3 mm from the margin. INFLORESCENCE erect, rai or longer than leaves; the peduncle 14-52 c long, ca. 2 mm diam.; spathe moderately sse bright red-orange (B & K Red 7/2.5), elliptic to ovate, 3. ey hk cm long, 2.4-6 cm wide, abruptly short tapex. cordate at b sometimes overlapping); spadix pale orange to red, 2-8 cm long, ca. midway, ta- pered at apex and usually y olet: Rowers square, 2 mm in both directions, the sides straight to + sigmoid, 2-3 flowers visible in the principal spi- ral, 5-6 flowers visible in the alternate spiral; 374 tepals glossy, sparsely and minutely papillate, lateral tepa mer- £e eo ents, g, 0.5 mm wide; anthers white, held at edge of 1. ca. 0.3 mm long, 0.5-0.7 mm wide; thecae ellipsoid, slightly di- varicate; pollen white. ESCENCE with orange to red berries. Fig. 171. The species is found in Costa Rica at eleva- tions of 1,300 to 2,100 m in premontane rain, lower montane rain, and montane rain forest life zones on the Atlantic slope ofthe Cordillera Cen- tral and the Cordillera de Talamanca. It has been collected most commonly in the Tapantí and La Hondura areas. Anthurium scherzerianum is a member of sec- tion Porphyrochitonium and is easily recognized by the large, showy, bright orange or red-orange to red spathe, and coiled spadix. The species is now commonly cultivated throughout the world and about 40 different forms are recognized in cultivation. The species has no close relatives. he type of A. scherzerianum was believed to have been collected in Guatemala; however since that time no other collections have been seen from that country. So far the species has been found only in Costa Rica, where it is plentiful, and it is likely that the type was collected there. CosTA RICA. ALAJUELA: edge of Lake Hule, Luteyn 3221. 3351 (DUKE) CAR : Burger & Stolze 5621 (NY, WIS); S of Cartago, Maxon 503 (NY); Stork 1913, 4534 (UC); Navarro Valley, Stork 1701 (MICH); between Quebrada Honda and o Sombrero, Luteyn et al. 4528 (DUKE); Tapanti, Croat 36077 (MO); Lent 836 (NY); Taylor 4468 (M ; Tessene 1296 (WIS); Turrialba, Brade 2504 (BR). HEREDIA: S of Cariblanco, Croat 35800 (MO); > of Vara Blanca, Croat 35583 (MO). SAN José: Santa María de Dota, Stork 1883 (MICH, UC); Tablo. Pittier 7909 (BR). Anthurium schlechtendalii Kunth, Enum. PI. 3: a. Anthurium schlechtendalii ssp. schlechten- dalii. Type: Mexico. Hacienda de la Laguna, Schiede & Deppe (not seen). Anthurium mexic Liebm., Vidensk. Meddel. Dans k ferui. fore. Kjebenhavn 1: 21.1849. ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 TvPr: Mexico. Veracruz: between Colipa & Mi- santla (not seen). —— kunthianum Liebm., Ann. Sci. Nat. Bot., V, 2: 372. 1854. TYPE: Mexico (no type cited). Adtodtioh brachygo onatum Schott, Prod. i wem Type: Mexico. Vera racruz: Cordoba, Kerber n. (B) (see discussion). Anbar tetragonum Hook. ex Schott, Prod. Aroid. 475. 1860. Type: Locality unknown (not seen). ee oe Engl., Bot. Jahrb. Syst. 25: 366. PE: Mexico. Veracruz: Fortin, Kerber 9 Anthurium tikalense C. Lundell, Wrightia 3: 161, Fig 55. 1966. Type: Guatemala. Petén: Tikal National Park, on temple in Group “H,” March 1, 1961, C. L. Lundell 18198 (LL). Usually epiphytic, sometimes epipetric; stem usually short, 2.5-5. am.; internodes an and leaf scars hidden by ie mass of thick, greeni roots, 3-8 mm diam.; cataphylls moderately thick, 7.5-16 cm long, caudate-apiculate at apex, drying brown, persisting as reticulate fibers LEAVES erect; petioles 12-23 cm long, 0.5-2 cm wide, trapezoidal, the abaxial margins some times sharply to obtusely ribbed, widest adaxi- ally, shallowly sulcate; geniculum 1-2 cm long. blades obovate-elliptic, thick, acute to short-acu- minate at apex, acute to obtuse at base, R cm long, 10-58 cm wide; broadest near or E middle; both surfaces matte to semiglossy; mi rib flat to weakly rounded at base above, coming more acute sometimes weakly ri diminished and sunken at apex, raised and A at base below, rounded at apex; primary o veins 15-16 per side, raised on both uw is departing midrib at 50? angle, straight near a margin, then arching toward apex, lesser be i ped visible; collective vein arising pe mary lateral vein in apical one quarter id : iip m from margin, sunken above low. INFLORESCENCE ple mei peduncle 33—43 cm long, 6-1 d mm shorter than | linear-lanceolate to lanceolate, 15-28 cm a —5 cm wide i , obtuse at we : : on strongly reflexed, inserted at TOE € duncle; spadix green, 8-29 cm long. diam. at base, 4-6 mm diam. at apex: square, 2.2-2.6 mm in both directione straight to weakly sigmoid; 7-17 flow in the principal spiral, 10-14 flowers visi 7 the alternate spiral; tepals green, matte, W! lish, lal ple punctations, the outer margins purpP™®™ ’ . res cas È b- 1.3 mm wide, the inor m m a - * | CROAT BUAL 1983] to 1 mm, dark purple; stigma linear, 0.8 mm long, with minute droplet ca. 4 days before sta- mens emerge, dry, black as stamens emerge; sta- mens emerging rapidly from the base, lateral sta- tract to hold anthers at side and inclined over pistil; anthers pale orange, 0.9 mm in both di- rections; thecae + oblong, ed divaricate; pollen pale orange fading to cre UC- TESCENCE arching-pendent; a brown, persistent; spadix up to 65 cm long, 5 cm diam.; berries bright red, lanceolate-elliptic 1-1.5 cm long; pericarp thick; mesocarp pulpy, white, with numerous raphide cells; seeds 2, greenish-white, d mm long, ca. 1.5 mm wide. Figs. 172 and 3. Anthurium schlechtendalii is in section Pachy- neurium and consists of two subspecies. The sub- species schlechtendalii ranges from Mexico (cen- tral Veracruz) to Nicaragua on the Atlantic aloes at an elevational range from near sea level t 1,600 m (most commonly below 1,000 m). This species is most closely related to A. schle- chtendalii ssp. Jimenezii from the Pacific slope of Mexico, which differs by occurring in season- ally drier habitats, by its generally smaller leaves, ES its frequently narrowly ovate, usually erect Anthurium | schlechtendalii merges into the range of A. salviniae only in a few areas in the Mountains of southeastern Chiapas, such as on Monte Ovando, where both species were seen &rowing, Specimens collected in the municipios of Siltepec, Escuintla, and Acacoyagua tend to Cataphylls of typical A. schlechtendalii. See dis- cussion following A. ma for separation n E these two : indi sing Mex aroids, Bunting (1965) C u "re an lets of a plant cultivated at Orel] University and reputedly collected by H. seenhouse ees this plant (Bunting 1579 and Nicolson 603) ú to confirm this. Anthurium crenatum so far V is known j is restricted to Puerto Rico and the igin Islands. Mes of populations of A. schlechtendalii in exico and elsewhere show most morphologica OF MEXICO AND MIDDLE AMERICA 375 characters to be variable. Petiole shape, although usually trapezoidal in cross-section with the an- gles moderately acute, may be broadly rounded abaxially, such as those illustrated by Schott for A. brachygonatum [see Schott drawing No. 473 (W)]. Although Bunting (1965) considered this distinct from A. schlechtendalii, it is believed to be a form of an unusually variable species. BELize. No other location, Schipp S-402 (MICH, MO, ui^ UC) BELIZE: Northern River, Gentle 1291 (F); no other location, im 11325 (MICH); Gracy Rock, Croat 23851, 23896 (MO); rice er & Dwyer 147 . CAYO DISTRICT: no other ek 53 cen MO); W of iib Highway, Spell- man & Newey 1984 (MO); vicinity La Flor, S no de Oro, Croat 23811 (MO); Waterhole camp near Vaca, Gentle 2567 (MICH). COROZAL DISTRICT: no other lo- ICT: 24297, 24333, 24431 (MO); Vanderveen 658 (MO); Moffredge procu Gentle 5372 (TEX); Río Grande, Puis 4723 (TE ATEMALA. ALTA VERAPAZ: La Tinta, Smith 1530a (US); W of Teleman, Croat 4 Ms 41551 (MO); be- tween Tucuráü and El Estor, Croat 41525, 41529 (MO). ESCUINTLA: SW of Palí 2420 (UMO). HUEHUETENANGO: between Ixcán & Finca San Rafael, Steyermark 49411 (F); below La Libertad, Steyermark 51171(MO). izaBAL: E of El Estor, Jones & Facey 3456 41839 (MO); vicinity Quiriguá Viejo, iani 24053 (NY, US); between Virginia and Lake Izabal, Steyer- mark 38769 (F). PETÉN: Dolores, Contreras 2254 (LL); El Paso, Herman 14664, 14671 (MICH); verc dd (MICH); La Libertad, Lundell 2873 (MICH); M canché Lake, Contreras 920 (LL); Santa son la dell 2708 (MICH); 2889 (MICH, NY); Tikal, Bern- MO); Contreras 1618 (LL); Croat peed 0 (TEX); Uaxactin, Bartlett 12295, 12296 (MIC «rio SAN MARCOS: Volcán Taj jumulco, Stey- ermark 37078 (F). HONDURAS. COPÁN: E of Copan, Croat 42508, 42521 2904 (MO). SANTA BÁRBARA: NW side of Lake Yojoa, Croat 42757 (MO). voRo: vicinity góc Yuncker et al. 8132 (MO, NY, US); 8174 (F, MO). MEXICO. CHIAPAS Matuda 19646 (DS); Angel Albino Corzo, Fon 3880 (MAD , DS); vicinity Pone, Roe as 1132 (WIS); Ton 258 7 (DS): vicinity rzo, Breedlove 26872 poe Croat 46437 MO); Uie 2646 (MEX U); SW of El Jocote, Croat MO); Escuintla, Croat 4387 : ae Matuda ; between Huixtla and Moto- zintla, Breedlove 28609 (DS); Croat 40766, 47225 i 723 (DS, MICH, NY); 1599 E IE p 413 "y: E y Las Margaritas, Breedlove 33199, |. 33446 (DS); esed Siltepec, Croat 47250 (MO); Ma- ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 Jechtendali 2 Anthurium schlechtendalii Kunth, Croat 40068.—173. Anthurium sch 5. Anthurium schottianum Croat & Baker, Croat 43247. 1983] CROAT tuda 18798 (MEXU); Ocosingo, ruins of Yaxchilan, Breedlove 33872 (DS); between Palenque and Oco- 982 (DS, MO); Breedlove 29004, 30358, 36555, 38332 (DS); Croat 40587, 40626, 40653 (MO); ee aea H261 Aa vicinity reos ue, "PR Hoov r 282 ucalco, Tenejapa, Breedlove 6242 (DS, F; MICH); 11731 (F, MICH, TEX); Ton 350 (DS); Tuxtla Gutiérrez, Guz- mán 13 (ME nde GuE UERRERO: Minas, Hinton 10125 ; S of Valle Nacional, Croat 39707, 39708, 39738, 39756 MO; TABASCO: Balan- cán, Valdivia 2028, 2052 (XAL); NW of Cardenas, Barlow 26/10 (WIS); E of Teapa, Croat 40109, 47894 (MO); S of Villahermosa, Conrad et al. 2866 (MO); Croat 40068 (MO on road to Jalapa, Dorantes 505 (CAS, F, MEXU); G astillo & Dillon 483, 715 (MO); Cordillera, aleotti 6055 (BR); Córdoba, Miranda 4894 (MEXU); Cui NM, Rosas 910 (XAL); Municipio de Dos Ríos, TMO); ; vicinity Ager abaa 3941 r 8940 (MICH, (er, is 689, 799, 870, 1038, 1045, 1115, 5, 1343, 1622 (XAL): 171, 236 (XAL, MO); 191, 128: 5 » 333 (MEXU); 222, 291, 421, 731, 1165, MEX 1674 (MO); 275, 667, 907, 931, 1196 (XAL, (ae U, MO); 458, 1094 (XAL, —— Vazquez 582 1. MO, XAL); vicini ityH 44000 ); Orizaba, Botteri : targets 1602 (P); Croat 39548 (MO); Muller pa ] 2 ; Palmasola, Nevling & Gomez-Pom- 74 (MO, XAL); Rancho Viejo, Purpus 15711 Andrés Tuxtla, Calzada 455 (MEXU); Dres- sler & Jones E (MICH, NY, UC); Fe nus & V So 52(MEXU); Moore & Ceno 6232 (BH, MEXU). oer al. 160.3 AL); above San José de Gracia, Qut 3961 1 (Mo) s Dx Tuxtla, Sousa 2158, 2373 & C, U); E slopes of Sierra de Los Tuxtla, Cochrane Ochrane 8625 (MO); Barra de Sontecomapan, Her- OF MEXICO AND MIDDLE AMERICA 377 nandez 608 (MEXU); Tapalapa, Gomez-Pompa 5 105 MO); Schott 638 (F); N of Muna, ae 2528 (DUKE); Córdoba, Bourgeau 1787 (MEX NICARAGUA. rada of Cuapa, Atwood & Neill (MO); Stevens 7377, 7383, 7386, 8736, 8753, 8755 (MO). b. Anthurium schlechtendalii ssp. jimenezii (Ma- tuda) Croat, comb. nov. Anthurium jime- nezii Matuda, Anales Inst. Biol. Univ. Nac. México 32: 147. 1961. Type: Mexico. Mé- xico: Barranca de Zacualpan on volcanic rocks, 1,300 m elev., Matuda 37245 (MEXU, holotype). Usually terrestrial or on rocks, usually 0.5-1 m tall; stems less than 30 cm long, ca. 3-4 cm diam.; roots moderately thick, directed down- ward, sometimes curved upward; cataphylls co- riaceous, 6—7 cm long, acute at apex with minute subapical apiculum, drying reddish-brown, re- maining + intact in upper nodes, ultimately de- ciduous. LEAVES erect-spreading; petioles 2-15 cm tong, 8-9 mm m di iam., MIbqua adieu broadly and sharply 5cm long; blades oblanceolate to lanceolate, gradually i ap ), obtuse to narrowly roundid at base, broadest at middle, 36-104 cm long, 6-32 cm wide; both surfaces semiglossy, lower surface sometimes having a bluish-green color; midrib slightly raised at base, more acutely raised midway and sunken at apex above, quadrangular and prominently raised be- low, becoming convex at apex; primary lateral veins 9-14 per side, departing midrib at 50? an- gle, raised above, usually raised below, arcuate- ascending; interprimary veins weakly sunken above, prominulous below; lesser veins obscure above, visible below; collective vein arising from one of the primary lateral veins in the lower half of blade, 2-4 mm from margin, flat above, prom- inulous below. INFLORESCENCE erect, shorter than leaves; peduncle 10-54 cm long, 3-5 mm diam., terete, longer than petioles; spathe sub- coriaceous, green weakly tinged with purple (B & K Yellow-green 6/5), 4-8.5 cm long, 1-2.5 cm wide, lanceolate, oblique and narrowly acumi- nate at apex, subcordate at base, broadest just above base, inserted at 30° angle on peduncle; 378 spadix usually sessile, greenish tinged with pur- ple (B & K Yellow-green 6/5), 3.5-14 cm long, 5-10 mm diam. at base, 2-5 mm diam. at apex; the flowers rhombic to sub-4-lobed, 2.7-2.8 mm long, 2.3-2.5 mm wide, the sides straight to jag- gedly sigmoid; 11-12 flowers visible in the prin- cipal spiral, 7-8 flowers visible in the alternate spiral; the tepals matte, obscurely and sparsely punctate with numerous, irregular droplets at an- thesis, lateral tepals 1.5 mm wide, the inner mar- n broadly rounded, the outer margin 3-4 sided; ro weakly emergent, medium green; the stig- a0. 5mm long, oblong-elliptic; stamens emerg- the base; de anthers creamy white, leaning over edge of tol, 0.8 mm long, 1 mm wide; thecae oblong- diat ind divaricate; pollen white. IN- FRUCTESCENCE pendent; berries red, oblong- ellipsoid, mune. at apex, 11 mm long, 9 mm wide; mesocarp with raphide cells dense and nu- merous throughout; seeds 2, oblong-ellipsoid, tan, flattened, 6 mm long, 4 mm wide, 2 mm thick with pale, punctiform raphide cells, encased in a slender, sticky, gelatinous, apically elongated envelope. Fig. 180. Anthurium schlechtendalii ssp. jimenezii is en- demic to Mexico and is restricted to Guerrero and southern Oaxaca in seasonally very dry for- ests at 250 to 1,350 m elevation. ubspecies jimenezii is recognized by its ex- clusively terrestrial or rupicolous habit, its thick, oblanceolate leaf blades with usually free-ending primary lateral veins, its quac spadix, green, lanceolate spathe, and red berri Th olated from the typical subspecies of A A. alee which is found in northern Oaxaca and also on the east- ern side of the Isthmus of Tehuantepec but not in eastern Oaxaca. Subspecies Jimenezii differs from the typical subspecies in being generally smaller, by occurring exclusively on rocks and in seasonally much drier habitats, and by havin usually thicker leaf blades. Matuda (1961) stressed that the taxon has peduncles relatively much longer than typical in A. schlechtendalii. While it is generally true that the inflorescences of ssp. Jimenezii are commonly as long as or longer than the leaves they may also be much shorter than the leaves. orago i Jimenezi ig Ln similar to A. hal- iC Mori spathe and mature Posi that are pale ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 greenish- yellow. Also similar is A. nizandense, in having proportionately much longer peti A f blong on onthe paren stig 5 s nizandense the blades are 1.3-2.5 times longer than the petioles where- as in A. schlechtendalii ssp. jimenezii the leaf blades are usually 6 to 10 times longer (rarely only as little as 3 times longer). MEXICO. GUERRERO: El d between Tierra Co- , Crisman & Willis m pulco to Chilpancingo, Moore H Acahuizotla and Agua de Obispo, Moore 5120 A UC); highway above Acahu ae Mi ore 6957 (B Tierra Colorada, eee 5T ids (MICH, US). oaxaca: N of Pu fMir 45926 (MO); between ‘Oaxaca aaa Pochutla, S o uatlén, Croat 46093 (MO); N of Pluma Hidalgo, C K 46142 (MO); “a Santa Rosa, 195 on Oaxaca-Puerto Escondido, Rzedowski 19592 (MEXU). pea vignes Croat & Bake p SW of Limón; steep ravine above tarata, 50-100 m elev., Croat 43247 (MO- 2584489, holotype; CR, F, K, SEL. US, iso- types; Live at MO). 0cm 24 de- Large terrestrial; stems usually less than 5 long, 3-5 cm diam.; leaf scars conspicuous, cm wide; roots few, green, ca. 5 mm diam. rá scending, often short and blunt; -—— 4 riaceous, es purple, 9-12 cm long. E at apex, d brown with tinges O ud acepta LEAVES with petioles eret to spreading, 40-150 cm long, 5-8 m j ee shallowly sulcate to flat adaxially, roun purple a few faint ridges abaxially, hear dark P blades from the base; geniculum cm long; ovate, gradually acuminate at apex (the 2 A long apiculate, turned down), lobed at base. int ear po owly Uf" the sinus db narrowly triangular, closed, acute at apex; the upper surface semiglossy, lower surface semiglossy; acutely ra a and sunken, prominently ral ~~ $ m oM | CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA n P af a P. FiGURES 176-179. 176. jos itn Croat & Baker, Croat 26976 (Type). —177-178. Anthurium *eibertii Croat & Baker, Croat 48568.— Anthurium seleri Engl., Croat 46505. dg » * SON Lm » Nt: L4 >. Ld » Z Z > e n © "n1 = a m Z n o o C & w le) «a. » A M. N3QGOGIVO TVD Poom me 180. Anthurium schlechtendalii ite aie nezii anam Croat, Croat 45926.—181. Anthurium seleri Engl. (Cultivated at NAYOX X im Anthurium silvigaudens Standl. & Ste oat 411 CROAT CRVA ILI 1983] veins 5-8 pairs, not coalesced, raised above and below; primary lateral veins 9—20 per side, de- parting midrib at 35°-50° angle, slightly arcuate- ascending, faintly loop-connecting, sharply raised above and below; interprimary veins very few, slightly raised near apex of leaf above, raised below; secondary veins mostly flat above, in part uous, slightly sunken above, slightly raised be- ow, mm from the margin. INFLORES- CENCE erect, much shorter than the leaves; peduncle sometimes absent or to 20 cm long, 4- 5 mm diam., dark purple, terete; spathe coria- ceous, dark purple with green splotches (B & K Red-purple 2/7.5), ovate-lanceolate, 6.5-15 cm long, 1.6-3.5 cm wide, broadest at base, abruptly acuminate t apex, rounded t l date at base, stiffly curled back lengthwise, frequently twisted at anthesis, inserted at ca. 45° angle on peduncle; spadix purple (B & K Purple 2/7.5), 4-14 cm ong, 7-9 mm diam. at base, 4-5 mm diam. at apex; flowers square, 2.6-3 mm long, 1.5-3.5 mm wide, the sides nearly straight; 4—6 flowers Visible in the principal spiral, 5-8 flowers visible in the alternate spiral; tepals matte, densely punctate, minutely papillate, the lateral tepals to 1.9 mm wide, the inner margin straight to con- Bee held flat against pistil; pistil violet-purple, minute and barely exposed 1 or 2 spirals ahead ofemerging stamens, gradually larger as stamens ‘merge; stigma a dry, linear slit when stamens e emerging; stamens emerging slowly from the ase, exserted on short filaments which retract ü hold anthers in a tight cluster around the pistil; 0 aments flattened, milky, the exposed part ca. ia — long. 0.6-0.7 mm wide; anthers yellow, : d above pistil, 0.5-0.6 mm long, 0.7-0.9 mm Wide; thecae ellipsoid, scarcely divaricate; pollen dario (B & K Yellow 9/5), drying pale yellow. FRUCTESCENCE pendent; spathe usually rien spadix to 35 cm long, to 2.5 cm wide, ai apical one half to one third often not devel- at "e berries; berries obovoid, acute to rounded the Pex, pinkish with a metallic sheen, darker in 6 apıcal third, 9-10 mm long; seeds 2, tan, to Mes m long, 5 mm wide, flattened, ca. 2 mm thick. '8s. 174 and 175 p The Species is known only from the region of uite wet forest-basal belt transition in reme southwestern Costa Rica at elevations OF MEXICO AND MIDDLE AMERICA 381 of less than 100 m. It will no doubt be found elsewhere in Costa Rica and Panama on the At- lantic Slope. 4 sad L L member of section Pachyneurium and is distinguished by its large, thin, ovate leaf blades with scarcely any posterior rib (i.e., all basal veins are free to the petiole) and by its short peduncle, its purple spadix and its ovate-lanceolate, often purplish, frequently twisted spathe. CosTA RICA. LIMON: near the Río Catarata between Bribri on the Rio Sixaola and the Caribbean costal plain, Baker & Burger 69 (MO); Burger et al. 10391 (F, MO); NE of Bribrí, ca. 40 mi SW of Limón, Croat 43247 (CR, F, K, MO, SEL, US); N of Bribri on banks of Rio Catarata, Utley 5500 (DUKE). Anthurium seibertii Croat & Baker, Brenesia 16 (Supl. 1): 85. 1979. Type: Panama. Chiriqui: wooded slopes on Cerro Horqueta, 1,650 m, Croat 26976 (MO-2251853, holotype; CR, F, NY, PMA, US, isotypes; Live at MO). Epiphyte; stems thick, short, 3-6 cm diam.; leaf scars mostly hidden, ca. 2 cm wide; roots numerous, descending, ca. 6 mm diam., green with crusty tan epidermis; cataphylls coriaceous, 12-21 cm long, caudate-acuminate at apex, the apiculum ca. 5 mm long, drying tan (B & K Yel- low-red 8/5), splitting at base, the apex remain- ing + intact. LEAVES held stiffly erect-spread- ing; petioles 10-55 cm long, 6-10 mm wide, + quadrangular, 5-ribbed with 3 ribs on the abaxial side, sometimes with 2 additional ribs on the lower corners; geniculum 1.5-3 cm long, ribbed as the petiole; blades strap-shaped, gradually acuminate at apex, obtuse to slightly rounded at base, 45-85 cm long, 4-16 cm wide, broadest at middle, the margin flat to slightly undulate; up- per surface semiglossy, lower surface matte to semiglossy; midrib convexly raised above, di- minished at apex, sharply below; primary lateral veins numerous on each side, departing midrib at 50°-60° angle, sunken above, sharply raised below, arcuate-ascending, prominently loop-connecting; interprimary veins numerous, slightly sunken above, prominulous below; secondary veins flat above, flat or slightly raised and darker below; collective vein arising from near base, about as prominent as the pri- mary lateral veins, sunken above, prominently raised below, 2-7 mm from margin. INFLO- RESCENCE erect or spreading at anthesis, ca. two thirds as long as leaves; peduncle 30-60 cm 382 long, 6-10 mm diam., terete with a single dorsal spadix violet-purple (B & K Blue-purple 2/ E 10-20 cm long, 7-213 mm diam. at base, 5-6 mm diam. at apex; flowers 4-lobed, 2-2.5 mm in both directions, the sides jaggedly sigmoid, 7-9 flow- ers visible in principal spiral, 7-10 flowers visible in alternate spiral; tepals matte, the lateral tepals ca. 1.5 mm wide, the inner margin straight; the pistil emergent, orange-yellow; the stigma ca. 0.7 mm long, mm wide, with conspicuous stigma droplets iun several days before stamens emerge; stamens emerging + rapidly from the base, the lower fourth with stamens emergent while apical flowers still are producing droplets on the stigmas; anthers pale orange, held just above tepals in a tight cluster, ca. 0.5 mm long, 1 mm wide; the thecae ellipsoid, scarcely divar- icate; pollen white. INFRUCTESCENCE pen- dent; spathe often missing; spadix usually 22-30 cm long, ca. 2 cm wide; berries orange, obovoid, flat to rounded at apex, 10-12 mm m long; meso- carp orange, mealy, sweet tasting but pungent; seeds 2, + oblong with rounded corners, ca. 5 mm long, 4 mm wide, 2 mm thick. Figs. 176, 177, and 178. Anthurium seibertii is known from higher al- titudes in Chiriquí Province (1,300—2,000 m) and from one collection at 1,040 m in San José Prov- ince of Costa Rica (Skutch 2626, MO, US). The species is a member of the section Pachy- neurium and is very similar to A. protensum. They share similar strap-shaped leaves, pale vi- olet-purple spadices and orange berries. Anthu- rium protensum differs in having a nearly terete petiole while A. seibertii has petioles that are prominently ribbed abaxially and rectangular. Anthurium protensum also has leaves that are usually pendent and are generally smaller than those of A. seibertii. COSTA Mee SAN JOSÉ: El General, 1,040 m, Skutch 2626 (F, MO). Anthurium seleri Engl., Bot. Jahrb. Syst. 25: 459, 1 . TYPE: Guatemala. Huehuetenango: near Chaculá, Seler 2643 (B, holotype). ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 Anthurium ——— Matuda, Anales Inst. Biol. Pul ue ico 36: 108. 1966. Type: Mexico. pes Lovene, Munic. Miahuatlán (N lof m Hidalgo), 1,200 m elev., Mar 384 (MEXU, holotype; F, isotype Terrestrial; internodes and leaf scars obscured by cataphylls and roots; roots numerous, medi- um thick, descending; cataphylls subcoriaceous, 6-11 cm long, sometimes faintly tinged reddish at apex, sometimes with long hair-like subapical apiculum, drying reddish brown, splitting at base, remaining intact at apex and persisting. LEAV with petioles erect to spreading, (7)14-48 cm long, 4-6 mm diam., sharply and shallowly sulcate; geniculum 1. Put 6 cm long; blade moderately thick, narrowly triangular to oblong-lanceolate or oblong-lanceolate with hastate posterior lobes, gradually acuminate at apex, usually broadly lobed, sometimes deeply lobed or + rounded to truncate at base, (9)24—50 cm long, (5)10-50 cm wide, broadest at base; anterior lobe (8. 5)25-41 cm long, margins convex to straight, posterior lobes 6-18 cm long, from sinus to outermost point; the sinus usually parabolic, sometimes a! cuate with petiole decurrent, broader than deep. rounded at apex; the upper surface matte to semi- alesced to 8 cm; posterior rib straight, ye turned up on outer margin; primary s EC 9-11 per side, departing midrib at 45°-50" ane 6 mm from margin. spreading, longer than leaves; peduncle 2 me cm long, terete, sometimes faintly ting purple; spathe moderately thin, lanceolate, green, tinged reddish at base, 5-9 cm long e, long-acuminate at apex, a 0°70" rounded at ee inserted on peduncle at 1:2 angle; stipe 4-20 mm long in front, l- ^ long in back; spadix green, tinged red vit a inner margins of tepals, 6-13 cm long, noni diam. at base, 2-4 mm diam. at apex 335 rhombic to sub-4-lobed, 3.5-3.7 mm ee mm wide, the sides + straight to grad i iral, moid; 4-5 flowers visible in the principal 4—10 flowers visible in the alternate d m wide, the inner margin iit 4 mn. vex, + transparent; pistils emergent to CROAT KULAI 1983] white to pale green with minute purple puncta- tions; stigma linear, ca. 0.5 mm long; stamens emerging slowly from the base in a + complete sequence, lateral stamens first followed by alter- nates, the basal spirals complete as lateral sta- held in a circle at edge of pistil; anthers tan, 0.5- 1 mm long, 0.7-1.2 mm wide, inclined over pis- til; thecae ellipsoid, slightly s pollen white to pale yellow. IN E pen- dent; berries red, ates ra mm long, 4 mm wide. Figs. 179, 181, and 184 Anthurium seleri is known primarily from a relatively small geographic area ranging from eastern Mexico (central Chiapas) to adjacent areas of west central Guatemala in the Department of Huehuetenango at elevations from 700 to 1,900 m. The species has also rarely been collected in southern Oaxaca in the Sierra Madre del Sur, and while these appear to be outlying popula- tons, both regions probably consist of vegetation that is “bosque pino-encino" on the Mexican Vegetation maps, which are generally quite arid. Plants are usually found growing on rocks or on Steep cliffs in shaded canyons but have also been found in open almost deforested areas. The species is a member of section Belolon- chium and is recognized by its coriaceous blades, which are either oblong or narrowly triangular with usually little or no development of posterior lobes or with the posterior lobes directed out- ward almost at a 90? angle from the midrib. An extreme development of the posterior lobes is found in some collections from the northern and Western part of the range of the species and a few collections such as Carlson 1 523, Reko 6203, and MacDougall 686 have posterior lobes of the blade almost as long as the anterior lo Ani interesting feature exhibited by A. seleri i is that while most collections dry matte and green- ish, some collections, such as Croat 46505, dried womi with the surfaces goy Since no other Srenc is assumed that these differences were a matter of differences in drying. Anthurium seleri collections with well devel- Nen Posterior lobes can be confused with col- gus. of A. berriozabalense. See the discussion owing that — for ways of distinguishing the s ES AILS des ted type of A. andresloviense Me ML sin represents a mixed collection. Pecan from MEXU, numbers 59608 and 8 correspond to the type description and to OF MEXICO AND MIDDLE AMERICA 383 the photograph, respectively, presented by Ma- tuda. Another MacDougall collection of 384 has been assigned the number 3844. It is actually A. ovandense. As mentioned in the introductory paragraph, Ma had often recollected what he be- lieved to be the same plants again; when he did, he had the unusual practice of assigning them the same number as he did the first time he col- lected the plant. In this case the plants he re- collected were in fact of some other species, even though they both have similar leaves and were both collected at San Andrés Lovene. Mexico. Cultivated Miami, Florida, origin Mexico, Croat 44760 (MO). CHIAPAS: Aguatenango, Breedlove 6612 (DS, MICH); Arriaga, Breedlove & Thorne 30579 pendencia, Breedlove 33460 (DS); Las Margaritas, Breedlove 34732 (CAS, DS), Ocozocoautla de Espi- nosa, Breedlove 28997 (DS); Breedlove 29187, 32876 (CAS, DS); Carlson 1523 (DS, UC); Croat 40541 (MO); MacDougall H260 (NY); Pueblo Nuevo Solistahua- cán, Thorne & Lathrop s.n. (LL); 40392 (DS, RSA); vicinity Teopisca, Croat 46505 (MO); Breedlove 26201 (C LL, MICH, MO); Venustiano Car- ranza, Laughlin 1 1 947 (CAS, DS); Breedlove 10564 (CAS, DS, MICH); vieil T (CAS, DS, EL); eee oe 20104 (CA S, LL). OAXACA: ochutla, Reko 6204 (GH); ie ud Lovene, MacDougall 384 D. Anthurium beth cea Standl. & Steyerm., Publ. Field Mus. Nat. Hist., Bot. Ser. 22: 69. 1940: Type: Guatemala. Alta Verapaz: near Tactic, elev. ca. 1,500 m, Standley 70505 (F, holotype). Epiphytic or usually terrestrial to 120 cm tall; stems green, 2 cm diam.; internodes 1—2 cm long; leaf scars 1.4 cm wide; roots medium thick, pale green, directed downward; cataphylls subcoria- ceous, 6-13 cm long, the apex round with a mi- nute apiculum, drying tan (B & K Yellow-red 9/ 10), weathering to linear fibers and persisting. LEAVES usually erect; petioles 23-54 cm long, 5-6 mm diam., terete; geniculum 1.5-2 cm long; blades ovate, moderately thin, gradually acu- minate at apex, lobed at base, 22-48 cm long, 13-31 cm wide, broadest just below middle; an- terior lobe 17-33 cm long, the margins convex; posterior lobes 8.2715 cm long, directed upward; the sinus hippocrepiform to spathulate, round to acute at apex; both surfaces matte (cells on lower surface glistening and weakly papillate); the mid- rib sharply raised above, diminishing at center, FiGures 184-187. N. E Brown, Croat 44309. ANNALS OF THE MISSOURI BOTANICAL GARDEN 184. Anthurium seleri Engl., Croat 44760.—185. HEVE silvigaudens St Anthurium spectabile Schott, Croat 1-26-78N.—187. Anthurium spat [Vot. 70 andl. & hiphla! 1983] CROAT sunken at apex, bluntly raised below; basal veins 4-7 pairs, the first and second (sometimes third) free to base, the third to seventh coalesced 1-3.5 cm; posterior ribs naked, upturned; primary lat- eral veins 4—6 per side, departing midrib at 45°— 55? angle, sunken in valleys above, raised below, straight or arching to collective vein; lesser veins prominulous above and below; collective vein arising from first or second basal vein, 4-10 mm from margin. INFLORESCENCE erect, much shorter than leaves; peduncle 3.5—10 cm long, 6— 7 mm diam., terete, shorter than petioles; spathe subcoriaceous, medium green (B & K Yellow- green 7/10), ovate, 5.5-8.5 cm long, 4.3-6 cm wide, broadest just above point of attachment, abruptly acuminate at apex, cordate at base, in- serted at 40? angle on peduncle; the stipe 7-15 mm long in front, 4 mm long in back, tapered at base of spadix, medium green; spadix medium green (B & K Yellow-green 7/2.5), 2.5-5.2 cm long, 1.4-2 cm diam. at base, 5-8 mm diam. at apex; flowers rhombic to sub-4-lobed, 3.5-4 mm long, 3.6-4 mm wide, the sides sigmoid to jag- gedly sigmoid, sometimes broadly rounded par- allel to the spirals; 5 flowers visible in the prin- cipal spiral, 7 flowers visible in the alternate spiral; tepals matte to semiglossy, lateral tepals 2.2-2.5 mm wide, the inner margin broadly rounded and turned up against the pistil; the pistil emergent to ca. 0 mens emerging throughout before the alternates cinerge, the third and fourth stamens following *Xserted, 0.5 mm long, translucent; anthers pale Yellow, retracted to upper edge of tepals, 0.7-0.8 mm long, | mm wide; thecae oblong-ellipsoid, slightly E pollen creamy white. IN- FRUCTESCENCE erect; spathe persistent; spa- dix 4-6 cm $C 2.5—4 cm diam.; berries ob- ovoid, vermillion-coral (see Steyermark 51703), rounded at apex, 11-13 mm long, 6-8 mm diam.; 7 1s PA oblong-obovoid, weakly flattened, 5.5- E. mm long, 3.5-4 mm wide, 2-2.5 mm thick. igs. 182, 183, and 185. Anthurium «lv; J, f di Guatemala * vicinity of Purulhá and Tactic as well as in th the * Department of Huehuetenango in the Sierra OF MEXICO AND MIDDLE AMERICA 385 de los Cuchumantes. It is certainly expected to be found in the Department of Quiché as well. In Honduras it was collected near Lepaera. The species is characterized by its moderately thin, ovate, prominently lobed leaf blades with tionship with that species except for having a lade of similar shape. It is unlikely that the species is even in section Ca/omystrium. It is apparently not closely related to any other species in Central America. GUATEMALA. ALTA VERAPAZ: road to El Estor, E of Highway 14 to Cobán, pios 41440, iila (MOJ; M cinity Tactic, Standley 70505 (F). BAJA VERAPAZ other location, Molina & Molin na 27728 (P: Krk age 1825 (US); NE of Purulhá, Croat 41306 (MO); of Purulhá, Croat 41173, 41753 (MO). HUEHUE- TENANGO: E of Las Palmas, Sierra de los Cuchumantes, Steyermark 51703 (F) HONDURAS. LEMPIRA: vicinity Lepaera, Molina 24135 Anthurium spathiphyllum N. E. Brown, Gard. Ch 52. 1877. Type: Origin unknown: Hort. Kew, Nov. 8, 1876, N. E. Brown s.n. (K). Epiphyte; stems ca. 10 cm long, ca. 2 cm diam.; covered by numerous, moderately thick, fuzzy roots; cataphylls chartaceous, 7-8 cm long, cus- pidate at apex with long, subapical apiculum ca. 1.5 cm long, pale tan (B & K Yellow 9/7.5), weathering into coarse, longitudinal fibers and persisting. LEAVES spreading; petioles usually triangular or sharply 3-5 ribbed; geniculum 7- 10 cm long, conspicuously bulging adaxially; blades linear to linear-lanceolate, subcoriaceous, 46-80 cm long, 5.5-12.5 cm wide, broadest at middle, the margins broadly undulate; both sur- faces semiglossy; the midrib scarcely raised and broad at base, narrowing and convexly raised at middle, acute nearly to apex, sunken at apex above, sharply raised below; primary lateral veins 20-30 per side, departing midrib at 40°-45° an- gle, straight to collective vein, sunken above, raised below; interprimary veins flat to scarcely sunken above, prominulous below; collective vein 386 arising from near the base, 3-6 mm from the margin, sunken above, raised below. INFLO- RESCENCE spreading, shorter than leaves; pe- duncle flattened, 2-4 ribbed, 33-48 cm long, 1- 1.5 cm wide, 2 times longer than petioles; spathe coriaceous, green, lanceolate to broadly lanceo- late 6.5—8 cm long, 1.7-3 cm wide, broadest be- low middle, short acuminate at apex, rounded at base, inserted at 50° angle on peduncle; spadix pale yellow to creamy white, (1.7)3-4 cm long, 0.9-1.1 cm diam. midway; flowers 4-lobed, ca. 1.4 mm long, 2.2 mm wide, the sides sigmoid; ca. 12 flowers visible in principal spiral, 15 flow- ers visible in the alternate spiral; tepals matte, mens scattered throughout before alternates ap- pear, exserted on fleshy filaments, held over and obscuring pistil; thecae oblong, not divaricate; pollen pale yellow fading to cream. INFRUC- TESCENCE unknown. Fig. 187. Anthurium spathiphyllum is known from Cos- ta Rica to Panama from sea level to 380 m (with the exception of Croat 35571, which was col- lected at 1,350 m elevation) on the Atlantic slope in wetter parts of tropical moist forest and in premontane wet and tropical wet forest life zones. The species can be recognized by its epiphytic habit, chartaceous leaves, usually triangular, sometimes 3-5 ribbed petiole, lanceolate to broadly lanceolate cucullate spathe, and short, scarcely tapered, pale yellow spadix. Anthurium spathiphyllum is in section Pachy- neurium and is most closely related to 4. cunea- S W of Tortuguero, Davidson & Donahue 8954 (MO): Pandora, Ocampo 1987 (MO): W of Dos Bocas, beside Rio Parismina, Lent 2444 (F). NICARAGUA. ZELAYA: N of Limbaika, Stevens 8255 (MO); between Rosita and Puerto Cabezas, SW of Rio Kukalaya, Stevens 8504 (MO). ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 Anthurium spectabile Schott, Oesterr. Bot. Z. 8: 181. 1858. Type: Costa Rica. Cartago: Vol- can Turrialba, Wendland 506 (GOET, 2 sheets). Anthurium longispathum Schott, Prod. Aroid. 520. 1860. Type: Costa Rica. Turrialba, Oersted (Pho- to of Schott Aroid Drawing #707, NYBG Neg. #N.S. 3952). Epiphyte; cataphylls subcoriaceous, 17-38 em long, obtuse at apex with subapical apiculum ca. 2 mm long, drying light tan (B & K Yellow-red 9/10), splitting at base, the apex remaining intact, persisting. LEAVES spreading; petioles quadran- gular, 14-53 cm long, 7-10 mm wide, br sulcate adaxially, with a faint convex rib abax- ially; geniculum 1.3-1.9 cm long; blades oblong: lanceolate to oblong or oblong-triangular, m erately thick, long-acuminate at apex, ro to truncate at base, 37-135(160) cm long, © 35(45) cm wide; the upper surface semiglossy o en at apex, convexly raised below becoming mor acute at apex; primary lateral veins numerous departing midrib at 45? angle, raised above - be one half of the blade, sunken above, raised bes 2-3 mm from margin. INFLORESCEN spreading, shorter than leaves; the peduncle 25 cm : rowly ovate to oblong-lanceolate, green or Vi purple inside, green outside, 10.5-3 ed; spa spiral, ca. 9 flowers visible in the alternate tepals matte to semiglossy, 2-2.5 mm lone 2.7 mm wide, the inner margins convex: " da. emergent, green, violet-purple at apex; ac 1 mm long, elliptic; stamens emerging jete Se- throughout length of spadix in a comp Ero quence; anthers held around dem cae narrowly ovoid, slightly divarice™: waaa INFRUCT ESCENCE with s dix to 50 cm long; berries oblong-in®™ ^ 3.5 ably orange at maturity, 10-14 mm long, 7,34 mm diam., tapered to a blunt apex; seeds ^ oe — (o promptly 7 AMI prob — B, cbe a agp me CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA HONDURAS wt 88. dietus siandien m & Baker (Cultivated at KEW). — 189. Anthurium stand- a los 43431. . Anthur subcordatum ssp. aparte it (Standl. & L. O. Wms » Allen et al. 6075 (Type).— th AUAM PAR gea ener mek Schott, Croat 42: ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo.. 70 m Schott |, CF FIGURES 192-195. 192. Anthurium subovat Croat 36455.— 194. Anthuriur 47089. — "atum Matuda, Croat 39815.— 193. Anthurium subsignalu n subsignatum Schott (Cultivated at KEW).— 195. Anthurium tenerum Eng e a o » m P A P X c x — S Ss O Tl z m ^ O O > Z J z Z2 Jg E m > < m a O > Ficures 196-199. 196. Anthurium subcordatum Schott, Croat 42500.—197. Anth urium spectabile Schott, Croat 46968.—198. Anthurium sub- ovatum Matuda, Croat 48066.—199. Anthurium testaceum Croat & Baker, Croat 26646 390 .6 mm long, ca. 1.6 mm wide with minute, triangular appendage at base. Figs. 186 and 197. Anthurium spectabile occurs in Costa Rica at 500 to 900 m in premontane wet forest and trop- ical wet forest. The species is recognized by its elongate, pen- dent leaf blades with numerous, prominent, pri- mary lateral veins supported on an erect, 4-sided petiole. Also distinctive is the inflorescence with its short peduncle, large, narrowly ovate, often violet-purple spathe and yellow spadix. Anthu- rium spectabile has been placed in section Pachy- neurium Anthurium spectabile is easily confused with A. ranchoanum, which has similar leaves (es- pecially on dried herbarium material) and a sim- ilar Anthurium , how- ever, has a subterete petiole and a leaf blade held more or less erect or in line with the petiole. Anthurium spectabile also has an elongate, more or less spreading spathe, while the spathe of A. ranchoanum is shorter and usually stiffly erect. pisi: MODO cbe. Z ar to A. pseu- dospectabile (Croat ined.) from Panama, but the latter has longer blades with more undulate margins, terete petioles, and occurs in premon- tane rain forest. Costa RICA. ALAJUELA: Cariblanco, Lent 3543 (F); NE of Villa Quesada, Croat 46968 (MO); Molina et al. 17287 (F). LIMON: Guapiles, Smith 4982 (US); S of Siguirres, Croat 43325 (MO) Anthurium standleyi Croat & Baker, Brenesia 16 (Supl. 1): 88. 1979. Type: Costa Rica. San José: ca. 5 miles SW of Canaan along gravel road from Rivas; steep slopes, ca. 900 m elev. 14 August 1977, Croat 43439 (MO- 2582987, holotype; CR, F, K, NY, PMA, SEL, US, VEN, isotypes; Live at MO). Terrestrial, ca. 1.75 m tall; stems to 90 cm long, 3-5 cm diam.; internodes ca. 1 cm long; leaf scars 2—3 cm wide; roots dense, tan, fuzzy, descending, ca. 1.2 cm thick at stem, tapering slightly; cata- phylls curled into a tubular cylinder, 15-30 cm long, somewhat coriaceous, short-acuminate and apiculate at apex, the apiculum ca. 1.5 cm below apex of cataphyll, drying tan (B & K Red 9/ 10), weathering into long, subpersistent fibers. LEAVES with petioles erect to spreading, 43— 120 cm long, 7-16 mm diam., terete; geniculum 2-4.5 cm long; blades ovate, moderately thin, obtuse to usually short-acuminate at apex, lobed ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 at base, 40-75 cm long, 25-56 cm wide, broadest at point of petiole attachment or slightly above; anterior lobe 33-55 cm long, the margins broadly convex; posterior lobes 8-21 cm long; sinus spathulate to obovate, often overlapping, the in- ner edges turned up, acute at apex; upper suriace semiglossy to glossy, lower surface semiglossy, vu mrs uM mute timinishintil ward apex, convexly raised below; basal veins 4-6 pairs, 4 or 5 of them coalesced 1-4 cm, raised above and below; posterior rib slightly curved, naked, the outer margin sharp and often rolled up; primary lateral veins 5-9 per side, departing midrib at 45°-65° angle, straight then arcuale- ascending, usually only the uppermost loop-con- necting, convexly raised above and below; 1n- terprimary veins obscure or occasionally raised in valleys above, raised below; secondary veins slightly sunken in posterior lobes above, flat and lective vein - | em rers | * | 1 th SAS tLy Vai nw arising from one of the primary lateral veins In the upper half of the leaf, 4-11 mm from margin. slightly sunken above, raised below. INFLO- RESCENCE erect to spreading, much shorter than the leaves; peduncle 29-76 cm long, 4-9 mm diam., terete; spathe moderately thick, gre? (B & K Yellow-green 6/7.5), oblong-lanceolil 10.5-18 cm long, 2-4 cm wide, broadest Ca. cm above point of attachment, narrowly pi minate at apex, obtuse to rounded at base, ™ serted at 60° angle on peduncle; stipe 1 1-70 mm long in front, 6-60 mm long in back, ca. 5 p wide, medium green; spadix green before an : sis, soon heavily tinged with violet-purple, ja ing yellowish-brown, 10-30 cm long, 7-108 a diam. at base, 4-6 mm diam. at apex e rhombic to weakly 4-lobed, 3.5-3.6 mm 77 2.9-3.3 mm wide, the sides nearly straight P^ rpendicult princ! spiral at droplets, lateral tepals 1.9-2 mm wide, the ! margin broadly rounded; 4-0.7 mm medium green; the stigma linear, 0.40. on long, a slight slit briefly filling with pare days before first stamens appear, dry and Pu darkened when stamens emerge; j] 4 sta ning to emerge from or near " Bet he held i i i ter 2 E mens in a spiral opening soon a acuta pistil: 0.6-1 m7 ange, fading creamy whi j 1983] CENCE arching; spathe reflexed; spadix to 30 cm long, 2 cm wide; berries broadly ovoid, rounded at apex, orange; mesocarp pulpy, bitter tasting; seeds 2, white, to 5 mm long. Figs. 188 and 189 The species is in section Pachyneurium and is known for certain only from Costa Rica in the Cordillera de Talamanca. It was first collected by Paul Standley at two sites, La Verbena and San Sebastian south of San José, in what is prob- ably premontane wet forest. It has been collected more recently in the region northeast of San Is- idro del General. The species ranges from 900 to 1,900 m elevation and may be restricted to premontane wet forest. A cultivated collection from the University of Hawaii was reportedly based on a Dressler col- lection from Achiote, a region of tropical moist forest in the Isthmus of Panama. The species has never been collected in Panama to our knowl- edge, and the collection cultivated in Hawaii is probably mislabelled. Costa RICA. SAN José: Cerro Tapezco, Jimenez 627 pu %2 mile above turnoff to Canaan at Rivas, Croat an (MO); 43439 (CR, F, K, MO, M S, VEN); vicinity La Verbena, Standley 32243 (US); Vicinity San Sebastian, Standley 49296 (US); along n p of Río Buena Vista, S of Buena Vista, NE of San p ro del General, Luteyn 3287 (DUKE); Río Her- : ura, NW of Canaán, Burger a Ted 7093 (MO, S); Talamanca, Tondus 8721 Anthurium subcordatum Schott, Oesterr. Bot. Z. 8: 181. 1858 a. Anthurium subcordatum ssp. subcordatum. TYPE: Guatemala. Quetzaltenango: Las Nubes on slopes of Volcán Zunil, 8,000’, Wendland 330 (GOET). Anthurium MESS dae Standl. ex we Field Mus. Nat. Hist., Bot. Ser. 17:316, Pl. 4. 1938. ridge nie El ie . Yuncker et al. Ens (F, holotype). _ m coibioni Standl. erm., Publ. Field Mus. Nat. Hist., Bot. Ser. 23: 209. 1947. TYPE: ci Zaca a: Sierra de Las Minas, Stey- ermark 42212 (F, holotype; US, isotype). Type: js ond uras. jeepers ulate at apex, drying light brown, dilacerating at a ' Persisting; LEAVES spreading; petioles TPly and shallowly sulcate, 16-44 cm long, OF MEXICO AND MIDDLE AMERICA 391 6-9 mm diam.; geniculum 1.5-2.2 cm long; blades ovate to narrowly triangular, moderately thick, gradually acuminate at apex, rounded, truncate or shallowly lobed at base, 28-62 cm long, 10.5—25 cm wide, broadest well above base; anterior lobe 28-40 cm long, the margins + straight; posterior lobes 4—8 cm long; sinus ar- cuate, to parabolic, rarely hippocrepiform, rounded at apex; both surfaces semiglossy, the lower surface sometimes glaucous; the midrib convexly raised above and below, sunken toward apex above; basal veins 2-3 pairs, usually free rt above; intermediate and secondary veins sunken above; tertia , flat; col eae arising ERE the first basal vein, ne 1 mm from margin. I RE NCE erect-spread- ing, shorter than leaves; peduncle 10—38 cm long, fused with violet-purple (B & K Yellow-green 7/ 10), lanceolate to narrowly ovate, 2.5-7.6 cm long, 1.4-2.7 cm wide, abruptly acuminate at apex, truncate to rounded at base, inserted at 30° angle on peduncle; stipe to 17 mm; spadix dark violet-purple or green heavily tinged violet-pur- ple, 3.5-13.5 cm long, 5.5-9 mm diam. at base, 3-6 mm diam. at apex; flowers rhombic to sub- 4-lobed, 2.5-3.5 mm long, 3-3.5 mm wide, the sides straight to sigmoid, difficult to discern; 5— 7 flowers visible in the principal spiral, 8-11 Fera visible in the alternate spiral; tepals matte to semiglossy, minutely papillate, lateral tepals 1.3-1.8 mm wide, the inner margins + straight; pistils emergent, green; stigma oblong-elliptic to linear, 0.4—0.9 mm long; stamens emerging in a moderately rapid sequence from the base, the laterals emerged from one fifth to two thirds the length of spadix before third to fourth emerge at base, held above and obscuring pistil then re- tracting and opening; anthers tan or orange- own, | mm long, 0.8-12 mm wide; thecae ob- long-ellipsoid, scarcely divaricate; pollen cream to yellow. INFRUCTESCENCE erect to spread- ing; berries obovoid, red, ca. 1 cm long, 1.2 cm wide; seeds 2, ca. 6 mm long, 5.5 mm wide. Figs. 191 and 196. The species ranges from western Guatemala in the Departments of Huehuetenango, Quiché and Alta Verapaz to northern El Salvador and western Honduras at elevations from 1,100 to 392 2,800 m. The species occurs in moist and wet forest. Anthurium subcordatum is in section Belolon- chium and, like a number of Mexican and Cen- tral American species, is quite variable, charac- terized by its thick, truncate to subcordate leaf blades, sharply sulcate petioles, violet-purple spadix (or sometimes green heavily tinged violet- purple), and bright red, ovoid berries. The species is probably most closely related to A. subovatum from the Sierra de Juárez in northern Oaxaca (Mexico) but differs in having the blade generally broadest well above the base (even on cordate blades as the posterior lobes are turned inward) with most of the tertiary veins obscure, in having the basal veins usually fewer in number (1-3) and usually widespread, sometimes having the second basal vein running to the apex or to the margin well above the middle, Anthurium sub- ovatum has 3-4 basal veins, usually tinged with red and these are generally directed toward the apex at a more acute angle than those of A. sub- cordatum. In addition, the berries of A. subcor- datum are bright red whereas those of A. sub- ovatum are orange. In his discussion of A. subcordatum in Flora of Guatemala, Steyermark (Standley & Steyermark, 1958) referred to A. quinquinervium Kunth of South America as a close relative and believed that the two species might be synonymous. Although I have not seen live material of the latter species, it is unlikely that the two are synonymous. Anthurium subcordatum consists of two sub- species. Subspecies subcordatum extends throughout the range of the species. Subspecies chlorocardium is endemic to the slopes of Cerro Santa Bárbara (Dept. of Santa Bárbara) in west- ern Honduras. Populations of A. subcordatum from south- western Honduras were described as A. basiat- tenuatum by Standley based on the more acute leaf base, but they are well within the expected range of variation of ssp. subcordatum and are connected to the more typical populations from El Salvador and Guatemala by another Hon- duran collection (Molina 22322). It is somewhat unusual that both extremes of from A. basiattenuatum, with well developed posterior lobes. VADOR. CHALATENANGO: E slope of Los Eses- miles, "Tucker 1048 (F, UC, US). SANTA ANA: Cerro ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 Montecristo, Molina et al. 16836, 16914 (F, NY, US); Wilbur et al. 16358 (DUKE, MO); Croat 42395, 423%, 42500 (MO); Molina & Molina 12680 (F); Cerro Mi- ramundo, Carlson 940 (F) UATEMALA. ALTA VERAPAZ: W of San Cristobal, Standley 897 50 (F); San Juan Chamelco, Wilson 40926 (F). CHIQUIMULA: near Aur of El e Steyermark 30935 (F). EL PROG as Minas, N of Finca Piamonte, Steermark 43608. (F, NY, US). HUEHUETENANGO: San Juan Ixcoy, Steyermark 50017 (F); 50041 (F, NY); pat Mateo Ixtatan, Breedlove 8786 a, Steyermark 3281 4 (F). QUEZALTENANGO: enango, Standley 86557 (F). QUICHE: e pect heh ux ther location, Aguilar 1276 (F); Hye (GH, K, US); Nebaj, Skutch m (F, GH). ZACAP i Río Sitio Nuevo, near Santa Rosalía, Steye rma (F, US); Sierra de Las Minas, Steyer mii 29877, 42 (F). HONDURAS. COMAYAGUA: El d Yuncker et b: 6268 (F, GH, MICH, MO); 6272 (F, GH. e OCOTEPEQUE: vicinity Nueva Ocotepeque, Molin 22322 (F). b. Anthurium subcordatum ssp. chlorocardium A. chlorocardium Standl. & L. Ceiba 3: 103. 1952. TYPE: Honduras. p Bárbara: upper rocky slopes and E Cerro Santa Bárbara, alt. 2,750 m, April 6, 1951, Allen et al. 6075 (US, F). to eA es fibers. LEAVES w P 44 cm long, 6-8 mm diam.; iens ovate, w erately thick, 26-37 cm long, 18-24 cm i ko broadest just below middle, acute at apex, ? at base; the anterior lobe 24— m lon terior lobes 6.5-8 cm long; the sinus e round at apex; the midrib drying + flat a : raised below; basal veins 3—5 pairs; wes eral veins 6-8 per side, drying raised on p surfaces; lesser veins drying prominulous on 3 surfaces; collective vein arising fro m the ve second basal vein. INFLOR ESCENCE vi duncle 13.5-39.5 cm long, 4-5 m diam.; i green, thick, 5-12 cm long, 1.5-6.5 a p narrowly ovate to ovate, e cst a cordate at base, broadest just above base 4-12 cm long, 1.5-2 cm diam.; flowers ioni 5-6 mm long; 4.5-6 mm wide, 6-7 fov ible in the principal spiral, 6-8 flowers e : the alternate spiral; lateral tepals l. 6 wide; pistils raised; stigmas — peru on fi long, drying brown; stamens € aper thecat ments ca. 0.5 mm long, ca. 0.9 mm ellipsoid. INFRUCTESCENCE wi Ee cutis oi Taaka A EE | -— Low 1983] berries, ca. 6 mm diam., immature berries yel- low. Fig. 190. Anthurium subcordatum ssp. chlorocardium is endemic to Honduras on Cerro Santa Bárbara (Department of Santa Bárbara) at elevations of 2,200 to 2,750 m. Subspecies ch/orocardium is a member of sec- tion Belolonchium and is distinguished by its narrowly ovate to ovate spathe and its geograph- ical isolation. Honpuras. SANTA BARBARA: summit of Cerro Santa Barbara, Allen et al. 6075 (F, US); E of Lake Yojoa, Clewell & Hazlett 3916 (MO, US). Anthurium subovatum Matuda, Bol. Soc. Bot. México 24: 36. 1959. Type: Mexico. Oa- xaca: road between Cerro Pelón and Yetla, elev. 1,900 m, January 12, 1959, Mac- Dougall 406 (MEXU, holotype). Terrestrial or epiphytic; stem to 2.5 cm diam.; leaf scars ca. 2.3 cm wide; ro numerous, ling at base, remaining intact at apex. LEAVES with petioles erect-spreading, 20-51 cm long, 5— 7 mm diam., shallowly and broadly sulcate, the margins sharp, rounded abaxially; geniculum 1- ? cm long, usually tinged red-violet; blade nar- rowly ovate-triangular to ovate, moderately thick, acute and apiculate at apex (acumen 2-5 mm long), weakly to prominently lobed at base, 27- ^4 cm long, 10-26 cm wide, broadest at base or Ken above point of petiole attachment; anterior 56, narrowing and sunken at apex, convexly dit almost to apex below; basal veins 3—5 pairs, e first free, second to fifth coalesced 1-3 cm; Posterior rib naked; primary lateral veins 3-7 Mei departing midrib at 30°-50° angle, loop- necting to collective vein, sunken above, “a below; secondary and tertiary veins clearly ri "a weakly sunken above, raised below; mid- re idi asal, and primary lateral veins usually tinged violet on lower surface, sometimes faintly CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 393 tinged on upper surface; collective vein usually arising from the first or second basal vein, sunken above, raised below, 4-7 mm from the margin. INFLORESCENCE spreading to pendent, ca. as long as leaves; peduncle 28-49 cm long, 4-6 mm diam., sometimes weakly glaucous or tinged red- violet; spathe moderately thick, lanceolate, green sometimes suffused with red-violet, matte on both surfaces, 6-9.5 cm long, 1.5—2.7 cm wide, grad- ually acuminate at apex, rounded to obtuse at base, inserted at 40°-50° angle on peduncle; spa- dix green turning dark violet-purple, 10.3-19 cm long, 7211 mm diam. at base, 3-4 mm diam. at apex; flowers rhombic to 4-lobed, 3.7-4.2 mm long, 3.2-3.9 mm wide, the sides weakly to jag- gedly sigmoid; 5-7 flowers visible in the principal spiral, 6—12 flowers visible in the alternate spiral; tepals matte, minutely punctate, with sparse droplets as stamens emerge, lateral tepals 1.5— 2.6 mm wide, the inner margin broadly rounded, usually turned up against pistil; pistil green, tinged violet-purple, weakly emergent; stigma linear, ca. 0.8 mm long; lateral stamens emerging slowly from the base, one at a time, the alternates fol- lowing by 1-3 spirals; faint yeasty odor as sta- mens emerge; anthers creamy white, held in tight circle, retracting slightly as pollen disperses; the- cae ovoid; pollen pale yellow to creamy white. UCTESCENCE spreading-pendent; spathe green, persistent; berries bright orange, some- times developing only in the basal one half, broadly obovate, rounded to truncate at apex, 6-9 mm long, 7-9 mm wide; mesocarp thick, juicy with numerous raphide cells; seeds | or 2, pale green, obovoid-ellipsoid, weakly flattened, 4.5-5.6 mm long, 3.5-5 mm wide, 2.8-3 mm thick, with sticky appendages at both ends. Figs. 192 and Anthurium subovatum is restricted to Mexico in the Sierra de Juárez of northern Oaxaca at 1,500 to 2,100 m in cloud forests in what appears to be a premontane rain forest life zone. The species is a member of section Belolon- chium and is characterized by its sharply sulcate petioles, thick subcordate to truncate leaf blades with sharply ascending reddish basal veins, nu- merous clearly visible tertiary veins, violet-pur- ple spadix, and obovoid, more or less truncate, bright orange berries. It is similar to A. subcor- datum in having a similar leaf shape, but the blades of A. subovatum are generally broadest at the base with the posterior lobes often directed 394 1 somewhat upward on well d ped plants. An- thurium subovatum also generally has the spadix sessile or nearly so and has orange berries, where- as the spadix of A. subcordatum is often con- spicuously stipitate and the berries are red. For additional characters see the discussion follow- ing A. subcordatum. Anthurium subovatum is also similar to A. nakamurae, a species apparently endemic to eastern Chiapas. That species differs in having blad ypically oblong and usually broader toward the apex. Considering the variability in Anthurium it is possible that A. subcordatum, A. subovatum and A. nakamurae may ultimately be proven to be no more than subspecifically dis- tinct. MEXICO. OAXACA: highway between Cerro Pelón and Yetla, MacDougall 406 (MEXU); N of Ixtlan de Juárez on the road to Tuxtepec, Rzedowski 30615 (MO); along road between Valle Nacional and Oaxaca, McAlpin 1047 (DUKE); between Tuxtepec and Oaxaca at Valle Nacional, Croat 39815, 39847, 48066, 48082 (MO); S of Valle Nacional, Rzedowski 33574 (ENCB). Anthurium subsignatum Schott, Bonplandia 9: . 1862. Type: Costa Rica. Limón: Ped- regal, Wendland 919 (GOET). Epiphytic creeper; stems green, 10 cm or more long; internodes 1.5—11 cm long; roots pale green, 2-4 mm diam.; cataphylls subcoriaceous, 3.5— 10 cm long, rounded at apex, with thick sub- apical apiculum ca. 4 mm long, drying reddish brown, weathering into longitudinal fibers, per- sisting at nodes. LEAVES with petioles spread- ing, 14.5-56 cm long, 3-7 mm diam., terete, sometimes sharply and narrowly sulcate; genicu- lum 1.5-3 cm long, obscurely sulcate, blades ovate-deltoid, short-acuminate at apex, broadly lobed to arcuate at base, 16-48 cm long, 11-52 cm wide, broadest at base; the anterior lobe 1353- 38 cm long, the margins + straight to convex; posterior lobes 4-21 cm long; sinus arcuate to parabolic, obtuse to rounded at apex; upper sur- face matte to semiglossy, lower surface semi- lowermost primary lateral veins, sunken above prominulous below, 3-5 mm from margin. IN- ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 FLORESCENCE erect, shorter than leaves; the duncle 7.5-28 cm long, 3—6 mm diam.; spathe moderately thin, yellow-green, oblong-lanceo- te, 5-25 cm .3-4 cm wide, long-acu- minate at apex, truncate at base, inserted at 45" angle on peduncle; spadix yellow to yellow-green, 5.5-26 cm long, 6-9 mm diam. at base, 3-5 mm diam. at apex; flowers 4-lobed, ca. 2 mm long, 2.2 mm wide, the sides jaggedly sigmoid; 4-7 flowers visible in the principal spiral, ca. 9 flow- ers visible in the alternate spiral; tepals glossy, lateral tepals ca. 1 mm wide, the inner margin convex; pistil minutely emergent, white, glossy; stigma ca. 2 mm long, brushlike, with very small droplets for 5—6 days about 7-8 days before first stamens emerge, dry, white to tan when stamens emerge; stamens emerging from base in mod- erately rapid sequence, the second lateral stamen emerging 1 or 2 days after first stamen; anthers yellow-orange; pollen pale yellow-orange. n FRUCTESCENCE with spadix to 30 cm long berries obovoid, violet-purple at apex, white al base, acute at apex, ca. 5 mm long; seeds 1-2, white, 1.5 mm long. Figs. 193 and 194. o north- levation rest life Anthurium subsignatum is endemic t ern Costa Rica at sea level to 200 m € in tropical wet and premontane wet fo zones. This species is a member of section Se phyllium and was mistakenly included as . onym of A. trilobum André in “Anthur pe st Costa Rica” (Croat & Baker, 1979). It 15€ ji j related to A. garagaranum Standl. of api and Colombia, which has a similar bright YF: n (1978) included A. garagaranum as a e "i of A. trilobum with admitted misgivings, ditional field work is needed 1n outh Chilamate, Leon 1784 cus) Bape Grove 34 (MO); Croat 4423 ; i along road near Río Sucio, Croat 357 oon Tra icinity Guapiles, Standley 37458 (US); ‘pill dson & pezco-Hda. La Suerte, W of Tortuguero, Donahue 8366, 8654 (MO). ICML y t. 25: i t. Jahrb. SY Anthurium tenerum Engl., Bo ;oquidt 377. 1898. Type: Colombia. Ant! dense hi" ro El Plateado in western Andes, "n 735 4 mid forest, 1,800-2,200 m, Lehma , ‘ | | CROAT LARVAL 1983] Scandent epiphyte; stems elongate with inter- nodes 1.5-7.5 cm long; roots moderately thin, usually 2-4 at each node; ETAO 4—6 cm long, thin, acute at apex, drying tan, persisting intact at nodes. LEAVES Cae petioles 1-3 cm long, 2-3 mm diam., terete, sheath extending 1.5-2 cm onto petiole, minutely free at apex; geniculum absent; blades oblong or lanceolate, moderately thin, 9-15 cm long, 2.5—4 cm wide, broadest at middle or just below, acuminate at apex, obtuse to rounded at base; the upper sur- face semiglossy to glossy, lower surface semi- he . INFLORESCENCE wipes short- er than Eves peduncle 2.5-4 cm long; spathe green, ovate, 1.5—2 cm long, 2-9 mm wide; spa- dix green (B & K Yellow-green 6/2.5), ca. 3 cm long, scarcely tapered at apex; flowers 4-lobed, the sides sigmoid; 4—5 flowers visible in the prin- cipal spiral, 3-4 flowers visible in the alternate spiral; tepals semiglossy, punctate, the inner Margin convex; pistils green, emergent, + mounded; stigma elliptic, brushlike. INFRUC- TESCENCE pendent; spathe absent; spadix to 3.5 cm long; berries subglobose flat at apex, greenish-white i color un- o probably 5 mm diam. at maturity. Fig. Anthurium tenerum is known from Costa Rica and Colombia (probably in Panama as well) at elevations of 800 to 2 ,000 m. In Costa Rica the Species is known from lower montane rain forest at 1,300 m. It is apparently also rare in Colombia since few collections have been made. The species is in section Xialophyllium and is most easily confused with A. scandens because of its similar habit, size, and leaf shape. Anthu- rlum tenerum can be distinguished by its lack of leaf punctations and by its long slender peduncle Which is, at least initially, borne in the conspic- uous petiole sheath. catia CARTAGO: vicinity of Cachi, Croat 47089 y Rio Naranjo, Lent 1583 (NY). HEREDIA: "Skutch 3658 (MICH). Cost (MO); vici Vara os Anthurium testaceum Croat & Baker, ocn 16 (Supl. 1): 92. 1979. Type: Costa Alajuela: Atlantic side of Alto de las Hi mas, 1,900 m, Lent 1820 (F, holotype). Epiphytic or terrestrial; stems elongate, ca. | OF MEXICO AND MIDDLE AMERICA 395 m long, 1-1.5 cm diam., drying irregularly grooved on 1 or more Sides; internodes | .5-2.5 cm long; leaf ca m wide; roots few, scattered at the nodes; cattle moderately thin, 6-15 cm long, green, sometimes tinged purplish, apiculate at apex, drying greenish-yellow to brown, dilacerating and often persisting, ultimately deciduous. LEAVES spreading, scattered in upper part of stem; pet- ioles (8)14—48 cm long, 3-6 mm diam., terete; geniculum 5-15 mm long; blades oblong-lanceo- late to lanceolate or oblong-elliptic, moderately thin, gradually acuminate to cuspidate-acumi- nate at apex (the acumen 1—3 cm long), obtuse to rounded (rarely acute) at base, 18—50 cm long, 4—12.5 cm wide, broadest usually below the mid- dle, the margin + straight; upper surface matte to semiglossy, lower surface matte; midrib weak- ly raised in a valley above, prominently raised below; primary lateral veins 10—16 per side, de- parting midrib at 60°-80° angle, weakly sunken above, raised below, loop-connecting to collec- tive vein; lesser veins obscure; collective vein usually arising from the first to third primary lateral vein, 1-8 mm from margin. INFLORES- CENCE spreading, usually shorter than leaves; peduncle 14—40 cm long, 3-5 mm diam., terete, three quarters to equally as long as petioles; spathe + thin, green, lanceolate, 3-7 cm long, 7-14 mm wide, acuminate at apex, slightly clasp- ing and rounded at base; spreading, inserted at 60° angle on peduncle; stipe 2-8 mm long in front, 1-7 mm long in back, ca. 3 mm diam.; spadix green, sometimes tinged with violet-pur- ple, 4-12 cm long, 4-6 mm diam. at base, 2-4 mm diam. at apex; flowers rhombic, 1.8-4 mm long, 1.5-3 mm wide, the sides straight to slightly jaggedly sigmoid; 2-4 flowers visible in either spiral; tepals matte to semiglossy, densely and minutely papillate, turned up against the emerg- ing pistils, lateral tepals 1.8-5 mm wide, inner margin convex, thin, pale, the alternate pair with inner margins convex; pistils early emergent, green; stigma ca. 0.5 mm long, elliptic; stamens held at or above the tepals on translucent fila- ments 0.5 mm long, 0.3 mm wide, lateral sta- mens developing throughout, followed quickly by peter anthers orange, 0.5-0.8 mm long, ca. 1 mm wide; thecae ellipsoid, slightly divar- icate; k orange fading to yellow. INFRUC- TESCENCE with spadix 7-14 cm long; berries oblong-ellipsoid, rounded at apex, bright red, 9- 13 mm long, 5-7 mm diam., steer exsert- ed before maturity, m gelatinous; seeds 2, oblong-ellipsoid to seas ebd at both — 396 ends, tan, scarcely flattened, 3-7 mm long, 3-4 mm wide, 2.5-3.4 mm thick, an appendage en- veloping the seed, extending beyond both ends and at least one side. Fig. 199. The species is found in Costa Rica and Panama at 800 to 2,000 m elevation in premontane wet and premontane rain forest life zones. In Costa Rica the species ranges from Monteverde in Puntarenas and Alajuela Provinces to San Vito de Java near the Panamanian border. In Panama the species is known only from western Panama near Costa Rica. Anthurium testaceum can be distinguished by its elongate, narrow stem, long, thin, persistent cataphylls, bright red, prematurely exserted ber- ries, few-flowered spadix spirals, and lanceolate leaf with the midrib and primary lateral veins drying very pale tan. The species is probably closest to A. cerropir- rense Croat (ined.) from Darién Province in east- ern Panama, but that species differs in having 5- 7 flowers per spiral and tepals that are not at all elevated. It also has shorter internodes and pur- ple berries. Live material of this species has been confused with specimens of A. pallens. Anthurium pallens differs, however, by its smaller, yellow berries, the midrib and primary lateral veins not signif- icantly different in color from the blade, and the secondary veins more numerous and prominent than in A. testaceum. The veins of A. testaceum probably appear more striking because of the great contrast between the rich green lower leaf surface and the light, almost bleached appear- ance of the veins. In “Anthurium in Costa Rica” (Croat & Baker, 1979) it was suggested that it would best be placed in section Leptanthurium, it is, however, best placed in section Y. ialophyl- lium. Costa Rica. Junction of Alajuela, Guanacaste, Pun- tarenas, Burger et al. 10840 (MO). PUNTARENAS: along Rio Cot s, near Cotán, N of La Unión, Croat 26646 (MO). SAN José: above Rio Cascajal, Lent 4040 O). Anthurium tilaranense Standl., J. Wash. Acad. i. 17: 245. 1927. Type: Costa Rica. Gua- nacaste: Quebrada Serena, southeast of Ti- arán, 700 m, Jan. 1926, Standley & Valerio 46310 (US, holotype). Anthurium latihastatum Engl. ex K. Krause, Notizbl. ot. Gart. Berlin-Dahlem 11: 612. 1932. ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 SvNTYPES: Costa Rica. near San Ramón, 1,300- 1,400 m, May 1913, Tonduz 17708 (not seen; near San Ramón, 1,200 m, Tonduz 17704 (not 1 ] seen). E Epiphyte; stems to 1 m long; internodes short; i cataphylls 7-17 cm long, round at apex with su petioles spreading, (8)16—45 cm long, 4-6 n diam.; blades subcoriaceous, 3-lobed, 20-46 lobed at base; sinus usually arcuate, usual rounded to truncate at apex; both surfaces sem glossy; midrib of the median lobe convexly rà to just below middle, diminished and sunken apex above, prominently and acutely raised low; basal veins 1-3 pairs, 2 of them coalesce raised; primary lateral veins 7-10 per si f parting midrib at 40°-60° angle, flat above, = er edge of posterior lobes to the apex 0 dian lobe, 3-4 mm from the margin. CENCE - erect, shorter than leaves; (7)10-25(30) cm long; spathe subcoriaceous, pale — green, broadly lanceolate to oblong or oblong elliptic, 5.5-11 cm long, 1.7-3 cm wide, abruP. acuminate at apex, acute at base; spadix Á green to white (B & K Yellow-green 8/ 0 14 cm long, ca. 8 mm diam. midway; flowe! rhombic to sub-4-lobed, 2.4-2.8 mm long, E 3.3 mm wide, the sides straight to jagged al moid; 7-8 flowers visible in the principal al ; a ; peduncle i and 5-7 flowers visible in the alternate tepals smooth, glossy, the lateral tepals 1.4-" mm wide, the inner margin obtusely ang id outer margin 5-sided; pistils weakly eme f 0.3 mm; stigma round, 0.2 mm diam., b F B stamens emerging in a rapid progres b ning in the apical half of spadix, d net exserted throughout before the alternates * aja- scarcely exserted above tepals On d: n ments; anthers white, ca. 0.3 mm long, o wide, held close but not contiguous, thecae ellipsoid, slightly divaricate; pollen long CTESCENCE with spadix to 22 €! berries red (at least apically), obovoid, Q nw e » í m. 5 Š m K i & O "1 z m E Q O > Z J = J ie) E tri » z m G 0 > FiGures 200-203. 200. Anthurium tilaranense Standl., Croat 363 18.—201-202. Anthurium tonduzii Engl., Croat 37247.—203. Anthurium trinerve Mia Croat 36357 398 long, 5 mm diam., obtuse at apex; mesocarp clear, gelatinous; seeds 2, greenish, ca. 3.5 mm long, 2.5 mm wide, 1.5 mm thick. Fig. 200 The species is known from Costa Rica and Panama at 400 to 1,500 m elevation in premon- tane wet, tropical wet, and lower montane forest. It is a member of section Semaeophyllium and is not close to any other species in the section. It is distinguished by its deeply 3-lobed leaf blade, with the lateral lobes falcate, persistent cata- phylls, and pale green to white spadix that is slightly tapered at both ends. In Costa Rica it can be confused only with A. subsignatum, which has blades only weakly 3-lobed and a bright yel- low spadix. In Panama it has leaves similar to A. madisonianum Croat (ined.), which has a longer, reddish-violet spadix and prominently dilacerated cataphylls. Costa RICA. ALAJUELA: Cafias-Upala Road, NNE of Bijagua, Croat 36257, 36318 (MO); between Naranjo and Que REDIA; vicin- ity of Vara Blanca, Skutch 3735 (MICH). PUNTARENAS: Monteverde, Jiménez s.n. (MO). Anthurium titanium Standl. & Steyerm., Publ. Field Mus. Nat. Hist, Bot. Ser. 23: 211 1947. Type: Guatemala. San Marcos: south facing slopes of Volcán Tajumulco, above Finca El Porvenir, along Río Cabuz to with- in 2 mi of Cueva de las Palomas, alt. 1,300— 1,500 m, March 16, 1940, Steyermark 37963 (F, holotype). Anthurium tiep ueber Matuda, Revista Soc. Mex. Hist. Nat 1949, Matuda 18644 (MEXU, holotype: DS. isotype ). Epipetric or terrestrial; stems short, 2-4 cm diam.; leaf scars 2.5-2.8 cm wide; roots thick, tan, 5-6 mm diam.; cataphylls subcoriaceous, 4.5-12 cm long, round to weakly emarginate at apex with a ponspicaóus subapical apiculum ca. 2mmlon y 1 (B & K Yellow 4/7.5), pop into coarse fibers. LEAVES with petioles erect- _ 31-90(112) cm long, 5-10(15) mm diam., D-sha ped, sometimes weakly sulcate and glaucous geniculum 1.5-4 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 cm long; blades moderately thick, ovate to broadly ovate or nearly round, abruptly to grad- ually acuminate at apex (the acumen apiculate 2-7 mm long), deeply lobed at base, 27-120 cm long, 29-80 cm wide, broadest at middle or just below, the margins sometimes markedly undu- late, weakly revolute; anterior lobe 19-100 cm long, the margins broadly rounded; posterior obes 10-40 cm long; sinus obovate to hippo- crepiform, acute to rounded at apex; both sur- faces semiglossy; midrib convexly raised above, more acutely raised below, narrowing and minishing, + flat at apex above; basal veins 1- 10 pairs, the first and second free to base, those remaining coalesced 2—4(7) cm, raised above and below; posterior ribs conspicuously upturned, naked; primary lateral veins 3-5 per side, de- parting midrib at 45° angle, + straight to collec- tive vein, loop-connecting to collective vein, weakly raised above and below; interpri veins conspicuous, flat above and below, lesser veins visible; collective vein arising from the 5 of the primary lateral veins in the la weakly sunken above, raised below, 7-15 mm from margin. INFLORESCENCE erect-spread- ing, as long as or hee than leaves; Pe duncle 26-47 cm long, 7-15 mm a terete, spathe subcoriaceous, green long-acuminate at apex, oe at base, inserted at 25°-30° i peduncle, held at ca. 45 M from We s back; spadix olive-green (B & K 5/5), sometimes weakly to heavily violet, (6)1 1—24 cm long, 1-3 cm 5-10 mm diam. at apex; flowers rh d 4-lobed, 3.7-5 mm long, 3.8-4 mm W! sis d sides + straight to weakly sigmoid; 7-10 4 ae visible in the principal spiral, 6-13 flowe utely ible in the alternate spiral; tepals pe min papillate, lateral tepals 1.8-2.9 m inner margin broadly rounded, thin, "pale 07 pistil green, emergent, raised; st igma lineal, 3 spirals before first emerges in next ane over pistil, retracting to edge of tepals € opening; filaments weakly exsert as wide as anthers; anthers pale stow, 0941 mm long, 1.2-1.4 mm wide; thecae € wide, he -— c 1983] CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 399 & Steyerm., Croat 47253.—206. Anthurium ,, FIGURES 204-207. 204-205. Anthurium titanium Standl. (Tineru, : z ‘ ~ J , & inerve Miq., Croat 34371.—207. Anthurium trisectum Sodiro, Croat 25633 400 scarcely divaricate; pollen cream. INFRUCTES- NCE probably with orange berries. Figs. 204 05. Anthurium titanium is endemic to a small area in southeastern Chiapas in Mexico and adjacent western Guatemala in the Department of San Marcos from 900 to 1,800 m, mostly from 1,300 to 1,800 m. The species is a member of section Belolon- chium and is characterized by its broadly ovate to subrounded blade with prominent posterior lobes and a weakly glaucous petiole, D-shaped in cross-section. The species is probably most closely related to A. montanum, with which it is sympatric and shares a weathered cataphylls, and petioles that are usu- ally D-shaped in cross-section and somewhat glaucous. The latter species differs in having a narrowly ovate blade which is generally smaller - than A. titanium. The type specimen is unusually large (perhaps twice as large as other collections in most aspects) but otherwise compares well with the material from Chiapas described by Matuda as A. xan- thosomifolium. The species is by far the largest Anthurium in Mexico. ); 18639 (NY, UC); 18644 (DS, MEXU, MO); 18799, 38603 (MEXU, MO); 19668 (MEXU, MO, ). > Anthurium tonduzii Engl., Bot. Jahrb. Syst. 25: 376. 1898. Type: Costa Rica. San José: near San Marcos, in forest near Alto de Pito, 1,400 m, Tonduz 7683 (B, holotype). Epiphytic or terrestrial, usually creeping; stems elongate, rooting at nodes; the internodes 2-21 cm long; cataphylls to 8 cm long, early deciduous. A Treo P quee Aa Tur" 1 I terete, nar- rowly and obscurely sulcate, 1.5-12 cm long, 2- 3 mm diam.; geniculum 1.4-2.3 cm long; blades à 3 , rving as the collective vein; primary lateral veins 3—6 per ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 side, scarcely more conspicuous than the inter- rimary veins, sunken above, raised below; col- lective vein arising from the base, 4-5 mm from the margin. FLORESCENCE erect-spread- ing, longer than leaves; the peduncle 3.5-14 cm long; spathe lanceolate, green tinged reddish, subcoriaceous, 1.5-3 cm long, 0.5-1 cm wide, abruptly acuminate at apex, acute at base; spadix green to red, scarcely tapered, 2.5-3 cm long, 4 5 mm diam. midway (dry); flowers rhombic to 4-lobed, ca. 2.2 mm long, 1.9 mm wide (dry) the sides straight to weakly sigmoid; 4-5 flowers visible in the principal spiral, ca. 7 flowers visible berries white, turning purple at maturity, oblong, ca. 1 cm long; mesocarp clear, gelatinous, 9 2. dark violet-purple, ca. 4 mm long, 2 mm Ww and 1.5 mm thick. Figs. 201 and 202. Anthurium tonduzii occurs in Costa Rica and Panama at 800 to 1,500 m elevation in premon- tane rain and lower montane rain forest. It has been collected on Cerro Pirre in Darién Pro of Panama and is to be expected in Colombia the Department of Chocó. 2 no The type collection was made in 1893 and collections have subsequently been made in ps ta Rica but the species has been collected times between 1975 and 1977 in Panama n Tetraspe™ m like A. 5&0 punctate leaves and scandent habit. It both by its longer internodes and en pore uous cataphylls. The cataphylls of A. sc pe and A. trinerve soon weather into fibers an sist, densely covering the stem. 3. Anthurium trinerve Miq., Linnaea H 67. t Anthurium scandens war. trinerve i Mé Matuda, Anales Inst. Biol. Univ. Nac. xico 15: 189. 1954. Type: Surinam. Awara, Forbe 120 (U). Anthurium brachyspathum C. Koch Koch, Allg. Gartenzeitung 25: 233. Surinam? D Bouché m — — & 1857. TYE i | CROAT KVALI 1983] Anthurium trinerve var. obtusum Engl., Bot. Jahrb. Syst. 25: 8. 1898. Type: Ecuador. Valle Palla- nga, B dira 2 (not seen). A trinerve var. angustifolium K. Krause, Bot. rb. Syst. 54 (118): 123. 1916. Type: Bolivia. \mahuacana at Espiritu Santo, 750 m, Buchtien p- de codajas ii G. M. Barroso, Arch. Jard. B Río de E 15: 97. 1957. Type: Brazil. Ate zonas: Codajás, Logos de Badajós, R. L. Froes 26590 (not seen). Epiphyte; stems elongate, creeping, usually less than 50 cm long, 6-9 mm diam.; internodes 1- 3.5 cm long; roots numerous, 2-3 mm diam., grayish-brown; cataphylls moderately thin, to 6 am ong, drying brown, weathering to longitu- dinal tem. LEAVES spreading; petioles 1.3-6 cm long, 2.5-3.5 mm diam., + D-shaped, broadly and sharply sulcate; geniculum 3-5 mm long; blades moderately thick, elliptic to ovate-elliptic to lanceolate, 8-17 cm long, 3-7.5 cm wide, acute at apex, obtuse to attenuate at base; the upper surface semiglossy, lower surface glossy and conspicuously punctate; the midrib acutely raised above, convexly raised below; usually 2 pairs of basal veins, the inner- most serving as a collective vein, 5-12 mm m from CENCES erect, shorter than leaves; peduncle 2- 6.5 cm long; spathe thick, oblong-elliptic, green- ish-white, 2.53.5 cm long, 0.5-1.1 (1.9) cm wide, abruptly acuminate at apex, obtuse at base, in- serted at 35° angle on peduncle, held erect at anthesis; spadix dark lavender or green to white, -4 cm long, 5-7 mm diam. at base, 2-3 mm diam. at apex; flowers square, 2.5-4.2 mm in both directions, the sides straight to gradually sewn, white to pale green; stigma line ar, ca al long, exserted, brushlike with conspicuous a droplets 2-3 spirals ahead of emerging " hos ns; stamens emerging slowly from the base : Prompt and complete sequence, exserted on ‘anslucent filaments ca. 1 mm wide, 0.5 mm Nae anthers white, held against pistil at same «i às emerged pistil; thecae ovate-ellipsoid, Gui divaricate; pollen white. INFRUCTES- A with spadix 4.5-6.5 cm long; berries or sometimes pale lavender, globose, 5-7 OF MEXICO AND MIDDLE AMERICA 401 mm diam.; seeds 4-10, oblong. Figs. 203 and The species is found in Guatemala and Belize to the Guianas and central Brazil at elevations of 0 to 800 m, in tropical moist, premontane wet, and tropical wet forest life zones. In Middle erica the species has been collected only on the yin slope but it occurs on both slopes in Costa Rica and Panama. Anthurium trinerve is easily confused with A. scandens. It is generally found at lower elevations (to 800 m) than A. scandens. Other differences are discussed under A. scandens. It is a member of section Tetraspermium. BELIZE. Eos DISTRICT: Mountain Cow Ridge, Bente 3522 (MICH 2785, 2829 Micra TOLEDO DISTRICT: Temash vend Dwyer 12841 (MO). Costa Rica. No other location, Pittier 9511 (BR). ALAJUELA: Llanuras de San Carlos, Brade 2546 (BR); Villa Quesada, San Carlos, Smith 2528 (MO). GUA- NACASTE: NW of Lake Arenal, Croat 260 (MO). L between Bribrí & Bratsi along Río al. 10421 (MO); Limón, Kuntze 1954 (NY) RENAS: Golfo de Nicoya, Stork C523 (UC); Terraba, Pittier 3865 (BR); Tinoco Station, Allen 5472 (DS). HONDURAS. ATLÁNTIDA: Lancet ora ee 72, 110 (MICH); Yuncker 4557 (MO). G AS A DIOS: Río Plátano yer village of Ras, mme et as 7561 (MO). NICARAG RÍO SAN JUAN: San Juan del Norte gepas “Smith 35 (MO). ZELAYA: Boca Rio La Tig- a, Stevens 8925 (MO); small Cano at Salto ity Quebrada El Toro, along (MO); N or NW of Colonia San aisle Stevens 9042 (MO); near El p Atwood & Neill AN228 (MO); n road to San Jerónimo, Stevens 7575 (MO); road to Mina Nueva Pai N of E mis ema d 8373 (MO); road between Rosita Ca s, SW of ipee reed ps 8507 (M O); nea Rosita, Neill 4476 (MO); along Río Sucio, E of dnt Neill 4027 (MO). Anthurium trisectum Sodiro, Anales Univ. Centr. Ecuador 20: 100. 1905. Type: Ecuador. Es- meraldas: en la orilla del Río San Antonio, Sodiro 8/904 (B, holotype). Usually terrestrial, sometimes epiphytic, creeping; stems elongate; internodes 1—6 cm long; cataphylls thin, 2-9 cm long, weathering into fine longitudinal fibers, ultimately deciduous. LEAVES with petioles erect, 18-36(40) cm long, weakly sulcate; blades 3-parted, moderately thin, leaflets sessile or on petiolules to 2 cm long; me- dian leaflet + elliptic to ovate, 12-25 cm long, 4-11 cm wide, long-acuminate at apex, obtuse at base, lateral leaflets inequilateral ca. two-thirds 402 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 as long as median leaflet, short-acuminate at apex, inner margin attenuate at base, outer margin rounded; the midrib in the median leaflet sunken n above, raised below; departing ENCE erect, shorter than leaves; pe- duncle 9—22 cm long, terete; spathe lanceolate or ovate, green becoming white or yellow, 4-6 cm long, 1.2-2.5 cm wide, acuminate at apex, round- ed at base; spadix green to yellow, scarcely ta- pered, 3.7—5 cm long, 4-6 mm diam. at base; the flowers 4-lobed, ca. 1.7 mm in both directions, the sides sigmoid; ca. 6 flowers visible in the principal spiral, ca. 5 flowers visible in the al- ternate spiral; lateral tepals 0.9-1 mm wide, the inner margin convex; pistils iio cue tic; ip spadix, held at edge of ioa rai ey NFRUCTESCENCE with spadix to 6.5 cm long, berries violet-purple, globose to ovoid, with a short beak. Figs. 207 and 208 The species occurs from Costa Rica to Ecuador at elevations from sea level to 500 m in tropical wet forest life zones. In Costa Rica the species is known only from the OTS La Selva research area, 100 to 200 m elevation in Heredia. In Panama, the species is known only from 300 to 700 m elevation in Coclé and Veraguas Provinces, prin- cipally on the Atlantic slope. Anthurium trisectum is in section Dactylo- phyllium and is distinguished by its usually ter- restrial habit and thin, 3-parted leaf blades with conspicuously impressed veins. No other Costa Rican Anthurium has completely 3-parted blades and A. trisectum is not closely related to A. sub- signatum, the only other species tending to have 3-lobed blades. CosTA RICA. HEREDIA: La Selva (field Station of Or- Laien for Tropical Studies) (previously Finca La Selva), Folsom et al. 8825, 9171: 1; Grayum 1833 mel 8274, 10457, 11496 MacDougall 999: Damon Smith 133; Sperry 834 . 1083; Todzia 1171; Wilbur 28111, 30024, 30363, 33542 (DUKE ). € umbrosum Liebm., Vidensk. Meddel. Dansk Naturhist. Foren. Kjobenhavn 1: 21. 1849. Type: Mexico. Oaxaca: near Mirador, m 13790 (C. holotype; F, US, iso- pes). Antari Temenni Schott, Bonplandia 7: 165. 1859. E: Mex o. Oaxaca: near Teotalcingo, 5 ,000', peste ©. Terrestrial to 1 m tall; stems 2.5-3 cm diam.; internodes very short; leaf scars to 2 cm wide; roots few, moderately thin, descending; ca phylls coriaceous, sometimes tinged with purp 5-9 cm long, emarginate and long-apiculate apex, drying dark tan (B & K Yellow 6/7 weathering into longitudinal fibers and pe ing. ba: mm ROG arching, subtere 20-43 cm long, 12-27 cm wide, broadest at or midway between base and point of pe attachment, the margin broadly undulate; an. rior lobe 15—32 cm long, margin broadly con posterior lobes 4—14.5 cm long; sinus hippo convexly raised above, raised and conspicue more acute below; basal veins 4—6 pairs, P alesced 2-3.5 cm, posterior rib straight to We curved, almost completely naked, the outer ms gin turned conspicuously upward; p primary i eral veins 3—5 per side, departing midrib at € 50° angle, straight to weakly arching to colle vein, sunken above, flat to prominulous 9€ lesser veins visible, + flat above and below; © : lective vein arising from the first basal | sunken above, raised below, 8-12 mm from © margin. INFLORESCENCE usually longer" leaves, spreading to pendent; peduncle 2 e long, 2.5-3.5 mm diam., terete, at least ; long as petioles; spathe coriaceous, green. times tinged with violet-purple, AB apex, narrowly inrolled to 2.5 cm long. ro : to subcordate at base, inserted at 4 5 on peduncle; spadix green turning il y purple to brownish at anthesis, 4.8-12 5-12 mm diam. at base, 3-6 m m di i flowers rhombic to sub-4-lobed, 3.3-4. 2.5 mm wide, the inner margins b s re anie pistils green, weakly exserted just r stigma elliptic, conspicuously ratse®, : s ma drop apparent vei before emo Stamens from base in a reg o x oO » T x E by x e E: -— S = o "T" = m a Q Ó » Z © = J J E^ m » S m = e » RES 208-211. 208. Anthurium trisectum Sodiro, Croat 27362.—209-210. Anthurium umbrosum Liebm., Croat 39787.—211. Anthurium (Type). etus nse Croat & Baker, Croat 36342 404 the laterals first, soon followed by alternates, complete in basal one quarter before the laterals emerge just above midway; anthers cream, held in Mig circle around pistil, ca. 1 mm wide; thecae ovoid, esu di- FRUCTES CE ange with numerous raphide cells; seeds 2, obo- void, 6-7.5 mm long, 4.5-6 mm wide, 3-4 mm thick, depressed at apex. Figs. 209 and 210. Anthurium umbrosum is endemic to Mexico and known only from northern Oaxaca, princi- pally in the Sierra de Juárez, south of Valle Na- cional on the Atlantic slope at 1,200 to 1,800 m elevation. It occurs in wet cloud forests on usu- ally steep slopes. The species is characterized by its terrestrial habit, moderately thick, usually narrowly ovate, long-acuminate leaf blades, green, coriaceous, lanceolate spathe, which curls under along mar- gins, spadix green in early stages, becoming vi- olet-purple, and orange, 4-ridged berries. Anthurium umbrosum is in section Belolon- chium and is closely related to 4. /ancetillense. It is distinguished from that species in having shorter, more flexible peduncles, longer, more pendent inflorescences, somewhat smaller, dis- tinctively 4-ridged berries, and a thicker spathe. Anthurium umbrosum occurs at much higher el- evations, A. /ancetillense having been collected only in the Lancetilla Valley in Honduras at el- evations of 10 to The species is also similar to A. ovandense from southern Oaxaca and Chiapas, but that species has a much thinner blade that is minutely undulate along the margin and that usually has anterior lobes concave along the margin. Anthurium umbrosum is also similar to A. le- zamae from northern Oaxaca, but that species differs in having a minutely alveolate epidermal pattern (at least on drying). MEXICO. OAXACA: Mirador, Liebmann 15790 (C, F P, US); Río Mono Blanco, MacDou gall 7040 (MEX. Teotalcingo, Liebmann 22349 (Neg. # Oaxa Moore & Bunting 8896 (MEXU), S of Valle CAMPER ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 Rzedowski 33844 (ENCB, MEXU, MO). vERACRUZ Plan de Cedino, Munc. pede Ventura 4665 (MICH). Anthurium upalaense Croat & Baker, Brenesia 16 (Supl. 1): 97. 1979. Tyre: Costa Rica juela: along road between Cañas (Guidi caste) and Upala, near Río Zapote, 1.8-2.7 km south of Río Canalete, ca. 100 m elev., disturbed margin of primary forest, 25 June 1976, Croat 36342 (MO-2381219, holo- type; CR, F, isotypes; Live at MO). p less than 20 cm long, 1.5-2 cm Miam. best scars hidden by persisting cataphylls and roots; roots moderately thick and long, 2-4 mm diam., brown, moderately dense and numerous; cataphylls coriaceous, 11-22 cm long, narrowly acuminate at apex, we tinged lightly with red, turning brown o remaining firm and intact, ultimately acidi LEAVES erect-spreading; petioles (6)10-39 cm long, 7-10 mm diam., subquadrangular, broadly and sharply sulcate, prominently 3-ribbed abaxially, the sides flat to somewhat concave geniculum 1-1.5 cm long, sometimes reddish abaxially; blades narrowly oblong-ellipti long-oblanceolate, moderately thick, gra acuminate at apex (the acumen apiculate, down- turned), broadly to narrowly acute at base, ( d 106 cm long, (4.5)8-23 cm wide, broadest at middle or slightly above; both surfaces semi- glossy; the midrib flat at base above, narrow : and becoming triangular toward apex, promi nently raised below, 3-ribbed near base; primary lateral veins more than 20 per side, midrib at ca. 60° angle, raised above an a low, + straight to collective vein, loop id ing in upper half; the interprimary veins pe slightly raised above, flat and slightly darker EC surface below; lesser veins less conspicuo ene lective vein arising in the upper half, flat 3 RES- and below, 3-5 mm from margin. INFL est CENCE erect to spreading or arching-pen peduncle (18)40-48 cm long, 4-9 mm fai terete or 1-ribbed below opening ord ( gh at Rosulate, thick, yellow-green (B & K with 7/2.5, 7/5), sometimes heavily ting colai olet-purple, de -lanceolate to lane dest ? cm wide, br dix sessile, sin to uiris (B 1 6/5), sometimes red, 9-30 cm long, 6- ) | 1983] CROAT diam. at base, 4-6 mm diam. at apex; the flowers square to rhombic, 2.3-3.4 mm long, 1.7-2.8 mm wide, the sides straight to weakly sigmoid; 7-15 flowers visible in the principal spiral, 10— 20 flowers visible in the alternate spiral; tepals matte, inconspicuously pale punctate, with drop- lets usually after 2 or more stamens emerge, lat- eral tepals 0.8-1.8 mm wide, the inner margin straight to slightly convex; the pistil raised ca. 0.5 mm, green to purplish; stigma elliptic, 0.4— 1.1 mm long, green, brushlike, exserted ca. 0.1 mm, with droplets for 5-15 days beginning 12- 22 days before the first stamens emerge, dry and spadix, the apical stamens usually preceding the basal ones, exserted on short greenish filaments, ca. 0.2 mm long, 0.5-1 mm wide, which retract to hold anthers in a tight circle around pistil or separately at sides of pistil; anthers pale orange, 0.5-0.6 mm long, 0.7-1 mm wide; thecae ellip- soid, scarcely divaricate, opening flat; pollen or- ange, fading to pale orange or lavender, white or tan. INFRUCTESCENCE pendent; the spathe reflexed; the spadix to 55 cm long, 6 cm diam.; berries oblong, red, tapered to both ends, 9-15 mm long, acute at apex, emerging + uniformly throughout spadix, well above tepals before fully mature; mesocarp mealy, sweet, white; seeds 2, oblong, to 4 mm long, 3 mm wide, 2.5 mm thick, creamy white, obliquely truncate at apex with a 5 appendage, greenish at base. Figs. 211 and E species is known from the Atlantic slope Icaragua and Costa Rica at elevations from near sea level to usually less than 600 m in re- ions of tropical wet forest, premontane wet for- : Cartago Province, ortheast of Río Grande de Orosi at 1,300 m is apparently also this species. mim upalaense is a member of section o c dp and is apparently most closely re- differs ; A. concolor K. Krause from Panama but edil T having the inflorescence pendent, the Alles: ong-tapered and more or less green at Nec IS, the spathe longer and recurled, the fresh scarce! pale oran the major lateral veins es y ami often forming a collective vein velis I the middle or above, the reticulate Ee clearly visible when fresh, weakly raise Prominulous when dried. Anthurium con- nge, OF MEXICO AND MIDDLE AMERICA 405 color, on the other hand, has the inflorescence stiffly erect, the spathe usually less than 1 long, usually not recurled at apex, the spadix not tapered, dark purple-violet, and usually less than 7 cm long at anthesis. Its flowers form minute globules of nectar on the tepals and the pollen is pale violet-purple, its major lateral veins are prominently raised and seldom form a collective vein except near the apex, and the reticulate veins are more or less obscure. Anthurium concolor is known only from the Atlantic slope in central Panama. CosrA Rica. Cultivated at Kew, from Costa Rica, N. E. Brown s.n. (K); Bull 467/1873 (K). ALAJUELA: W of Fortuna, NW of New Volcán Arenal, Taylor & Tay- lor 1170 (NY); San Juan de Lajas, Austin Smith 1498 (NY); near hie and Methodist Rural Center, NE sada, Molina et al. 17264 (EAP, F); vicinity Los Chiles, Rio Frio, Hom & Iltis 771 (NY); U); road between Cañas pala, N of R Zapote, Croat 36342 e 2: MO): 36380 (MO); road between Cafias and Upala, NNE of Bijagua, Croat 36319 (MO); S of esc near the Rio Zapote, and aker 1001 > 18194, 18195 (CR). c Orosi at Tapanti, Utley 5062 (MO); Forest of Tuis, pee Turrialba, Pittier 12720 (US); Lake Bonilla, Pa- o 92 (F). T ANACASTE: El Arenal, Standley & Val- je 45328 (US). HEREDIA: Finca va, Sanh Viejo de Sarapiqui, ja 4297 (MO); near Puerto Viejo along roa Rio Sucio, Croat 35708 (MO); SW of e on a trail paralleling the Río Carbón, ; Taskí, Talamanca Valley, Tonduz 9507 (BR, CR) Pittier 9509 (BR); Forests of Shirores, Talamanca, wy. (MO); vicinity of Bribri nea (MO); general y gee ween & Burger 8446 e at Cahuita, Baker & Burger 17A Nic ARAGUA. wee Cerro Musün, W & NW A above Salto Grande of Quebrada Negra and in alley of Río Regu Near 1782 (MO). RÍO SAN JUAN Neill 1552(M A: along stream N of settlement Las eae de P Vincell 328A (MO). Anthurium utleyi Croat & Baker, Brenesia 16 (Supl. 1): 100. 1979. Type: Costa Rica. Ala- juela: 2 km N of Ángeles Norte de San Ra- món, 4,000', Luteyn 3695 (DUKE-233270, holotype). Epiphyte; stems short, densely covered by 406 ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 4 niaan upalaense Croat & Baker, Croat 43305.—213-214. FIGURES 212-215. E 21 Anthurium "© lidifolium K. Krause. € roat 108 215 . Anthurium verapazense e Engl., Boutin 5122. CROAT BUAL 1983] cataphyll fibers; roots few, moderately long, de- scending, 2-3 mm diam.; cataphylls 3-6 cm long, apiculate at apex, drying tan, weathering into dense, tangled, persisting fibers. LEAVES 8.5-24 cm long, 2-5 mm diam., subterete, shallowly sulcate to weakly flat- tened adaxially, rounded abaxially; geniculum to oblong-elliptic or oblong-oblanceolate, acu- minate at apex, acute to obtuse or narrowly rounded at base, 15-35 cm long, 3.3-7.5 cm wide, broadest near middle or just above; both surfaces punctate, conspicuously black punctate on lower surface; midrib convexly raised near base above, narrowed toward apex, prominulous below; pri- mary lateral veins 14—20 per side, departing mid- ril at ca. 45° angle, + obscure above, scarcely visible below; collective vein arising from the base, scarcely visible above and below, 2-7 mm rom margin (often more remote from margin in lower half of blade than in upper half). INFLO- RESCENCE usually erect-spreading, equalling or shorter than leaves; peduncle 10-27 cm long, 2-3 mm diam., as long as or longer than petioles; spathe green to purple, narrowly lanceolate, 3.4— 6 cm long, 1-1.5 cm wide, broadest at base, acu- minate at apex, obtuse to rounded at base, in- Wee at 45° angle on peduncle; stipe 1-6 mm n Iront, to 3 mm in back, ca. 2 mm diam.; spadix a to brownish-red or purplish-green, 2-15 hs ong, 7-9 cm diam. at base, 3-5 mm diam. apex; flowers rhombic to 4-lobed, 2.1-2.5 mm long, 2-2.5 mm wide, the sides sigmoid; 4-5 visible in the alternate spiral; tepals glossy CE weakly and minutely papillate, lateral hi $ 1.2-2.5 mm wide, the inner margin con- ellip gna emergent 0.3-0.5 mm, white; stigmas be 1C, ca. 0.4 mm long with small droplets be- "iid i emerge; stamens emerging rapidly Mes P base, lateral stamens preceding alter- ca. 0.5 y only 1-2 spirals, exserted on filament antc mm long, which quickly retract to hold "Y = at edge of tepals; anthers ca. 0.4 mm long, polis m wide, thecae white, dish-like, ellipsoid; e n white. INFRUCTESCENCE with orange, void berries. Fig. 216. E. " Species is known from Costa Rica in for- "E Vicinity of San Ramón and near Monte- * at elevations of 760 to 1,690 m. nium utleyi is in section Porphyrochito- m and is most closely related to A. friedrichs- OF MEXICO AND MIDDLE AMERICA 407 thalii, but differs from it in its shorter and broad- er leaves and its beaked berries. The species is named in honor of John and Kathy Utley, who made many important collections of Araceae in Costa Rica. Costa RICA. ALAJUELA: S of Balsa de San Ramón, Lent 3519, 3520 (F); Utley & Utley 1851 (F); Los An- geles de San Ramón, Brenes 14844a (NY); N of An- à KE Ram tarates de San Ramón, Brenes nR , beyond Ángeles Norte, Luteyn et (DUKE). GUANACASTE: near E edge of Mon- teverde Preserve along road between Santa Elena and Monteverde, ca. 2.5 miles from Santa Elena, Croat 47135 (MO). HEREDIA: S of Cariblanco, Croat 35795 MO). ALAJUELA & PUNTARENAS: Monteverde, Palmer 90 (NY); Dryer 1337 (Fy, Luteyn 3398 (DUKE); upper drainage of Río Peñas Blancas below Monteverde Cloud Forest, Lawton 1165 (F). Anthurium validifolium K. Krause, Notizbl. Bot. Gart. Berlin-Dahlem 11: 607. 1932. TYPE: Panama. Chiriquí: forests along Río Ladri- llo, above El Boquete, 1,200-1,300 m, Pit- tier 3052 (US). Epiphyte; stems to 16 cm long, 12 cm diam.; leaf scars 2.5 cm wide; roots numerous, green, descending, 1-1.5 cm diam.; cataphylls subco- riaceous, ca. 20 cm long, acuminate at apex, drying tan (B & K Yellow-red 9/10), persisting + intact. LEAVES erect-spreading; petioles 20- 27 cm long, 7-10 mm wide, + quadrangular, broadly and shallowly sulcate, geniculum 3-3. cm long; blades lanceolate, coriaceous, 47-69 cm long, 10-19 cm wide, acuminate at apex, round- ed at base, broadest at middle, the margins broadly undulate; both surfaces matte; midrib broadly and convexly raised at base above, more acutely raised toward apex, diminishing and sunken at apex, raised and square at base below, more acute toward apex; primary lateral veins numerous, departing midrib at 60°-70° angle, + straight, arching near margin, sunken in valleys above, raised below; interprimary veins scarcely visible; collective vein arising from one of the primary lateral veins in the apical one quarter of the blade, 5-6 mm from margin, weakly sunken above, raised below. INFLORESCENCE spreading, shorter than leaves; peduncle 33-39 cm long, 9-12 cm diam., subterete, twice as long as petioles; spathe coriaceous, dark green (B & K Yellow-green 5/10), turning yellow, ovate-lan- ceolate, 16-17 cm long, ca. 4 cm wide, acuminate at apex, rounded at base, inserted at ca. 45° angle Le Z Z > = n oO T = i» m <= Un n am w C x w A = > A O > ^ » zx J m a FIGURES ae 9. =2%s6. Anthurium utleyi Croat & Ba saa <= = 47135.—217. Anthurium validifolium K. Krause, Croat 10667.—218. Anthurium verapazense En . Croat 41683.—2\19. Ant thurium verapazens .. Croat 406 $ 1983] on peduncle, reflexed and held parallel to pe- duncle; stipe 7-13 mm long in front, ca. 3 mm long in back; spadix pale lavender to violet-pur- ple (B & K Purple 6/5 to 5/2.5), green at apex in early stages, 15-23 cm long, 1.2-3.5 cm diam. at base, 5-10 mm diam. at apex; flowers rhom- bic, 3-3.5 mm long, 3.1—3.6 mm wide, the sides straight to weakly sigmoid; 7-8 flowers visible in the principal spiral, 8—10 flowers visible in the alternate spiral; tepals matte, lateral tepals 1.5— 1.8 mm wide, the inner margin broadly rounded; pistils scarcely emergent, green; stigma elliptic, ca. 0.6 mm long, droplets sas for ca. 1 week ae aay held inward and partially cover- ing ee at level of tepals, ca. 1 mm long, 1.2 €; thecae oblong-ellipsoid, scarcely di- vile em yellow-orange (B & K Yellow- red 8/5). INFRUCTESCENCE pendent; spathe twisted; spadix 24-26 cm long, 2.3-2.5 mm diam.; berries bright orange, obovoid, flattened at apex, 0.8-15 mm long, pericarp thin, + trans- parent, orange; mesocarp fleshy, orange with mi- es io cells; seeds 1—2, ovoid, pale yellow, él brown at apex and base with somewhat we appendage at both ends, encased in a Bs T dry, sac-like structure. Figs. 213, 214, M up validifolium is a member of section Chir €urium. It is known for certain only from m Cruces Botanical Garden near San ü Ri àva near the Panamanian border in Cos- the may be a native to the area according to Owner, Mr. Bob Wilson. Most collections have been made in lower montane wet forest and Thi : E S further substantiates that A. validifolium is 50 native to Costa Ric A nthurium vidi is characterized by its l i rge, moderately thick, lanceolate leaf blade, avender and thi those species in having broader icker leaf blades. Its petiole is subquad- CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 409 rangular and somewhat intermediate between the terete petiole of A. protensum and the sharply quadrangular petiole of A. seibertii. Although this suggests A. validifolium may be a hybrid between A, pridem and A. seibertii, ss me pe port tionally shorter blad oun of these species. CosrA RiCA. PUNTARENAS: Las Cruces Botanical Gardens, Croat 44445 (MO). Anthurium verapazense Engl., Pflanzenr. IV 23B: 191. 1905. Type: Guatemala. Alta Verapaz: Cubilhuitz (Finca Cubilhuitz 15?40'N, 90925'W) elev. 350 m, H. von. Tuerckheim 7788 (US, holotype). Usually epiphytic, sometimes terrestrial; stems 4—8 cm long, 5-6 cm diam.; internodes and leaf scars hidden by persisting cataphylls and thick, descending, dark brown roots; cataphylls sub- coriaceous, 6—9 cm long, acuminate at apex, sometimes tinged red-violet, subapical apiculum ca. 2 mm long, drying dark brown (B & K Yellow 4/7.5), persisting as reticulate fibers. LEAVES with petioles erect-spreading, 26-61 cm long, 6- ally acuminate at apex, prominently lobed at base, 24—60 cm long, 14—27 cm wide, broadest at base or sometimes at point of petiole attachment; an- terior lobe 24-49 cm long, convex to almost straight; posterior lobes 8—17 cm long; sinus hip- pocrepiform to spathulate to sometimes ovate, apex; both surfaces semi- rib acutely raised above, diminished and sunken from just above the middle to the apex, convexly raised below; basal veins 4—6 pairs, the first free to base, those remaining coalesced 2.5-4 cm, sunken above, raised below; primary lateral veins 7-9 per side, departing midrib at 45°-55° angle, loop-connecting to collective vein, sunken above, prominulous below; lesser veins less conspicu- ous; collective vein arising on the first basal vein, sunken above, raised belo mm from he margin. INFLOR RESCENCE arching-pen- dent; peduncle 53-90 cm long, 4-6 mm spathe moderately thin, oblong-lanceolate, green (B & K Yellow-green 7/10), sometimes tinged violet-purple at margins, 8.5-15 cm long, 1-2 cm wide, broadest near base, narrowly acumi- nate at apex, obtuse at base, inserted at 45? angle on peduncle; spadix violet-purple (B & K Red- -= 410 purple 2/5), 10-25 cm long, 6.5-8 mm diam. at base, 3-5 mm diam. at apex; flowers rhombic, 3.6-3.8 mm long, 2.3-2.8 mm wide, the sides straight parallel to the spiral, gradually sigmoid perpendicular to the spiral; 6—7 flowers visible in the principal spiral, 8—10 flowers visible in the alternate spiral; tepals matte to semiglossy, ob- scurely and sparsely punctate, minutely papil- late, the lateral tepals ca. 2 mm wide, the inner margin + concave and turned up against pistil; pistils weakly emergent, violet-purple slightly paler than tepals; stigma slit-like, oblong, ca. 0.6 mm long; stamens emerging in a moderately rap- id sequence from the base, the laterals first, fol- lowed shortly by alternates, the leading stamens preceding third and fourth by several spirals, briefly exserted on purplish filaments, withdraw- long, 0.6 mm wide, held in a close, tight circle around pistil; thecae ellipsoid, divaricate; pollen lemon-yellow fading to white. INFRUCTES- CENCE pendent; spadix 15-34 cm long, 1-2.8 cm diam.; spathe often withered or deciduous; berries subglobose to obovoid, rounded at apex, developing in the basal one third to one half, ca. 7 mm long, ca. 6.7 mm diam.; mesocarp juicy, with numerous whitish raphide cells; seeds 1—2, 3.5-4.2 mm long, 3-3.6 mm wide, 2-2.5 mm thick, broadly ellipsoid to subrounded, flattened, emarginate at apex, pale purplish-violet with nu- merous, clear punctations, the surface Sticky, the hypocotyl small. Figs. 215, 218, and 219. Anthurium verapazense ranges from Belize to Guatemala along the Atlantic slope in principally moist forest habitats at elevations of 100 to 800 m. It is easily confused with A. lucens, which is more wide ranging, occurring throughout much of Chiapas and northern Guatemala as well as in northern Honduras. The latter is generally found at higher elevations, from 350 to 2,000 m. Anthurium verapazense is in section Belolon- chium and is distinguished by its usually long, narrow spathe, much less than half as long as the spadix at anthesis. In addition the inflorescence is usually 1.5 to 3 times longer than the longest petioles. Anthurium lucens differs in having a relatively short spadix with the spathe usually over half as long as the spadix at anthesis, by the relatively shorter inflorescence which is shorter than the leaves or rarely more than 1.5 times longer than the petioles, and by the red, obovoid berries with 2 relatively much larger seeds. Leaf shape in 4. verapazense is greatly vari- ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 able, ranging from deltoid to ovoid to deltoid- hastate. The lobes may be directed downward, sharply outward, or even somewhat inward. Even plants within a single population may show great variability. Neither A. verapazense or A. lucens show correlation with regard to leaf shape al- llecti from the same area show sim- ilar trends in variation. Worthy of mention is à collection of A. verapazense made by Dr. R. Tejada (307) purportedly collected in Maza- tenango, a town on the Pacific slope. I believe it is unlikely that the species ever occurred there and perhaps there was (in 1914) another village called Mazatenango located on the Atlantic slope or perhaps the collection was cultivated in Ma- zatenango. Other species were also collected by Tejada at Mazatenango that are likewise not be- lieved to have ever occurred there. It is possible that additional work will prove that the taxon as constituted here consists of two species. A live collection made by Hans Wiehler near Purulhá in Alta Verapaz (Guatemala) and cultivated at Selby Botanical Gardens po subglobose, red-orange berries that had 6-9 s ; seeds. This has been vouchered as Madison 37 3 . A second Madison collection (6965) from a live plant purportedly also from Purulhá had sim : berries with 2 large seeds. These were similar the berries of other collections studied, n: Jones & Facey 3389 from the department 0 : abal and Boutin 5122 from Belize. The pm live at the Missouri Botanical Garden. e drastic differences are not easily explain. have not been exhibited in other species © e thurium. The problem warrants addition vestigation. thouoh e : ivated BELIZE. BELIZE DISTRICT: S of Gales Point A at Kew, Hunt 7043 (K, MO); Gracy Rock, Peck 1 & Liesner 1512 (MO); Croat 23852 (M DISTRICT: no other location, Lundell 6 j Flor hunting camp, S of Grano de E Odi MICH. (MO); Mo i ley 3254 (MO); between San Luis and cal, Croat 23533 (MO); Waterhole Camp € stant Gentle 2445 (MICH). STANN CREEK DISTR d ; Creek Railroad, Schipp S177 (F). jor Schloss Maya Mountains, Union Camp, Boutin ^ 5122 (MO). on 8% GUATEMALA. ALTA VERAPAZ: Chamá, Jone hurc (UCLA); along road to Finca Trece Ag i HE vicinity in Senahú, Luteyn & Almeda 3523 (DUE uon of Laguna Sapalau, SW of Sibicté, Steyer se road (MO); 4-9 miles up road to Oxec, May which turns N off Highway 7E between m MENS MET 4 1983] CROAT LCAVAL Estor ca. 6 km NE of Panzós, Croat 41605, 41683 ZABAL: vicinity Exmibal, NW of Izabal, Jones & Facey 3389 (LL, NY). PETEN: Poptán, Molina 15613 (NY). M XICO. CHIAPAS: Finca Prusia, Mapastepec, NW of Huixtla, MacDougall 337 (CAS, DS, MEXU); Lagos de Montebello, 3 mi W of Dos Lagunas, Croat 46656 y; la de Espi , Breedlove 27489 (DS); N of Ocozocoautla de Espinosa on gravel road to Ap- itpac, Croat 40628 (MO). vERACRUz: Lake Catemaco, Soteapan, Beaman 6142 (XAL). Anthurium watermaliense Hort. ex L. H. Bailey, Stand. Cycl. Hort. 1: 303. 1922. TYPE: Co- lombia. Not seen (introduced into horticul- ture via Watermall, Belgium). Terrestrial; stems to 25 cm long; cataphylls thin, 4.5-9 cm long, rounded at apex, weathering into coarse fibers. LEAVES with petioles erect- Spreading, 12-88 cm long, + terete, weakly flat- tened or sulcate; geniculum 1-1.5 cm long; blades e » Posterior rib mostly naked, weakly turned "-— lateral veins 4—8 per side, departing bu at aer-45° angle, prominulous in valleys vl , prominulous below; lesser veins less con- uous; collective vein arising from the first or Nisi vein, raised above and below, 3-6 ue m margin. INFLORESCENCE spread- Fi > er than leaves; peduncle 12—66 cm long, MEC gd : m., spathe lanceolate-triangular, cm log np violet-purple, subcoriaceous, 5-21 E. m .3-8.5 cm wide, ong-acuminate at te c use to rounded at base, inserted at 7 5° white tc peduncle; stipe 0.3-3 cm long; spadix 5i o z e yellow or purple, 7-10 cm long, the i lam. at base, 2-5 mm diam. at apex; 8-3 E rhombic to 4-lobed, 1.5-3.4 cm long, ee ee ree ee spiral PA. flowers visible in the principal tepals matt Owers visible in the alternate spiral; cr edges €, green with purple on inner and out- mens Ee with sparse droplets when sta- the nes resh, lateral tepals 1.5-2.2 mm wide, i ing gin convex to + straight; pistil dark stamens x raised, weakly exserted just before mm lon merge; stigma narrowly elliptic, ca. 0.5 8, with papillae not exserted, a small stig- OF MEXICO AND MIDDLE AMERICA 411 ma droplet apparent 2-3 days before stamens emerge, dry and gray as stamens open; stamens emerging rapidly from base, or sometimes scat- tered throughout, exserted on short translucent filaments, 0.3-0.4 mm long, 0.6 mm wide, which retract holding stamens in a tight cluster over pistil; anthers creamy white, 0.9 mm long, 0.9- 1 mm wide; thecae ellipsoid, divaricate; pollen white. INFRUCTESCENCE with spadix to 22 cm long; berries yellow to orange, ovoid to obo- void, beaked, 1—1.5 cm long. Figs. 224 and 225. The species is known from Costa Rica to Co- lombia, usually from 750 to 2,400 m elevation in premontane rain and lower montane rain for- est. Some populations also occur near sea level in tropical moist forest and premontane wet for- est. The species is extremely variable, especially in leaf size and shape, and spathe and spadix color. It is strange that the species has not been found in tropical wet forest or at elevations be- tween 100 and 700 m. Perhaps further study will show that the lower elevation populations are a distinct species. Anthurium watermaliense is an atypical mem- ber of section Pachyneurium and is not generally confused with any other species in Central Amer- ica. It can be recognized by the ovate-triangular to sub-3-lobed leaf blades, the broad, frequently dark purple spathe, green to purple, stipitate spa- dix with long-exserted stamens, and yellow to orange berries. The species has long been known in horticul- tural circles, being first introduced from Colom- bia to Watermall, Belgium, by an unknown col- lector; although the name was validated by a short note by Bailey in the Standard Cyclopedia of Horticulture (1: 303. 1922), no type was des- ignated. I have as yet seen no Colombian ma- terial of the species and therefore decline to name a lectotype at this time. Costa Rica. No other location, Lankester 349-49 (K). ALAJUELA: Monteverde Biological Reserve, Ken- nedy & Guidon 3799 (MO). CARTAGO: Camino Raiz de Hule, SE of Platanillo, Croat 367 13 (MO), vicinity & Wilbur 4376 (MO). HEREDIA & 1548 (MO). LIMON: bet rí and Bratsi, along Río Sixaola, Burger et al. 10444 (MO); vicinity Bribrí, Río Catarata, Croat 43224 (MO); N of Limón, Lent 347 (MO). Anthurium wendlingeri G. M. Barroso, Bol. Soc. Venez. Ci. Nat. 26: 151. 1965. TYPE: Costa ANNALS OF THE MISSOURI BOTANICAL GARDEN * md FIGURES 220-223. 220. illiam K. Krause, Croat 36628.—222-22 Mies wendlingeri G. M. Barroso, Croat 34295.—221. Anthurium V! Anthurium yetlense Matuda, Croat 39768. FIGURE > Croat 3 732 s 224-227. t R +e ee eee 8 WOC [rw s v wy f | H 3 | t : Inthurium watermaliense Hort. ex L. H. Bailey, Croat 38129.—226-227 ae ` LX LOUP B La LAC S ^ nac e? LAG * E " > WANO . Anthurium wendlingeri G. M. Barroso, [£861 VODBIJNWV TIGGIN ANV OOIXAN IO WAPIOHLNF —LVOMNO tiv 414 Rica. Cartago: Turones (Pavones), Turrial- ba, 700 m, Wendlinger s.n. (RB, holotype). Pendent epiphyte; stems ca. 20 cm long; en internodes 1.5—2 cm long, except near apex; ro moderately edge — cance wnt ick. 5-15 cm lon we ES into coarse dein longitudinal E bers. LEAVES lip ip a teen petioles te- rete, pos 30 cm long, 2-4 m m.; geniculum 1-1.5 cm long; blades oblong sedere thick, 32-80 cm long, 3-11 cm wide, acuminate at a apex, round to une. at base; upper surface weakly velvety, obscurely and sparsely glandu- lar-punctate, lower surface a sparsely glandular-punctate; midrib convexly raised at base above, more acute iu: apex, diminish- ing and flat at apex, raised at base below, flat from ca. midway to apex; primary lateral veins 11—24 per side, departing midrib at 30-40? an gle, raised and nearly obscure above, flat below; lesser veins obscure; collective vein arising from the lowermost primary lateral vein, weakly sunk- ately Bie pale green tinged purplish, + oblong, 7-11 long, 8-13 mm wide, gradually Heute at apex, inserted at ca. 45? angle on peduncle, withering and curling horizontally or deciduous; spadix pale green to white or grayish- white, straight when young, becoming spiralled in age, 12-80 cm long, 3-5 mm diam. at base, narrowly-tapered to apex; flowers + 4- lobed, 3- 3.1 mm long, 2.4-2.6 mm wide, the sides smoothly sigmoid; 4-5 flowers visible in the nutely papillate, the lateral tepals 1.6-1.7 m wide, the inner margin broadly rounded; pistils weakly emergent, pale green, glossy; stigmas lin- mens soon following with no apparent pattern; anthers cream, held at edge of tepals € sides of pistil, ca. 0.7 mm long, ca. 0.6 m wide; thecae ellipsoid, not divaricate; ar e creamy- white. INFRUCTESCENCE pos ndent; berries reddish, + oblong or globose, ca. 5 mm lo flattened at apex with a central dede Figs. 220, 226, and 227. Panamanian plants considered 4. wendlingeri range from Veraguas in the west to Coclé, Pa- ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo. 70 nama, and Colón Provinces on the Atlantic slope and along the Continental Divide in tropical wet forest and premontane rain forest at elevations of 350 to 900 m. Anthurium wendlingeri was described from live material collected in Cartago Province of Costa he type description reports the inflores- cences as being ‘‘atropurpureus” and no other flowering material has been seen for Costa Rica. Panamanian collections closely matching this species in vegetative condition, habit and all oth- er € epicuri have spadices that are yish-w tanthesis. More material n from Costa Rica to determine if those a h really do have purple spadices and to determine if the Panamanian plants arè the same or if they should be considered distinct. One collection from Coclé (Croat 37329) was reported to have a reddish spadix but it has con- sistently flowered with a white spadix in culti- vation cies is in section Porphyrochitonium and is distinguished by its coriaceous, oblong, glandular-punctate leaf blades, very elongate. whitish, corkscrew-like spadix and reddish ber- ries. The blades are weakly velvety and obscure punctate above and below. It is most easily con fused with another Panamanian species. A. pe 4 dens Croat (ined.), which occurs in similar hà Provinces. That le to by its stouter, relatively short, green nish-purP brownish spadix, which never coils. . 830 Costa RICA. HEREDIA: S of Cariblanco, Croat 35 MO). ;zbl. Bot. Anthurium williamsii K. Krause, Notizbl. Gart. Berlin-Dahlem 11: 610. 193 a Panama. Darién: near Cana, Willia (not seen) : ria- Terrestrial; stems short; cataphylls sabes spreading, 23-61 cm long, 5-8 mm á terete, sometimes faintly 2-3- or more minate to cuspidaté at à pee po bed at he ; 48 cm long, 19.5-36 cm wide, brosse o | dle; anterior lobe 21-35 cm long, the looo I CNL LPR | gi7 | 1983] CROAT—ANTHURIUM OF MEXICO AND MIDDLE AMERICA 415 broadly rounded; posterior lobes 9-16 cm long from apex of sinus to outermost point; sinus tri- angular, sometimes closed, acute at apex; matte to semiglossy on both surfaces; midrib weakly and convexly raised above, prominently raised below; basal veins 6—7 pairs, usually all free to base; posterior ribs absent; primary lateral veins 8-10 per side, departing midrib at 35?—50? angle, sunken in valleys above, prominently raised be- low, basal and primary lateral veins drying con- spicuously yellow; lesser veins scarcely visible; collective vein arising from one ofthe uppermost basal veins, 2-3 mm from the margin. INFLO- RESCENCE erect; peduncle (16)40—53 cm long, 4-5 mm diam., shorter or equalling petiole; spathe membranous, pale green (B & K Green 7/5), sometimes withering early, lanceolate, 7212.5 cm long, 1.5-2.8 cm wide, caudate-acuminate at apex, subcordate at base, inserted on peduncle at 25° angle; stipe 3-5 mm long in front, 1-3 mm long in back; spadix violet-purple or green tinged violet-purple, (7)) 1-20 cm long, (6)7-10 mm diam. at base, 4-5 mm diam. at apex; flowers Square, 2.5-3 mm long, 2.1—3 mm wide, the sides weakly sigmoid; 4—5 flowers visible in the prin- cipal spiral, 7-9 flowers visible in the alternate Spiral; tepals glossy, lateral tepals 1.7-2 mm wide, the inner margin convex; pistils dark purple, scarcely emergent befi th ge; stigmas minute, scarcely visible; stamens emerging rap- idly from the base, the leading stamens emerging pi as third and fourth are emerged two- iiia the length of spadix, exserted on white, OPENA flattened filaments, ca. 0.4 tm long, es to red-violet 0.4—0.5 mm in both directions; key ellipsoid; pollen orange (B & K Yellow- ds 3). INFRUCTESCENCE pendent; berries it red-violet, obovoid; mesocarp gelatinous loc poen raphide cells; seeds 2, pale yel- E reddish brown at apex, green at base, » lattened, 4.6 mm long, 3.3-3.7 mm wide, ar thick, encased in a sticky substance. Fig. uu E from Bear sea level to 1,800 m ele- prettintar ^re a Rica it has been collected in collca ne rain forest. In Panama it has been forest in premontane wet forest, tropical wet » and premontane rain forest. en williamsii ranges from Costa Rica Mus Species is placed in section Polyneurium for is m to A. cuspidatum (see that species "Ierences). One collection from Costa Rica (Croat 36721) is somewhat intermediate be- tween A. williamsii and A. cuspidatum and may be a hybrid Costa RICA. CARTAGO: vicinity of Moravia, Croat 36628, 36721, 36790 (MO). Anthurium yetlense Matuda, Anales Inst. Biol. Univ. Nac. México 32: 151, Fig. 7. 1961. Type: Mexico. Oaxaca: near Yetla (San Ma- teo Yetla?), January 14, 1959, MacDougall 410 (MEXU, holotype). Terrestrial: stems ca. 30 cm long, 3.5 cm diam., green; internodes short; roots thick, tan, pubes- cent, directed downward; cataphylls subcoria- ceous, 7-9 cm long, tinged red-violet, rounded at apex, acumen apiculate, subapical apiculum 2-5 mm long, drying brown, persistent, splitting at base, the apex intact. LEAVES erect-spread- ing; petioles 28-45 cm long, 4-7 mm diam., te- rete, heavily tinged red-violet in early stages; ge- niculum 2.3-2.5 cm long, weakly sulcate; blades broadly ovate to rounded, moderately thick, gradually long-acuminate at apex, deeply lobed at base, 20—44 cm long, 15.5-29 cm wide, broad- est at point of petiol I tor slightly above; anterior lobe 15-29 cm long, the margins round- ed; posterior lobes 7-15 cm long from apex of sinus to outermost point; sinus obovate to cam- paniform, sometimes closed, acute at apex; both surfaces matte to semiglossy, lower surface much paler; midrib convexly raised above, flat at apex, acutely raised below; basal veins 4—5 pairs, first and second free, third to fifth coalesced 2.5-3 cm, raised in valleys above, raised below; pos- terior ribs + straight, inner margins sharply rolled up; primary lateral veins 4—6 per side, departing midrib at 40°-50° angle, curving toward apex and below; lesser veins scarcely visible; collective vein arising from the first basal vein, sunken above, raised below, 7-12 mm from margin. INFLO- RESCENCE erect-spreading, shorter than leaves; peduncles 18—20 cm long, 4—5 mm diam., terete, green heavily tinged with red-violet, one third to one half as long as petioles; spathe co- riaceous, green tinged with red-violet, broadly lanceolate, 7.5-9 cm long, 2.5—3 cm wide, broad- est at base, acuminate at apex, subcordate at base, inserted at 45? on peduncle; stipe 2-2.5 cm long in front, 1.5-1.8 cm long in back, 5 mm wide, green tinged with red-violet; spadix green (B & K Yellow-green, 6/7.5), 8-8.2 cm long, 1.2-1.5 416 cm diam. at base, 7-9 mm diam. at apex; flowers rhombic to 4-lobed, 3-3.2 mm long, 4-4.5 mm wide, the sides sigmoid; ca. 5 flowers visible in the principal spiral, ca. 7 flowers visible in the alternate spiral; tepals matte, weakly punctate, densely and minutely papillate, lateral tepals ca. .3 mm wide, the inner margin straight; pistils emergent, weakly raised, green; stigma linear, nearly obscure; lateral stamens emerging from base to apex, rapidly followed by third and fourth, exserted on short, translucent filaments; anthers yellow, ca. 0.5 mm long, 0.3 mm wide, held in a contiguous circle over pistil then retracting and opening at edge of tepals; thecae ellipsoid, scarce- ly divaricate; pollen yellow, fading to white. IN- F CENCE pendent; berries dark orange, ovoid, 1.2-1.5 cm long, 1.2-1.4 cm diam., rounded at apex with 4 faint ridges; seeds 2, green, 1-1.3 cm long, 7-8 mm thick, markedly flattened on one side. Figs. 222 and 223 Anthurium yetlense is endemic to Mexico in the Sierra de Juárez in northern Oaxaca. It is known only from the area south of Valle Na- cional from 600 to 1,500 m in wet cloud forests. The species is a member of section Belolon- chium and occurs with and is similar to 4. um- brosum. Both species are terrestrial; however, A. yetlense differs from A. umbrosum in that the inflorescence is much shorter than the leaves and the leaf is broadly ovate rather than ovate-tri- angular. The spadix of A. yetlense is green at anthesis whereas the spadix of 4. umbrosum be- comes violet-purple. MEx OAXACA: near Yetla (San Mateo Yetla), Mac Dougall 410 (MEXU), between Valle Nacional and Oaxaca, W of Valle Nacional, Croat 39768 (MO). SPECIES EXCLUDED Anthurium eggersii Engler. This species is a distinct species known only from Pacific Coastal Anthurium crassinervium (Jacq.) Schott. This species is restricted to bns Cones Cordillera of Venezuela. Specimens cited EON F*Viv5Sdlv A. salviniae or á. sehlechténdali Anthurium um Schott. This species is restricted to northern Colombia in the moun- tainous region south of Santa Marta. Material ANNALS OF THE MISSOURI BOTANICAL GARDEN --— OOS 23B, Heft 21, [Vor. 70 with which it has been confused in Central Amer- ica belongs to A. huixtlense. LITERATURE CITED BERLIN, B. & P. KAv. 1969. ters c Color Terms, Their on. University of Cali- Commentary on Mexican Ara- ceae. Gentes Hd 9. Cornell University, Ith- aca 1 Croat, T. B. 1976. Studies in Araceae I: Section Leptanthurium Schott, the Anthurium gracile- fra complex of Central and South Amer- ica. per a 1: 357—364. 979. The distribution of Araceae. In K. Lar- son P L. B. Holm-Nielsen, editors, Tropical Bot- any. Academic — Lon l owering behavior of the Neotropical genus Anthurium Pesce’ Amer. J. Bot. 67: 888- 904. 1978. Studies in Araceae Il: Anthurium Section Polyphyllium. Selbyana 2: 230- 238. . The genus Anthurium (Ara- ceae) in Costa Rica. Brenesia 16, Supl. 1, De- partment of Natural History, Museo Nacioldl De Costa Rica. TING. 1979. Standarta o Anth a2: 15-25 ENGLER, A. 1 ie die morphologischen Verhältnisse der Aracea ie Theil. Über Blattstellung und Sprossbe berhaltni der Araceae. Nova Acta der Ksl. Leop.-Caro un Deutschen Akademie der Naturforsher 39: 232. pere. Das Pflanzenr. IV. 1-330. Fores, G., L. Jim ae X. peer GAL, R. MONCAYO & F. TAKAAKI. 1971. Mapa de Tipos de en tacion de la Republica Mexicana 1:200, retaria de Recursos Hidraulicos. HOLDRIDGE, L. R., W. T. LIANG & J. A. TOSI. ments in Tropical Life Zones. Pergamm N Minem, u F. 1936. Araceae: Flora of Peru. Fie t. Hist., Bot. Ser. 13, 1(3): 428-48 9 m compatibility in the genus Anthurium on Tetra with a taxonomic revision of the seo B of mie mium. Doctoral dissertation, Uni waii 1980. , W. L. THEOBALD & H. KAMEMO ro Tax xonomy of Anthurium scandens ( Aroideana 3: 86-93. 3. STANDLEY, P. C. 1944. Araceae. Part Il. r^^ di RE ser Jr. & R. W. "UE Panama. Ann. M . & Lh x. 1958 363. of Guatemala. Fieldiana, Bot. 24d) 1 304-3 cm E E LLLLLL ,LLL vein an uii ra aie I sa RR INDEX OF LATIN NAMES Numbers in boldface type refer to descriptions; numbers with ! refer to illustrations; names in italics refer to synonyms Acorus 211 Anadendrum 211 Anthurium 240 —sect. Belolonchium Schott 220, 223, 246, 256, 260, 267. 268, 270, 272, 279; 282, 284, 309, 317, 28,027, 331; 336, 342, 355, 363, 364; 366, 367, 385, 392, 393, 400, 404, 410, 416 —sect. Calomystrium Schott 215, 216, 222, 250, e E 297. 202.1207, 305, 306;:329; 335, —sect. Cardiolonchium Schott 216, 222, 284 —sect. Chamaerepium Sc —sect. Dactylophyllium Schott 220, 277, 308, 402 —sect. Digitinervium Sodiro 216, 220, 311 —sect. Episeiostenium sep 281 —sect. Gymnopodium Engl. 219 —sect. Leptanthurium Schott 219, 222, 242, 300, 323, 396, 416 —sect. Oxycarpium Schott 223, 349, 351 215; 216; 219, 221, 01, 302, 323, 324, 336, 356- 358, 360, 369, 375; 381,382, 386, 390, 391, 405, 409, 4 —sect. Polyneurium Schott ci vat 342, 415 245, en. 252. 292. 299, 300, 310, 320, 341, 351, 359, 374, 407, - Sect. Ütfiroplációm Schott 220, 343, 356 —sect. caer Schott 219, 394, 398 —sect. Tetraspermium Schott 216, 219, 220, 372, 400, 401, 416 —sect. Uro ospadix Engl. 2 —sect. Xialophyllium seno n 223; 261; 265, 289, 307, 326, 340, acutangulum p 213, nin 234, 235, 242, 243!, 245, 251, 359 acutifolium Engl. 221, 240, 243!, 244 aemulum Sch meses Croat & Baker 220, 234, 243!, 359 , 245, --— Standl. 309 methystinum Croat & Baker 323, 324 coe num Kunth 325 andicola emen; 213, 214; 220; es 245, 246, 247!, 83 angustispadix Croat & Baker 223, 238, 247!, 248 angusturense En (340 eniense Croat 21 1, 213, 222, 231, 2471, 249, 251, 253, 306 aureum Engl. 310 cono Croat & Baker 220, 234, 248!, 251 i bakeri Oo f 213, 214, 220, 233, 248!, 252, 2531, Nnm K. Krause 325 n rnt Standl. ex Yuncker 391, 392 M Standl. & L. O. Wms. 213, 222, 231, 254, berriozabalense Matuda 213, 214, 220, 224, 231, 2531, 256, 258!, 259, 312, 383 bochilense Matuda bogotense Schott var. concinnatum (Schott) Engl. 279 tt 348 z4 [S] o brachyspathum C. Koch & Bouché 400 brenesii Croat & Baker 221, 238, 255!, 258!, 259, 357 brevipetiolatum Engl. 342 brownii Mast. 223, 224, 227, ns 260 burchellianum (Engl.) Macbr burgeri Croat & Baker 223, nu n 260, 261 campii A. D. Hawkes 245, 246 caperatum Croat & Baker 223, 227, 2571, 261, 262 carnosum Croat & Baker 223, 235, 258!, 263, 264!, 265, 307, 349 caucanum Engl. 222, 289 tuda 221, 228, 229, 233, 264!, 265, 266!, 267, 270 cerropelonense Matuda 213, 223, 230, 231, 266!, 267 vieler Croat (ined.) 396 cham ssp. pase Pica Matuda 213, 221, 228, 229, 231, 233, 246, 267, 268, 269!, 270, 271, 282, 329 ssp. oaxacanum Croat 211, 213, 221, 229, 264!, i chiapase ssp. aa Standl. 213, 220, 225, 269!, 271, 321 4! chlorocardiu a & L. O. Wms. 392 chochotlensis Matuda 24 circinatum Croat 320 clarinervium Matuda 214, 222, 231, 3s 275,312 clavatum Croat & Baker 222, 225, 274 276 clavigerum Poepp. & Endl. 213, 220, eed 2 276, 277, 278! ssp. prie aap Croat 211, 220, 225, 272. 2135 arsi Standl. 221, 225, 277, 278!, 279 codajas . Barroso coiblonii Standl. & gag 391 pe ae Engl. 3 colonicum K. Krause a 362 — Schott 223, 226, 244, 263, 278!, 279, concolor x Krause 405 consobrinum Schott 221, 235-237, 280!, 281, 292, r. cuneatissimum Engl. 2 PAS tok ER Mu 221. 230, 278, 282, 2 3!, corrugatu Sodiro sg ea Engl. to T poked 221, 226, 280!, 282, 283!, 84, ssinervium (Jacq.) Schott 416 crenatum (L.) Kunt 5 cruxiphyllum Matuda 271 418 cubense Engl. 213, 214, 221, 235, 283!, 284, 285, 286! cucullatum C. Koch cuneatissimum (Engl. . LA 211, 221, 239, 285, 286!, curvilam: minum Croat (ined.) 301 cuspidatum Mast. 221, 226, 238, 286!, 287, 288, 415 cuspidatum Matuda 271 cuspidifolium ae 336 davidsoniae Standl. 222, 225, 288, 289, 291! ecurrens Poepp. 349 donnellsmithii Engl. 2 durandii Engl. 219, 220, 234, 235, 251, 289, 290!, 92, 359 eggersii Engl. 314, 416 enormispadix Matuda 369 eryt tachyum Yee (ined. eximium Engl. 221, 222, 237, 240, 290!, 292 fatoense K. Krause 221, 236, 239, 281, 291!, 293 filiforme Engl. 359 firmum Engl. 309 flavescens Poepp. 263, 336 flex ssp. flexile Schott 214, 221, 225, 290!, 291!, 293, 296! ec muelleri Croat & Baker 221, 225, 294, 295! rmosum Schott 218, 222, 227, 2961, 297, 298! fortinense Hook. ex Schott 374 raternum Schott 276, 306, 385, 416 Fidei sehe Orr em 219, 220, 234, 297, 2981, 299, 300, 4 ragaranum ghiesbreghtii Linden = Schott 355 eum Matuda 3 glaucotitanioides n anak (Rudge) Lindl. 213-215, 219, 222, 236, 292, 296!, 299, 30 var. r. friedrichsthali (Schott) Engl. 297 gracilens Standl. gracilescens erdum a ymnopus Grise aeina Engl. 213, 220, 233, 296!, 298!, 300, 301 halmoorei Croat 211, 214, 221, 236, 239, 301, 302, !, 304!, 378 helleborifolium Schott 347 hoffm por Schott 222, 227, 276, 302, 303!, 3041, ; 335 huixtlense tees 213, pe e 230, 231, 251, 303!, 304!, 305, ae hypoleucum Standl. interruptum Sodiro 213, 214, 223, 235, 263, 303!, , 315!, 349 ison erviu um Sta ndl. & L. O. Wms. 359 Peete N. E. Brown 277 rwinskyi Schott 348 rdi Matuda 331 kunthianum Liebm. 374 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 kunthii "ead & Endl. 220, 224, 304!, 307, 308, lactiflorum Engl. 297 lan wort Crom 211, 213, 223, 230, 232, 308, 309, 314 lancifolium he 220, 234, An 310, 314! var. albifructum pes (ined.) 310 lapathifolium Schot latihastatum Engl. ex K. Krause 396 lentii C rnt Baker 220, 224, 310, 311, 314!, 315! leuconeurum Lem. a 222, 231, 311, 312, 318 315! lezamae Matuda 214, 222, 231, 312, 313, 3161, 366, liebmannii Schott 402 linearifolium Engl. 297 iis e Engl. 289 audatum Engl. 288 i oa acti Matuda 213, 220, 224, 3161, 317, 319! longipes Matuda 299 longispathum Schott 386 louisii Croat & Baker 220, 233, 318, 320! lucens ds ex vni 220, 225, 271, 319!, 320, , 3221, 323, 364. Pire Matuda dm 246, 282 arpege ais tuda 221, 237, 3221, 323 bum 12 Hort. mattia nianum Croat (ined.) 398 mapastapecense Matuda 320, 321 ; membranuliferum Schott & Ohlend. ex Engl. 35 mexicanum Engl. 294 mexicanum Liebm. 374 mexicanum Macbr. 294 michelii Guillaumin 221, 238, 240, 319!, 323, PU microspadix Schott 213, 214, 222, 225, 235, 19!, 322! leui mirandae Mat monserratense eA 268, 270 Vesti iiigeorirs Steyerm. 277 montanum Hemsl. 213, 221, 230, 231-233, 282, 322!, 326, 327, 328!, 329, 4 329 Wie rici Croat & Baker 222, 227, 305, mon ntico olum motozintl yp tuda 268 myosuroides (HBK) Endl. var. angustifolia Engl. E^ weg 213, 214, 221, 228, 237, 330. 3321, nane alense ia 325 nelsonii Croat 211, 213, 221, 229, 330, 332! rvatum Croat (ined.) 360, . nah ei Matuda 213, 214, 221, 236, 239, 32 3325 333, 334!, 378 nymphiifolium C. Koch & Bouché 276 oaxacamonticolum Matuda 245, 246 oaxacarupicolum Matuda 282 obovatum Gleason obtusilobum Schott 213, 222, 226, 333, 3 ochranthum C. Koch 213, 223, 228, 334!, 355 341, 335 335, 336. ocotepecense Matuda 339 38! oerstedianum Schott 221, 225, 235, 336, 3 oreophilum Sodiro 288 orteganum Engl. 287 ————————— ———— —"—————ÉHEÉÉREREREERE————ÉEERRRA —————————— u—À—Áá——nnam- l—————— Vl — eÁ—— A7" ———————— d 1983] INDEX ovandense Matuda 213, 223, 232, 338!, 339, 366, 3 ovatifolium Engl. 311 pallens Schott 222, 235, 237, 325, 337!, 339, 340, 341, 39 paludosum Engl. 220, 234, 292, 337!, 340, 359, 416 d.) 262 var. burchellianum E panduriforme Schott 223, 224, 226, 337!, 338!, 341 n thum Hemsl. 213, 221, 237, 330, 338!, 342, pedatifidum Regel & Linden 343 pedatoradiatum ssp. blicberifollum (Schott) Croat 211, 220, 224, 343, 3441, 346!, 347 ssp. pedatoradiatum Schott 213, 214, 220, 224, 275, 312, 343, 3441, 346!, 347 pendens Croat (ined.) 414 pentaphyllum var. bombacifolium poete Madison ans, 214, 220, 224, 344!, 346!, 348 pittieri var. fogdenii Croat 223, 237, 345!, 350 var. pittieri Engl. 223, 235, 237, 263, 307, 345!, , 350 pluricostatum Croat & Baker 223, 227, 336, 3461, 351, 3541, 355 podophyllum (Cham. & Behien. ) Kunth 213, 214, 220, 224, ds UN 355, 3 polytomum Scho 355 porrectum Nu 325 porschianum K. Kraus Prolatum Croat & Baker 221, 238. 3521. 356, 357 protensum Schott 219, 221, 238, 3531, 357, 382, 409 icum. Standi. & Steyerm 307 monense Engl. ex K. Krause 213, 220, 234, 242, 353!, 354!, 359 ranchoanum Engl. 221, 226, 284, 354!, 360, 361!, 362, 390 ravenii coe A Dikar 213, 222, 227, 361!, 362, 363 recussatum Schott 285 retiferu y m Standl. & Steyerm. 213, 221, 237, 342, 361!, 363 rigidulum Schott 371 oe Matuda 220, 225, 364, 365! ede Matu uda 221, 229, 232, 364, 365!, 366 Esca, 221, 228, 230, 232, 364, 365!, Mivadorense Croat 221, 236, 239, 367, 368!, 369, sib Hemsl. 214, 221, 236, 368!, 369, 371, 375, dens SSP. pusillum Sheffer 220, 233, 370!, 373 Ssp. scandens (Aubl.) Engl. 213, 214, 220, 233, a d TAE 395, 400, 401, 416 erve (Miq.) Matuda 400 Scherzerianum Schott 220. 233, 3701, 373, 374 419 schlechtendalii ssp. jimenezii (Matuda) Croat 211, 213, 214, 221, 239, 302, 333, 375, 377, 378, 380! ssp. schlechtendalii Kunth 213, 214, 216, 221, 236, 39, 323, 369, 371, 374, 375, 376!, 378, 416 schottianum Croat & Baker 221, 225, 376!, 378, 381 colopendrinum (W. nth 299 c Standl. & L. O. Wms. 244 seamayense Standl. 320 idee: Croat & Baker 221, 240, 357, 379!, 381, [^] S 382, 409 seleri s 213. 221: 224. 278,.232. 237, 258. 239, 379!, 380!, 382, 383, 384! silvaticum Sodiro 325 "-— Standl. & Steyerm. 223, 230, 380!, 383, sinuatu um n Benth. ex Schott 277 peni fen M E. Brown 213, 221, 239, 287, 384!, 385, 386 spectabile Schott 221, 226, 240, 360, 384!, 386, 390 standleyi a 221, 226, 227, 387!, 390 subcordatu ssp. asec A (Standl. & L. O. Wms.) Croat 211, 221, 228, 229, 230, 387!, 392, 393 ssp. subcordatum Schott 221, 228—230, 238, 267, 342, 3871, 389!, 391, 392-394 subovatum Matuda 213, 221, 228, 229, 238, 388!, 389!, 392, 393, 394 subsignatum Schott 219, 224, 225, 388!, 394, 398, talamancae Engl. 2 redii tebe ed 4325 nerum Engl. 223, 235, 388!, 394, 395 pets Engl. 326 testaceum Croat & Baker 223, 235, 237, 389!, 395, tetragonum Hook. ex Schott 369, 374 tikalense C. Lundell 374 tilaranense Standl. 219, 224, 396, 397! titanium Standl. & Steyerm. 213, 221,229, 230, 317, 394 trinerve Miq. 213, 214, 220, 233, 372, 399!, 400, var. angustifolium K. Krause 401 var. obtusum Engl. 401 sectum Sodiro 220, 223, 399!, 401, 402, 403! turrialbense Engl. 252 wae rapes ae 223. 232; 259, 309. 327, 2, 403!, 404, 4 upalaense Croat * wine 213, 221, 239, 240, 403!, 404, 405, utleyi Croat & m. 220, 234, 405, 407, 408! valerii Standl. 360 validifolium K. Krause 221, 239, 406!, 407, 408!, 4 verapazense Engl. 216, 220, 224, 225. 271,272 317, 323, 406!, 408!, 409, 41 vinicolor Standl. & L. O. Wms. 360 420 ANNALS OF THE MISSOURI BOTANICAL GARDEN hapa neris Schott 371 rma angustifolia Kunth 371 a latifolia ae 37 eee Sch waterm alid ed ex Bailey 221, 226, 411, 413! wendlandii Schott 27 wendlingeri G. M. Barroso 220, 233, 411, 412!, 413!, 414 williamsii K. Krause 223, 226, 288, 412!, 414 xanthosomifolium Matuda 317, 400 yetlense Matuda 214, 223, 230, 412!, 415, 416 yunckeri Standl. ex Yuncker 320 Dracontium repens Descourt 371 1-371 es aig 299 laciniatus M. Martens & Galeotti 355 vi SERA Swartz 371 [Vor. 70 ds foa iati He eitis pinta Y pig dà ain at EN as "b mik it ae ——À an — Professional Dr. amateur? OPERA BOTANICA OPERA BOTANICA is a supplement to the NORDIC JOURNAL OF BOTANY, containing more comprehensive papers Within the same fields. Recent titles are: Anders Theler, The genera Dirina and Roccellina (Roccellaceae); Gunnar Sei- denfaden, Orchid Genera in Thailand XI. Cymbidieae Pfitz: Lauritz W. Olson, Al- lomyces ~ a different fungus; Ole Matts- Son, The morphogenesis of dimorphic Pollen and anthers in Tripogandra am- plexicaulis. Tübscription price (1984): 30 D.Cr. per 6 printed pages, plus postage. Subscription and sale Payment is by invoice. Orders and in- quines should be placed with, The Sec- bail Nordic Journal of Botany, Inst. for et, @. Farimagsgade 2D, K-1353 Copenhagen K, Denmark. — you'll be satisfied with the NORDIC JOURNAL OF BOTANY The NORDIC JOURNAL OF BOTANY is an international journal on primary re- search within taxonomy, geobotany and anatomy of higher plants, fungi, lichens and algae. In addition, the journal con- tains contributions from important inter- national symposia, book reviews and ews. Being the product of an amalgamation of old Nordic journals, nestor among which the Botaniska Notiser from 1835, the NORDIC JOURNAL OF BOTANY is edi- ted by Nordic scientists. Editor-in-chief is Professor Morten Lange. Subscription price (1984): 650 D.Cr. (c. US$ 65) per year, plus postage. Sections The contents of the NORDIC JOURNAL OF BOTANY are divided into seven sec- tions: 1. General botany and taxonomy of higher, holarctic plants, 2. Taxonomy of higher, tropical plants, 3. Geobotany, 4. Structural botany, 5. Mycology, 6. Li- chenology and 7. Phycology. Articles within a section are available to individual subscribers as first class off- prints, distributed three times a year. Individuals subscribing to a section may buy the corresponding issues of OPERA BOTANICA at a reduced price. Subscription price (1984): 135 D.Cr. (c. US$ 13.50) per year, plus postage. Annual rates: ANNALS: | ANNALS SSOURI BOTANICAL CARDEN 1983 NUMBER 3 CORNUS KOU: CONTENTS BIOGEOGRAPHICAL RELATIONSHIPS BETWEEN TEMPERATE EASTERN ASIA AND TEMPERATE EASTERN NORTH AMER- ICA: The Twenty-Ninth Annual Systematics Symposium Gerrit Bue ee E 42 Eastern Asian-Eastern North American Paoa raphin Relation- ships—A History from the Time of Linnaeus to the Twentieth Century D. E. Boufford & S. A. Spongberg 23 Cretaceous and Cenozoic History of the Northern Continents | Warren Hamilton : 440 Holarctic Landmass EE Goak Events, and Cenozoic Terrestrial Organisms Malcolm C. McKenna 459 Late Cretaceous and Cenozoic Vegetation in China, Emphasizing their Connections with North America Jen Hsü Vegetation of China with Reference to its Geographical Distribution 509 Hsioh- Yu Hou Quaternary History of Deciduous Forests of Eastern n l and Europe Margaret B. Davis | mee BOTANICA The Tibetan Plateau in Relation to the Vegetation of China avic H. S. Chane A Comparative Study ee V | in the Carolinas of the United States North America p^ 2 1 egetation in Hubei Province, C WUL | 984 Zhong Cheng 71 OF = outside the Garden will also be accepted. Authors should writ Editor for information conrerning SE for pu the ANNA ALS. "EDITORIAL COMMITTEE NAN Y Morin, Editor re Mni MP ue Garden — _ MarsHALL R. CROsBY —— - Missouri Botanical bog DER T DAVIDSE Missouri Botanical e fl — ANNALS MISSOURI BOTANICAL GARDEN VOLUME 70 1983 NUMBER 3 BIOGEOGRAPHICAL RELATIONSHIPS BETWEEN TEMPERATE EASTERN ASIA AND TEMPERATE EASTERN NORTH AMERICA: THE TWENTY-NINTH ANNUAL SYSTEMATICS SYMPOSIUM! The Twenty-ninth Annual Systematics Sym- posium was held at the Missouri Botanical Gar- den in St. Louis on 23-25 September 1982 and consisted of two days of contributed papers and one day of invited papers. All previous meetings In this Popular symposium series have been one- oe each with a program of invited aa primary reason for the change of Bom 1s time was to take advantage of the of Asia n = the symposium of a large delegation Pais z pus especially the large group (16) this * People’s Republic of China. Because Mea mathe largest delegation of Chinese bota- States sin fe in the United Pii ce 1949, 1t was an excellent opportunity a ann to exchange ideas and infor- Bepalen to renew or establish contacts for a ive scientific research and interchange. ex MI relationship between temperate east- in bio and North America is a classical topic geography. The striking biological similar- T 1 : . Asins sliminary arrangements for this symposium, and particularly for the pa were mada h du itt ised of B. Barthol ity between the two continents was first com- mented upon by Linnaeus in 1750 and was later, beginning in 1850, much elaborated upon by Asa Gray. The papers presented in this symposium provide a broad review of the geological, pa- leontological, zoological, and botanical aspects ofthe problem Both cl ical d cladisti th ods were used to analyze these biogeographical relationships. A special effort was also made to present a broad overview of the vegetation of China. Invited speakers were the following: Daniel I. Axelrod, University of California, Davis; Mar- garet B. Davis (evening speaker), University of Minnesota; Warren Hamilton, U.S.D.I., Geo- logical Survey; Hou Xue-yu, Institute of Botany, Beijing; Jen Hsü, Institute of Botany, Beijing; Malcolm C. McKenna, American Museum of Natural History; and Wu Zheng-yih, Institute of Botany, Kunming. Because of the length of this symposium, it is Haue tani Ber] 3 rs 7 se iversi Nine ey (currently California Academy of Sciences); D. E. Boufford, Harvard University Herbaria; uri Botanical Botanical Garden TB, and S. A. Spongberg, Harvard University rDaria, D, reviewed and e Vegetation th e L4 (Mi : ; udo au Botanical Garden) assisted with additional preliminary editing. The contribution of these people to & nf . pa pa PET ae aia 4 . 6.1 1 1 dg A Th a? of China (1 :] .400 NOD) 41 : } Nan Hsioh-Hu Hou (Hou Xue-Y), Insti f Bot D dissi Mi iA Beijing People’s Republic of China. — €Y R. Morin, editor. ANN. Missour: Bor. Garp. 70: 421-422. 1983. 422 ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 being published in two parts. The second part will dition, Mark Twain Bancshares and anonymous appear in the following issue of the ANNALS OF donors helped defray travel expenses THE MISSOURI BOTANICAL GARDEN. This symposium was supported by National Science Foundation Grant DEB 78-10180. In ad- Gerrit Davidse, Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166. NOTE ADDED IN PROOF: Since this symposium issue went to nep an extremely important paper by Zhang et al., which discusses the complicated tectonic evolution of Chin appeared. The complete reference is: Zh. M., J. G. Liou, and R. G. Coleman. 1984. An Outline of m Plate Tectonics of China. Geological Socie of America Bulletin 95:295-312. Reprints may be obtained from the authors, Department of Geology, Stanford University, Stanford, California 94305. EASTERN ASIAN-EASTERN NORTH AMERICAN PHY TOGEOGRAPHICAL RELATIONSHIPS—A HISTORY FROM THE TIME OF LINNAEUS TO THE TWENTIETH CENTURY! D. E. BOUFFORD AND S. A. SPONGBERG? ABSTRACT EF. The "p ER and | study of eastern North American eastern Asi the time ofI Thomas ; Nuttall's ioni cba overlooked ebion to this pea is discussed, and an annotated spine of the taxa that he attributed to eastern Asia an merica is presented. Charles Darwin’s influence on the thoughts and writings sa Gray is also discussed, based on published and unpublished letters and man PO in the Archives of the Gray Herbarium of Harvard University; the role that this correspondence played in the development of Gray’s phytogeographical ideas an rwin’s theory of evolution is considere d. A brief summary of the writings of Adolf Engler, and other hx century i hemisphere is also given. The sign making known the full extent of disjunc ns the basis for further work beyond the sry and theoretical stages of this fascinating pattern of plant disjunction. The discontinuous distribution of the same or — related taxa of plants between eastern America and eastern Asia is but one of nee patterns of disjunction that become ev- ident when the flora of the temperate Northern Hemisphere is c d (Raven, 1972; Thorne, 1972; Wood, 1972; and others). Many of these patterns have been termed Tertiary relict dis pene Of these, however, the *'classic" east- orth American-eastern Asian pattern is un- chan the best known and most often cited example of the disjunct occurrence of closely re- ated taxa on two continents separated by thou- sands of kilometers. Despite the fact that most of the shared taxa in the two regions have been Sca to be distinct from one another (the re- d M are primarily closely related species " * same genus, or closely related genera of * same family), other factors noted by bio- Th € similarities of the forests of Japan, central CMM ‘We th os g Shins ofa other literature that we were lac ANN. MISSOURI Bor. GARD. 70: 423-439. 1983. kin arvard University Herbaria, 22 Divinity Avenue, ‘Cambridge, Massac China, and the southern Appalachians in ap- pearance as well as in ecological associations are in many instances so great that a sense of déja vu is experienced by botanists of one of the re- gions visiting the other. Another important factor that has served to emphasize the floristic relationships between eastern North America and eastern Asia is the fact that this particular disjunction pattern was the first to be recognized by botanists. Moreover, edi y and nce of this pattern fig- wed n in n due discussions surrounding Darwin's theory of evolution, and it has been discussed not only by plant geographers but by botanical and scientific historians as well. The purpose of this paper is to trace briefly the historical aspects ofthe recognition ofthis distribution pattern and the accumulation of facts pertaining to it, and to summarize the major contributions to the study of this pattern that have been made from the time of Linnaeus to the beginning of the twen- tieth century. Special mention of the contribu- tion and observations of Thomas Nuttall is in- n of the manuscript, O. T. Solbrig, Director of the Gray Gray material in the Archives of the Gray Herbarium, and iren for ed patience and “ti We also thank , be this anu and for their comments, and P. ies non. Lastly, we are saad to B. Bartholomew st the library of the C providi y of Miquel's paper of 1868, and d PH , and P. F. H. Rav ven, who suggested | that we attempt A iences Raven and the Missouri Botanical Garden for husetts 02138. 424 cluded inasmuch as his penetrating insights into the relationships between the eastern Asian and eastern North American floras, as well as his phytogeographical observations, have to our knowledge hitherto passed unnoticed. EARLY RECOGNITION OF THE FLORISTIC SIMILARITIES BETWEEN EASTERN NORTH AMERICA AND EASTERN ASIA— HE LINNAEAN ERA The history of the early recognition of the flo- ristic similarities between eastern Asia and east- ern North America has been well outlined by Graham (1966, 1972b) and Li (1952, 1955). While credit for the discovery of the floristic re- lationships has freq ly been given to Asa Gray, it was pointed out by M. L. Fernald (1931) tha the first published reference to these similarities appears in a Linnaean dissertation, Plantae Camschatcenses Rariores, that was published in 1750 by Linnaeus’s student, Jonas P. Halenius. As discussed by Graham (1966), the dissertation, including the listing of nine species then thought to occur in both Kamchatka and North America, was probably written by Linnaeus himself and not his student, who was apparently only re- quired to defend the dissertation in Latin. In 1751, the year after the initial publication of Halenius’s dissertation, Pehr Kalm, another Swedish botanist, who had traveled extensively in North America during the years 1747 to 1 749, published a “Short Account. . .” (see Larsen, 1939) of the plants he had collected in North America, seeds of which he had taken back for culture in Swedish gardens. In both this publi- cation and in his Travels into North America (published in Swedish in the years 1753 to 1761, appearing in a first English edition in | 770), Kalm mentioned ginseng (Panax quinquefolium L.) and noted that it also grew in China and Korea. Kalm also noted the reputed medicinal value attached to the plant by the Chinese and recounted that the French collected the plant in Canada for ex- port (via France) to China. The dissertation defended by Halenius was re- printed in 1751 and again in 1787 (Graham, 1966); this last edition utilized Latin binomials and appeared three years after Thunberg’s Flora Japonica (1784) had been published. Thunberg, so a student of Linnaeus but by 1784 an es- tablished botanist and explorer in his own right, also made brief mention in the Preface to his flora that many Japanese plants also Occurred in ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vou. 70 Europe, America, and the East Indies, but par- ticularly in the northern and adjacent, vast Chinese region. Among the plants treated are several that give direct evidence that Thunberg considered some plants in Japan and North America to be the same. Included in this group are? Sanicula canadensis L. Viburnum dentatum L. Sambucus canadensis L. Lilium canadensis L. Juniperus virginiana L. Pinus strobus L. Amorpha fruticosa L. Adiantum pedatum L. in thé No comments were made by Thunberg !n ivan ntrast- Thunb. is distinguished from A. negundo trifidum Thunb. is separated from A. pe icum L., and Spiraea callosa Thunb. 1s CO ith S. tomentosa L. HE s ni evid that Thunberg was awa? ofthe Halenius dissertation, but the next bo gr to comment on the floristic affinities be two regions was not only aware of TE Flora Japonica but compared it with ST” vius's (1739, 1743) Flora Virginica. ee pe^ glioni used a comparison of these two e ril evidence of the similarity of the cim two regions in his two volume work n negli Stati Uniti, that was published in i ws Thur *The nomenclature and taxonomy follor 03008 berg p that apply to these plants in both many are obvious. “regions today: but | 2 —— 1983] 1790, three years after his return to Italy from the United States, where he had spent three years (1785-1787) traveling. In the account of his trav- els in the United States, he noted the affinities of many American plants, which he had ob- served growing, with those of Japan. In the sec- ond volume (p. 156) he wrote, *It is moreover to be observed that the plants are very similar to those of the eastern coast of northern Asia, over, with a lengthy section (pp. 169—402) en- titled *Osservazioni sui Vegetabili piu utili degli Stati Uniti" in which numerous North American species were described with notes on their dis- tribution and uses. Castiglioni’s work appeared in a German translation in 1793, and the original Italian edi- tion was apparently reissued in 1845 (Stafleu & Cowan, 1976). It remained, however, for C. E. Faxon (1891) to point out the significance of Cas- üglioni's Observations in relation to the biogeo- graphical relationships between Asia and North America. Faxon's appreciation of Castiglioni's Work also remained relatively unknown until Li (1955) rediscovered it, after his 1952 survey of the history of the recognition of the relationships had been published. FREDERICK PURSH ene Pursh (1814) was apparently the next lanist to comment on eastern North Ameri- cu ern Asian floristic relationships. Pursh's d to the knowledge of plant disjunc- em Asi tween eastern North America and east- mi me however, was relatively slight and pri- the Poi imited to the single sentence describing in " herbarium, which was quoted by Con- d? (1972) and is repeated here: **This exten- * collection, now in the possesion of A. B. mbert, Esq., was highly useful to me, in com- rae the plants of North America with those * north of Asia, to some of which they have pe ons a and others are common to both Es PE The few other instances where Pursh mes note of eastern Asian-eastern North A *rican disjuncts in his Flora are indicated in Ppendix 1. BOUFFORD & SPONGBERG — HISTORY 425 THOMAS NUTTALL In twentieth century reviews of early botanical literature dealing with the disjunct distribution of plants between eastern Asia and eastern North America, Asa Gray has generally been given the major share of the credit for being among the first to point out this pattern (Fernald, 1931; Hara, 1952, 1972; Li, 1952; Graham, 1972; Good, 1974). Because ofthe detail of Gray's (1840, 1846, 1856, 1857, 1859) studies, the attention they re- ceived at the time, and their contemporary im- portance in relating to Darwin's ideas on evo- lution, Gray has also been credited with being among the founders of the field of biogeography (Good, 1955; Li, 1952). The few papers that pre- ceded Gray's (Halenius, 1750; Thunberg, 1784; Castiglioni, 1790; Pursh, 1814) were in no way as detailed, nor as critical in their analysis of this pattern of disjunction as was his “Diagnostic Characters" paper of 1859. Each of these papers tended to point out that some of the plants from North America were the same as plants in dis- tantly separated parts of the world, particularly those in eastern Asia. None of them discussed the implications of populati fa single speci being so widely separated and, with botany in most of the Northern Hemisphere still in the exploratory stage and with vast areas still bo- tanically unknown, it is unlikely that they should have. The explorations that continued through the early part of the 1800s in the American West, in eastern Asia, in Canada, and in other parts of the world led to the discovery of large numbers of new species and made available numerous specimens to help fill in gaps in plant distribu- tions. The floristic works mentioned above that were based on these explorations took into ac- count specific geographical regi d were often based on very limited numbers of specimens so that it was difficult to make direct comparisons with other parts of the world except through the literature. Even so, the influx of specimens con- tributed to the growth of the large private her- baria that were acquiring specimens at the time and eventually made possible the kinds of com- parisons of the world's flora made by Gray. One exceptional work, however, that preceded Gray's papers, and one that has apparently been overlooked despite its importance to North American taxonomy, is Thomas Nuttall's (1818) Genera of North American Plants. Nuttall's Gen- era was the culmination of nearly ten years' study ofthe plants of North America, particularly those 426 of the area that was then the United States and its territories. Nuttall was unusually well pre- ared to write the Genera for he was equally familiar with many North American plants in the field and in the herbarium, and had become familiar with others that were growing in the gardens of Philadelphia, particularly those in Bartram’s Garden (Graustein, 1967). His studies before writing the Genera included field work across much of the northern United States from the East Coast nearly to the Rocky Mountains, travel down the Mississippi River to New Or- leans, field work in the eastern coastal region, especially around Philadelphia, short excursions into New York state and into Canada around Niagara Falls, and a fairly long trip through the southeastern United States that retraced parts of the routes of several earlier botanical explorers including those of Bartram and Michaux (Grau- stein, 1967; Pennell, 1936). From the time of his arrival in Philadelphia in the spring of 1808 Nut- tall had ready access to the largest herbaria in the United States through his connection with Benjamin Smith Barton and other Philadelphia botanists and was later able to study specimens in the large private herbaria in England before preparing his Genera. Nuttall was also apparently aware of the works of earlier and contemporary botanists. In the in- but it was more than two years before he could return to the United States for he was caught in England when the War of 1812 broke out. He made good use of his time by contacting the leading botanists in En- ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo.. 70 gland and in examining the holdings of Joseph las in Siberia and eastern Asia. Nuttall's famil- iarity with the plants ofthe northern United States and the chance to examine specimens in the Lambert and other herbaria, where he no doubt saw many North American plants that were still unknown to him, prepared him well for a trip to the southeastern United States when he returned to Philadelphia in 1815. He was able to retrace the routes of many ofthe first botanical collectors to that part of the country as he traveled through the mountains of Virginia, Kentucky, Tennessee, and the Carolinas and across the Piedmont and Atlantic Coastal Plain provinces of the Carolinas and Georgia (Graustein, 1967). Nuttall was also able to examine specimens in the herbaria of several noted contemporaries who lived in that part of the country and saw additional plants from the southeastern United States in their col- lections. One particularly useful collection was that of William Baldwin, who had collected spe imens in western Florida during two years pos there while in the Navy (Graustein, 1967). = lowing these field and herbarium studies, Nu t spent most of the years 1816 and 1817 -— the Genera, which was published in 1818 (Pen nell, 1936). The eie important aspect of the Genera = a phytogeographical standpoint was Nuttall's clusion of comments on “A view of the = graphic distribution of each genus .--> pion ways sufficiently accurate for the pee of the science, and the rapid progress e » discovery" (Nuttall, 1818, p. vii). These tie on geographic distribution were Nutta fate tempt to explain the relationships and e: 2 tion of North American plants on à wor scale. Nuttall mentioned the geographi cal range and ; Jants While he did not limit his observations e exhibiting an eastern Asian-easte f interest North American genus was also found ee Asia (here taken to include the Indian tinent and the Himalayas), but in other signif" seems clear that he was well aware of dely tP icance of disjunction between the two W! — 1983] arated regions. Nuttall's comments (pp. 274, 275) under Chimaphila, a genus in which he recog- nized two species, C. umbellata and C. lata, are especially insightful: "Indigenous also to the North West Coast of enzies. Probably both species of the genus will be found also in East Asia and Europe." The prediction that plants peculiar to one of these regions mi; be expected in the other was later made by Gray (1859, p. 416) and discussed by Wood (1972), but the fact that Nuttall had earlier suggested this has apparently been overlooked Although Nuttall did not restrict n — nte to plants in those two geographical regions, we have listed in Appendix 1 examples from the Genera where he indicated disjunction between eastern North America and eastern Asia. The few instances where Pursh (1814) had previously commented on the same genera are also indi- cated. It must be remembered that taxonomic concepts of the early 1800s were different from our own. Nevertheless, Nuttall’s ideas on phy- togeography and the relationships of the North American flora were remarkably modern in many rospects and deserve to be accorded their rightful place in the progression of knowledge of plant 1860; Ewan, 1971; Graustein, 1967; Jones, 1937; Pennell, 1936; Smith & Thieret, 1959; Stuckey, 968) have mentioned it. We hope that by in- dicating these examples of his awareness and in- terest in plant distributions that future biogeog- Taphers will acknowledge particularly his contribution to the study of eastern North Amer- ican-eastern Asian plant disjunctions. ASA GRAY — indicated aboye; Asa Gray was not the first een. f the east- s coe American and eastern Asian floras, cite s series of papers on this topic did focus € attention on this disjunct distribution Me m. Moreover, Gray fully examined and nin ved the problem utilizing the materials an € to him and in so doing laid the foun- on which future work was possible. Gray's YPothesis of the causes of the eastern North em Asian disjunctions, more- mi a rought the observed pattern into the sci- € realm of biogeographic discussion and ysis on a aa A level. FVL OUL BOUFFORD & SPONGBERG—HISTORY 427 The first evidence that Gray was aware of the relationship appeared in his reviews (1840, 1846; see Graham, 1972a, b, and Stuckey, 1978, for reprintings) of Siebold and Zuccarini’s Flora Ja- ponica and Florae Japonicae Familiae Natu- rales, in both of which he listed genera and species that were either similar, or common in both east- rn North America and Japan. Gray apparently put aside this topic for about ten years, since the next mention he made of the relationship in print was in 1856 in the preface to his list of deter- minations of specimens collected in Japan by S. W. Williams and J. Morrow. Both of these men had been attached to Perry’s expedition to Japan and the China seas. In his introductory remarks, Gray again mentioned me floristic relationships dind ] that he ho them again once Charles Wright’ s collection of Jap- anese plants had been studied and named. Wright had been one of the botanists on the North Pa- cific Exploring Expedition under the command of Commodore Rogers and had returned home in 1856, during the time that Gray was involved in the preparation of his “Statistics of the Flora of the Northern United States." It was in this paper (1856, 1857) that Gray laid the framework of comparison that he was to utilize again in his now classic paper (1859) based on Wright's col- lections. It is of interest to note that Gray had previ- ously received and reviewed (Gray, 1854) J. D. Hooker's dissertation concerning the New Zea- land flora (Hooker, 1853), which had been pub- lished in 1853 as an “Introductory Essay" to his Flora Novae- Zelandiae. The format of that essay in comparing the distribution and affinities of New Zealand plants was basically the same that Gray followed in his “Statistics” and “Diagnos- tic Characters” papers. It is also of interest to note that in his introduction Gray mentions that he was “requested by an esteemed correspondent _to exhibit, in a compendious and convenient £e C out any doubt, Charles Darwin, with whom had begun to correspond in 1855,* and whom he rwin's letters to Gray are preserved in the Ar- chives of the Gray — of Harvard University, but most have been publ . Likewise, many of Gray's letters to Darwin, the originals of i are n in the Cambri Library, h published. Gray (1893); and F. Darwin (1887, 428 mentions elsewhere in the “Statistics” paper. From reading Darwin’s letters to Gray during the period 1855 to 1859, and from studying and comparing Hooker’s New Zealand essay with Gray’s subsequent papers, one is given the dis- tinct impression that Darwin’s questions and ee prodded Gray into action, and that arwin’s and Hooker’s ideas strongly influenced Gray’s idoneis concerning the geographic affin- ities of the North esi flora and the analysis of these relationshi In his response Q2 Mey 1855) to Darwin's first letter, Gray wrote that he **. .. had already in- tended, when the Flora of N. America should be finished to work up the geographical and climatic relations of this flora, and to compare it critically with the N. European and N. Asiatic floras." He continued by stating that he welcomed Darwin's questions, “and, if you will kindly give me hints as to what is needed, and how to do it, I will undertake the comparison of the plants of this moderate area (bounded by the Atlantic Coast, ew Brunswick, St. Laurence, Great Lakes, Mis- sissippi, and Potomac or Chesapeke Bay) with the General N. Amer. flora and with that of the northern part of the Old World, and print the result either in the volume itself? or in some Journal, as you suggest. So, you see, far from taking your suggestion amiss, I respond to it by asking you to tell me din egeo how to do it, so that it may be of u In his second letter to ies (8 June 1855) Dar- win asked specifically concerning the geograph- ical distribution of North American plants and their affinities and total ranges. Darwin wrote, “The ranges of plants, to the East and West, viz. whether most are found in Greenland and West- ern Europe, or in E. Asia appear to me a ve interesting point as tending to show whether the migration has been Eastward or Westward,” and he expressed pleasure that Gray was thinking of “drawing up sone notice on geographical dis- tribution . In his third letter to Gray (dated 2 May, prob- ably 1856?) Darwin wrote, “I would give a list of temperate plants, if any, found in Eastern Asia, China, and Japan, and not elsewhere. Nothing would give me a better idea of the Flora of U.S. than the proportion of the genera to all the gen- 5 The volume referred to by Gray was the new, ond edition of his Manual of Botany ipo is Northern United States on which he was then worki ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 era, which are confined to America, and the pro- portion of genera confined to America and East- ern Asia with Japan; the remaining genera would be common to America and Europe and the rest of the world; I presume it would be impossible to show any especial affinity in genera ... be- tween America and Western Europe; America might be related to Eastern Asia, (always ex- cluding Arctic forms) by a genus having the same species confined to these two regions; or it might be related by the genus having different species, the genus itself not being found elsewhere." After receiving “sheets” of the first section of Gray’s “Statistics” paper, Darwis wee d i pisi tober 1856) that than the greater generic and specific affinity be- tween the eastern North American flora and the eastern Asian flora, an affinity greater than that between the eastern id western Amerika s ras, and he asked i aini or was it one of the “many utterly inexplicable problems" of botanical geography. In an another, undated letter written in 1857, Darwin eec Gray to contribute further to geographical pro i lems despite the fact that Gray complained 0 his incomplete knowledge of plant distribution. Darwin urged Gray to prepare reports as the Wo ed progressed, and wrote, “So do pray look at pe^ side of the question, and let us have hk paper or two like the last admirable ones 2 While Gray did mention the relation tween eastern North America and eastem dh in his “Statistics” paper, these relationships not developed fully, undoubtedly ee fact that those comparisons he did m iul based on literature reports and only a re with few specimens. Thus, in a section dalt widely dispersed species that “reappear n js bot the Himalaya, or some part of northern en are not European," Gray listed only the fo lineart 13 species: Brasenia peltata Pursh, Sium £2.25 gistylus (Torrey) DC., Aralia qunm Decaisne & Planchon, Viburnum or" Michaux, Monotropa uniflora L., Phry f, Til tostachya L., Smilacina mee (L.) e yl lium erectum L. var., tosorus a 1859. (L.) Link, and pa ages however, Gray had had the opportunity wright and name the rich collections of Charles from Japan, and the list of shared species alone tween eastern North America and : included 134 species! Gray now had =— ~ 1983] mens that would allow for the floristic compar- ison with eastern Asia that Darwin had requested in his second letter. Gray's (1859) published analysis of Wright's Japanese collection took the form of an essay in which he compared the Japanese flora with the North American and other north temperate flo- ras. In it Gray described only the new taxa from among Wright's collections (these proposed in footnotes), indicating (p. 377) that a detailed ac- count of the entire collection was **. . . intended to form a part of Mr. Wright's general report upon the extensive and interesting botanical col- lections made by him during the whole cruise of the Expedition. As this report, and that of the important scientific results of this Expedition in various other departments, may not be published for some time, I am permitted and requested, by the Commander of the Expedition, —ever con- siderate of the interests of science, — briefly to make known the principal novelties which have been discovered in this field. The discoveries. . . also relate to the detection of known species in à region where they were not known before, and to the identification and elucidation of many ob- Scure species. It will be best, therefore, to take a cursory general notice of the more interesting plants of this collection, adding any remarks Which it may be worth while now to make upon their geographical distribution of their charac- ters, and appending the diagnoses of new genera and species in the form of foot-notes.” It is both in the text of the essay (where he elucidated previously described species and Milit his needing on the similar- ities of Ja orth and in the discussion of ear “Tabular View of the Distribution of Japanese Plants and Their carest Allies in the Northern Temperate Zone" (PP. 424-436) that the disjunct distributional pattern between eastern North America and east- s ta (Japan) became vividly apparent. Some inia ray’s comments and observations concern- IS pattern and the floristic similarities are Quoted here Es c: discussion of the Berberidaceae (pp. 380, Tay e, “But perhaps the most inter- esting and most ip tie discovery of the ex- PP uter that the scientific results of the expe- t um of Harvard University BOUFFORD & SPONGBERG— HISTORY 429 pedition is that of two strictly Eastern North American species of this order, —each the sole representative of their genus, —viz. Caulophyl- lum thalictroides, and Diphylleia cymosa, of Mi- ha . Supposing these two plants to be siti iih identified as to species,” are we to regard them as the descendants of a common stock, though now separated by one hundred and forty degrees of longitude? í Or are we to suppose them i two such widely distant regions?” And i in the discussion (p. 416) of his newly proposed genus Heloniopsis, Gray wrote, “It may be briefly described as a Helonias with few flowers, a single and slender style sur- mounted by a depressed-capitate stigma, and the seeds appendaged only at the hilum. Two things t:— 1. Its con- formity to the rule, if it: may be so called, that peculiar Eastern North American types have their counterparts in Japan. For the original and only true Helonias—one of the rarest plants in the United States—is found only in a few localities in New Jersey, the adjacent parts of Pennsylvania and Delaware, and in Virginia." Numerous other examples could be quoted, some dealing with identical species, others concerning closely allied species, or, as in the last example above, closely allied genera. In his Bises View,’ ” Gray listed about 580 Japanese rel in other asa distant parts of the northern tem- perate zone” (p. 422), and listed in parallel col- umns the corresponding taxa represented in the floras of Europe; Central and Northern Asia; Western North America; and Eastern North America. Identical species shared between the flora of Japan and another area(s) were set in italic type. In analyzing this data, the following summary was presented (p. 437 ati VES The whole number of Japanese entries is about 5 80 The whole number of Asiatic entries is about 44; in italics 274 The whole number of European entries is about 282; in italics 214 The whole number of W. American entries is about 216; in italics 162 The whole number of E. American entries is about 356; in italics 223 7 The Japanese and North American species of both Cadapan and Diphylleia were subsequently shown to be disti 430 Based on this analysis Gray was able to con- clude (p. 437) that “It is intersting to notice that, urope. Also, —showing that this difference is not owing to the separation by an ocean, —that far more Japanese plant I ted i st- ern North America than in either. . . . If we regard the identical species only, in the several floras, the preponderance is equally against Western as compared with Eastern North Amer- ica, but is more in favor of Europe. For the num- ber of species in the Japanese column which like- 157." But, considering the overall similarity at all levels, Gray concluded that of the 580 Jap- anese entries, 48 percent have corresponding Eu- ropean representatives and 27 percent are iden- tical species; the comparative figures for western North America were calculated as 37 percent for corresponding taxa and 20 percent for identical species, while for eastern North America, the percentage of corresponding taxa was an over- whelming 61, and that for identical species was 23; In summarizing the floristic data that he had presented concerning the New Zealand flora and its affinities with the floras of Australia and Tas- mania and temperate South America, Hooker (1853, p. xxxvi) wrote that “Enough is here given to show that many of the peculiarities of each of the three great areas of land in the southern lat- itudes are representative ones, effecting a botan- ical relationship as strong as that which prevails throughout the lands within the Arctic and Northern Temperate zones, and which is not to be accounted for by any theory of transport of variation, but which is agreeable to the hypoth- esis of all being members of a once more exten- sive flora, which has been broken up by geolog- ical and climatic causes.” It was with a quotation of this passage that Gray concluded his review of Hooker’s essay in 1854, and in attempting to awnila- +h Ascot + Cia . 13 . and the otherwise strong floristic affinities be: tween the Japanese and eastern North American floras, Gray built a similar argument to explain the observed facts. Gray wrote (1859, p. 440), *. . . it will be al- most impossible to avoid the conclusion, that there has been a peculiar intermingling of the Eastern American and Eastern Asian floras, which ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 demands explanation." On page 444 he wrote, “The discovery of numerous closely related species thus divided between two widely sepa- rated districts might not, in the present state of our knowledge, suggest former continuity, mi- gration, or interchange; but that of identical species peculiar to the two inevitably would." Gray suggested further (p. 442) that the inter- change between “the temperate floras of the western part of the Old World and the New had taken place via Asia," an idea that he attributed to Bentham. Gray (1859) credited Bentham with pointing out that interchange between all of the Old World, including western Europe, and the New World had taken place via Asia. Gray thought the statement may have been made in conversation or in correspondence between the two for Gray did not believe these thoughts had been published. Gray cited as evidence to sup- port this theory the fact that such large No American genera as Aster L., Solidago L., and Eupatorium L. (Compositae), Euphorbia L. e phorbiaceae) and Solanum L. (Solanaceae) ".-- are represented in eastern Asia by a small num- ber of species, which gradually diminish or al together disappear as we proceed westward to- ward the Atlantic limits of Europe, whilst the types peculiar to the extreme west of Europe - cluding of course the Arctic flora) are wholly de ficient in America." Bentham (1858, P. 35, mn his “Synopsis of the genus Clitoria, € published his view that there had at one tim been an "ancient continuity of territory jue America and Asia, or at any rate with à clima i more meridional than would be effected ar : junction through the chains of the Aleutian Kurile Islands." old It should be recalled that Gray was writing prior to the publication of Darwin’s de igin of Species, although Gray was aware o cep : : CONE : epee oreovel: of their views (p. 443, in footnote). e p Darwin and Gray had carried on a cons od prior frequent correspondence during the pe ^. to the publications of Gray's “Diagnostic acters” (see Good, 1955). Darwin had of his ideas on Gray, anxious for bot innean c - oO on his knowledge of the American flora. Hook; ering the influence of both Darwin "m ncluded it is perhaps not surprising that Gray CO ns with his analysis of Wright's Japanese ws an attempt to explain the revealed dis) ee ~ —— en P ——J—————————————— 1983] utilizing paleobotanical evidence and geological and glacial hi d) coupled with an “evolutionary” outlook that em- braced “common descent.” Thus, the concept of a disrupted, once more widely distributed North Temperate flora was proposed as a working hy- pothesis to explain the observed facts. Gray had not only succeeded in fully pointing out the flo- ristic similarities between eastern North Amer- ica and eastern Asia, but he had (with due credit to Darwin, Hooker, and his geologist friend and colleague, Professor J. D. Dana) also proposed the working hypothesis of the causes—a hypoth- esis that remains valid in many ways today. Gray’s hypothesis gave support to Darwin’s theory of evolution but refuted the theories of the Harvard naturalist, Louis Agassiz (Gray, 1860). Dupree (1959) has devoted several pages to the conflict between Gray and Agassiz and to the debate that both Gray’s “Diagnostic Char- acters” and Darwin’s Origin excited. That Gray’s paper and Gray’s influence surrounding the ac- ceptance of Darwin’s theories in North America were important undoubtedly brought Gray’s Diagnostic Characters” paper to the attention ofa wider scientific public than might otherwise hes - expected. It is understandable, there- Nee. at Gray has generally been considered the : tanist to recognize the eastern Asian-east- m North American pattern of disjunction. bae Coe of the major share of the credit baa Sa JTày 1n connection with the floristic re- lonships between eastern North Ame f. 4L d y (as they were then d ) bens concerning this topic that he delivered Ness e appearance of his “Diagnostic Char- Pass = in 1859. Thus, in 1873, in his ad- "€ " the American Association for the Ad- Pine of Science (Gray, 1873a), Gray Se is floristic comparisons to include a Nea em of Asia (N ortheastern Asia, Japan the oe and the Himalayas) and to compare Pesci cca of this area with western North Gavai (Oregon and California). In this paper the is en a list of species common to iu ed States and Europe, a list which was reprinted in England (Gray, 1873b). And in wu n a lecture presented to the Harvard Uni- ty Natural History Society entitled “Forest ie and Archeology,” Gray outlined the the. tan ies and dissimilarities of the forests of Pérate zone and compared those of the BOUFFORD & SPONGBERG—HISTORY 431 *Japan-Manchurian" region with those of Eu- rope and eastern and western North America. In an effort to explain the similarities and dissim- ilarities he again emphasized that current distri- butions were the direct result of past history. EUROPEAN CONTEMPORARIES OF GRAY F. A. W. MIQUEL Soon after the publication of Gray’s “Diag- nostic Characters," several European botanists turned their attention to the problem of the dis- junct distributional pattern described by Gray, the affinities of the floras involved, and the hy- pothesis offered by Gray to explain biogeograph- ical data. F. A. W. Miquel (18672, b), a Dutch botanist, published two papers in 1867; one in which he discussed the affinities of the Japanese 1867b). While Miquel accepted the distribution- al pattern described by Gray and did not question the geological evidence of a once more wide- spread temperate flora in the Northern Hemi- sphere, he was unwilling to accept the Darwinian concept of common descent to explain the pres- ence of *analogous" species in Japan and eastern North America. In his paper on the relationships of the Japanese flora with that of North America, Miquel tabulated 27 woody species common to Japan and all of North America, and 76 herba- ceous species, not including analogous or similar species pairs. A year later (1868) he considered 81 species to be shared between eastern Asia and all of North America. While he agreed to the single (vs. the multiple) origin of a species, he maintained that the analogous species in each area were “created” independently and were not descendants from a common stock. ADOLPH ENGLER Aware of Darwin's, de Candolle's, Gray's, Hooker's, and Bentham's ideas and the work they had done in this area, Adoph Engler (1879) used evidence from a number of fields in his attempt to interpret the history of the world's vegetation since the Tertiary and to explain present-day dis- tribution patterns. Engler was convinced that the distribution of extant plants was the result of disruptions in ranges, changes in climate, migra- tion, and evolution, and he placed the greatest importance on past geological and climatic events as the causes. He believed that numerous extant 432 species had histories that extended back to the Tertiary, and he was aware that during the Ter- tiary the vegetation of the eastern part of North America extended westward at least as far as the Rockies and had occurred there into the Neo- gene. Engler cited fossil evidence based on the work of Lesquereux to support this hypothesis. He noted that many eastern North American taxa that were not now extant in western North America were represented there by fossils, but that only three western genera. Prosopis L., Cas- sia L., and Ziziphus Miller, were known as fossils in the East. Parallel forms in eastern and western North America were considered to be remnants of a once more widespread floristic element. En- gler postulated that the Rockies had been lower in the past, the prairies less extensive, and that a mostly continuous, or at least similar, flora extended from the East to the West Coast, and it Pee Bre AA UCET, ming more widespread flora still remained in such taxa as Ptelea angustifolia Bentham, Aesculus cali- fornica Nuttall, Acer negundo L. (as Negundo californicum Torrey & Gray), Staphylea bolan- deri A. Gray, Rhus diversiloba Torrey & Gray Cercis occidentalis Torrey ex Gray, Amelanchier alnifolia Nuttall, Calycanthus occidentalis Hooker & Arnott, Rhododendron californicum Hooker, and Styrax californicum Torrey (Engler, p. 12) in the West. All of these plants have dis- junct counterparts in eastern North America. to the Himalayas. e gave examples of plants (Liquidambar E: Platanus L., Ostrya Scopoli, Castanea Miller) ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 known as fossils in Europe, and believed that these genera had been richer in numbers of species, and were much more widespread in the past, as were many other genera now restricted to the warmer parts of North America and east- ern Asia. Engler believed that the plants now restricted to eastern Asia and eastern North America had a much wider range, were even cir- cumboreal, during the Tertiary because ofa pe rc B 25 4 4L ant portant aspect f Engler's work was that he pointed out that many of the eastem Asian-eastern North American disjuncts consid- ered by Gray to be “identical” were already being shown to be distinct. He noted that Miquel (1868) had reduced Gray’s (1859) list of 134 “identical species to 81 taxa from Japan and North America that Miquel considered to be the same. Grise- bach (1872, pp. 520-522, 602, 603), as pointed out by Engler, further reduced Miquel’s list t0 only two examples, “Elodea petiolata” and Car- ex rostrata Stokes, of plants restricted to Japan and the eastern United States (21 plants were automatically excluded from consideration by Grisebach (1872) because they also occur" x Canada, and 41 were excluded on the basis 0 their occurrence in western North America). x gler (1879, pp. 25, 26), however, provided a of 140 plants he considered to be shared between jority were forest plants, and only very fi found in open habitats. There are som distribution data in the list, for example, d occidentalis Torrey & Gray and Mahoni a folium Nuttall are both credited to Atlante ed America while Melothria pendula L. 1s pee to Pacific North America. In general, ho Gray's the list is more complete and refined than ont (1859) and indicates that many plants that» had considered to be identical are actually inct d's vege Je use of m In explaining the history of the —- tation Engler (1879) made ne in the fossil evidence, especially of fossil flo d surface arctic regions. He believed that the lan sable cli- great, perhaps resulting in a more uen mate that accounted for both a widesprc? tuations mesophytic forest in present day arctic a species and the opportunity for an exchange ier North across a bridge in northeast Asia-north | i L í [| Í | H 1 i | | —— — eat 1983] ward as the climate cooled. These plants mostly became extinct in western North America be- cause of the markedly different climate there, but still occurred in Japan and eastern North Amer- ica where the climates remained similar. Those plants that survived in the West underwent evo- lutionary change in response to different edaphic factors, thereby resulting in different species of “Negundo” Boehmer, Rhus L., Cercis L., Os- morhiza Rafinesque, and Platanus L. in the east- em and western United States. The less drastic climatic changes in Japan accounted for the rel- atively higher number of genera, particularly for the higher number of monotypic genera, surviv- ing there. Engler was apparently the first to refer toextant floras wh t abundantly represented by fossils from the Tertiary at high Andes. This reference to the Andes suggests that Engler recognized an eastern North American temperate element in the flora of Mexico and Central America. This is the first indication of an awareness of this additional pat- tern of disjunction, an awareness that has gen- erally been credited to Watson (1891; see Gra- ham, 1973), The extension of the recognition of floristic affinities to a wider area of the Asian continent — hand in hand with the development of flo- "istic knowledge of China and other areas of mainland Asia, Thus, by 1873 H. F. Hance, who had been occupied with Chinese plants for many years, was able to list 21 species of the Chinese us occurred or were represented by closely Fur Species in the Caucasus region of southern » a and western Asia, a region now known T Upport a flora rich in examples of Arcto- x elements. The first installments of Forbes from n sley's “Enumeration of All the Plants addi 18a... a ared in 1886 and 1887, eitis the Systematic treatments the Chinese and "lerritorial distribution was given for each a While this series was not completed until ne the wide circulation of the parts as they tivi published stimulated further collecting ac- Y in remote areas of the Chinese interior BOUFFORD & SPONGBERG— HISTORY 433 (Thiselton-Dyer, 1905). The valuable collections made by Augustine Henry in central China were acquired by and distributed from the herbarium at Kew as a direct result of this stimulation. The continued discovery of botanical novelties in in- terior China, many of considerable botanical and horticultural interest, also increased interest in the botanical exploration of the Chinese empire. It was essentially in response to a strong desire to introduce these newly discovered ornamental species into cultivation in England that the Veitch nursery firm sent E. H. Wilson to China on two expeditions. Wilson not only carried on the work begun by Henry (and the earlier French mis- sionaries), but brought the Chinese flora to the attention of late nineteenth and early twentieth century botanists. DIELS AND THE FLORA OF CENTRAL CHINA As mentioned above, the gaps in the flora of central China that had been noted by Engler (1879) began to be filled within a few years as the collections of French missionaries and other European botanists began to reach Europe. Enough new material had been amassed and new info ti d ilable since Engler’s (1879) paper to enable Diels (1900-1901, 1905) to write a definitive work on “Di ora von Central- China” in which he discussed the history of bo- tanical exploration, the geography, ecology, vegetation, and the relationshps of the flora, in addition to providing a list of the known taxa. The area treated by Diels took in nearly all of the present-day province of Sichuan as well as the adjoining parts of the surrounding provinces. Coincidentally, this happens to be a region con- taining a large number of taxa with North Amer- ican relatives. Diels’s comparison of the flora of central China with those of other parts of the world showed that a greater number of taxa were shared with Japan to the east than with the much closer Himalayan region to the west, although the vegetation in central China and Japan was somewhat different (Diels, 1900-1901). Diels showed that the Asian distribution of the ele- ments in the flora exhibiting disjunction between eastern Asia and North America could be cate- gorized into four types: 1) genera with members reaching to Malesia; 2) genera reaching from the Himalayas to Japan; 3) genera reaching only to Japan; 4) genera restricted to central China and North America. Examples of each are: 1) Ari- saema Martius, Aletris L., Magnolia L., Schi- 434 sandra Michaux, Illicium L., Polygonum L. sec- tions Cephalophilon Meisner in Wallich and Echinocaulon Meisner in Wallich, Vitis L., Par- thenocissus Planchon, Gordonia Ellis, Nyssa L., Clethra L., and Saururaceae; 2) Phryma leptos- tachya L., Panax L., Adiantum pedatum L., and several others; 3) Caulophyllum Michaux, Ha- mamelis L., Stewartia L., Diphylleia Michaux, plus others; 4) Cypripedium arietinum R. Brown, Calycanthaceae, Liriodendron tulipifera L., De- cumaria L., and Gymnocladus Lamarck. Diels clearly based his classification of eastern Asian- regard to the distribution of the North American counterparts. Mention must also be made of the Forest Flora of Japan, a volume Charles Sargent (1894) wrote after his first-hand observation of the Japanese flora during his journey there in 1892. Continu- ing in the tradition of Asa Gray, his mentor and friend, Sargent accounted for the ligneous flora of Japan and compared the genera with those that occur in eastern North America, essentially updating Gray’s “Forest Geography and Ar- cheology.” Charles Sargent should also be given credit as one of the first Americans who realized that the climatic and floristic similarities be- tween eastern Asia and eastern North America would allow for the successful cultivation of nu- merous Asian species in eastern North America. As first director of the Arnold Arboretum of Har- vard University, Sargent had been successful with The history of the botanical exploration of China (for which see, in particular, Bretschnei- der, 1898) is not within the scope of this paper, but Wilson’s activities both for the Arnold Ar- boretum and earlier for the Veitch Nursery firm in England during the period 1899 to 1910 great- ly increased knowledge of the Chinese flora and extended the knowledge of the eastern Asian- eastern North American distribution pattern. Thus, Sargent (1913) was able in an Introduction to Wilson’s A Naturalist in Western China, to ANNALS OF THE MISSOURI BOTANICAL GARDEN compare the ligneous flora of North America with that of China, again pointing out the similarities but also noting the differences between the floras and emphasizing the richness of the ligneous flora of China. Further comparisons, particularly of the ligneous flora, were subsequently made by H. H. Hu (1935, 1936) and were the first con- ies of taxa exhibiting this distribution pena and in 2 ° £ +know edge. Most of these contributions are well known (1952, 1956), Li (1952, reprinted 1971 with ^ ditions), Wood (1971, 1972), Raven (1972), an Thorne (1972), as well as papers published * this symposium volume, should be consulted ences not included here. Moreover, edited by Graham (19722) contains a coUs’ of pertinent articles based on symposia ss conjunction with the XI International Wen Congress in 1969 and a special Japan" : to Gray arwin wrote nei work." is impor stood and assessed. D alee 1857 that “Undoubtedly careful discrim of species is the foundation of all 8 and Wood (1972, p. 122) concluded ge tant paper by. stating. that "Maru unctions. and, consequently, to the study of dis) It is perhaps appropriate to conclude pen: historical survey by reiterating these of the ments—our knowledge and understand! oristic eastern North American-eastern Asian and be connection will in large measure pie determined by our knowledge of the sys of the taxa involved. ET | [Vor. 70 this brief | | | j 1983] LITERATURE CITED ANONYMOUS. 1860. Biographical sketch of the late Thomas Nuttall. Gard. Monthly & Hort. Adver- tiser 2: 21-23. BEIDLEMAN, R. G. 1960. Some biographical nodu on Thomas Nuttall, 1786-1859. Proc. Amer. P los. Soc. 104: 86-100. 58. nos of the genus Clitoria. n. Soc. Bot. 2: 33-4 BRETSCHNEIDER, E. 1898. tees of European Bo- tanical Discoveries in — mpson Low, Mar- ston & Co. Ltd., Lo CASTIGLIONI, L 1790. negli Stati Uniti dell’ Ameri ik Setxaditeiphale: vita negli anni 1785, 1786 e 1787 con alcune osservazioni sui vegetabili piu tili ConsTANcE, L. 1972. Patterns in the distribution of Japan -American lLImbhelliferae Pp. 93-99 in Graham (editor), Floristics and Paleofloristics of Ms me: one tern North America. Elsevier Publ. erdam 1859, On the Origin of Species by Means of Natural Selection, or the Preservation of Favoured Races in the Struggle for Life. John Murray, London. =e A R W. cE. 1858. On the nee: species to form varie ties; and on the perpetuati of varieties and “sig s by natural EU ote se- lection. J. Proc. Linn. Soc. ., Zool. 3: 45-62. Darwin, = (editor). ge The Life and Letters o Charles Darwin. 2 Volumes. D. Appleton & Co., New P rk. ~ (editor). 1903. More Letters of Charles D win: A Record of His Work in a Serie Me iore Unpublished Letters. 2 Volumes. D. Appleton & rk 0., New York. Ditis, : 1900-1901. Die Flora von Central-China. t. Jahrb. Syst. 29: 167—659. “Sree: — zur Flora des Tain ling shan und andere Zusi: ral-China. -1 38. Gray. Harvard University Bs " bridge. RAND, E. 1860. Biographical notice of the late Thomas Nuttall. Proc. Amer. Philos. Soc. 7: 297- Since A. 1879. Versuch einer Entwicklungsge- Schichte der Pflanzenwelt, inbeso rengebiete seit der Tertidrperiode, 1. Die extra Pischen Gebiete der nordlichen Hemispháre. ican Bo : : editor), Classica Botan- pja Americana Vol. 7 7. Reprint Edition of Nuttall’s “Genera of North American Plants.” Hafner Publ. Finn New York. ON, C.E. 1891. Castiglioni's travels in the United tates. Gard. Forest 4: 11 0, 111. 1931. Specific segregations and iden- xe in some floras oraa eastern North America and Old w ra 33: 25-63. (Reprinted as . HEMSLEY. 1886-1905. An enumeration of all the plants known from China Proper, Formosa, Hainan, Corea, the Luchu Ar- BOUFFORD & SPONGBERG—HISTORY 435 chipelago, ; and the Island of pines tied aging with thei Bot. 23: 1-521; 26: 1-592; 36: n Goop, R. 1955. Plant geography. Pp. 747—765 in E. L. Kessel ipie A Century of Progress in the tural Sci 1853-1953. California Acad. Sci., a San uetus 1974. The Geography of the Flowering Plants. d. 4. Longman, London. 1966. Plantae Rariores Camschatcen- ses: a translation of the dissertation of Jonas P. Halenius, 1750. Brittonia 18: 131-139. (editor) 1972a. Floristics and Paleofloristics of Asia and — North America. Elsevier Publ. Co., Amste 1972b. Dutline of the origin and historical recognition of floristic affinities ven Asia and eastern North America. Pp. 1- A. Graham (duh Santor and ue of Asia and Am Easte rth America. Elsevier Publ. Co., Am- enne 1973. History of the seus temperate element in the northern can Biota. Pp. 310-314 in A. Graham (editor), ‘Vegetation in Vegetational T of Northern Latin Am Elsevier, Amster GRAUSTEIN, i E. 1967. | es Nuttall, Naturalist. Explorations in America, 1808-1841. Harvard University Press, Cambridge 1840. Dr. Siebold, Flora Japonica; sectio view). Amer. J. Sci. 39: 175-176. (Reprinted in Graham, 1972b, and in Stuckey, 1978.) 1846. Analogy between the AA of Japan and that of the United States. Am . Sci. Arts II. 2: 135, 136. (Reprinted in Gaan. 1972, and in TPA 1978. 854. ienesa tris in Dr. Hooker’s ra of New Zealand: V mer. J. Sci. Arts IE 17: v d 334-350, Chis constitutes a re- view by Gray of Hooker’s 1856. List of dried Manu collected i in Japan, by S. Wells Williams, Esq., and Dr. James Mor- vy, by Le a the Gov- ernment of the United States, Peste: 1857. € of the id of the rn United States. Amer. J. Sci. Arts II 22: rei 1856; 23: 62-84, 369-403. 1857. (Re- panee “4 jacet 1978.) ostic characters of new species p Char Exploring ning irt (Published by reques Captain Jam odgers, mmander of the Ex- pedition. ) With th the Japanese flora to that of North drei and Northern oem mpera ne ; : —452. Reprinted in part in c. s Sargent (editor), Scientific oA Gray 2: 122-141. 1889, and in Soit. 978.) 436 that of Central and Northern Asia, Europe, a i s 4: 135. (This article constitutes the minutes of the ies ON od be Academy, 11 January 1859, which Gra sented the second half of his "Diagnostic Characters" paper Baird 1859). 7 The and Agassiz.) : 1873. Address of vidit of the As- sociation. Proc. Amer. Assoc. Adv. Sargent (editor), Scientific papers of Asa Gray 2: 142-173. 1889, and in Stuckey, 1978.) 1873b. Plants of the United States and Eu- rope. J. Bot, 11: 173. . 1878. Forest geography and archeology. Amer. J. Sci. 16: 85-94, 183-196. (Reprinted in C. S. Sargent (editor), Scientific papers of Asa Gray 204—233. 1889, and in Stuckey, 1978. This article was also appended in reprinted form to Gray and . D. Hoo vd s “The en 0s of the Rocky Mountain n mparison with that of other "e of the Wound. = " Bull. U.S. Geol. Geogr. urv. Territories 6: 1-77. 1880.) Gur). LE 1893. Letters of Asa Gray. 2 Volumes. Houghton, Mifflin & Co., The Riverside Press, mbridge. GRISEBACH, A. 1872. Die oo der Erde. Wil- helm Engelmann, Leip Gronovius, J. F. 1739, 1743. Flora Virginica. Lei- den. HALENIUS, + 1750. Plantae Rariores Camschatcen- ses. Thesis, Univ. Uppsala, Uppsala. HANCE, i F. 1873. Ona Chinese maple. J. Bot. 11: 168-172. a, Hara, H. 1952, 1956. Contributions to the study of variations in the Japanese plants closely related to those of Europe or North America. Part 1. J. Fac. Sci. Univ. Tokyo III. 6: 29-96. Part 2. J. Fac. Sci. Univ. Tokyo III. 6: 343-391. 1972. Corresponding taxa in North America, Japan and the Himalayas. Pp. 61-72 in D. H Valentine (editor), prono, Pia and Evolution. Academic Press, London RVEY, Dina OF Paaa SIONAL & SERVICE PROJECTS, WORK PROJECTS pen MINISTRATION. 1939, Calendar of the Letters of arles Robert Darwin to Asa ray. ith an eto di by : St en Historical Rec- un The — of the Antarctic Voyage of H.M. Disc ry Ships Erebus and Ter- ror, in the Years 1 839-1 845, Underthe rry) of Captain Sir James Clark Ross, Kt., R.N., & L.S., etc. II. Flora Novae-Zelandiae pi L Flowering Plants. T Reeve, London. (Intro- ductory Essay, pp. i-xxxix.) Hu, H. H. 1935. papier of the ligneous flora of China and me North America. Bull. Chin nese Bot. Soc. 1: 79-97, ———. 1936. T "geras and affinities of Chinese flora. Bull. Chi webs Soc. 2: 67 Jones, G. N. 1937. uttall . . . botanist, or- nithologist, geologist. Linde ‘Gard. 8: 6, 25. ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 1789. Genera Plantarum. Pari n Travels into North America. 2 Vol- mes rrington. Printed by W. Eyres. ae "s glish Gite E reprinted with a JUssiEU, A. L. DE. KALM, P. troduction by A. B. Benson by Wilson-Eri pete Inc., New York, 1937, and by Dover Pübliotiedi Inc., New York, 1966.) LARSEN, E 1939. Peter Kalm's short account of article into English.) LLH. LE: 1952. RN eee mes between eastern Asia and eastern — rica. Trans. Amer Philos. Soc. 42: 371 9- (Repaid. = For- ward and ad one pos asa 1971.) boretum fe sae . 11955 d aid asa amens in -— geography o dee introduction. Proc. Amer. Philos. Soc. 99: 51-56. 1955. MiQUzL, F. A. W. 1867a. Sur les affinités de la si du Japon avec celles de l'Asie et de l'Améri Nord. Adansonia 8: 132-153. (Also ublished i in Arch. Néerl. y dti 1867.) oe tbey e caractère et l'origine de la flore ed n. Adansonia 8: 204-263. (Also published SAR N Néerl. 2: 289-384. 1867. 18 Mi Verwantschap der Flora von Japan E Azie en Nord America. Verslagen Wein: Afd. Plants and a Catalogu 1817. Printed for the cater by D. Heartt des 1936. Travels and emet collec- Plants of North America. 2 Volumes. White, Cochrane, London Raven, P.H. 1972. Plant species disjunctions: a SUM- mary. Ann . Gard. 59: 234-246. SARGENT, is S. orest Flora of Japan & Com , Bost : ——. 1913. Todido In E. H. Wilson, — — in Western China. Vol. 1. Doubleday, LIT ore Nuttall iat C. E.& W. THIERET. 1959. Thomas by. (1786-1859) an evaluation and bibliograp Leafl. West. Bot. 9: 3 Ge- SPRENGEL, K. 1802. Anleitung zur Kenntnis der wachse. 3 Volumes. Halle. STAFLEU, F. A. & R. S. COWAN. 1976. T Literature, ed. 2. Vol. 1, A-G. Regnum 978. Essays on North American y a "dace from the Nineteenth Century. Arn Dd Dim W.T. 1905. Historical note. /- Soc., Bot. 36: v-xi. em — 1983] THORNE, R. F. 1972. Major ener in the geo- Rev graphical ranges of seed plants. Quart. . Biol. 47: 365-411. es Cc me Flora Japonica. I. G. Mul- Leip: Saag S: 1891. Ale relation of the Mexican — to that of the United States. Proc. Amer. Ass Adv. Sci. 39: 291-292 . Some floristic relationships een the Southern Appalachians and w North America. Pp. 331-404 in P. C. Holt (edi- tor), The Distributional History of the Biota of the Southern Appalachians. Part II. Flora. Virginia Polytech. Inst. State Univ. Res. Mono: ogr. 2. . 1972. Morphology and phyto ogeography: the classical approach to the study of Files . Missouri Bot. Gard. 59: 107- APPENDIX 1 Taxa indicated by Nuttall (1818) in his Genera of North American Plants as exhibiting disjunc- lion between North America and Asia (occa- sionally also occ urring disjunctly in other parts ofthe world). Comments by Pursh (1814) on the same genera are in brackets; “No” indicates that ` h made no remarks. Our notes are in paren- eses. PA GE VOLUME 1. may perhaps prove distinct. There is another roe a ed Veronica Sibirica, inhabiting Dau- E ngth of the — " [No] (Now placed in 10-1 rem Fabriciu Elytraria. ‘ “Species à E. caroliniensis. There 80-81 are also 2 other species in India.” [No] “ere “A genus of India and America on n the tropics, consisting of 2 species." [No] um is credited to North America by 90. lechen “An American genus, with the exception ES L. verticillata of India.” [No] (Five species 2e ted to N erica by Nuttall. : Cep phalanthus £ iar to Ni are 4 in p place of 5.” [No] (One speci pecies of Ceph- ‘a alanthus, C. 5, accidentalis, recognized by N eq “Thi cmd divided between India and the tropical "en of America, The genus Hedyotis to which bu ott is inclined to refer the O. glomerata rën equally divided betwixt India and South erica.” [No o] a. (The fruit of this Ee bitter aey jesponi is eaten of the Mi issouri, from whence it and appears Ss “gs ee BOUFFORD & SPONGBERG—HISTORY 437 again in Siberia. )... Of this genus there are two Europe, and 2 which are common ica." [No 127. Pole ium i S appears as yet to con- tain but 2 genuine species the other is common and Asia." [No] to Europe 152. Claytonia. “A No rth American genus, with the exception of C. sibirica; C. lanceolata a of Pursh extends into Siberia, an 173. - rolea. “A genus of six species, (Sec. Persoon) indigen ous to the tropical or warmer regions of , with the exception of H. Zeylanica of India." 174. Heuchera. “A North American genus, with the exception of H. confinis discovered also in Kamschatka (sic) by the late professor cadi [“In the herbarium of A. B. Lambert, E specimens of a Heuchera, collected by Pallas in Kamtschatka, which appear to be the same with the present species." Pursh (p. 188), speaking of Heuchera Deco Pursh.] 76. Panax. ' quinquefolium. Gin-seng. Indige- nous also to Lei (sic)." (Nuttall's concept = Panax was much broader than rth is s Oey a fag ~~ ae the West Indies, and a uds s, from New Holland, in the genus. ) [No] 201-202. Fine "Of this genus there is another species on the border of the Caspian sea." [No] (Now considered to be a Logs genus of the southeastern United Sta 203. Rhus. **5. Vernix. also ai aw E [No] American — E 205. Aralia. Speaking of the extra-Nort taxa Nuttall said “The remaining species of this genus are indigenous to the [o 9 merica; there are also 3 species in Japan an [No in Chin na. 212. Mahonia. “A third species of this genus is indi- cated by Mr. Pursh as enne! vi the kingdom of Nepaul (sic) in India; moun- tainous country." [Of B. psec aquifolium al h and B. nervosa Pursh, ys: “The preceding species is cue in every other re- spect, and together with another in the collection of A. B. Lambert, Esq., collected in Nepaul (sic) by Mr. Buchanon, forms a new division of the en otal a new genus, whenever the fruit is per- tement I have given of it ing the two American species recognized.) 223-224. Erythronium. “Of this genus there is but a single species out of America indigenous to Si- beria and the south of Europe.” [No] (Nuttall recognized three species in North America and suspected that a fourth, confused with E. amer- icana, remained undescribed.) 438 224. Uvularia. * A North American genus with the ex Thun these is now treated as a Tricyrtis, a species of Disporum, the third is a Fritillaria; it is not known which two Japanese vu- laria. Uvularia i is now considered to be endemic ) pus Convallaria. “1. C. ma jjalis. Common also o Europe; flowers fragrant. There are but 2 other abri of this genus as . is now constituted, in- digenous to Japan.” 226-221. Orontium. “Of mus genus there is a secon berg; this is now placed in the genus Rhodea (Lil- arat a genus with a peculiar aroid-like inflo- rescence. 221. lahat “Common to Europe and North Amer- ca, there is also a second species in China.” [No] 239-240. Trillium. *A North Am in Kamtschatka according to Pallas," [*T. cam- tschaticum. Pallas in herb. Lambert. In Canada near Montreal. . . . The specimens in the herbar- ium of A. B. Lambe species of Trillium are now known from Asia; no Asian and American species are con i 242. = cals “A North American genus with the ex- of AE. Hippocastanum of northern Asia." 274-215. Chimaphila. (two species recognized, C. umbellata and C. maculata) oo also by the North West Coast of America.—Menzi Probably both species of the genus will be found also in East Asia and Europe.” [No r a. “An 285. Tiarella. - "Species. 1. cordifolia. 2. Menziesii. 3. trifoliata. 4. biternata. A North American genus, of icu Nos. 1 and 3, are also indigenous to rn Asia." [No ] 285-286. Mitella. “A North American genus with the exception of M. nuda of Northern Asia." [No] (Five species recognized by Nuttall. 292. Penthorum. “According to Mr. Pursh there is a second species of the 2 nus in China, ec eg by Sir G. Staunton.” [*In the Herbari of A Penthorum nse, mplici tereti, foliis elongato- incasi Sue eae inaequaliter ser- S, S erminalibus, seminibus ovatis co 307. Spirea. “A genus Reina!» equally divided betwi ixt Siberia and North Ameri ” [No] (Nuttall rec- ognized 11 species in North A America, one of which ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 is now treated as Aruncus dioicus (Walter) (Fer- 309. Dalibarda. “2. Fragarioides. Flowers yellow. This species is also found in Siberia uals the cur- rently ait Memes a) [Dryas trifoliata. Pallas in Herb. La p. 350] 309. Geum. “... 8. rain 9. ee PH. Radical 1 HU y 23.5 3. 4322 | A Coum Anemonoides. Also indigenous to Kamtschatka as well as No. 8 American genus of which there are also 6 species in Europe, | in t 311-312. Calycanthus. “A North A 8 the exception of C. praecox of Japan. Nol VOLUME 2. 10. Actaea. “Of this small genus — is 1 speci indigenous to Europe and another 17-18. Illicium. “An American ge pai s ea ception-of I. anisatum of Japan and a 18. Magnolia. ‘ Ud aen divided between China and the Un! Sta ite es; there is also 1 species in tropical Ameri- ini 18-19. Liriodendron. “Of this goana there are 2 y species in China and 1 in the mountains boina.” (Amboina = island, I es 22-23. Trollius. “Of this genus there species, 1 ig and the other in Siberia.” nesia) are 2 other digenous to there is ters 0 of India rth Ameri ca “Pi entape- Tis being "doubt. talus" and “Reniformis, [No i 35. Dracocephalum. “Principally a sna v [No] (Five species listed as being in ica by Nu ) 39. —— “Of this genus there is pen cies in Cochinchina." [No o] (Three ro aed for Nort gnonia. 44. Bi f 60 or more cies, is with a few exceptions in I emer ; and Japan, exclusively indigenous to regions of America." pecies, 98. Oxytropis. “A genus ming pe eret principally indifctónét to o Siberia; T 7 ane tion of a few species in Europe an : State us of near 98-100. havens L. “This vast gent ot a a (Eleven species recognized for Nuttall.) 107-108. Lespedeza Michaux. “A genus Pf 1983] North America, there are, however, about 3 tta. ing np se ana, eub rest of this xclusively indig- enous to tropical America, the only exceptions are d species in India and anotheri in China.’ is " [No] Wor Id Be wate R. pseudo-acacia L., however, i is widely ralized in Asia and elsewher re.) 133. Dia Schreber. “An American genus with the except ion of V. anthelmintica of India, the 10 to the tropical regions of America." [No] (S even species recognized for North America by Nut- 159-1 62. Solidago L. *Solidago is Bence A North merican genus, with the exception of 5 or 6 species in Europe, and 2 near Canton in cnt The arborescent inim of St. Helena and New BOUFFORD & SPONGBERG—HISTORY 439 Zealand will probably be excluded from this ge- nus, if ever carefully examined." [No] (Nuttall listed 51 species for North Ame rica.) 187. Elephantopus L. “A genus of 6 species eee to tropical America, with the exception of 1 o India 199. Cypripedium L. “Of this singular genus there are other species in Siberia, 1 in Japan, and 1 in Europe. " [No] (Six species listed for North Amer- ica by Nuttall. 219. Liquidambar. “ Styraciflua. From New En- co and extending to s shores of the Pacific this genus there is a second species in the Le- vant." [^New En vtm to | Florida, and in all the western countries." Pursh, p. ] 238-239. Gleditechia (sic L. “Of this genus there ap- pears to be another species indigenous to India and China." [No] 244. Menisperm um L. “M. canadense is also found in Siberia." [No] (Nuttall recognized three species in the genus, all occurring in North America.) CRETACEOUS AND CENOZOIC HISTORY OF THE NORTHERN CONTINENTS WARREN HAMILTON! The dispersal and migration of terrestrial plants and animals are controlled to a large extent by the distribution of land, sea, and orographic fea- tures and by climate. All these factors in turn are controlled in substantial part by the motions of the great plates that comprise the earth’s outer shell, its lithosphere. Continents stand above sea level because of their thick, light crust, but they are parts of plates that also include the ocean floors. Seven very large plates, and numerous mid- and small-sized ones (the concept of co- I t plates breaks d t the small-scale end), each typically 80 km or so thick, all move relative to all others. The plates tend to be internally rigid, and they interact mostly at their edges, although large, mostly continental parts of some plates undergo severe internal deformation. Ve- locities of relative motion between adjacent plates range upward to about 13 cm/year—an unim- posing rate by human standards, but equivalent to enormous displacements when continued for tens of millions of years. Spherical geometry requires that any motion between two parts of a spherical surface can be expressed as the rotation of one part relative to the other, defined by an angular displacement around an Euler pole of relative rotation. All trajectories of relative motions must be along small circles to that pole. (This pole is not to be thought of as having any mechanical signifi- cance.) The development of plate-tectonic con- cepts in the late 1960s was made possible by the demonstration that the relative motions that could be quantified fit these geometric require- ments. The mechanisms causing plate motions are much debated, but the reality of the motions is proved. ! United States Geological Survey, ANN. MiSSOURI Bor. GArp. 70: 440—458. 1983. tallize as intrusive igneous rocks. The ridge stands high because its crust and mantle are hot, and hence are low in density. A steady state develops the plat tinuet e apart: the material at the ridge is hot when newly formed, but it cools and contracts as it migrates away, and the sea floor sinks to greater depths. The oceanic crust and upper mantle formed at spreading cen- ters retain petrologic characteristics that can be recognized after subsequent events have brought continental plate (as along the west side of South America), along the east side of the Mariana island arc). either case, a magmatic arc—volcanoes surface, intrusive complexes at depth tends form in that part of the overriding plate tha is about 100 km above the top of the subducting plate. Compositions of the magmatic rocks 2 quite different in continental and oceani Pa Sediments and other materials scraped o [e) h : snow- t f wngoin late accumulate, p of the downgoing p e a har- i ch about the ¢ tionary wedge can reveal mu succ bas acter and history of oceanic lithosp e m been subducted beneath the overriding P Collisions result where light crustal any accretionary-wedge material 1 pe "ee oft wo. Collisions are suture between the tw 1 ae rarmation of on r M ENS m m nvergence con or both of such plates. Converg ch as su continues after a collision, but inasmus or impossible, subduction stops at th which oceanic lithosphere had previo subducted, and a new subduction system ent #8 through on the outer side of the no mâ- enlarged by the aggregation of the colli Denver, Colorado 80225. Eanna a —————— —— ~ m "O 1983] terial. Actual collisions are commonly oblique rather than head-on, and they progress in one direction or another with time. Strike-slip faults bound plates that slide past one another. Any strike-slip fault transfers spreading or convergence at one end to spreading or convergence at the other end. Many actual strike-slip boundaries are complex in detail and have obliquely convergent and divergent parts and also zones of distributed motion. Plate motions are often illustrated by two-di- mensional cartoons, many of which incorporate the false concept that the subducting plate is rolled over a platen fixed in the mantle and is injected down a fixed slot. All or most hinges actually migrate into the subducting plates as the over- riding plates advance. Boundaries between ad- Jacent plates tend to end at triple junctions, where three plates meet; such areas can display great variety and complex evolutions. I have summarized the evidence for, and evo- lution of, these concepts elsewhere and have il- lustrated the development ofa number of actual convergent-plate features (Hamilton, 1979). late tectonics has operated throughout at least the last 2,500 million years—all of Proterozoic and Phanerozoic time. Continental rifting is re- corded by truncations of basement terranes and y the deposition on the truncated margins of continental-shelf-and-slope stratal wedges. Sub- HAMILTON — NORTHERN CONTINENTS 441 mined primarily by the sun, and they have not varied widely within the last 3,800 million years— the age of the oldest rocks known for liquid water has always been present, and continental glaciation has occurred intermittently during the last 2,000 million years at least. uring Archean time, 2,500 to 3,800 million years ago (Ma), continents were formed by magmatism more intense than that of later time, in ways that are much disputed, but that many of us regard as the result of rapid motions of small lithosphere plates. Even then, however, the earth must have lost heat primarily through spreading oceans and by subduction-related magmatism, because the crustal geothermal gra- PU js DOO SRM | pee EES Ca in Pees | eroded Archean rocks are much like those in young terranes. Before 3,800 Ma, whatever crust the earth possessed was so thin, and de- that no rocks then formed have been recognized. No beginning in time can be seen for plate motions. The earth’s lithosphere has always con- sisted of rifting, drifting, subducting, and collid- | ing plates. All continents are aggregates of frag- ments, d together at times past. Continents have been deformed complexly by strike-slip faulting and by oblique and orthogonal rifting and compression, and the process continues. Re- constructions of prior positions of land masses 1 of the same structural and magmatic indicators that characterize modern tectonic systems; by aggregations through collisions of distinct con- tinental masses and island arcs; and by paleo- magnetic, paleoclimatic, and paleontologic in- dicators. The details of the processes, however, ve changed with time. Thus, most of the earth’s oo had been cycled out of the mantle and € continents before Paleozoic time; Pha- nerozoic granitic rocks are both voluminous and ADEN only where their magmas have risen ough Precambrian crust, or through volumi- nous sediments derived from such crust. Meta- oe TOcks of blueschist facies, products of ios stallization at high pressures but relatively w temperatu 24]. Det ids ee systems are voluminous in Mesozoic and Cenozoic terranes, oe less common in older Paleozoic Mio rare in late Proterozoic terranes, and un- eii in older tracts; the earth probably is cool- nd losing heat at a progressively decreasing Tate. Surface temperatures, however, are deter- Tes considered increases. Although valiant attempts ave been made at reconstructions for all of Phanerozoic time (e.g., Smith et al., 1981, and Ziegler et al., 1979) and even for the late Pre- cambrian, such reconstructions ignore most pre- constrained primarily by approximate latitudes inferred for some of the bits, and are schematic at best. Even for Cretaceous time, major uncer- tainties exist regarding the geometry of eastern (Central America, Caribbean, Mediteranean, Al- pine Europe, North Africa, middle East, Hima- layas, Southeast Asia, Indonesia, Melanesia); of onsho d offshore East Asia; of the Scotia- West Antarctica region. Many of the ambiguities are more likely to be resolved by paleobiogeo- graphic studies than by geologic and geophysical , o =] The biogeographic consequences of plate mo- tions and interactions must be enormous. For example, a small landmass on a moving plate 442 can carry a terrestrial biota derived partly from a biota shared with some previously adjacent landmass, and derived partly by fortuitous im- migration, both components evolving more or less in i i the land ith ti through different climatic TR: » E us a TE aS. | zones. The configu- d, hence, the amount and distribution of land above sea level, can change radically with time. When two land- masses collide, their biotas mingle and compete. Pre-collision fossil biotas of the two lands are different, whereas post-collision biotas are shared. This paper briefly summarizes some tectonic factors that may be relevant to the biotic rela- tionships of the northern continents during Cre- taceous and Cenozoic time. Little mention is made of the biotic features themselves. NORTH AMERICAN AND EURASIAN PLATES North America north of Mexico, and Eurasia north of the Alpine-Himalayan mountain sys- tem, behaved as internally coherent continental plates during Cretaceous and Cenozoic time, al- though they had complex histories of pre-Cre- taceous sundering and aggregation, and although they were much deformed internally during the Cretaceous and Cenozoic. The two megaplates began to separate in middle Cretaceous time, and the Atlantic Ocean has been widening ever since. Relationships between Siberia and North Amer- ica throughout this time are poorly understood. The following presentation is concerned with some internal and marginal features of the two megaplates, and not with relationships between the megaplates. NORTH AMERICAN PLATE Tectonic accretion of western North Ameri- ca. Much of the western part of the North American Cordillera consists of material added ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 conveyor-belted many thousands of kilometers to their present North American sites. The ag- gregate width of the accreted materials generally increases northwestward along the continental margin, from 50 to 100 km (much of that off- shore) in Baja California and southern Califor- nia, through 500 km in Washington and Canada, to 1,000 km in Alaska. Throughout Cretaceous and Cenozoic time, the oceanic and continental- margin materials have had northward compo- nents of motion relative to interior North Amer- ica, whether the specific motion has recorded oblique subduction, oblique extension, or strike- slip North America. Conversely, the fragments for which distant origins are likely are largely per small islands before collision. I know of no biotic interchanges that can be ascribed to the accretion of any of the fragments, although juxtaposed fos- sil assemblages are grossly disjunct in many places. Western Interior seaway. The Western In terior of the United States and Canada—a region that is now at generally high surface altitude 1? such provinces as the High Plains, Colorado Pla- teau, and much of the U.S. Rocky mown a was inundated by a shallow sea that conne" the Gulf of Mexico and the Beaufort Sea o middle Cretaceous time (McGookey d al, 197^ transgressing south from the Arctic ie from the Gulf of Mexico joined in late Al (latest Early Cretaceous) time, and t remained more or less continuous pa early Campanian (middle Late Cretaceous) There was then widespread regression of the " in Canada, but the Western Interior ofthe pe States remained broadly inundated through -4dle of the rest of Late Cretaceous time. The - Cretaceous seaway was continuou Gulf of Mexico and the Canada * ge Arctic Ocean and must have much hinde a west migration of terrestrial organisms. : a marine flooding of the continental m relatively minor. : Central America-Gulf of Mexico Cari Sea. From the end of Triassic time Un pe in Pliocene time, North and South ATHE wi separated by water, although tectonic 27 —————— — 1983] canic lands may have connected them at times, and although continuity may have been main- tained via Europe and Africa during much of Mesozoic time. The shapes and positions of landmasses within Mexico, Central America, and the Caribbean have changed greatly with time and are still only partly understood (Pindell & Dewey, 1982). Although subduction complexes have joined the continents at times, submarine gaps largely blocked migration of land organisms during Cretaceous and Tertiary time. The great intercontinental biotic interchange began only about three million years ago, when the Isthmus of Panama rose above sea level (Keigwin, 1978). Internal deformation of North America. Alaska, western Canada, and the western United States were much deformed by lateral motions during Cretaceous and Cenozoic time. In addi- tion to marginal tectonic accretion, an array of compressional, extensional, rotational, and strike- slip deformation (e.g., Hamilton, 1978) caused Deformation in. Eastern Canada, Greenland, and Alaska, is discussed in subsequent sections. EURASIAN PLATE Tectonic accretion of southern and eastern Eurasia. 7 gation of large and small subcontinents and h varied Oceanic debris swept up between them, as interveni Eeee a (e.8., Burrett, 1974, and Hamilton, 1970). North- em sutures are in general older than southern ones. The land masses south of the Alps, Car- Pathians, Lesser Caucasus, and Himalayas have collided with Eurasia during Late Cretaceous and Cenozoic time (e.g., Burchfiel, 1980; Dewey et al., 1973; Sinha Roy, 1978). The accretion is still going On in both the west, where Africa and Ara- 1a are in the process of closing with the northern continent, and in the east, where Australia has pecs colliding with island arcs migrating away paan Asia (Hamilton, 1979). Continental masses Ve Con«x beneath one or both of them. Bits of continents ve rifted, rotated, and re-aggregated. Island arcs have migrated moss ds PO WEST & is between ". * m , ging conti pes of collid- 2 masses changed as they were crushed togeth- * ch HAMILTON—NORTHERN CONTINENTS 443 Each of the arriving continental plates must have carried with it immigrant plants and ani- mals and an assemblage of fossils — both different from those of the newly adjacent mainland. Thus, the middle Tertiary contact between Africa and Eurasia allowed horses to walk onto Africa and lephants onto Eurasia. Smaller island thave carried limited biotas, evolved in varying iso- lation. Biotas were derived partly from the con- tinent from which they rifted away; other biotas were derived partly from overwater immigrants. Thus, the middle Tertiary collision of Australia with island arcs that migrated from the margin of Asia caused much of the limited exchange of attenuated and variably evolved biotas between those two continents (Hamilton, 1979). Seaways between Europe and Asia. During much of Late Jurassic and Early Cretaceous time and part of the early Tertiary, central Eurasia was covered by broad shallow seas that extended from the Tethyan ocean on the south to what is now the Arctic Ocean on the north (Grossheim & Khain, 1967; Vereshchagin & Ronov, 1968). During Late Jurassic and most of Early Creta- ceous time, a seaway west of the Urals connected what are now the Caspian and Barents Sea re- gions. By late Early Cretaceous time (Albian, possibly also Aptian), the sea had receded from the northern part of this region, and dry land has connected the Russian Platform and the Urals ever since. East of the Urals, another broad, shal- low sea g d southward during early Late Cretaceous (Cenomanian) time and covered what are now the West Siberian Lowlands through most of remaining Cretaceous, Paleocene, and Eocene time. Shallow seas also covered much of the Caspian-Aral region to the south, but be- tween about 50° and 54° present latitude, north- ern and southern areas were separated by land, except that a more or less continuous seaway connected between northern and southern seas during most of Eocene time. Both eastern Russia and the West Siberian Lowlands were emergent in Oligocene and later time. During much of the Early Cretaceous and again during most of the Eocene, marine water thus separated northern Europe and Siberia within what is now the USSR, although the sea may have regressed enough intermittently to provide land corridors between east and west. The Eocene seaway connected flooded continental shelves, beyond which were deep oceans, to the north and south, so migration of land biotas would have been severely hindered. The Early Creta- 444 ceous seaway similarly extended to deep ocean in the south and probably did so in the north Internal deformation of Eurasia. South of about latitude 45° in Europe and 55° in Asia, Eurasia has been deformed by Cretaceous and 'e i tensi - and strike-slip motions, in addition to the accretion of land masses to the south and east margins of the con- tinent (e.g., Alvarez et al., 1974; Burchfiel, 1980; Cohen, 1980; Dewey et al., 1973; Dewey & Sen- gor, 1979; Hamilton, 1979; Molnar & Tappon- nier, 1975; Tapponnier & Molnar, 1979). Dis- tribution and character of mountains, plateaus, basins, and deserts has changed complexly, but the general continuity of land in the main part of the continent, independent of accretion and rifting effects and the seaway just noted, has not. OPENING OF THE ATLANTIC OCEAN At about the end of Triassic time, North and South America, Africa, and western Europe were parts of a single large continent, the product of the suturing together of various lesser continents during Paleozoic time. Eastern Europe and Asia were bounded on the south by the broad Tethyan ocean. India and various lesser continental pieces of alpine Europe and the Middle East were not yet in contact with the northern continent. The subsequent opening of the North Atlantic is rea- sonably well constrained by the fitting together of ocean-floor magnetic anomalies (which are products of sea-floor spreading dated by deep- ocean drilling) and of the margins of continental crustal masses (Fig. 1). Following and accom- panying tensional thinning ofthe continents along what became the Atlantic, North America plus Europe began t I laway from Africa plus South America, in Middle Jurassic time, about 165 m.y.a. Europe and North Amer- ica remained attached to one another; Iberia, at the southwest corner of the European plate, slid past Africa when the Atlantic opened beween the eastern United States and northwest Africa (Scla- ter et al., 1977). Simultaneously, the Gulf of Mexico, plus oceanic lithosphere since overrid- den by the eastward-expanding Caribbean plate opened between the southern United States and South America. These motions continued until the end of Early Cretaceous time, about 95 Ma The configuration of Central America and the Caribbean region during this time, and indeed during Late Cretaceous and early Tertiary time, ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 4 PE 14 +E lithosphere did not connect North and South America, but tectonic or volcanic lands may have joined them at times. Dry-land connections be- tween the Americas, however, might have been maintained, at least intermittently, during Ju- rassic and Early Cretaceous time via Africa, Iberia, and western Europe. The motion pattern changed greatly at about the beginning of Late Cretaceous time, and the Atlantic Ocean began to open; North and South America have been moving mostly westward since then, with little relative motion between them, away from Africa plus Europe. Sea-floor spreading was simple in the South Atlantic, and in the North Atlantic as far north as Newfound- land and Britain: the ocean widened steadily at the center, although at varying rates (Sclater et 977). PLATE BOUNDARIES BETWEEN THE ARCTIC D ATLANTIC OCEANS North of the latitude of Ireland, the spreading history of the Atlantic Ocean has been complex, patterns have changed with time, and transform faults have stepped spreading into the Arctic. Between the start of spreading in the middle Cre- taceous and about 60 Ma in the Pakt spreading was concentrated west of Greenland, in the Labrador Basin; from about 60 to 38 Ma, spreading occurred both east and west of Green- land, although that to the west was very slow after about 50 Ma; since 38 Ma, SP nland (Jackson has taken place only east of Greenla 1977; Scla- Perry et al., 1981). Spreading between pee and Greenland has jumped from rift to 1 ? :41 rift con- has not been confined to a single m al rift tinuous in time. ARCTIC ATLANTIC i Atlantic E The Arctic sector of the ee as Greenland and Norwegian Seas—0P Greenland moved away from Norway: along spreading system ends at transform faults, ental and southwest of the submerged Mm margin of Spitsbergen and the Barents int These faults step the spreading into the E ch the Basin of the Arctic Ocean, within W isi spreading axis trends to the Siberian co” margin near the delta of the Lena ane north Until about 38 Ma, Spitsbergen an a 1983] 165 MA HAMILTON—NORTHERN CONTINENTS 445 125 MA 95 MA 65 MA 36 MA PRESENT N : a ~~ u f see P Me ry FR, F, and R represent , IB, ia, Caribbean, Iberia, Africa, Flemish Cap plus Orphan Knoll, ntour around each continent is shown with a light line; heavy 1 1 4 = c tg P R eproduced niversity of Chicag “ast Greenland were sliding past one another on and à transform fault trending north-northwestward, Continental crust was continuous between them (Talwani & Eldholm, 1977; Vogt, Bernero Com » 1981, fig. 2). That this slip had a small Pressive component is shown by the Paleo- wal Permission, from Sclater, Hellinger and Tapscott (1977, fig. 3), Journal of Geology, v. 85, p. 517, copyright Oo. gene deformation of southwest Spitsbergen. At about 38 Ma, the slow spreading between Greenland and North America ceased, the mo- tion of Greenland became northwestward rela- tive to Norway, and oblique extension began be- tween Greenland and Spitsbergen; the result was Dc LAA =, >S — ~~ ETT EJ = LS ah fr i ~ a S Jj Wa SENS = Se NE = fj "08, A Y p : b yf /// \ ii E ! Lp : | d^ n d x f | \f J A K 2/7 Fy UM " (e. a. € nland-Norwegian Sea and the Eurasian Basin of the Arctic Ocean (right), and paleobathymetry in late i times. Contour interval 500 m, labeled in 100s of meters. Derived by hing magneti li depth curves for subsidence of oceanic lithosphere. Reproduced, with permission, 9vt N3Gq3vo TVOINV.LOS IANOSSIN JHL AO STVNNV 0L 10A] ee — ——QAX 1983] a shallow seaway on thinned continental crust between them, followed by a narrow but deep and widening oceanic rift from about 30 Ma (Talwani & Eldholm, 1977; Vogt, Bernero et al., 1981). nd connections. Dry land thus could have connected Greenland and Spitsbergen through- out Cretaceous and early Paleocene time (before most spreading occurred in the Arctic Atlantic) and during much or all of the rest of Paleocene and Eocene time (while continent-against-con- nent transform faulting was underway). Fur- ther, great basalt plateaus were built during Ter- tiary time on oceanic crust between Scotland and Fast Greenland, and these might have provided an intermittent land bridge before some time in the Miocene (Eldholm & Thiede, 1980). pitsbergen is now separated from Norway by the shallow sea of the continental Barents Shelf. Seismic-reflection data indicate that the strata beneath this shelf are mostly of pre-Cenozoic age (Eldholm & Talwani, 1977), so the shelf may mà: been emergent during much of Cenozoic e BAFFIN BAY AND LABRADOR SEA The early, western center of North Atlantic SPreading between Labrador and Greenland is represented by the oceanic Labrador [Sea] and ffin [Bay] Basins. Magnetic anomalies, drilling data ded bate: ine indicate that d ocean floor was forming in the Labrador sin between Greenland and Labrador by a lit- tle before 75 a Labrador until about 60 Ma; then, the no a motion slowed and changed to about betes “northeastward, and spreading accelerated eats Norway and Greenland. Motion be- 38 m bland and Labrador ceased about Ns. history of the northern part of the rift, lis een Greenland and Baffin Island, is less well m 9wn. Baffin Basin is floored by oceanic crust; «i «fend anomalies trending north-northwest- Mee indicate sea-floor-spreading of possible Pa- d e and Eocene age in the deep, central part 996 basin (Jackson et al., 1979; Keen et al., ). A Late Cretaceous age can be extrapolated HAMILTON —NORTHERN CONTINENTS 447 for the margins of the oceanic rift, and for ex- tensional thinning and faulting ofthe continental shelves prior to complete rifting. Upper Creta- ceous marine strata are present on the Baffin Is- land continental shelf (MacLean & Falconer, 1979). Ifthe Baffin Basi li identified rectly as of Paleogene age, then the width of Pa- leogene ocean formed between Greenland and North America changes little within the Labra- dor and Baffin Basins; the Euler pole of relative rotation is distant, and northeast-trending trans- form faults step spreading from one basin to the other through the Davis Strait region. Reconstructions between Greenland and Can- ada are complicated by crustal extension and ba- Iti gmati long tl gins of Baffin Bay and the Labrador Sea. Continental margins are thinned by extension accompanying rifting, and wedges of strata that prograde across them main- tain continental shelves. Something like half the width oftl tinental slope shall than 2,000 m along a rifted continental margin may typically represent extension of continental crust. The uncertainties of such extension are quantitatively most important for reconstruc- tions of narrow ocean basins, such as those of Baffin Bay and the Labrador Sea. These uncer- tainties are further complicated by new crust built of very thick lower Tertiary basalts of mantle origin, onshore and offshore of west-central Greenland between 68? and 73°N, and on east- ernmost Baffin Island and the shelf east of south- eastern Baffin Island (Keen et al., 1974; see also Johnson et al., 1982). Davis Strait, between the Baffin and Labrador Basins, is only 500 to 1,000 m deep and has crustal characteristics of a basalt plateau (Keen et al., 1 The Late Cretaceous magnetic anomalies of the Labrador Basin indicate an Euler pole of rel- ative rotation between Greenland and North America somewhere near Alaska, and from this itcan be predicted that the amount of Cretaceous 1. P Baffin Bay. This prediction is not fulfilled: south- eastern Baffin Bay is too narrow, even if much of the width of its flanking continental shelves be assumed to represent spreading. Other plate boundaries northwest of the Labrador Basin must accommodate some of the Cretaceous spreading. The problem is alleviated but not eliminated if the central Baffin Bay magnetic anomalies are partly of Cretaceous age. Interpretations devel- oped here are illustrated by Figures 3 and 4. 460° 448 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 80° 80° 80° | Conf X P d FACTS: SIGNS) UREN. Wty ~ T fne l E Oo | | ae | ee: | "X (S } 7 06 NORTH AMERICAN PLATE GREENLAND PLATE LABRADOR ^ END wed 100° FIGURE 3. Schematic illustration of Cretaceous and Cenozoic plate motions inferred Canada-Greenland region; Atlantic and Arctic Ocean features are not included. p - Į by circles: NA/G, G: and mainland Canada. Other abbrevations: CS, Cumberland Sound graben; FB, : HS St t o st Passage fault. Reconstruction using : i fault relative to Northwest Passage of the northern Arctic Islands is assumed to be the product of the resultant compression. Only requires counterclockwise rotation of Nares Strait adian shoreline is shown. ROTATION OF BAFFIN ISLAND Rotation ofan internally fragmented Baffin Is- land lithosphere plate, counterclockwise relative to both Greenland and mainland North Amer- ica, provides the likely explanation for much of the problematic geometry. Although Baffin Is- land generally has been assumed to be part of rthwe 60° | northem verall Double arrows Skone gulet Baffin Island | within the these i Te foldbel rtiary l the mainland ) fault; the i | plate, its int - | rwise (as BP | Canada pi | the rigid North American marginal geology indicate othe Shade, 1982, recognized). Mainland shallow Hudson Bay are separat Island plus shallow Foxe Basın widening graben systems of Fox son Strait. These grabens contain 1983] in a region otherwise surfaced now mostly by Precambrian basement rocks (Sanford, 1974). Similar grabens containing Paleozoic strata form Frobisher Bay and Cumberland Bay, in eastern Baffin Island (MacLean & Falconer, 1979); young normal faults cross southern Baffin Island at least along the Frobisher Bay trend. Normal faulting requires crustal extension. The amount of fault- ing, and hence probably also of extension, de- creases northwestward, so Baffin Island appar- ently has rotated counterclockwise relative to mainland Canada. The Euler pole of that relative rotation might lie along the boundary between the Baffin and Canadian plates, in which case crustal compres- sion should exist beween the plates along their boundary to the north of that pole, or the Euler pole might lie beyond the north limit of the boundary (here taken to be the Northwest Pas- sage fault), in which case the boundary should be entirely extensional. Published geologic data appear to permit either interpretation. Cambrian to lower Tertiary strata are broken by numerous faults and monoclines, mostly trending north- westward, in northern Baffin Island and adjacent areas. Although the faults have been termed nor- mal by all geologists who have mapped them, the strata of the downdropped sides rise toward , 1980). This geometry, and the presence of bpon folds away from faults, Permit the inference that the Cenozoic defor- ‘ive deformation of the Boothia uplift, west of Northern Baffin Island, is possibly of Cretaceous s early Cenozoic age. The straightness and con- "m of some of the Phanerozoic faults (e.g. censos et al., 1968) of west-central Baffin Is- however, permit an ete of domi- nantly strike-slip faulting ther Ps. Tétaceous and Paleogene caries on the or- s E = km might reasonably have : produced (Rice, us Shade, 1982, suggested 80 km.) An Euler Pole in the region of 69°N, 86°W, near Melville HAMILTON—NORTHERN CONTINENTS ^e N NEC ouo " aw *, SEET or ag puan 1008 o (T » o "ot 3 D d ®© Ny S L^; ma 45° + 3 adl FIGURE 4 1 east C lines ‘a mainland Canada, northeastern eng Island, of Arctic-island pond middle C nt shore- 68°N, 173°W, after the compilation by V (1981). pegase lines are 15° apart and soleh the equal-area projection. Plate boundaries marked by heavy lines: LBsc, Labrador Basin-Baffin Basin spread- ing center; NWf, Northwest Passage transform fault; NSf, Nares Strait transform fault; Bc and Bd, conver- = and divergent parts, respectively, of boundary be- een Baffin Island and North American plates. Circles pus Euler poles of subsequent grt ee of plates: NA/G, North America and Greenland; Noe North anpra and Baffin Island plate; RÀ A/OE, N rth America and Queen Elizabeth Islands plate. Mi oca and Atl ean features, and Eurasian, Alaskan, and Cordilleran landmasses, are not depicted. Peninsula, would account for the normal faulting of the southeastern Baffin region by orthogonal extension; for strike-slip faulting in west-central Baffin Island; and for crustal shortening in north- ern Baffin Island and any young component of shortening in the Boothia uplift— if these dis- parate elements are all of the same age and are characterized correctly in this scenario. Adop- tion of such an Euler pole, plus the assumption of 100 km of extension in the southeast, would predict about 5° of rotation, and hence about 70 km of total crustal shortening between north- east Baffin Island and the mainland west of the Boothia uplift. If, however, the Boothia compression is entirely older, and the north Baf- fin structures are extensional rather than com- pressional, then the Euler pole lies farther north- west, beyond the Northwest Passage. Such a pole would require a smaller angular rotation of the Baffin plate. Either interpretation can be inte- 450 grated with the motions defined for the opening ofthe Labrador Basin to account for the observed progressive northwestward decrease in the net motion between Baffin Island and Greenland. SEAWAY The history of rifting, oceanic circulation, and paleoclimates was deduced by Gradstein and Sri- 10(09/»x 1 TY E I T = iy *1 n lis vastava (1 ISU) Dy ay oi well on the shelves flanking the Labrador Basin, an stu and again in the early and early middle Eocene, subtropical Gulf Stream waters reached about as far north as pres- ent latitude 50°N (paleolatitude about 35? or 38°); water was cooler during the Paleocene and again during the middle Tertiary, but there probably was a general northward flow of water toward a nonglacial Arctic Ocean. By late Miocene time, the cold, south-flowing Labrador Current was in existence and glacial conditions presumably ex- isted in the Arctic. TRIPLE JUNCTION AT THE NORTH END OF BAFFIN BAY The 450 km or so of spreading of northern Baffin Bay represents plate motions that could not have stopped at the north end of the bay, and that must have been accommodated on oth- er plate boundaries to the north of Baffin Island and Greenland. The various proposals for these boundaries in the literature include a ridge-trans- form junction, a ridge-ridge-ridge triple junction, a ridge-trench junction, a pivot between spread- ing and compression, and rejection of plate tec- tonic concepts. None of these proposals fits both the geometric requirements and the known geo- logic and geophysical parameters, and I do not discuss them here. Instead, I note evidence of a triple junction between the Baffin Bay spreading center and two left-slip transform faults. One of these faults follows narrow Nares Strait north- northeastward from the north end of Baffin Bay, separating Greenland and Ellesmere Island. The other fault trends westward from the north end of Baffin Bay, along the aligned straits of the Northwest Passage. Fault, fault, and spreading center meet at mutual angles of about 120°. Nei- ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 ther fault is itself in a transform (small-circle) relationship to any likely spreading direction be- tween Greenland and Canada; but by postulating simultaneous motion on both faults, the sepa- | ration of Greenland from mainland Canada can | be reconciled with Arctic geology. The linear trend of Nares Bay, between Ellesmere Island and Greenland, continues beyond Ellesmere as the north edge of the submerged continental margin of North Greenland (Figs. 3 and 5). The strait was as sumed by Wegener (1920), and by mo ya td ag zh du of plate tectonics, to mark a strike-slip fault hav- ing about 250 km of left slip, related to the open- ing of Baffin Bay. I believe that such a fault does indeed exist Many geologists disagree. A symposium on ev dence for and against the fault was convened in 1980, and its results were published in à volume edited by Dawes and Kerr (1982). In gen symposium authors concerned wit structions showed that no geometric alternative to a strike-slip fault has ł recognized, whereas most of the geologists familiar with the er on one side or the other of the strait argued 10r little or no faulting. The latter group ¢ presented an impressive number of argumen ERES anlogic zones Taulting Dy D & - TOJCUUMS and lines across the strait with little or no offset, as summarized by Dawes and Kerr n ec paper in the book. Having examined the "d s geologic literature on both sides of the S reconnaissance 1:250,000. geologit am various reports and monographs ‘reget more limited materials for Greenland— none of the anti-fault arguments as D pec hence, regard their sum also as unconvin The critical components at issue are E p^ resenting changes in sedimentary facies p sides of deformation in Paleozoic strata 0n E of the strait. Some such lines are draw? a offset across the strait by the anti-fault 7 po but are unconstrained on one side oT st and hence are irrelevant. Others coul dappea! strained by data on both sides but insted trs to be drawn schematically where they me " : 0 continental shelf Paleozoic strata, nasi appe of shelf strata, and various other ele ollectively ts eral, | h plate recon- = i] —QÀ 1983] HAMILTON—NORTHERN CONTINENTS -08L y: VN (ej 4. y pum Tz Ve d S7 FIGURE 5. Bathymetry of the Arctic Ocean. Contour interval 1 km, wit p me to accord with left slip of about 250 km ae the strait, just as Wegener inferred long o. tata oreat Nare« Ennis . “rcle. It is not a small circle to any likely Euler Pole of relative rotation between Greenland and nada, and hence must have changed in shape, Position, Or ar . Ta’ é as al ters A twe; 1 P PEZ tO op e ading 5 pr so rthwest Passage fault. The aligned sounds, to 100 km wide, of the Northwest Passage curve Westward from the north end of Baffin Bay , with an additional contour at 0.5 km. Reproduced, with permission, from Taylor, Kovacs, Vogt and Johnson (1981, fig. 1a). and separate the northern and southern islands of the Canadan Arctic Archipelago. The midline of these sounds approximates a small circle to a pole near 85°N, 95°W, so the passage may mark a strike-slip fault along which the northern is- lands have slipped relative to the rest of Canada. Published data on distribution of facies in the widespread Paleozoic platform strata are com- patible with a left slip of as much as 200 km, but not with any right slip (cf. Daae & Rutgers, 1975, and Kerr, 1980; these authors, however, as- at lin) I lu 4i aA s P variably thick Cretaceous and Paleogene sedi- 452 mentation in the passage are shown by geophys- ical data (Daae & Rutgers, 1975), and exten- sional faulting (which I assume to be oblique) is indicated in the eastern sector (Kerr, 1980). Kerr (1974, 1977) argued against young strike- the compressional **Cornwallis foldbelt" trends directly across the passage and is of Devonian age. Although structures north and south of the passage are indeed approximately in line, I in- terpret the geologic maps of Bathurst and Corn- wallis Islands (Kerr, 1974; Thorsteinsson & Kerr, 1968), north of the passage, to indicate that the structures there are in fact of Tertiary age, and hence do not constrain strike-slip faulting. For extensional) consists of an asymmetric syncline of the young rocks, against the steep, west limb of which is faulted a complex anticline of lower Paleozoic strata. Such geometry characterizes compressional, not extensional, structures. South of the Northwest Passage, north-trend- ing compressional faults of the Boothia uplift break basement as well as cover rocks, and can be dated directly only as postdating Lower De- vonian strata and as predating highest Creta- ceous strata (e.g., Miall & Kerr, 1977). That an clastic Upper Silurian and Lower Devonian sed- iments shed from it both to east and west (Miall & Gibling, 1978). No analogous upland is shown by correlative strata north of the passage, where dominantly cart } lated during this stratigraphic interval, platform conditions to the southeast giving way northwestward to con- tinental-shelf ones. The lack of present conti- nuity may be a result of post-Devonian faulting along the Northwest Passage, and the anticipated northern continuation of the upland and deriv- ative clastics may now be hidden beneath youn- ger strata and the sea farther west. 131mo RECONSTRUCTION OF CANADA AND GREENLAND magnetic-anomaly patterns of the Labrador Ba- sin, the northern part of West Greenland is ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 I brought against Baffin Island in a reasonable fit (Fig. 4). The total motion of the south tip of Greenland, relative to mainland Canada, thatis , indicated by this reconstruction is about 800 km — in northeastward direction. The motion reversed — corresponds to a counterclockwise rotation of | about 9? of Greenland relative to mainland Can- ada, about an Euler pole near 65°N, 140°W ina } Canadian reference frame. This reconstruction of Greenland, Baffin Is- / land, and Canada is possible only if there is left slip on both Nares Strait and Northwest Passage faults. The specifi tion requires about 300 km of left slip through Nares Strait (com- pared to the best geologic value of about 250 km), and about 200 km of left slip along tht V Northwest Passage. The reconstruction also re- quires that the Northwest Passage and Nares Strait fault boundaries of the northern Arctic Is lands (Queen Elizabeth Islands) have rotated to- ward one another about 5°, about the point of ! their intersection; the predicted north eMe en up by the northwest-widening belt of bed ceous and early Tertiary folding in the islan TECTONICS OF THE ALASKAN REGION The west limit of continental crust, as MC l by the presence of continental-shelf stra by other criteria, that was part of i | southern Yukon, thence northward thro gh | northern Yukon to the Arctic Ocean. A of now to the west of this, including nearly ! Alaska, probably was added tectonically yer | continent during Devonian and youn ht Diverse rotations and strike-slip motions accompanied the tectonic accretion. The plexity of the collage increases northwes and comprehension of its evolution d 1979, | correspondingly (e.g., Jones & Silber i ppe Jones et al., 1982; Monger et al., eae » et al., 1981). A large part of the co uc sembled (although with grossly different or uration than it has now) before the ge a | Cretaceous time. I will treat it here n © | sory fashion. OPENING OF THE CANADA BASIN The oceanic Canada Basin, north pe | and western Canada (Fig. 5), was dos gcomett l sozoic spreading, although the specifi 1983] and timing of that spreading are poorly con- strained. Interpretations have varied widely in detail but in general have advocated either that northern Alaska rotated counterclockwise away Eo (unlike ^" M s patterns of the Canada Ba g the tidy, symmetrical anomalies of the Eurasia Basin; Fig. 6), and can at the present level of knowledge be fitted to either rotation or sliding models (Vogt et al., 1982). The onshore and off- shore geology and geophysics of northern Alaska and northwestern Canada indicate that rifting premonitory to opening of the Canada Basin had begun by Early Jurassic time, and that an ocean of undefined width was present at least as early as late Early Cretaceous time (e.g., Grantz et al., 1979; Miall, 1979; Young et al., 1976). The mag- netic anomalies of the basin were inferred by Vogt et al. (1982) to date the ocean floor as hav- ing formed largely within Late Jurassic and Early Cretaceous time. . The Canada Basin apparently is largely or en- urely older than the Late Cretaceous and Ce- nozoic spreading and transform faulting by which the Eurasia Basin, Arctic Atlantic, Labrador Sea, and Baffin Bay were opened. ROTATION OF NORTHERN ALASKA The Paleozoic and early Mesozoic geology of northern Alaska —the Brooks Range and North Slope—and adjacent northern Yukon Territory (c£, Dillon et al., 1980; Grantz et al., 1979; orris & Yorath, 1981) is compatible with the morphologically reasonable rotation of that re- blage Consists of Mississippian through lower e esozoic strata (in the Canadian Arctic, the Mesue Basin fill) lying unconformably on old- complexes. The other two shared assemblages àre pre-Mississippian. The southern one in the HAMILTON—NORTHERN CONTINENTS 453 Figure 6. Magnetic anomalies of the Arctic Ocean and the northern Greenland-Norwegian Sea, showing > ath 14 + th fhil LY) » less ain 1 than (white) the global magnetic reference field. The discontinuity near 160°W—60°E presents different data- ductio: hods for the Soviet dat the Eurasi side, and the generally more detailed U.S. Navy and adian data. R uced, with permission, from Vogt, Bernero, Kovacs and Taylor (1981, fig. 2). Canadian Arctic is a tract about 150 km wide of contorted and low-grade-metamorphosed upper water ctrata Wall olidla (e.g., Trettin et al., 1979). This terrane lies north £l. —— dde oe | 1 147 1.7 rc 7 . OIL D shallow-water materials, deposited on North merican crust. I interpret the deep-water ma- continental rise, and abyssal plain, crumpled to- gether while oceanic lithosphere beneath them was being subducted beneath an advancing : vai dispu age acci avs northwest Yukon region are exposed in the Ro- manzof and other uplifts and are known also in the subsurface along the north coast. The north- 454 ern terrane in the Canadian Arctic is an assem- blage of continental and oceanic crystalline and sedimentary rocks, of late Precambrian to De- vonian ages (e.g., Trettin, 1982), which record poorly understood Devonian and older plate convergence and tectonic accretion. Equivalents of this northern Canadian Arctic assemblage ap- pear to be represented by the Devonian and older terrane of the southern Brooks Range and nearby Yukon Territory. CENTRAL AND SOUTHERN ALASKA Alaska south of the Brooks Range consists of diverse terranes, tally i d continental, that | 1 long dist with regard to the Brooks Range, mainland North America, and each other. The terranes are var- iously of Paleozoic, Mesozoic, and Cenozoic ages, but their present juxtapositions and configura- tions are primarily products of Cretaceous and enozoic subduction, strike-slip faulting, and oroclinal folding. Many terranes, sutures, and deduced; but any attempt at statewide palin- spastic reconstructions for Cretaceous and early Dalang. A 4 T3 4 4 1 quite speculative. A large part of central and southern Alaska was nevertheless structurally part of North America by Late Cretaceous time. The suturing of the large Yukon-Koyukuk terrane, apparently an oceanic island arc, to the south edge of the Brooks Range occurred late in Early Cretaceous time. Before then, various terranes now farther south in Alaska were attached to the continent, but at positions farther southeast along the Cor- dillera; they slid northwestward to their present positions to form an increasingly wide Alaska The Aleutian Trench marks the present south margin of the Alaskan plate. Pacific Oceanic lithosphere is moving relatively northwestward and subducting beneath Alaska and the Bering Sea. Alaska is widening as oceanic materials are scraped off into the accretionary wedge at its leading edge. TECTONICS OF NORTHEASTERN EURASIA The northern part of mainland Eurasia west of present latitude 162°E was assembled by the end of Jurassic time, although both major inter- nal deformation and tectonic accretion contin- ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 ued in the east during the Cretaceous and Ce- nozoic (Churkin & Trexler, 1979; Dickinson, 1978; Fujita & Newberry, 1983; Hamilton, 1970; Takahashi, 1983; Tapponnier & Molnar, 1979). It is likely that since about the middle of Cre- taceous time, continuous crust of continental and Alaska, although much doubt clouds specific palinspastic restorations. PACIFIC MARGIN Kamchatka Peninsula, the Koryak Highlands to the northeast of it, and the region northwest of the Koryaks, consist of a complex collage of accretionary and magmatic-arc terranes, 700 km wide, gan early in Cretaceous time, for Lower and Up- ar Cret tie racks lie across the magr truncated and poorly understood con 5s older Mesozoic mainland terranes west ‘om oryaks. Jurassic and Cretaceous fossils sparsely in the accretionary-wedge melanges that oic rocks ogy with northwestern North America. 4 arc rocks tend to become younger toward -e Pacific, indicating a general sequence © pe accretion in that direction. Middle he Paleozoic components of mega- ing fragments of seamounts and ato spread in the Pacific half of the Ko f this a and as it is unlikely that any seafloor 9 : have ? Plate-tectonic interpretations of Soviet geology from outside the Sov! indt USSR 6 pa : is : lications. The following brief on own redit the references cited above and from ps northeast of the Soviet literature regarding the SSR. j anmenn f tectomi — i ———— 1983] was still unsubducted in Late Cretaceous time, tectonic assembly elsewhere is likely. The C eous magmatic arc of the Asian the matching accretionary terrane is in Sakhalin, far west of the middle and late Cenozoic Hok- kaido-Kuril-Kamchatka arc system. Dickinson (1978) inferred from regional geologic and geo- physical relationships that a small continental crustal mass, represented now by the shallow part of the Sea of Okhotsk and perhaps by exposures in western Kamchatka, collided with the conti- nental- gi bducti y in about Eocene time, and that subduction then broke through on the Pacific side of that added mass. Migration of the arc systems, and the opening of the Sea of Japan and the deep-water part of the Sea of Okh- otsk, has occurred since. ARCTIC MARGIN The modern spreading center of the Nansen- Gakkel Ridge trends south-southeastward to the north edge of Eurasia, north of the Lena River delta (Figs. 5 and 6). The amount of Late Cre- laceous and Cenozoic spreading decreases to- ward Eurasia but is still about 600 km at the Eurasian continental slope. One or several plate undaries must connect this spreading center lo past and present plate boundaries around the Pacific Ocean. A zone of diffuse seismicity con- üinuing southeastward through the Lena River region to the Sea of Okhotsk presumably rep- resents the modern boundary along which ex- lension decreases southeastward between Eur- asian and North American plates (Chapman & RE 1976). Spreading within the region of Seismic zone cannot be great enough, how- ever, to account for more than a small part of the total motion needed, for this mainland region E aei e ically high and is not conspicuously : One or more other plate boundaries, distinct t diffuse and now inactive seismically, must lie bie. the onshore or offshore region of Arctic he Lena River and about 160°E, mark "aas m. HEN P te from . tancian HAMILTON—NORTHERN CONTINENTS 455 the modern seismic zone (cf. Chapman & Sol- omon, 1976). BERING SEA REGION The crust beneath the deep southwestern half of the Bering Sea, inside the Aleutian island arc, is oceanic, but that of the northeastern half is of continental thickness. Although now covered by shallow water, much of this latter half undoubt- edly stood above sea level during much of its history. Two broad geologic provinces can be recog- nized from geophysical and geologic data from the Bering Sea continental shelf and its margins. rence Island to northeastern Chukotsk Peninsula (e.g., Patton & Tailleur, 1977). In the south is the broad terrane of accretionary geology and magmati ting southwest Alaska and the northeast Koryak-southwest Chukotka re- gion, probably consolidated mostly in Creta- ceous and early Paleogene time (Csetjey et al., 1971; Marlow et al., 1976; Moore, 1972; Patton et al., 1976; Pratt et al., 1972). The accretionary and arc-magmatic history of the present shelf region ended in latest Cretaceous or very early Tertiary time, when subduction northward be- neath the shelf region ceased. The modern boundary between the Pacific and North American plates south of the Bering Sea is the Aleutian trench, which curves from south- west mainland Alaska to a trench-trench junc- tion off Kamchatka. The Aleutian island arc has been active only since early or middle Eocene time; oceanic lithosphere between the Aleutians and the Bering continental slope probably is a trapped bit of Mesozoic ocean floor (Cooper et al., 1977) but might instead have formed in early Cenozoic time behind migrating island arcs (Langseth et al., 1980). Curving from the north side of the central Aleutian arc to the Koryak region are the Bowers and Shirshov submarine ridges — segments of a fossil arc beneath which subduction was southward (Cooper et al., 1981). Various explanations (e.g, Ben-Avraham & Cooper, 1981) have been proposed for the evo- lution of this complex submarine region. The most promising is one communicated to me by M. Worrall, who integrated offshore geo- 456 physical data with onshore geological informa- tion. Worrall inferred that a northward-migrat- ing, north-facing Bowers-Shirshov island arc collided with a south-facing Bering-Koryak Highlands continent-margin trench system in early Eocene time. He also inferred that both of these systems were abandoned when the Aleu- tian subduction system broke through the ocean- ic plate to the south, trapping Mesozoic oceanic crust to the north of it. LITERATURE CITED ALVAREZ, WALTER, TOMASO Cocozza & F. C. WEZEL. 974. Fragmentation of the Alpine orogenic belt by microplate dispersal. Nature 248: 309-314. BEN-AVRAHAM, Zvi & A. K. Cooper. 1981. Early evolution of the Bering Sea by collision of oceanic rises and North Pacific subduction zones. Bull. Geol. Soc. Amer. 92: 485-495. BLACKADAR, R. G., W. L. Davipson & H. P. TRETTIN. 1968. Geology, Erichsen Lake, District of Frank- BuRRETT, C. F. 1974. Plate tectonics and the fusion of Asia. Earth and Planetary Science Letters 21: 181-1 189. CHAPMAN, M. E. C. SOLOMON. 1976. North pe merican-Eurasian plate boundary in northeast a. Jour. Geophys. Res. 81: 921-930. wu MICHAEL, JR. & J. H. TREXLER. 1979. ment on ‘The Siberian connection—a case for Precim brian separation of ay North American and Siberian cratons.’ Geo ogy 7: 467—469. COHEN, C. R. 1980. Plate tectonic ins for the Oli- go-Miocene evolution of the western Mediterra as Tectodopibaies v 283-311 Cooper, A. K., M. S. MARLOW & ZvI BEN-AVRAHAM. 1981. Multichannel indir evidence bearing on the origin of er ts Ridge, Bering Sea. Bull. Geol. Soc. Amer. 92: 84. — & x Y SCHOLL. 1977. The e Bering Sea—a multifarious marginal basin. Amer. Geo- ph CSETJEY, BELA, JR. PATTY E 19 gale ea plutonic is of St. Lawrence Island, Alaska—a preliminary report. Profess. Pap. U.S. Geol. Surv. 750-D: 68-7 Daag, H. D. & A. T. C. AULA 1975. Geological history ofthe Northwest Passage. Bull. Canad. Pe- trol. Geol. 2 x ue Dawes, P. R. & j. RR (editors). 1982. Nares Strait and the D Ben of Greenland—A Conflict in Plate Tectonics. eae 8: p^ : DEWEY, J. F., W. C. x N II, W. B. F. RYAN & J: BONNIN. 1973. dio irae and the Mor of ong dien -— Bull. Geol. Soc. Amer. 84: 3137-3180. .SENGOR. 1979. A egean and surround- ing regions.- complex multiplate and continuum tectonics in a convergent zone. Bull. Geo Amer. 90: 84-92. DICKINSON, W. R. 1978. Plate tectonic evolution of ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 _— Pacific rim. Jour. Physics Earth 26, Suppl: l- — L T., G. H. PEssEL, J. H. CHEN & N. C. VEACH. 1980. Middle Paleozoic magmatism and e ks esis in the Brooks Range, Alaska. Geology 8: 338- | 343. | ELDHOLM, O. & M. TALWANI. tribution and structural framework of the Barents & J. T. NEWBERR nary beg and tent evolution of no no ria. Pp. 4 M. e Siberi 2575 Hashimoto & H Uyeda (editors), Accretion Tectonics in the Cir cum-Pacific s. Terra Sci. Pub. Co., Tokyo GRADSTEIN, F. M. S. P. SRIVASTAVA. 1980. Asp ofC the Labrador Sea and Baffin Bay. Palaeogeogr. Pe laeoclimatol. Palaeoecol. 30: 261-295. ental margin north of Alaska. Tectono physics 59: eei GROSSHEIM, V. KHAIN (ogi 1967. At las maps of the VOR yes F. ternary. USSR Ministry of Geology. HAMILTON, WARREN. 1970. The Uralides TT Geol. "n Amer. 81: 2553 -2516. Mesozoic a of the westem Pa- United Se In Pacific Section Soc. hi » leontologists and Kimera i Pacific r Symposium 2: 33-70 9. Tectonics d the Indonesian Region. eol. Surv. 1078. 345 pP VASTAVA, D. Geophysical transects of the Labrador n $ rador to southwest Greenland. Tectonop 151-183. Jackson, G. D. & A. Davipson. 1975. bes Map-area, District o in. Geo. ada Pap. 74-29: P Um RKH rU &K 1979. New ea irs a i tral Ba evolution of the Labrador Sea. Pap. 82-1B: 7-20. Jones, D. L. & N. J. SILBERLING. Stratigraphy—The Key to o Tec Southern and Central Alaska. U Open-File Report inis 37 x i: "n iacere rag HE G. PLAFKE A TETE y ipa pM of Alaska Geol. Surv. Circ. 844: ick 1974. Keen, C. E., M. J. Keen, D. I. “Ross & M. gear lo Baffin Bay — small ocean basin ue by PY pull. 5 spreading. Amer. Assoc. Petr 1089-1108. Keicwin, L. D., JR. 79. Mesoro i Ana ^ f the 1978. Pliocene closing ° 1977. Sediment dis- , } | 1983] Isthmus of Panama based on biostratigraphic evi- dence from nearby Pacific ean and Caribbean 30 4. Geology of Bathurst Island Group and Byam Martin Island, ~~ c Canada. Geol. Surv. Canada Mem. 378: 1 1977. Cornwallis fold ie and the mecha- m of vore uplift. Canad. Jour. Earth Sci 14 1374-14 1980. SESS Framework of Lancaster em mb Arctic Canada. Geol. Surv. Cana 131 KRISTOFFERSEN, Y. & M. _ TALWANI. — . — 1977. Seen motion of Greenland relative to North peur Bull. Geol. Soc. Amer. 88: 1037-104 , M. A. HOBART & Ki-rri Horai. 1980. Heat flowi in the Bering Sea. Jour. Geophys. Res. 85: 3740-3750. Le PicHon, X., R. D. HYNDMAN & G. PAUTOT. 1971. Geophysical study of hio opening of ees Labrador 743. 24-4 . 1979. Geolog- geophysical studies in Baffin Bay and Scott bid dy Gulf and Cape Dyer-Cumberland und areas of the Baffin Island Shelf. Geol. Surv. da Pap. 79-1B: 2 4. Marrow, M. S., D. W. ScHoLL, A. K. Cooper & E. C. |: 1976. Structure lex Mee of Bering Sea shelf south of St. nce Islan Amer. Assoc. Petrol. Geol. Bull. gy 161- 183. McGookey, D. P., J. D. nuy L. A. HALE, H. GoopeLL, D. G. McCussin, R. J. WEIMER & G. 90- Island, onto ic Canada. Geol. Surv. Can- E Mem. 387: 235 p Co HE BAL KL. & W. S. Hopkins, JR. 1980. Cretaceous and Tert ry Sediments of Eclipse Trough, Bylot E Thina: ri Arctic Cana nd eir Regional Setting. Geol. Surv. Cana da Pap. 79-23: 20 pp. ^ —-&M.R. GLG. 1978. The Siluro-Devonian Clastic wedge of Somerset Island, Arctic Canada, and some regional pin duis implications. Sedimentary Geol. 21: & R 1977. Pid Gio stratigraphy t Island a: and sedimentology of Somerse nd north- thia Peninsula. Geol. Surv. Canada Pap. Mo, tA: 99-106. pes R, PETER & PAUL TAPPONNIER. 1975. Ceno- Onics o z ucl "Ss lision, E ieu 419—426. NGER, J. H NX D. J. TEMPEL- MAN Kin. 1982. Tectonic accretion and. the n in the Canadian Cordillera: Geology 10: 70- Pl CORE, J. C. 1972. Uplifted trench sediments— uthwestern Alaska-Bering Shelf edge. Science 5. Na 175: 1103-110 jum D. K. & C. J. YonarH. 1981. The North Mm plate from the Arctic Archipelago to the 9manzof Mountains. The Ocean Basins and Margins 5: 37-103. Plenum, New York. HAMILTON—NORTHERN CONTINENTS PATTON, W. W., JR., M. A. LANPHERE, T. P. MILLER & R. A. SCOTT. cance of volcanic rocks on St. M Bering Sea, Alaska. U.S. Geol. Surv. Jour. Res 67-73. & I. L. TAILLEUR. 1977. Evidence in > Ber- ing Strait region for differential movem - tween North America and Eurasia. Bull. bol. Soc. Amer. 88: 1298-1304 PINDELL, JAMES & J. F. sida "s Permo-Triassic reconstruction of w a and the evolu- tion of *» Gulf i Seo or dia region. Tec- tonics 1: 179-2 TI ,M. : Tu F. W. WALTON & J. A. BUSCHUR. 1972. niega of Alaskan struc- ol Bay, Bering Shelf, 994—4999, Reflection seismic interpretation and seafloor dn of Baffin Bay. ian Soc. Petrol. Geol. Mem. 8: 245-265. 1966. Lower Cretaceous Soc. P SANFORD, B. V 4. Pélpoacic geology of the Hud- son Bay region. un Surv. Canada Pap. 74-1A: 144—146. SCLATER, J. G., LLINGER & C. TAPSC 1977. The EAS a or of the Atlantic Ddin from the Jurassic to the present. Jour. Geol. 85: 509- SINHA Rov, SuBIMAL. 1978. Eastern Tethys and mi- SMITH, A. G., A. M. Heus & J. C. BRIDEN. Phanerozoic oo World Maps. bridge University Pres SRIVASTAVA, S. 1978. Chains of the Labrador and its bearing on the early evolution of the North Atlantic. Geophys. Jour. Roy. Astron. Soc. AG 1981. Cam : 313- TAKAHASHI, MAS. 1983. Re me in io of late Saten to early Cenozo em: Asia and its tectonic esca, "ov. 69-88 n M. Hashimoto & S. Uyeda (editors), n MOON in the Circum-Pacific region. Terra Sci. 5 TO kyo. TAPPONNIER, "PA UL & PETER MOLNAR. 1979. Active faulting and Cenozoic tectonics of the Tien — Mongolia, and Baykal regions. Jour. Geophys. R: 84: 3425-3459 AYLOR, P. T., L. C. Kovacs, P. R. Tt & G.L. JoHNSON. 1981. Detailed ois investi- gation of the Arctic Basin, 2. Jour. Geophys. Res 86: 6323-6333 THORSTEINSSON, R. & J. W. Kerr. 1968. Cornwallis Island and Adjacent Smaller Islands, Canadian c Archipelago. Geol. Surv. Canada Pap. 67- 64: 16 pp. Tipper, H. W., G. J. WooDSWORTH & H. GABRIELSE. 1981. Tectonic assemblage map of the Canadian illera. Geol. Surv. Canada M : Trettin, H. P. 1982. Innuitian Orogen and Arctic Platform. Geol. Soc. Amer. D-NAG Special Publ. 1: 49-55. 458 ——,, C. R. BARNES, J. W. Kerr, B. S. NoRFORD, A. d, Di Geol. Surv. Canada Pap. 79-1B: 269-279. VERESHCHAGIN, V. N. & A. B. Ronov (editors). 1968. Atlas of the Lithological-Paleogeographical Maps of the , III, Triassic, Jurassic and Creta- u: ini Geology of Oceans Sym , R. K. PERR v. R. H. Fip Coe ROREM egian G. Stehli (editors), The Ocean Basins and Margins 5: 493-598. Plenum, New York. ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 — —— , P. T. TAYLOR, L. C. Kovacs & G. L. JOHNSON. 1982. The Canada Basin—aeromagnetic con- straints on structure and evolution. Tectonophys- ics 89: 295-336 WEGENER, ALFRED. 1920. Die Entstehung der Kon- tinente und Ozeane. Vieweg & Sohn, Braun schweig. WiLLIAMS, G. D. & C. F. Bunk, Jr. 1966. Upper Cretaceous. Pp. 169-189 in R. G. McCrossan & R. P. Glaister (editors), Geological History of ) Mage Canada. Alberta Soc. Petroleum Geolo- A ur T. G., D. W. Myru & C. J. YonarH. 1976. Geology of p sete eri Basin. Geol. Surv. p. 76-1 ZIEGLER, A. M., C. rn qoe ae s. "McKERROW, M. E. JOHNSON & R. K. BAMBACH. 1979. — paleogeography. Annual Revi view Earth and Plan tary Sciences 7: 473—502 und ilius. seinen HOLARCTIC LANDMASS REARRANGEMENT, COSMIC EVENTS, AND CENOZOIC TERRESTRIAL ORGANISMS MALCOLM C. MCKENNA! ABSTRACT Rearrangement of landmass configurations has profoundly affected the evolution of mammals and other organisms in the Northern Hemisphere. At the north end of the Atlantic Ocean, evidence is very strong for a now sundered landmass of Euramerica. This landmass existed toward the end of the Paleocene but ipi es to separate into at least two landmasses in eed resian ids ie the AME rn until the end of pene time. Thus at the end of the Paleocene and until some point vus pn in Pom Eocene, two land bridges may have simultaneously bec uns North America with sep parat e > parts of ride across the sd end of ne Atlantic: one dom tegen: the other by »» early knees of the Icelandic enne Possibly a isis ion from Asia and Europe to the northeastern end of North America existed at about 37 million a years paa (Ma) through the combined activities of the Yermak hotspot, waning compression between Green- land and Svalbard, and the dryi rying of Eurasian seaways. Throughout nearly all of the RA a land connection between Asia and Alaska existed at the Bering Strait, but the —— of the Bering Bridge was initially high and climate must have exerted a strong filtering action on mammals an other terrestrial o isms. It is possible, but only weakly supported on geological grounds, that a the Kula Ridge was subducte r rearrangements that have strongly affected the composition of northern mammalian faunas and Lipsio would have iei ipis [o organis ell have n the reconnection of S u ri North America in ,the niga of India with m polar y Cen ozoic were the result of a lower mean spin axis obliquity of the Earth than the present 23.5°, no one has explained adequately how mean spin axis obliquity can be perm enone changed significantly i in the €nozoic, where the necessa sary force would come from to shift it, how such a force would be coupled to the Earth, or how the resulting crustal heat would be dissipated. At at sot known high latitude early Cra environm mtus of paper and other organisms are not well understood, but asi ignificant change of Earth's mean spin obliquity since the early Cenozoic is evidently out of the n midi ion. A catastrophic theory of Apollo-class asteroid impact at supposedly 35 Ma has recently urn-over. However, major turn-over in both - marine and terrestrial realms occurred well before 35 Ma pe N America, as did the “North American tectite strewn field." Several such fields exist, however. Under a stable continent riae » concepts of corridors, filters, and sweepstakes routes were developed s A ihe ae mpso y O. G. added E then, su oah's arks and grounded iei funeral = ; for colliding « or ene terran vicariance bioge ography, with of various areas, has an ‘born and is undergoing growing Pai ns. To these concepts may be added the notions af whi ereby of a biota can be reconciled with their presence on young mid-ocean-ridge islands like Iceland or mid- Plate Folaat edifices such as those of the Hawaiian-Emperor chain. The retreat from, and return to, co for biological voyages to nowhere and return, as would happen if the aa fe fr g its parent continent, remained for a while i in isolation, and then once more con Garay € continen Shores of the North Pacifi , as carriers of Parias biota, now endemic to a foreign land Before acne "e a block possible carrier of a terrestrial or even shallow water Marine biota, it is i o determine if the block possesses continental or oceanic tectonic basement. Was the surface s ud block always above sea level? Were its organisms transported only pues s NEN ‘ Department of se Paleontology, American Museum of Natural History, Central Park West at 79th Street, New York, New York 10024. ANN. MISSOURI Bor. GARD. 70: 459-489. 1983. 460 ANNALS OF THE MISSOURI BOTANICAL GARDEN as fossils? Are the proposed vicariant sister groups at reasonable taxonomic levels in view of the time f tectonic transport of the block while in isolation? A parsimonious approach to all natural history, ect not merely biology, is necessary. Earth V +h 2 a 415. A | hoc hypotheses should be held to a minimum. The expanding Studies of mammalian diversity based on the fossil record have often been biased by tacit assumptions that biostratigraphic units have represented aliquot divisions of time. Calibration of mammalian hit tilth | : A 3 l M ppe 4 45 . Jet ‘ae biostratigraphy is under way, J should be taken with a grain of salt. 1 like thic nea Authors of invited pap this one are usu ally given the title of their paper, covering subject matter that the symposiarch hopes will somehow be dealt with adequately. However, the task usu- ally proves to be overwhelming and the belea- guered author, nearly always at the last minute, produces instead a manuscript entitled **Prob- lems of such and such.” The author is then off the hook and able to postpone the real work. The "definitive effort" can always be written some- time next year. However, this paper is frankly such a preliminary effort even though it is longer than I would like it to be. In it I try to get at some ofthe main points in a vast subject in which I have taken some interest over the years, but it is aimed at an audi I l led ce ge of ver- tebrate paleontology or modern geology is not necessarily very extensive and who might find a documented review useful. To me it seems ob- vious that some of the physical changes in the world that either have or might have affected the of the planet's biota. Our world has had a single history, which we try to decipher from a mixture of signal and noise coming from many sources of potential evidence. With regard to spatial dis- tribution, the signal has in part been generated by geological processes and the noise in part by i1: c Á i wnercal A MOBILE GARDEN OF EDEN Early attempts to discuss the relationships of the vertebrates of the Northern Hemisphere, like those concerning plants or invertebrates, wert necessarily based on the living biota alone and were made on the assumption that the geography of the Northern Hemisphere has always been more or less the same as it is now, give or take [Vo. 70 | | | ) a seaway here or there, a shift in a coastline, o | a cyclic change in climate (Dana, 1847, 1863). The Sclaters’ (1899) and other nineteenth cen- S Platnick, 1981). Gradually, however, the richness of the Recen as been supplemented by an im j ta ; knowledge of the vertebrate fossil record of vat ious parts of the Earth. The geologic history of the world, in part dated by the fossil record al in part dated by other means, has become > creasingly well known, spectacularly so 1n the last few years of intense effort associated plate tectonic revolution in perspectives have shifted radically. At first, rope was taken as an adequate me e i i rses, 10 i larctica. The evolution of ho idi ied it with thé. l i ontinuity rep 221 was believed to bea c aeoth rit the European record by Eocene Pal We Miocene Anchitherium and Hipparion 1872: nally Pleistocene and living Equus (Huxley. scent Kowalevsky, 1873). However, the princ: the E of equid evolution was elsewhere and tinuous: ropean record is now seen to be qier the result of repeated immigration to p closely related animals rather than ee s of our sparse temporal sampling of 4 same nil digenous evolving lineage. Much the and id of camels, elephants, rhinoce! many other groups of mammais. zu d 1983] Important as the European record was (and still is) for deciphering the Cenozoic vertebrate record of Holarctica, it suffers from several pro- found defects. For instance, there is no record of terrestrial Tertiary fossil tetrapods anywhere fi the whole Fennoscandian northwest end of Eu- rope, an area including Sweden, Norway, and the formerly subaerial Barents Shelf that lies in the Arctic between them and Svalbard. Another dif- + “sh +h. TT nm . PE PST ONA Thaler, 1977; Vianey-Liaud, 1979), there has been no record of medial or early Paleocene ter- restrial vertebrates in Europe. Starting in the 1870s, an explosive phase in the geological and paleontological exploration of the North American West began. This work soon produced an apparently much more continuous view of the evolution of various mammalian groups than had come to light in Europe other an for obvious European endemics such as doormice, certain kinds of artiodactyls, or di- mylid insectivores, to name a few random mam ry rocks d out to be more completely represented 1n Europe. Until recent discoveries, the ter- restrial early or medial Paleocene fauna was rep- resented solely in North America, where the land- laid deposits of the Rocky Mountain area of the United States and Canada supplied abundant and fairly well preserved fossil remains representing the earlier parts of the Paleocene. It was natural, then, that this North American terrestrial Paleo- to return. But as the North American terrestrial verte- brate record was filled in by the efforts of E. D. Cope, O. C. Marsh, W. D. Matthew, and their Successors, several important discontinuities ap- Peared and ^N eol i ae liances began to loom. Although a much more Continuous record of certain groups, such as the _— horse sequence, was established in North America, a break between the American Eocene and Oligocene became obvious from the mam- dl record (Wilson, 1980). Also, the source Several floods of latest Paleocene and/or early McKENNA —HOLARCTIC LANDMASS REARRANGEMENT 461 Eocene immigrants (Rose, 1981) into known fos- sil-producing areas of both North America and Europe from somewhere else became an increas- ingly vexatious problem. Furthermore, known North American late Cretaceous therian mam- mals did not appear to be ancestral to most Ce- nozoic forms. For these reasons, northern Asia came to be regarded as the likely true Garden of Eden, where, if the record of remote Central Asia could just be sampled, all would become clear. Professor Henry Fairfield Osborn of Columbia University and the American Museum of Natural History held the firm belief, as did most of his associates, that the northern Asian (“Central Asian") recor eld t to the origin of various northern groups, including our own ancestry among the higher primates (Osborn, 1923a, 1923b). Os- born's view was responsible in large part for the mounting of the Central Asiatic Expeditions, a ina 1 held the scientific spotlight in the field of natural history in the 1920s and whose spectacular labors filled in for the first time much of the Cretaceous, Eocene, Oligocene, and late Miocene terrestrial history of Mongolia and northern China. In Cen- tral Asia Osborn expected to find the ancestors of various rootless lineages of vertebrates known from Europe and North America. Given the as- sumptions of a stable-continent framework that prevailed in North American paleogeography in those days, what could be more natural than to find such links on the geographically interme- diate landmass of Asia, half way between North America and Europe? EARLY CENOZOIC HOLARCTIC REARRANGEMENT But, the northern Asian record for the early Cenozoic turned out to be a disappointment for ancestor seekers. Instead, as so often happens in science, a whole new series of problems arose (Szalay & McKenna, 1971; Fox, 1978). In place of ancestors, the Mongolian and Chinese early Cenozoic in many, even most, instances pro- duced a new endemic cast of characters wholly unexpected in the terrestrial deposits of what was supposed to have been a continuous corridor (Simpson, 1943, 1953) connecting North Amer- ica across Asia all the way to Europe. More recent collecting in the Paleocene of China has lent an even more peculiar aspect to early southeastern 462 Asian terrestrial faunas (Li & Ting, 1983). For instance, early primates, common enough in the North Am thus far completely absent from the Chinese Pa- leocene with only one possible but disputed ex- ception, Petrolemur. Primates apparently reached eastern Asia only at the very beginning of the Eocene (Dashzeveg & McKenna, 1977). Multi- tuberculates are also totally unrepresented in the known Chinese Paleocene, although they were previously abundant in the Mongolian Creta- ceous and occurred again in Chinese and Mon- golian assemblages at the very beginning of the Eocene. In place of these primates and multi- tuberculates, Chinese Paleocene faunas supply us with a flood of new taxa that most Western pa- leontologists have found bewildering in variety and difficult to place in the scheme of things. In facile retrospect, clearly something was wrong and the flaw surely could have been spotted in the 1920s or 1930s had not the stable-continent rationale prevailed (Simpson, 1953, 1965). at I refer to, of course, is that landmass interconnection within Holarctica in the early enozoic was arranged a bit differently from the familiar geography of the present day and of the Pleistocene (Wolfson, 1949; Nelson, 1973; Fennoscandian shield of northwestern Europe north of the Baltic were for a time separately connected to North America by parallel routes around the north end of the Atlantic Ocean. Geo- sistent with this reconst than one favoring an indirect connection by way of continuous land from northwestern North America to eastern Asia and then from western Asia to various parts of Europe. A continuous and habitable land con- nection from France and the British Isles to America via the Greenland-Scotland Ridge was a reality in the late Paleocene and earliest Eocene; most of the route, as far east as Iceland, may still have been intact above sea level as late as the Miocene (McKenna, 1975, 198 3) although at least one break in land continuity must have been present as early in the Eocene as some point in Ypresian time. North of the Greenland-Scotland Ridge, a second connection between Greenland and Fennoscandia may have been possible throughout much of the Eocene because of compression between Svalbard and Greenland during the Eocene (Lowell, 1972) and possibly because of the early activity of the Yermak hot- ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 spot beginning at about 40 million years ago (Ma) (Batten et al., 1981). Meanwhile, Asia was difficult to reach directly from North America or vice versa, not because — of the lack of a land connection, but because the — ancient land continuity between northeastern Si ' ri Alaska was then at very high paleo- — latitudes (Wolfson, 1949). Until the end of the ! Eocene, marine barriers existed between central | Asia and western E south of the Baltic, pre- venting terrestrial organisms of France, England, Belgium, Spain, and Portugal from reaching Asia — from the west and forcing any dry-land dispersal of such forms in tl te Pal early Eocene to take place via the Greenland-Scotland Ridge, the eastern Arctic, and the north shore of the ) Pacific by way of Alaska and eastern Siberia. Russian authors have sometimes claimed that the Turgai Strait broke down in the early Eocene and permitted Eurasian transfers to take place via the so-called Kustanai elevation (Novoé vorskaja & Janovskaja, 1977) at the oe | part of the Turgai Strait, but this connection, it existed, would not have been a route con: necting western Asia directly to western ec i south of the Baltic. Rather, a Turgat Bridge? — the Kustanai elevation in the early Eocene wou” | have connected Asia directly to Fennoscant® 1 m whence O terre E | el via | the western Europe | able to reach the rest of h pe I a Tn alanat —— g aivui'-auvuutta yrie = and proto-Iceland. Dispersal in the oppo rection over the same route might aie Re as well. In the mid-Tertiary, southwes f bs et « was apprently isolated from Europe (Jaco sited | different geography of the northern wo ee profound, although even now they are pe ail understood. No longer, however, can iw as, for instance, Darlington (1957: that fossil mammals support the notio manent continental allegiances and pr nist i oceanic barriers. Rather, patterns of en malia? i shown by early Cenozoic Holarctic n faunas are consistent with the paleog ectonió reconstructed by students of plate t (McKenna, 1975, 1983). ON THE EARLY Cenozoic PoLE POSITI er scant I have elsewhere summarized the € of the | published data bearing on the pos! of early Cenozoic rotational pole W! | 1983] TABLE l. Position of early — rotational pole itl McKenna, 19802). Lon Age tude N. tude E. Error Author 60 Ma 75.99 147.7 — Butleran Taylor, 1978 45 Ma 87° 169° 6° Simpson and ox, 1977 45 Ma 82° 170° 3" Irving, 1979 North America (McKenna, 19802). Various pre- vious Gupta yielded the results shown in Table etel these. = nip been upgraded by Harrison and Lindh (1982a: 1912, table 8; 1982b), ona a — ik conceived method of averaging and weighting data. Their results for à period covering the interval from about 80 to 20 Ma are shown in Table 2 Harrison and Lindh (19822) made a concerted effort to avoid inclusion of paleopolar data from displaced terranes of Western North America (Beck, B5 jon 1980). Their ise differ sig- nifi cerning the 4 paleopole of Briden et t (1981), but the Sins included data from the almost certainly rotated Twin Sisters dunite of the state of Washington (Beck, 1975). arrison and Lindh removed the Twin Sisters dunite data and, using the method of Briden et al., obtained results close to their own (Harrison & Lindh, 1982a: 1910). Thus, in the middie of the Eocene the averaged magnetic pole position, or in other words the Position of the Earth’s rotational axis, was rel- ob about 7? closer to the Bering Strait area ax at present. Put another way, since the me- e Eocene the lithosphere in the Bering area 5 moved away about 7? — to the mean Er of the Earth's spin a of El I5 also important to genis eu paleolatitude " esmere Island in the early Cenozoic. Eocene ssil invertebrates, vertebrates, and deciduous emis , 1981) and therefore these or- hozoic e yield inbena about the early Ce- * € environment there. At the time or times hs € early Cenozoic when these organisms were € = € on Ellesmere Island, the world's in, 1 97. re quite warm (Savin et al., 1975; Sav- mes Buchardt, 1978; Shackleton & Boers- 981 1) and poleward transport of heat by McKENNA —HOLARCTIC LANDMASS REARRANGEMENT 463 TABE 2. Position: of early Semion rotational pole son & Lindh 1982a, "1982b). Stu- Lati- Longi- Median Age dies tude N tude E Error 76.3 Ma 11 70.6° 195.1° 7.4° 67.9 Ma 18 76.6° 188.4° 5.6° 58.5 Ma 23 80.0? 183.8? 4.7? 48.3 Ma 26 83.1? 178.2* 3:5" 38.9 Ma 26 83.4? 165.4? 3:5" 30.4 Ma 29 84.7? POU 3.6° 19.9 Ma 25 85.9? 151:1* 3:6? oceanic currents and the atmosphere was evi- dently great. Ellesmere Island was at that time almost certainly located on the path between anada and either of two bridges that separately connected Greenland with lands to the east. The paleolatitude of terrestrial Eocene fossil verte- brate-producing sites on Ellesmere Island, whose present coordinates are here taken for conve- nience as 78.75°N, 277.25°E, can be ped for 48.3 Ma using the following formu Q + cos lat., long..)]. Paleolat. = sin~'[sin lat.,sin lat., -cos lat.,cos(long., — In this formula, s = site and p = position of former rotational pole. Such a calculation yields a medial Eocene pa- leolatitude for the terrestrial fossil-bearing sites: 75.9°, error 3.5°. This result is not significantly different from my previous paleolatitude calcu- lation for the Ellesmere sites: 77.5°, error 6° (McKenna, 1980a), nor from their present lati- tude. If the paleopole position for 48.3 Ma determined by Harrison and Lindh i is accepted, wand have lived at an angular distance of 14.1°, error 3.5°, from the spin axis, well within the Arctic Circle if the extent of the Arctic Circle was then the same as now. The path of polar relative motion between the rotational axis position and that of the Bering Strait area was approximately at a right angle to a line from the pole to the Ellesmere sites. For this reason the paleolatitude of Ellesmere Island has not changed significantly during the Cenozoic, while that of the Bering Strait area has decreased by about 7° since the medial Eocene and by a total of about 13° since the end of the Cretaceous. Thus, in the early Cenozoic, the Bering area was relatively much closer to the rotational pole than the Eurameri- 464 can connections at the north end of the Atlantic Ocean MEAN OBLIQUITY OF EARTH’S SPIN AXIS Several authors, notably Allard (1948), Wil- liams (1972, 1974), Wolfe (1977, 1978, 1980), and Xu (1980), have claimed on the basis of botanical and other indirect climatic evidence that the mean spin axis of the Earth was more nearly parallel to the axis of the Earth’s orbit in the early Cenozoic than its current mean obliqui- ty of 23?27'3". This would mean that the radius of the Arctic Circle has increased since the early Cenozoic. S tion is att i t first glance because it would have permitted nearly equal day d nights throughout the year in those times and would not have fatally disrupted photosyn- thesis. But there is a problem with the physics of this hypothesis. Mechanisms exist, of course, for moving a large part of the Earth with respect to another oppositely moving part or parts (Gold, 1955; Goldreich & Toomre, 1969; Jurdy, 1981), but a large extraterrestrial force operating over a long time would be required to affect the an- ular momentum vector Hx of the entire Earth (Fisher, 1974: 4044). Except for orbital reso- nances that are not expected to affect Earth for another couple of billion years (Ward, 1982; Harris & Ward, 1982), no such enormous force capable of affecting Earth significantly in the Ce- nozoic is known. Moreover, were it to exist, no way is known to couple it to the Earth. The law of conservation of momentum stands in the way. Collisions with large, high-velocity comets or large, relatively low velocity asteroids would, of course, produce instantaneous effects as well as long-term ones, especially at very early times in the Earth's history (Clube & Napier, 1982), but by Cenozoic times such effects would have be- come negligible. An effect large enough to shift the axis instantaneously as much as 5? (Nafziger & Dachille, 1965) would have totally disrupted the life of our planet as well as its geology. Av- eraged over a long time, essentially random ar- rivals of large extraterrestrial bodies would affect the spin axis's orientation if enough such arrivals occurred. But if a long-term gradual shift of the mean obliquity of the Earth's spin axis with re- gard to the ecliptic had in fact occurred since the end of the Mesozoic, then one would expect that in the much longer interval since the early Pre- cambrian, the mean obliquity of the Earth's axis as well as those of most other large bodies in the ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 solar system would have shifted so much that a random scatter of their mean inclination values would prevail. Such is not yet the case (Mc Kenna, 19802). Moreover, for the Earth at 850 Ma in the late Precambrian, obliquity calcu- lated from growth features of stromatolites at one site was about 26.5? (Vanyo & Awramik, 1982), close to the present value and therefore evidence for long-term stability since the late Precambri- an. True, only one site of that age has been in- vestigated in this way so far, but the matter can easily be tested further. One could also bring up the problem of how to dispose of the enormous amount of heat that would necessarily have been generated ina short time if a large extraterrestrial force was indeed applied rapidly, etc. I am thus unconvinced by all indirect argi- ments claiming modification of mean obliquity with respect to the Cenozoic ecliptic that do not deal quantitatively with the required physics 0 the process (Ward, 1982). If empirical evidence of such an event were found, then physics would have to explain it, but thus far the argum . about past climates that are in some cases them- selves suspect. de that rom the preceding arguments I conclu sonal, and that in general the North Ae nections would have been more hospitab " was the case at the north end of the Pacilic. THE GRANDE COUPURE The Grande Coupure (Stehlin, Me i con E break in western European fauna ae i torfian of some authors. See net, 1975, 1977; Cavelier, 1976; pe 1981; Hartenberger, 1983; Berggren : gocen? ss). E The western European terrestri Eocene, with many mammalian extin from els also many mammalian introducti where, mainly from Asia and southe rope after the drying-up of certain : ropean seaways and the former Tis (Obik Sea). The latter had until then 10 1909) is a strong 4 | | pr | | | 1983] i eae water barrier between Asia and Europe nearly continuously since the Jurassic Pei cai 1975; Heissig, 1979). A sharp decline in oceanic bottom water temperatures through- out the world occurred at the same time (Kennett & Shackleton, 1976; Corliss, 1979; Keigwin, 1980; Cavelier et al., 1981; Norris, 1982). Thier- stein and Berger (1978), as well as Miller and Curry (1982), have argued that this temperature of the world ocean, a sort of latter-day reincar- nation of the Gartner and Keany (1978, 1979; see also Clark & Kitchell, 1979) Arctic spill-over model in which cold, low-salinity water from the Arctic mixed with that of the World Ocean. By contrast, Kennett and Shackleton related the event to the opening of the more distant Tasman away as Australia separated from Antarctica. In western Europe the Grande Coupure is im- portant because on paleontologically parsimo- nious grounds it supports the existence of a large Scale basal Oligocene terrestrial dispersal event, mammals and those other organisms whose fos- sil record possesses high enough resolution to document its occurrence. In the marine Cenozoic rocks of the world, but especially in Tethyan Eu- rope, "rm and floral changes are marked at the ne-Oligocene boundary, although many taxa cross it with little or no modification (Beckmann et al., 1981; Cavelier et al., 1981; Hubbard & Boulter, 198 3). Seventeen western European terrestrial mam- ee genera whose last species became extinct ei e © Grande esc. between the Ludian and s abonian and the Sta mpian, sen lato), con- 1977 the following list (modified from Brunet, yaenodontidae Pte erodon McKENNA —HOLARCTIC LANDMASS ANGEMENT 465 Primates Amphidozotherium Pseudorhynchocyon Patriotheridomys At the same time, twenty previously unher- alded new occurrences in western Europe, most of which appear to have originated elsewhere, include species of the following mammalian gen- era (modified on the basis of Brunet, 1977): Perissodactyla Rhinocerotoidea Eggysodon (Ronzotherium and Cadurcotherium followed almost immediately thereafter, prior to the Ronzon Fauna Artiodactyla Entelodontidae Carnivora ricidae ossi Chiro erations nidae Myotis (but the family was in Europe earlier) Lagomorpha Asiatic form) ?Shamolagus Rodentia Aplodontidae Sciurodon Theridomyidae SAM (probably not from elsewhere, how- dw Gliri eer (probably not from elsewhere, how- ever) uridae Palaeosciurus ricetidae Eucricetodon Heterocricetodon 466 ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 Steneofiber Passing through the western European Grande Coupure with no more than species-level ad- justments were the following twenty-five lines of mammals (Brunet, 1977, with modification): Perissodactyla Palaeotherium Rhinolophi Rhinolophus s sensu lato Paraphyllop Palaeophyllophora Well along in the early Oligocene, at the end of the Sannoisian (= earliest Stampian, sensu lato), between the levels represented by the late Sannoisian Ronzon and the early Stampian Vil- lebramar faunas, another, less important event seems to characterize the western European Oli- gocene faunal sequence. Seven mammalian lines becoming extinct at this time were: Perissodactyla Pseudopalaeotheriu Plagiolophus minor t formerly long-lived species quite separate from Plagiolophus fraasi) Artiodactyla thracotheriidae P Caraiv Ciinidie Cynodictis (may have given rise to later Haplo- D c visité iiti Six mammalian immigrants appear to be the following: —€— uidae notées cpm (perhaps earlier) Anthracothe Anthracotherium eloci Lophiomeryx Hypertragulidae Bachitherium Carnivora Mustelidae Plesictis Nimravidae Nimravus (perhaps earlier) Passing through the event with little species adjustments were thirty-nine lian lines: more than mamma- Perimodariyia Amphicynodontidae Amphicynodon Nimravidae Eusmilus McKENNA—HOLARCTIC LANDMASS REARRANGEMENT 467 rop Rhinolophidae Rhinolophus, sensu lato Paraphyllophora Palaeophyllophora mballonuri Vespertiliavus lossidae Palaeosciurus Cricetidae ucricetodon Heterocricetodon Eomyidae Omegodus Castoridae Steneofiber It is interesting as well to compare western European mammalian faunal turn-over, or rath- ĉr the lack of it, at an Eocene event just before m —— Coupure. According to Brunet’s extin ) data, here slightly updated, there were few Peng or introductions in the time imme- Y preceding the Grande Coupure. Thus, be- ae the levels represented by Montmartre in A Cryptopithecus. In a similar but more detailed + Ln 1 LT TABLE3. Summary of and early Oligocene mammals. Ex- tinc- Occur- Unaf- Time Planes tions rences fected Earliest/Early 7 6 39 Stampian Grande Coupure 17 20 25 (38 Ma) Montmartre/St. 5 2 41 Capraise d'Eymet study of the Paleogene mammalian faunas of Quercy, France, Crochet et al. (1981) found no dramatic breaks until the Grande Coupure. The same progressive, rather than sudden, change ap- ars to have characterized the evolution of western European floras in the Eocene (Collinson et al., 1981). By “progressive” I do not neces- sarily wish to imply “gradual” or “smooth.” What I do wish to indicate is that nothing spectacular occurred until the Grande Coupure. These lists of western European late Eocene and early Oligocene mammals are summarized in Table 3. Note that in all three levels of discernible ern European mammalian lineages passing through the event in situ with little or no change than the number becoming extinct. Known pre- Grande Coupure, unheralded, new mammalian additions were limited in the late Eocene to Am- phidozotherium and Cryptopithecus, whose affin- ities are poorly understood but are not neces- sarily Asian or American except through remote ancestry. Of the new western European occur- rences at the time of the Grande Coupure, how- ever, nearly all had obvious Asian or American affinities or both. Exceptions or possible excep- tions are the artiodactyl Gelocus, the therido- myid rodent Sciuromys, and the glirid rodent Peridyromys. Most or all of the new occurrences at the somewhat later earliest/early Stampian event at the end of the Sannoisian in the Oli- gocene also ian and/or American affinities. These new occurrences may therefore legitimately be labeled arrivals. In the case of both the Grande Coupure and the later event at the end of the Sannoisian, the number of western European mammalian ex- tinctions Lad I t l to th be ofmam- malian arrivals in each example. The Grande 468 Coupure, therefore, can be regarded as a biolog- ical event related to the advent of a dry-land he event is similar to what happened to South America’s biota when connection to North nozoic (Webb, 1976; Marshall & Hecht, 1978; Marshall et al., 1982). In historical terms, it might be compared to the Oklahoma Land Rush of April 22, 1889, although I admit that fewer taxa were involved in the latter. All terrestrial organ- isms would have been affected, including, of course, plants, although one might suspect that many pl t d i ls were able to cross the narrowing water barrier somewhat before it dried up completely. Repenning (1967) and Ted- ford (1970) have recommended that such bio- logical invasions be made the basis of the begin- ning points of Mammalian Ages. Such a effects on various more distant marine environ- ments and on climate. Indeed, the event is cor- related by western European stratigraphers with the marine faunal turn-over between Priabonian and Stampian (sensu lato) faunas that is taken curred at the same time (Kennett & Shackleton, 1976), although in western Europe floristic evi- dence indicates that cooling of terrestrial envi- ronments had begun in the Eocene and was not particularly dramatic at the end of the Eocene (Collinson et al., 1981; but see Hubbard & Boul- ter, 1983). In the Arctic, however, the Oligocene temperature drop was more dramatic (Norris, 1982), suggesting as well that climatic zonation became more pronounced at the beginning of the Oligocene. CONNECTEDNESS OF THE BERING LAND BRIDGE AREA AND THE SPECTRE OF AN ALEUTIAN BRIDGE Although a brief stretch of shallow sea water presently interrupts land continuity between Asia and Alaska at the Bering Strait, a wide land con- nection formerly existed there (Scholl & Sains- bury, 1961), long before humans first used it in the Pleistocene. A dry land connection in pre- historic times has been a factor in anthropologi- cal dispersal theory since its original proposal by Fray Jose de Acosta in 1590 (Wilmsen, 1965). But, even before the Pleistocene, dry land had ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor 70 apparently existed in the Bering Strait area since the Mesozoic (Hopkins, 1967; Scholl et al., 1968; Churkin, 1972), with minor epicontinental transgressions occurring at the very end of the Cenozoic (Ostenso, 1968). For a time in the 1960s and 1970s late Miocene and even earlier Ceno- zoic seaways through the area were postulated (Mitchell, 1966; Durham & MacNeil, 1967), but further work has discounted such early incar- nations in the Cenozoic of the modern Bering Strait (McKenna, 1983, and references cited erein). Evidently, the utilization of this bridge by ter- restrial organisms has been controlled primarily by climate, such as warm interglacials or warm periods that extended poleward like that of the medial Miocene (Addicott, 1969) before glacia- ot E etaceous onward, th steadily retreated southward relative to the ro- tational pole. During the early Cenozoic ing area would have suffered longer winter dark- ness and a more extreme yearly climate than wid of the then more southerly placed bridges vd necting Canada to Greenland and parts yis rope. The Bering Bridge area has thus loe as a filter ti g 4 eometimes 4 whose action has always been controlled pf' marily by daylight length and climate. ms the Cenozoic, only toward the close of r p was a short Beringian water gap added m ia and een Asia twi has been ix. A possible second bridge betwee Alaska, via the Aleutian Island chain, inted | suggested by DeLong et al. (1978), he- r The latter may have jumped sou present position relative to the rest of 1976) the end of the Cretaceous (Cooper et al^ itio? or it may have had a more southerly d (Marlow & Cooper, 1983). According ~. y et al. (1978), the present Kamchatka Stra! | at the west end of the modern AJ the nated in the late Cenozoic. Unfortun4? * 1983] exact position or time of subduction of the Kula Ridge is not known, having been estimated as occurring either in the late Paleocene to early Eocene (Hayes & Pitman, 1970; Byrne, 1979) or in the mid-Cenozoic about 30 + 10 million years ago (Grow & Atwater, 1970; DeLong & Mc- Dowell, 1975). DeLong et al. (1978) opted for a time of emergence beginning about 42 Ma and ending at 15 Ma. Probably any possible conti- nuity ended before 15 Ma because the Meiji sediment tongue, directly south of Kamchatka Strait, began to form in the early Miocene (Scholl et al, 1977). Utilizing Harrison and Lindh’s (1982a) paleopole positions, I calculate that at 38.9 Ma the paleolatitude of both the eastern and western portals of the ancient Aleutian arc would have been at about 63°N, with the south- *rnmost, central part of the arc at about 58°N, provided that the arc was not farther south as suggested in one of Marlow and Cooper's (1983) models, If an Aleutian Bridge existed, it would have filtered any exchange of organisms between Alas- ka and Asia because its narrowness would have allowed few biotopes and because, to cross it, organisms would have required a distribution that included the restricted position of one of its "Indeed an Aleutian Bridge was a reality and blocked water exchange with the Pacific Ocean to the south. The Sea of Japan was briefly such à fresh water body near the end of the Miocene (Burckle & Akiba, 1978). Such a possibility in ates Basin could be tested by the drill. lv * moment, a mid-Cenozoic Aleutian Bridge ying to the south of the Bering Bridge is an in- ‘cresting but highly speculative possibility in need of further investigation. e body for a time in the mid-Cenozoic, TECTONIC FLOTSAM Eis tectonic history of the north end of the "IC Ocean and its continental borders is be- Coming known rapidly, largely through paleo- nanie studies and through application of tudies of the fates of the now-subducted Kula Plate and Kula Ridge as well as of the nearly McKENNA—HOLARCTIC LANDMASS REARRANGEMENT 469 swallowed Farallon Plate (Stone et al., 1982; Jones have been added to the coasts of the bordering continents or subducted beneath them (Alvarez et al., 1980; Irving et al., 1980; McElhinny et al., 1981; Barron etal., 1981; McWilli & Howell, 1982). The history of these collisions is largely a Mesozoic one, but chunks of Alaska and British Columbia have continued to move a thousand kilometers or more during the Cenozoic (Cowan, 1982; Marlow & Cooper, 1983) and various clockwise (Beck & Plumley, 1980; Bates et al., 1981; Globerman et al., 1982) as the result of the generally northward motion of the Pacific oceanic floor and plate consumption along the West Coast (Packer & Stone, 1972; Stone & Packer, 1977; Cox, 1980; Drake, 1982). Rota- tions have also been observed along the Pacific coast of Asia. Southwestern Alaska has been the site of tec- tonic unrest throughout the Cenozoic as various terranes have arrived from the south (Scholl et al., 1975). Essentially all of the terranes south of the Denali Fault were formed at sites many de- grees to the south of their present position and have been mashed against the North American Plate as the Kula Plate and part of the Pacific Plate moved north to destruction under North America and eastern Asia. In contrast, Paleocene volcanics just north of the Denali Fault are mag- netized with steep inclinations, indicating that they were formed in high latitudes (Plumley et al., 1982). Alaska north of the Denali Fault was in essentially its present position much earlier than terranes to the south (Sweeney, 1982). Thus, environmental information derived from Paleo- gene and Mesozoic floras south of the Denali fault system tells us not about ancient Alaska but about somewhere else. THROUGH THE TANK OF SuHEEP-DiP It is tempting to envision some of these far- traveled small blocks as Noah's arks (McKenna, 1973), each bearing a biota whose affinities lie with some distant homeland either along the western shore of North America or across vast oceanic stretches. Indeed, many of the blocks were in the southern hemisphere as late as Ju- 470 rassic times (Stone et al., 1982). However, their i in the north from an ultimate source in oceanic vol- canic arcs or mid-plate volcanic edifices that had somehow become populated from elsewhere, was in most or all cases apparently too early to have had any discernible effect on the composition of the same place. Finally, most of the blocks seem to rest on oceanic basement (Ben-Avraham et al., 1981; Jones et al., 1982) and thus those that do were never initially part of any continent. Possibly the biological effects of passive trans- transoceanic transfer on the fragments that have ultimately come to rest in Alaska, Siberia, Japan, or British Columbia and the western coast of the United States. Such an analysis should proceed according to the method set forth by Rosen (1978) and by Nelson and Platnick (1978, 1981). But, as I have elsewhere (McKenna, 1981) pointed out, such blocks, if embedded in oceanic crust, from a generating ridge. The amount of Jools sinking is generally proportional to the square root of the time involved since the site was gen- erated at the ridge crest (Sclater & Parsons, 1981), but for most “normal” ocean floor the depth with respect to time since plate generation is perhaps better represented by the formula: d = 6,400 — 3,200e-vezs where d is the depth in meters and t is time in millions of years. This formula, developed by Parsons and Sclater (1977), has the advantage of yielding results that are asymptotic to a limiting water depth of 6,400 meters. After 70 million years of lateral motion, rocks once at the mid- oceanic generating ridge subside to within about 1,000 meters of such a depth. Small blocks of continental crust, aseismic ridges, guyots, or anomalous crust such as that of the present-day Ontong Java or Manihiki plateaus (Heezen et al., 1965; Nur & Ben-Avraham, 1982; Rogers, 1982) would rise higher than “normal” oceanic floor and thus would require the use of different con- stants in the equation. However, if such small blocks are to be invoked as Noah's arks, it must be shown how their surfaces could have re- mained continuously above sea level for the en- ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo.. 70 tire journey in order to avoid drowning their terrestrial inhabitants. Even photic zone inhab- itants, for instance corals, would soon perish if in the course of their passive lateral transport they were carried to greater than shallow depths. At present, blocks like the Ontong Java, Mani- iki and many other oceanic plateaus are, after all, either mostly or totally far below sea level. 2 NOAH'S ARKS Large blocks of sialic crust are another matter: we call them continents! India's northward flight one example of a major Noah's ark transier was consummated some time either in the late Cretaceous or early Cenozoic prior to the medial Eocene (Maluski et al., 1982). The vicariant 1n- habitants of India that had resulted from it5 de- parture from other parts of Gondwanaland were largely swamped by dispersal on India’s contact with Asia. Patterns of endemism that remain 1 un Bingham, 1980; Boulin, 1981; Gupta & D 1981). Australia's and southern OS a . Tar- another example (Raven & Axelrod, 1975; A a ling, 1980; McKenna, 1980b; Vink, 1982) Oat | ES rt dis- smaller scale, the many jostlings over sho southernmost Spain, Italy, Crete, aD * how various stages of transfer from ra complete (Adrover & Hugueney, oe vs fiel, 1976; Leinders & Meulenkamp, 17% denberg, 1979; Cohen, 1980). [d EFFECTS OF TIME ON NOAH $ ARKS ^ traveled A problem with invoking small, = b tectonic blocks to explain Meets en : e i that ities or apparently embedded emics / 1983] such travel in oceanic isolation takes a long time, during which significant morphologic changes may have occurred in the passively transported biota. If one accepts that a special taxonomic similarity between transoceanic sister species, genera, or families indicates former continuity of two widely separated occurrences, and one finds that fragments of a once continuous land- mass upon which the ancestors of these taxa once existed are now allochthonously embedded in distant continents (Nelson, 1981), then any such taxon must independently be expected on gen- erally parsimonious grounds to have existed at the time ofthe original fragmentation. In the case of "Pacifica" or in arguments about differential Earth expansion, this time can amount to more than a hundred million years, depending on the rate of ocean floor spreading and the distances QST T travelled. Although it is generally gr the fossil record is imperfect, it is not likely that the fossil record is perverse, as some cladists would apparently have it. Thus the chances that, Say, 15 discerptible sequential character states of horse evolution lasting a million years apiece actually took place in the same sequence but 15 million years earlier in each case, without leaving lation of “Pacifica” as is adumbrated by Nelson and Platnick (1981: 542)! Parsimony is a factor In all science, not just in the preparation of clado- rams. With regard to vertebrates, I would think that only rather high-rank taxa, such as a class Or subclass, might be involved (if at all) in the Pacifica" or differential expansion scenarios. But know of no convincing evidence that any ter- Testrial vertebrate lineage has been brought alive t0 the shores of North America from the other ec of the Pacific on a fragment of “Pacifica.” This is not to say that far-traveled and once sub- Merged tectonic blocks, once they become imbedded in a new land, cannot be further trans- Ens along the edges of continents by trans- Meis faulting, etc., carrying ancient fossils and “i y acquired terrestrial biotas along with them new destinations. PLANETARY POPCORN? That the Earth has expanded is an idea that can be traced with certainty back to the works McKENNA —HOLARCTIC LANDMASS REARRANGEMENT 471 of W. L. Green in the nineteenth century (Jor- dan, 1966, 1971; Carozzi, 1970; Carey, 1976). Modern metaphysical underpinnings for a form peals to unknown mechanisms should be enter- tained only when all else fails, as was the case in the early days of continental drift theory. Dirac (1937, 1938, 1974) concluded induc- tively that there must be some connection be- tween the gravitational “constant” and time. He, and other students since then, concluded on the basis of his large numbers hypothesis" that the gravitational *constant" must be i 2 billion years. Current estimates of the age of the Universe are up to an order of magnitude greater and vary, depending on differing deter- minations of the Hubble constant, from at least about 10 to as much as 18 billion years (de Vau- couleurs, 1982). This gives the postulated ex- pansion process a longer time in which to operate than Dirac originally contemplated, drastically reducing the Phanerozoic component of what- ever total expansion might have occurred. That the Earth itself should have participated in Hub- ble expansion seems reasonable enough. There- fore, since mid-Paleozoic time, say 350 M a maximum of about three percent of the Earth's expansion would have taken place. This amount is smaller than the figure given by Carey (1975, 1976), partly because his value for the age of the Universe was only about a third or a quarter of that now accepted and partly because Carey be- lieved that the expansion accelerated exponen- tially with time. Owen (1976, 1981, 1983), on supposed empirical g ds that do not take into account the effects of propagating rifts, believes that the Earth's diameter has expanded by 20 percent of its modern diameter since 180-200 Ma. Steiner (1977) also believed in rapid ex- pansion, concluding that the Earth has expanded about 11 percent since Jurassic time, an argu- ment that precipitated a series of outcries and explanations in the June, 1978 issue of Geology (e.g., Kaula, 1978). Vink (1982) and Courtillot (1982) provided models of continental rifting that effectively ex- plain the spherical triangular apparent gores that 472 are the raison d’etre of attempts to match coasts or continental slopes on an Earth of smaller ra- dius than the present one. As a rift begins to invade a continent, ; breaking it spart into two continents, t dilate and also bend down as the rift propagates. Hence the apparent gores. Actually, what is required is not an explana- tion of Earth expansion. Expansion may have occurred, but, if the whole Earth expanded uni- formly, then there would be no crustal separa- tions to explain and paleomagnetic inclinations would not be expected to change. However, the Earth is not like a balloon in which the contained matter expands and the balloon itself does not. What a balloon does is to deform, thinning to the point of breakage. If a balloon truly expand- ed, its rubber would thicken instead of thinning as its area increased. If paleomagnetic studies were to be conducted on such a deformed, bal- loon-like crust, they would show modified field inclination vectors for one of two reasons. Either general crustal thinning would do the job or de- crease in curvature would provide a similar re- sult in case the continental fragments did not expand, more or less analogous to the fate of pieces of paper attached to the surface of the balloon. What is needed is an explanation of why, if general expansion has occurred and continents sar separated, continental crust should not also ave participated in the expansion process af- Feed. the whole Earth, thickening as well as increasing in area. Non-expansion of part of or of the continental crust thus is an ad hoc assumption in most scenarios. I have seen little quantitative discussion of this in the expansion- ist literature except that of McVittie (1969), who came to no conclusion other than that different parts of the Earth and other celestial bodies might possibly expand at differing rates at different depths. Usually, the problem is not mentioned at all. However, if one gives proponents of substan- tial differential expansion the benefit of the doubt and assumes that only the Earth's core and si- matic crust expanded and, as a further ad hoc assumption, that all expansion occurred between the eastern and Western shores of the Pacific, then a maximum of about 1,200 km of separa- tion there due to differential expansion would have occurred since the mid-Pale eozoic, when marine faunal similarities between Australia and North America began to drop sharply (Gill, 1958). That differential expansion was, if it occurred, ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 less than 1,200 km or was offset by other - is indicated by the calculations of McE al. (1978), who concluded on the basis of ee magnetic evidence from presumably non-ex- panding continents that the Earth has had a di- ameter within one percent of its present diameter since at least the Devonian. McElhinny and his co-workers also concluded that Mad Mell and the Moon have not expanded significantly since Precambrian time. Golombek and McGill (1983) obtain similar results for the Moon. In any case, on Earth the amount of allowed separation due to differential expansion alone would be a small fraction of the present width of the Pacific Ocean. Thus we may conclude on quantitative grounds that differential expansion, even if it were true, would not help much h in the field of paleobiogeography except to increase the degree of some ancient separations by a small amount. Far more important is the fact that the floor of the Pacific is decoupled from its SU rounding continents and the size of the Pacific Ocean itself has been shrinking, not oe created is eventually subducted. An area ocean floor equivalent to that of the present Pa- cific Basin has been subducted since early Cre- taceous time alone (Larson & Pitman, 1972), and there is ample evidence that the process was e going long before that. Such processes have “ recycling sediments derived from the continent into the mantle since Precambrian time Paolo, 1983). m Recently, however, Nelson and some T: students of vicariance biogeography have impressed b non-quantitatively asses a : models, evidently 1 » simply removing the Pacific i Bain Earth.on. "hid. ihe continenté somehow ei ssent y the the same area as now, covering © cially J whole surface of the Earth. This superfi tractive solution does not seat epi a | several questions: 1) where and h ocean water of the world was pei D is to be done with abundant g — 1983] that an ocean has existed in the Pacific area, for instance off North America’s west coast, since the early Paleozoic or Precambrian (Roberts, 1972; Cook & Taylor, 1975; Badham, 1978, 1979; Hoffman, 1979), 3) how 304-311 million year old ophiolites in California (Saleeby, 1982) or Ordovician ophiolites further north (Churkin & Eberlein, 1977) are to be explained, or 4) how evidence of pre-Pangaean Wilson cycles like that of the Iapetus Ocean are to be discounted (Glik- son, 1972)? Such an Earth would necessarily have expanded differentially at a much greater rate than is required by Dirac’s inductive insights, rn. If pansion envisioned would clearly be evidence that the Universe is much younger than the cur- rent consensus would have it. Thus, one might conclude that, if proponents of accelerating ex- pansion are correct, all physics should adjust it- self in some major way to their requirements (Kaula, 1978; Steiner, 1978). But, perhaps supporters of the supposed bio- &ographical explanatory power of the differen- tial expanding Earth hypothesis have erred in not going far enough. Generalized tracks (Croizat et al., 1974) are drawn between landmasses in such of the model, continents are gratuitously as- sumed to have remained of fixed area and thick- ness. If students of transoceanic Pacific vicari- ance are willing to stretch generalized tracks so that they fall on the Earth’s surface rather than iis the Earth, why are they unwilling to stretch € tracks so that they go around the Pacific rath- tr than through it? ; Against a backdrop of general parsimony uud I cannot take the accelerated expanding Earth model seriously as a source of discernible late Phanerozoic biogeographical pattern. For such Purposes, either linear or accelerated expansion has been investigated and has been found want- ng. Those who disagree should seek a flaw in the arguments presented by Van Andel and Hos- ri (1969), McElhinny (1973), McElhinny et al. (NA and Golombek and McGill (1983). Carey 75) has done so with regard to the arguments McKENNA — HOLARCTIC LANDMASS REARRANGEMENT 473 of Van Andel and Hospers, but, insofar as I know, McElhinny's work has not yet been seriously challenged, even by Owen (1981, 1983). IMPACTS OF APOLLO-CLASS ASTEROIDS AND COMETS Recently, the possible effects of impacts with the Earth of Apollo-class asteroids and comets have been much discussed (e.g., Napier & Clube, 1979; Emiliani, 1980; Alvarez, 1982; Alvarez et calcareous marine microplankton, there is as yet Apollo-class bolide on large terrestrial organ- isms, such as the last surviving dinosaurs, was apparently nearly synchronous at the power of resolution studied (Russell & Singh, 1978; Rus- sell, 1979; Hickey, 1981; Clemens et al., 1981; Archibald & Clemens, 1982; Archibald et al., 1982; and, especially, Alvarez, 1982, for statis- tical analysis). Later, supposedly at 34.6 + 4.2 Ma (Glass & Crosbie, 1982), another major impact, the cause of the North American tectite strewn field, has been widely discussed. But at 35 Ma there was no especially pronounced terrestrial vertebrate faunal turn-over. Prothero (1982) has noted a minor vertebrate faunal turn-over at 32.4 Ma, but this followed a period of stasis after a major faunal turn-over in the interval from 40 to 37 Ma Keller et al. (1983) give convincing evi- dence that the North American tectite strewn field occurred at 37.5-38 Ma in the latest Eocene, not later as suggested by fission track age eterminations or those based on glauconite. Keller et al. (1983) also found evidence of at least a e wo othert Oligocene boundary, but no mass extinction was associated with any of them. 4 4 at about 35 Ma with the Eocene-Oligocene boundary (Glass & Crosbie, 1982) or presumed equivalent boundaries between marine inverte- brate or North American mammalian biostrati- hi its h imply i d abundant high temperature K-Ar dates from volcanic rocks in- terlayered with fossil bearing sediments whose paleomagnetic stratigraphy can also be studied. These dates calibrate Uintan-Duchesnean- Chadronian faunal and floral turn-over (Leopold & MacGinitie, 1972; Wilson, 1980) in North 474 America as occurring close to 38 Ma (Berggren et al., 1978) or 36.6 Ma (Berggren et al., 1982; in press). Prothero et al. (1982) come to similar conclusions. Nor could an impact at a later date explain the flood of Asian and/or American ar- rivals that characterized the Grande Coupure in Europe. In Europe the Grande Coupure has been dated at 37-38 Ma (Gramann et al., 1975; Van Couvering et al., 1981). Iridium-rich layers incorporating undifferentiated Solar System matter are indeed associated with some (but not all) large Cenozoic impacts, but if one hypoth- esizes that large-scale effects on terrestrial biotas have occurred through the agency of a particular bolide, one should provide clear evidence of si- multaneity. This has not yet been produced. Me- teoric dust has of course long been known to be a component of the Earth’s sediments (e.g., Skol- nick, 1959). Moreover, more biol gical evidence than the extinction of five species of radiolarians (Glass & Crosbie, 1982; Ganapathy, 1982) needs o e 8 Q A A i 8 [^7] S Ó 2 o2 vr 3 pact that occurred in close association with the Cretaceous-Tertiary boundary (Emiliani et al., 1981), so it is unclear whether a causal relation exists. Microtectites and iridium anomalies should be sought tł gl he geol gi ] 1 before those few that are known now are specif- ically assigned far-reaching importance as evi- dence of extraterrestrial impacts of sufficient magnitude to have served as forcing factors in the biological world. Already, there appear to be 1 Aidat +} : 1 1 . . x uf EE terrestrial impact craters (Hughes, 1979) to be the sources of at least three different Late Eocene EFFECTS OF RADIOISOTOPIC CALIBRATION ON TEMPORAL DISTRIBUTION Most of this essay has been concerned with spatial distribution, but recent discussions of rates of extinction and of faunal turn-over (Flessa, 1975; Raup & Marshall, 1980; and other articles on the same general subject in the journal Pa- leobiology) prompt a few brief comments about the obvious but nevertheless seem necessary. A ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 first point is that compilations should be based on specific lineages, not on higher taxa. Tech- nically, it is not genera, families, orders, etc., that originate, evolve, or become extinct. For ex- ample, it is the perishing of the last species of a higher taxon that causes the higher taxon to per- ish. Discussions of the sudden extinctions of higher taxa may well mask a series of more grad- ual extinctions of various genetically isolated lin- eages within any particular higher taxon. Studies of extinction, to be convincing, need to discuss the relative abundance and presence or absence of taxa at the species level and to relate this in- formation to measured stratigraphic sections studied in the field rather than to derive their basic information from popular texts such as Ro- mer’s (1966) Vertebrate Paleontology. A related point has to do with rates of faunal change. For these to be calculated, some form of temporal calibration is required, of which the most effective one is a time scale measured in years. Such a calibrated time scale is not yet available for much of the fossil record, at least not at the power of resolution necessary for de- tailed study. What generally happens instead 1s that biostratigraphically-based time div VUE 1800! within a p nterval are treate x (Fig. 1). The abundance of taxa within these pe is assumed to be meaningful, but instead the " versity of a longer interval could be caused mixing the diversity of two or more unrecog? " levels within the unit. This has been demone ed, for example, in the case of the long No s American Wasatchian and Uintan Land e mal ages relative to the much shorter Bridg e Land Mammal Age (Berggren et al., 1978). bie apparently high diversity of some units g aliquot status in biostratigraphy may ee ý simply by their long duration and there er their temporally composite sampling jer " ing kaleidoscope of life (Raup & Marshall, (Fig. 2). i HIC EXTENSIONS OF SIMPSON’S BIOGEOGRAP PRINCIPLES 3, Simpson (1940a, 1940b, 1943, 1952, 195: 1965) developed the notions of p "2 corridors, filters, an ( of 1975). Basically, these constitute a spei permissiveness with regard to diserta m sess statistical signatures when $a m me geological record. Provided the reco pa quate, one should in principle be able to à 1983] C Hansa C onc PR B B E deme EE. A: A A 1.5 T" . deos ades to be given ae status. Although their temporal sequence is known, the temporal duration of E A, B, and C is not at first known but is later ibrated. ni sh Ze \ t tack € each ther bya a continuously Hobo e substrate, when the Substrate was partially a barrier, and when the barrier was more complete as in the case of ran- dom over-water transport of organisms to is- lands. Assuming the imperfect geological record of organisms to stop short of perversity, one can un me the Extinct biotas of various ; landmasses biological resemblances whether certain land an ’s biotas. Simpson t cue his arguments on the stable-con- tinent rationale and most of his examples are Wn from fossil mammals. However, his prin- ciples are adaptable to a mobilistic world and are still valuable in spite of criticism of past ex- cesses by Du Toit (1944), and in spite of simple avoidance of discussion of them by most vicar- iance biogeographers. There is no need to repeat all of Simpson's arguments here, but suffice it to say that terres- trial corridors are defined when biological resem- blances between two distant samples are ap- Proximately the same as would be found at not be long delayed. Ecologically balanced biotas Would be involved. At the other end of the spec- trum, sweepstakes dispersal would be character- 2 *d by random arrivals s, one-way transfers, eco- oin imbalance, and delayed introductions. In ween would be various types of filters that erbe some but not all organisms to pass. PSon discussed this spectrum as though its ja niponents represented separate processes, but id mobile world it is especially obvious, for nce, that whole oceans were once inches, McKENNA — HOLARCTIC LANDMASS REARRANGEMENT 475 Li ias E2. Before biostratigraphic subdivision, the biologi koc of a sequence of rocks is merely known n the sequence (left). Upon biostrati- dia subdivision into aliquot parts, the diversity of mir be found to differ (center), leading to estimates of dd E Wh ho | "i cr 1.42 is provided by vean calibration M CERD, the "ow he diversity of units of long dura y turn out t be better interpreted as the res si err Satine of se- oq diversities that. are themselves relatively con- stan y processes d may turn out to have been in error. then feet, then miles wide. At certain points in this progression the statistical wide re) Simpson's corridors, filters, and sweepstakes would each have helped to € < evolution- ary history of the biotas involved. Because of the stabilist iode implicit in Simpson's earlier work, Simpson believed one- way transfers of balanced biotas to be unlikely. To be fair, it should be noted De Viii did toy with the idea of one-way or “revolvi transfer by means of cedi. Miet and then subsided blocks, thus avoiding la horizontal displacements such as those contem- plated in present geological thought. However, under a mobile rationale such transfers are pos- sible passively, as the result of motions of parts of landmasses to join others. As have many oth- ers, I have elsewhere (McKenna, 1975) called the substrates on which such one-way transfers take h’s arks." A related concept dealing with fossils is that of “grounded Viking funeral ps" (McKenna, 1975). The idea here is that fossils, as well as living organisms, can be carried been demonstrated for Nort ., 1978). The difference is that host landmass do not stray from their resting places to infect the rest of the landmass's biota. Like transferred fender paint in an automobile collision, they reflect only former allegiances. 476 Such transfers are to be expected as the result of multiple Wilson cycles that open and close oceans and, on a smaller scale, along large transcurrent faults that cross water barriers. VOYAGES TO NOWHERE AND RETURN Another feature of mobilistic tectonics that aah Pe LL) 1 ; Me 2 Z" a. a. I y ot the theme that plate tectonics supplies mechanisms un- dreamed of in stabilists’ philosophy is the phe- nomenon of back-arc spreading. This results in the separation of an intracontinental volcanic chain, such as the Andes or Cascades, to become an offshore arc that may later be driven back against the same continent from which it origi- nally separated. Japan is undergoing this process (Uyeda & Miyashiro, 1974), and a similar pro- cess affected the southern Andes in the past (Dal- ziel et al., 1974; Dalziel & Palmer, 1979). Such separations occur because the cooling of sinking slabs of oceanic crust sometimes results in sea- ward jumps of subduction zones. This is because, in effect, the whole slab sometimes pulls away from the surf. ta point closer to its generating ridge than the point of its earlier subduction. The der an advancing continent would result in the arc's rejoining the continent unless another sea- ward jump of the subduction zone occurred in time (Uyeda, 1981). During the time that such an island arc was in isolation, both vicariance and dispersal followed by differentiation would produce peripheral isolates in abundance. Then these would be merged with the biota of the mainland. Much diversity could be created in this manner, but there could bea substantial time delay before re-introduction to the mainland. Voyages to nowhere and return are not the result of simple intermittent barrier erection and destruction such as might be envisioned in the case of coastal islands created by eustatic rises of sea level, followed by a fall. They involve I I phy ofa bio- ta to a significant distance ona moving substrate, followed by return and passive merger. ESCALATOR COUNTERFLOW ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 opposite direction, the child remains stationary with regard to the escalator’s foundations. What does this have to do with biogeography? Con- sider Iceland. Iceland lies at the north end of the Atlantic Ocean, astride the Mid-Atlantic Ridge. It isa volcanic edifice built above a hotspot that has produced abnormally large volumes of lavas ep- isodically and variably throughout most of the Cenozoic (Schilling et al., 1982). The oldest rocks on Iceland are known from coastal areas in eastern and northwestern parts of the island, where they are about 16 million years old (Moor bath et al., 1968). Fossil plants and insects are own from late Tertiary volcaniclastics in northwestern Iceland, but as yet no fossil man mals have come to light. Geologically, Iceland is a steady-state phenomenon, in that its spreads from the center above the hotspot and sinks below sea level after about 15 or 20 m years of lateral transport. Thus the main m aerial part of the Icelandic lava pile has cov more or less the same area of crust for oo time (Vogt, 1972). Early in the Cenozoic, " Iceland rose up to choke the whole North 4 lantic for a time, creating a land bridge p Greenland to Scotland. Later, as the North lantic widened, Iceland came to be y then wholly isolated. : Apparently, Iceland retained its conne with Greenland and thence to North ported species of Magnolia, glans, Betula, Acer, Abies, Ulmus, species-level relatives at the time were ch & S monarson, 1981). It is not yet kn rganisms first arrived in Icelan ey however, after the glacial events o d. At present pleisto- ing Eurasian and North American with the cause the Icelandic hotspot 1s unn ally plate boundary and because p to the volumes of volcanic products are b 1983] surface, Iceland is subaerial rather than just another part of the Mid-Atlantic Ridge. About 20 million years after any particular batch of new crust is created at the plate boundary over the hotspot, the crust will have been moved aside because of the creation of still newer crust at the boundary, eroding at the surface and cooling and subsiding to sea level and below as it spreads at arate of about one centimeter per year. After sinking beneath the waves, 20 million year old or older Icelandic volcanics on either side of Ice- land simply become parts of the submarine Greenland-Scotland Ridge. _ The land surface of Iceland at any particular time thus comprises two **down" escalators ar- ranged back to back, whose oppositely directed lateral components of motion start at the plate boundary in the middle and continue to the ocean on opposite coasts. To have escaped drowning, the terrestrial biota of Iceland merely needs to have moved over the land surface toward Ice- land's central volcanic zone at a rate of about one centimeter per year for as long as the hotspot has maintained Iceland as an island above sea level (F Ig. 3). Subaerial Iceland began to form at about the time of marine magnetic anomaly 24, at about 54 Ma, and appears to have remained con- hected to Greenland until some time in the Mio- cene. It might therefore be expected that Iceland SPersals, earlier than the oldest rock now pres- ent above sea level in Iceland (McKenna, 1983). Twenty million years from now, after Iceland's Present rocks have moved beneath the Atlantic, some of the members of the present Icelandic biota may still be there on Iceland's newest crust, slowly climbing up “down” escalators. Thus iż: Peer Oars E o as oe adv. State j kage iae er ai : tate islands lying over volanic spreading centers Can hawe Ks: PLA :onificantlv A than the oldest rocks now present either at e island's surface or located directly under- McKENNA —HOLARCTIC LANDMASS REARRANGEMENT 477 world. Obviously, taxa in passive isolation on evolving hotspots must be as old as the hotspot's isolation or must have arrived by dispersal at some later date. HOPSCOTCH ON THE ESCALATOR A potential mode of distribution that is in some ways intermediate between vicariance and dis- persal is apparently represented well by the Hawaiian Islands and perhaps by other mid-plate hotspot edifices (Axelrod, 1972). Vicariance, it will be remembered, results when a barrier arises after a distribution of organisms has been established, breaking it apart into iso- lated sister taxa (Croizat et al., 1974: 277-278; Nelson & Platnick, 1981). One way, but certainly not the only way, to cause vicariance is to pull apart the passive riders on two differentially moving tectonic plates. Dispersal, on the other hand, crosses pre-existing barriers to establish a new, expanded distribution, but processes such as the founder effect and difficulties of passin through the same barrier in the opposite direc- tion (if, say, oceanic currents involved) all differentiation of sister taxa after dispersal oc- urs. Further, dispersal is an active process whereas vicariance brought about by tectonic ti d t i tabove [e] the substrate on the part ofthe participants. Both processes may be responsible for the origin of biotas, particularly those of islands. The Hawaiian Islands are collectively a long term steady state phenomenon, like Iceland, but, in contrast to the latter, the Hawaiian chain is not located on a mid-oceanic ridge nor has it recently been connected to any continental mass (Jackson et al., 1972; Dalrymple et al., 1973; Dalrymple et al., 1981). Both places are the result of subaerially produced hotspot lavas; but in the case of the Hawaiian chain, the Pacific Plate, which currently underlies the chain, is moving with respect to the hotspot and thus carries away the volcanic piles generated by the hotspot. The hotspot itself has moved relatively little, if at all, with respect to Earth's spin axis since Cretaceous time (Hammond et al., 1979; Suarez & Molnar, 1980; Gordon & Cape, 1981). At present the sea floor of the Pacific Plate in the Hawaiian area is moving west-northwestward over the Hawaiian hotspot, passing it by (Wilson, 1963) and car- rying formerly tropical islands northwestward into cooler waters (Greene et al., 1978). How- 478 ANNALS OF THE MISSOURI BOTANICAL GARDEN t and spreads in RE 3. ocean ridge volcanic i Icanics as they nd such as Iceland. A: New crust rises from the underlying hotspo 5 8 = oO B an oO T En [e] Ò la} Nn O g au c 5 da S A - = a4 i - o e < o — & v 2 5 © "$ = = P Escalator Counterflow. This cartoon attempts to show two stages in the geologic history ofa mid- l but the biota could still be present, evolving in isolation and being ed by extinction as well by occasional arrivals from elsewhere. Thus, it is possible that some island biotas either have 0 dicpersal | y ary iana and by the E ever, the rising magma column is not constantly beneath a weak plate boundary as is the case with Iceland. Rather, the column must bend as the plate passes overhead, occasionally breaking and reestablishing a more direct route to the surface. Every so often magma breaks through the plate from below to form new volcanoes above, but these form generally at the southeast end of the Hawaiian chain, first as submarine volcanoes (Emery, 1955) and then massive subaerially pro- jecting structures as they grow in size. Indeed, one is forming beneath the surface of the ocean hoff et al., 1982). Some of the individual vol- canoes in the Hawaiian chain reach a height of 9,000 meters above the sea floor, projecting up to about 5,000 meters above sea level. Mean- while, the crust to the west-northwest cools as it departs the vicinity of the hotspot. Gradually it sinks and the older volcanic piles lying on it erode (Fig. 4). Eventually they sink beneath sea level (Vogt & Ostenso, 1967). The oldest islands of the Hawaiian chain are therefore located in the A 4 west-northwest, as J. D. Dana realized as eat as 1849 and as was finally demonstrated Ke tassium-argon dating more than a hundred "is later (McDougall, 1964). Several voi in the chain may have been present on the an ! Plate before their sites passed over the pec | however. Wentworth Seamount and od land have yielded Cretaceous K-Ar ages ( sed the 1975). Rotondo et al. (1981) have disc ese possible biological consequences of ad edifices to the Hawaiian chain. | The sporadic nature of the hotspot Men | results in gaps between the islands, n tat f not very great in most case ; s. Islands * the volcanic PF e f nlarge. The combined Hawaiian-Empe" (Bat contains 107 or more recognized volcan gar & Jackson, 1974). $ § jon this for a very long time. About 43 milli Plate ago the present WNW motion or fof began (Dalrymple et al., 1977; age ^... con , currently accepted decay and pie motion stants). During the early Cenozoic 1983] McKENNA—HOLARCTIC LANDMASS REARRANGEMENT 479 FiGURE 4. Hopscotch on the Escalator. Based on the model provided by the Hawaiian Islands, it is possible to envision a situation in which new islands arise at one end of an is. d to the ongoing sequence of islands could have oc land chain while others sink at the other Thi among inhabitants might be present that had never had to cross more than minor water barriers in order to be present On at least one island at any particular time. If the hotspot were old enough, it would probably once have been t trial 1 1 ight I ived across barriers | 1 1 tk those beneath a continent. Thus some usually postulated in the biogeographical literature. the Western part of what is now the Pacific Plate (Izanagi Plate of Woods & Davies, 1982) was More nearly northward because of the action of the Emperor Spreading System, but prior to about 67 Ma. Pacifi Ps aati (Farrar & Dixon, 1981; but see : ordon, 1982). The trace of the Hawaiian hot- i on the present-day Pacific Plate turns north- 3%), Yuryaku Seamount (Long. 172°E, Lat. id , where it continues as the Emperor Sea- r unt chain (Morgan, 1972). Basalt dredged from ca aku Seamount has an age of about 43 Ma id gue et al., 1975; age corrected). One of the St southerly of the drowned volcanoes of the bee (Clague & Dalrymple, 1973; age corrected) a may be slightly older than that on the basis fro Pi mperor Seamount chain extends as Finis. of drowned volcanoes as far as the pres- a edge of the Pacific Plate, where Cretaceous hs ers of Hawaii, now mere guyots, are being Sumed in a cusp between the two curved Oceanic subduction zones at the Aleutian and D i. mia i 4 & ol. J 1 a T BER ofthe p y p chain is about to be subducted there. The oldest fossils from sediments on Meiji Seamount (DSDP Site 192) are at least 70 million years old (Scholl et al., 1971) and are of near tropical organisms. By Oli- gocene time Meiji had moved north far enough o encounter subarctic conditions (Scholl & Creager, 1973). How long before the beginning of the Cenozoic “Hawaiis” older than Meiji Sea- mount were being created, transported, subsid- ed, and eventually consumed is not known be- cause of tl to obliterate direct evidence. Nevertheless, a large number of previous Hawaiian-Emperor Islands may have been subducted because the consum- ing arc systems have apparently become choked long enough to form the cusp between the arcs. The Hawaiian hotspot and its fossil trail of mid- oceanic lava piles could even have already been in existence before the beginning of its presence under the Pacific Plate. It is conceivable that truly ancient versions of Hawaii could have formed in contact with or even on top of a con- tinent that happened to be passing over the hot- spot 480 For this reason, it is possible to envision a the chain as terrestrial organisms were forced to play the children’s game of hopscotch in order to survive. As new islands formed at one end of the chain, others were sinking at the opposite nd. Thus, as with Iceland, the terrestrial fauna range over-water transport but rather at least in part from an early vicariant event followed by a series of short dispersals (Axelrod, 1960, 1972). The average distance between volcanoes along the Hawaiian chain is about 56 km; in the Em- peror chain the average is about 77 km. Such distances would not be difficult for many insects or flowering plants to cross, as for instance ap- pears to be the case in the Solomon Islands (Whitmore, 1973) or near Puerto Rico (Heatwole & Levins, 1972) at present. Although the rocks of any particular island still above sea level in the present-day Hawaiian chain are very young (Zimmerman, 1948), the ances- tors of some of the present inhabitants of the island may have been living in precarious mid- oceanic isolation for more than a hundred mil- lion years. Authors steeped in a stable continent rationale, like Zimmerman (1948), Carlquist (1974, 1982), or Carson (1982) were unable to appreciate this mobilist notion. This is because the Hawaiian chain, like Iceland, is the result of steady-state phenomena. As new Hawaiis are created, old ones sink beneath the waves to the northwest of the string of lava piles created by the Hawaiian hotspot on the Pacific Plate as the latter moves northwestward over the hotspot. Thus, arguments for rapid evolution of Hawaiian patie oe (e.g., Carson, 1976; Kaneshiro & 982) may in some cases overestimate Bikini in the Marshalls (Ladd, 1965) may indeed be real and permitted by the hotspot history of the areas. Even dispersals to the islands from far away could have been taking place over a much longer time span than that represented by the oldest presently exposed rocks of the Hawaiian Islands, thus helping to explain the differentia- tion of birds like the Hawaiian finches (Hawaiian honeycreepers), various flightless geese, rails, and an ibis, three species of an extinct genus of owls, ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 and still other endemic birds that occur in fossil deposits of Kauai, Oahu, Molokai, and Maui (Ol- son & James, 1982). Obviously, the longer the winnowing process went on, the greater would be the probability of extinction of any original vicariant lineages and the more erratic in taxonomic composition would be the remening sample i ong terresttil Í houpat idend ph ‘the Pacifi Ocean alone there | are many potential examples of the process, such as the Tuamotus and their symmetrically rela i Nazca Ridge counterpart, the Austral-Cook chain, the Easter-Sala y Gomez chain, the Juan Fer- nandez seamount chain and Juan Fernandez it- self, the Galapagos-Carnegie Ridge area and Co- cos Ridge. However, the Line Islands may no! have been generated by a hotspot of the Hawai- ian-Emperor type (Haggerty et al., 1982 argued here, however, that on geological grounds itis by no means mandatory to conclude that all inhabitants of presently isolated hotspot n groups arrived by relatively recent long-distance dispersal over water or necessarily had an cestors that did so. Rather, a series of short dispersal carried out over a long time, may explain. existence of some of the taxa present. of terrestrial immigrants and arriving products " over-water dispersal alike would be pon players of hopscotch on their moving subst ver à ia the moment they began residence 0 ET z pot. basspie in the case of some island groups. i, theory of island biogeography as acArthur and Wilson (1967)isin need of ification. Olson and James (1982) have ( tbe noted this with regard to the avifauna i tion Hawaiian chain because of the recent realiza that the extinct endemic avifauna was ™ ect of twice as diverse as the present one. Thee a winnowing action over a long pe sults as are generally attributed to the spo agencies of long-distance dispersal. CLOSING NOTE OF CAUTION me geologi n some e bio- e. I hav? In this paper I have tried to list so processes that might help to expla logical distributions in space and ti tried also to review some extraterres d cosmological causes that might influe distributions either directly OT indi result, I have been led to discoUn unt ). Ihave . developed by > ) 1983] mean spin axis stability change, the ‘Pacifica’ concept as an explanation of discernible Ceno- zoic biogeographic pattern, and accelerated dif- ferential Earth expansion as a similar explana- tion. I have also taken the view that “what is good for mammals must be good for everybody,” but if the conclusions based here on geology and on mammals are tested and found wanting on the basis of studies of other organisms, then pro- gress will have been made. Inasmuch as the paper was prepared for a botanical symposium, it will be interesting, at least for me, to see if the geo- logical processes discussed here lead to any new fa regarding the floras of various parts of e Earth. I must caution the reader, however, not to just accept the dictates of plate tectonics in place of analyzing geology and biology separately in order to preserve mutual testability. It is curious how willing some biologists are to believe the geol- ogists, yet there is a far smaller flow of infor- mation in the opposite direction. Few geologic Papers cite biological attempts to work out pa- leogeography, although Wegener did not hesitate to do so. Probably this is because the efforts of the biologists have not seemed to be very rig- orous or quantitative to geologists and geophys- cists. Mammalogists, geologists, and, yes, the tanists, need to analyze their data by means b: the cladistic methods developed recently by 1 Icariance biogeographers and also by the more assical and statistical methods of Simpson, the tter stripped oftheir stabilist shortcomings and amplified by the addition of some new effects Mime by mobilist geology. There are dangers circularity and non-testability if one incor- ap data too early that should first be used €sts, but, once that phase is completed, syn- thesis is in order LITERATURE CITED A PIcoTT, W, O. 1969. Tertiary climatic change in © marginal northeastern Pacific Ocean. Science A 165: 583- 586. E R. & M. HucuENEv. 1975. Des rongeurs ammalia) Africains dans une faune de l'Oli- Aen élevé de Majorque (Baléares, Espagne). "o E . Mus. Hist. Nat. Lyon, fasc. 13, suppl., A T >, H. n3 1948. Length of day in the climates ane E plant life. Pp. 101-119 in A. E. Mur- R. O. Whyte bamer ~ Vernalization and ca Botanica Co., Wal- McKENNA—HOLARCTIC LANDMASS REARRANGEMENT 481 ALVAREZ, L. W. 1982. Experimental evidence E an asteroid impact led dt en — of m species 65 million years ago. Law ward Lab. Rep. Ai 1-72. [Also i in Sedi, ex Na- A. ALVAREZ, W., gn H. V. MICHEL & L. W. ALVAREZ. 1982. Iridium anomaly approximately synchronous with terminal Eocene extinctions. Science 216: 886-888 , D. V. KENT, js PREMOLI SILVA, R. A. eather & 1980. Franciscan Com limestone deposited at 17° south preme Bull . Am —484 of Mammalia and B ry in Garfield County, Montana. Univ. Calif. Publ. Geol. Sci. gos 1-286. R, E. H. LiNpsAv, W. A. CLEMENS “1982. Upper "Cretaceous-Paleo R. F. BuT & a Dinaus. Creek and Tullock formations, northeastern in tana. Geology 10: 153-159. ks . A. CLEMENS. 1982. Late Cretaceous nctions. Amer. Sci. 70: 377-385. pe at Es, M. G. 1981. jeter t history of the region of "Wallace" s Line. Pp. 24-35 in T. C. Whitmore (editor), Wallace's Line and Plate Tec- tonics. Oxford Mons on Biogeography, Clarendon A Oxfo M. HUR , À. ey G. SmitH. 1981. Con- tinental Masak in the Mesozoic and Ceno- zoic. Pp. 9-23 in T. C. gg cd (editor), Wal- lace's Line and Plate Tectonics. Oxford Monographs on Pos ADhy, Clarendon Press, Oxford. AxELROD, D. I. 1960. The evolution of flowerin plants. Vol. 1, pp. cee n S. Tax ar Evolution after Darwin. Univ a hicago Press, Chi- cago. 1972. Ocean-floor spreading in relation bi ecosystematic problems. Pp. 15-68 in & F. C. James oton), A rem syste matics, aT BADHAM, J. P. N oceanic margin to we estern North America since Archean time? Geology 6: 621-625. . 1979. Reply [to P. F. Hoffman, 1979]. Ge- ology 7: 227-228. BANDY, O. L., E. A. BUTLER & R. C. WRIGHT. 1969. Alaskan upper Miocene marine glacial and the Turborotalia pachyderma datum plan Science 166: 607-609. BARGAR, K. E. & E. D. Jackson. 1974. Calculated volumes of — shield volcanoes along the Hawaiian-Emperor chain. U.S. Geol. Surv. Jour. Res. 2: 5 50. Barnett, C. H. 1962. A suggested reconstruction of the land masses of the Earth as a complete crust. 7-448. . A. Harrison, J. L. SLOAN II & W. W. Hay. "1981. — phy, 180 million um ago to the present. Eclog. Geol. Helv. 74: 70. G. K, JR. & R. F. BURMESTER. ; Tectonic rotation s in the Cascade Range n. Geology 9: 184-189. BATTEN, D. J., cen R. Dawes, A. K. HIG- 482 ANNALS OF THE MISSOURI BOTANICAL GARDEN GINS, B. E. Kocu, I. PARSONS & N. J. SoPER. 1981. Pe e volcanicity on the Eurasia Basin mar- gin. Nature 294: 150-152. BECK, M. E., JR. 197 5. Remanent magnetism of the "T Q. +, the tectonics of the Western Cordillera. Earth Planet. Sci. Lett. 26: 263-268. : _ 1980. Paleomagnetic record of plate- -margin f North America. a Geophys. Res. 85: 7115-7131. — &P UMLEY. 1980. Paleomagnetism of intrusive eo the Coast Range of Oregon: mi- croplate rotations in middle Tertiary time. Geol- ogy 8: 573-577. BECKMANN, J. P., H. M. Borui, K. PERCH-NIELSEN, F. PROTO DECIMA, J. B. SAUNDERS & M. TouMAR- KINE. 1981. Major calcareous nannofossil and ose events between the middle Eocene and ear iocene. Palaeogeogr. Palaeoclimatol. coral: oie EE 190. VRAHAM A. K. Cooper. 1981. Early evo- lution of d ie Sea by collision of oceanic d North Pacific subduction zones. Bull. Geol. Soc. Amer. 92: 485-495, . Nur, D. Jones & A. Cox. 1981. Conti- nental accretion: from oceanic plateaus to alloch- thonous terranes. Science 213: 47-54. BERGGREN, W. A., D. V. Kent & J. J. FLYNN. 1982. Cenozoic geochronology. 1982. Geol. Soc. Amer. Abstracts with Progra & "ts 4. Paleogene geochro- nol and chronostratigraphy. In N. J. Snelling (editor), virg pi reed and the Geolo ogical Rec- l. Soc n, Special Paper (in press). ————,. pe "EA McKENNA, J. HARDENBOL & J. D. OB- RADOVICH. 1978. Revised Paleogene polarity time scale. Jour. Geol. 86: 67-81. Bosma, A. A. 1974. Rodent biostratigraphy of the Eocene-Oligocene transitional strata of the Isle of den ‘koa t Micropaleontol. Bull. Special Publ. Ea ao D. York & R. A. F. GRIEVE. 1979. Possible source craters for the North American tectites—a geochronological i investigation. Trans. mer. Geophys. Union 60(1 8): 309. BouLIN, J. 1981. Afghanistan structure, Greater India concept and eastern Tethys evolution. Tectono- physics. p. 261-287. BRIDEN, J. C., A. M. HunLEY & A. G. SMITH. 1981. Paleomagnetism ka Mesozoic-Cenozoic paleo- iii maps. Jour. Geophys. Res. 86: 11631- BROOKFIELD, M. E. . H. REYNorps. 1981. Late Cretaceous emplacement of the Indus suture zone ophiolitic melanges and m . 1962. Wüchst die Erde mit Katastro- phen? Düsseldorf. "a BRUNET, M. 1975 grands mammifères chefs de file de P iniata n ere et le probléme de la limite Eocéne-Oligocéne en urope. Thése d'État, Fac. Sci. reri CNRS. ; no. AO 11 606. i- xii + [Vot. 70 1977. Le mammifères et le prote de la pecial 1: YEA BuCHARDT, B. 1978. Oxygen isotope palaeotemper- a: € the Tertiary ure 275: 121-123. es B. "C. 1976. Geology of Romania. Geol. BURCKLE, L.H. & F 8. Implications of late Neogene fresh-water sediment in the Sea o Japan. Geology 6: 123-1 But er, R. F. Ei I H. TavLomR. 1978. A middle ocene paleomagnetic pole from the Naci- miento Formation, San Juan Basin, New Mexico. : 495-4 7: 341-344. Univ. Tasmania, Hobart 19 Palaeomagnetism and Earth expan- sion. Chayanica Geol. 1: 1 5! 76. The Expanding Earth. Development "Geotectonics 10. Elsevier Sci. Pub. Co., erdam, Oxford, New York. Css ow; S. 1974. Island Biology. Columbia Univ. Press, New York and "as don aC The first a . Pp. 20-30 in Ad mon (editor), Hawaii: ian of Evolutio: Hist. 91(12). CARozzi, A. V. 1970. — in of the theory of continental drift. B oc. Amer. 81: 283-2 of origi a HE 76. Inference of the sy 395-38 m" of some Drosophila species. Nature 2 He 1982. An introduction. Pp. ! vation, Nat mon (editor), Hawaii: Showcase of Evo Hist. 91(12). ve ne-Oligocént. CAVELIER, C. ges Sci, Paris New historical ai 1976. La limite Éocè se gee Oligocen GN The geological ex events at i limatol. Palaco eogeogr. Palaeoc 8. North Am in Asia. Bull Geol. Soc. Amer. ; TX EX. dd Ancient to vorda terranes of the North Am € A relation and microplate passe es Amer. 88: 769- CLAGUE, D. A. dod The Hawaiian- mount chain: rigin, pe p tions for ee E Diss. 35(10): 4 -Emperor $% CLARK, D. L. & J. A. KITCHELL. d reply on ‘The terminal Creta Period in the North Sea the | 1972. pe boundary of =, — ~- | and impli — vel "Internat. P $ 1983} logic problem with sa oceanographic solution.’ Comment. Geology 7: 228. CLEMENS, W. A., J. D. ARCHIBALD & L. J. Hickey. 1981. Out with a whimper not with a bang. Pa- leobiology 7: 293-298. CLUBE, SS. V.M M. & W . M. NAPIER, 1982. The role of Earth Planet Sci. Lett. 57: 251—262. Plate tectonic ens for - ad of th go-Miocene evolutio e western Medi ectonophysics 68 ere CoLEMAN, P. J. 1980. Plate tectonics background to SK. JE R Floristic changes indicate a cooling climate in the oat of southern England. Nature 291: 315- Cook, H. E. & M. E. TAYLOR. 1975. Early Paleozoic continental margin sedimentation, trilobite bio- facies, and the thermocline, western United States. Geology Oct., 1975: 559—562. Cooper, A. K., D. W. SCHOLL & M. S. MARLow. 1976. Plate tectonic — for the molti de the east- n. Bull. Geol. Soc. Amer. 87: 1979. Response of deep-sea benthonic oraminifera to development of the psychro- Sphere near the Eocene/Oligocene boundary. Na- ture 282: 63-65. CourTiLLor, V. 1982. Propagating rifts and conti- z nental breakup. Tectonics 1: 239-250. WAN, D. S. 2 iiem ias evidence for post-40 m.y.B.P. large-scale northwestward displacement of pat of soutien Alaska. Geology 10: 309- Cox, à enn. Rous of vi green in bine rth America. Pp. 305-32 rain Geol. Assoc ada Special ‘ pecial Pap. 20. Chocuer, $E. J HARTENBERGER, J.-C. sd A. Remy, B. Sick, J. S LIAUD. rieures as la "Grande Coupure" découve Quercy. Bull. Mus. kien Hist. s “Ser. 4, 3. Sect. C(O): 945-26 IZAT, L., G. NELSON & D. E. ROSEN. 1974. Cen- ters of origin and related : concepts. Syst. Zool. 23: DALRYMPLE, G. B., D. A. CLAGUE, M. O. Garcia & S. 1981. Petrology and K-Ar ages of mples from Laysan Mn and egi ampton Paar volcanoes , Hawa Ridge, an Evolution of the wie M chain. "ws er. Amin 884-93 1977. — volcanic c » E. A. SILVER & E. D. JACKSON. 1973. Origi DA ofthe Hav Hawaiian Islands. Amer. Sci. 61: 204-308. i L. Ee . J. be Wit & K. F. PALMER. : 4. ~ marginal basin in the southern An- H oes 250: 291-294. 77 & K. F Parmer. 1979. Progressive defor- McKENNA —HOLARCTIC LANDMASS REARRANGEMENT 483 ity or the pec. Bull. Geol. Soc. Amer. 90: 259- 280. DANA, J. D. 1847. A general review of the geological effects of the earth's cooling from a state of igneous fusion. Amer. Jour. Sci. Arts, Ser. 2, 4: 88-92. anual of Geology. Theodore Bliss & Co., Philadelphia, m d ai Co., London DARLINGTON, T. eography: The Geographi ical uen eth ofA Naas John Wi- ley and Sons Inc. w Yor DASHZEVEG, D. & M. C. i id 1977. Tarsioid primate from the early Tertiary of the Mongolian People's Republic. Acta Palaeontol. Polon. 22 7 2. Ihe Accidental Universe. , Cam Zoogeography of the Land and Inland T aca and Jackson, Lon- x & F. W. McDoweELL. 1978. Subduction of the Kula Ridge at the Aleutian Trench. Bull. Geol. Soc. Amer. 89: 83-95. F. W. McDowELL. 1975. K-Ar ages from the Near Islands, western Aleutian Islands, Alas- ka: indication of a orte gl thermal event. Geology Dec., 1975: DeEPAOLo, D. J. 1983. no mean i life of continents: t Nda nd don DELONG, S. E., P LFO . Geophys. Res. Lett. 10: 705-708. DE pesce aped G. 82. Five crucial tests of the smic ora scale using the y as funda- tandard. Nature 299: 303—307. 1937. The cosmological constants. 1938. A new basis for cosmology. Proc. Roy. Soc. London A165: 199-208. 197 4. Ses iier aig snag xd the ur don A33 —— 1982. Tectonic apuse of the Oregon n. Or oe l. 44(2): 15-21. DURHAM, J. W. & F. S.M nedum "1967. Cenozoic migrations of marine E cae oron the Bering Strait region. Pp. 312-325 in D. M. Hop- kins a. bad Bering Land Bridge. Stanford niv. Press, Stanford. Du Torr, A. L. Te Tertiary mammals and con- ee drift. A roger to George G. Simpson. r. Jour. Sci. 242: 145-163. | 1963. Tu elements and affinities of and D. Love (editors), North Atlantic Biota and Their mposium D. » Univ. Iceland, n Press, Oxford. Miis K. O. eastern end of Hawaiian chain. Bull. Geol. Soc. Amer. 66: 1648. — C. 1980. Death and renovation at the end e Mesozoic. Trans. Amer. Geophys. Union ire 505-506. ——, E. B. Kraus & E. M. , SHOEMAKER. ce Sud- rth Planet Sci. Lett. 55: 317-334. 484 FARRAR, E. & J. M. Dixon. 1981. Early Tertiary and ica. Geol. Assoc. Canad. Special Pap. 18: 577- eee W. L. 1966. Zür Geologie von Brjans- lae r (Nordwest-Island) unter besonderer eese ag fossilen Flora. Sonderveróff. Geol. Inst. Uni 109. gee, 1981. Die fossile Flora Islands Zeugin der Thule-Landbrucke. Spektrum S. Oct., 1981: 22-31. Gidran R. 1982. Evidence for a major meteor- ite she ag on the Earth 34 million years ago: im- plication for Eocene extinctions. Science 216: 885— 886. GARTNER, S. & J. KEANy. 1978. The terminal Cre- taceous event: a geologic problem with an ocean- ographic solution. Geology 6: 708-712. 1979. Comment and reply on ‘The terminal Cretaceous event: a geologic problem with an oceanographic solution.’ Reply. Geolo; ogy 7: 229. & WwW GATES, O. . GiBSON. 1956. Interpretation of the configuration of the MK Ridge. Bull. Geol. Soc. er. 67: 127- GL E D. 1958. Aosta lower Devonian pa- laeobiology in is, to the concept of continen- tal drift. Pp. n S. W. Carey (editor), Pads tinental Drift, a Suid: Geol. Dept., Uni Tasmania, fa Grass, B. P. & R. CROSBIE. 1982. Age of Eocene/ orig boundary ased on extrapolation from o merican microtectite layer. Bull. Amer ssoc. Petrol. Geol. 66: 471—476. 1972. Early Precambrian evidence and island nuclei of e. Bull. Geol. Sine: Amer. 83: 3323- 3344. GLOBERMAN, B. R., M. E. BECK, JR. & R. eee 1 : Paleomagnetism and tectonic eere of Eocene basalts from the Black Hills, Washing. ton Coast Range. Bull. Geol. Soc. Amer. 93: 1151— 1159 GobpFRIAUX, I. & L. THALER. 1972. Note sur la dé- continental du Hainaut (Belgique). Bull. Acad. Roy. Sci. Belgique, Cl. Sci., Ser. 5, 58: 536- 54 GOLD. T. 1955. Instability of sie earth's axis of ro- tation. Nature 175: 526-529. bisce P. & A. Toomre. 1969. Some remarks on polar wandering. Jour. Geophys. Res. 74: 2555- GOLOMBEK, M. P. & G. E. McGıL. 1983. Grabens, 1. d +, + ` and th ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 & C. D. Carr. 1981. Cenozoic latitudinal shift of the Hawaiian hotspot and its implications for true polar wander. Earth Planet. Sci. Lett. 55: 37- 47. GRAMANN, F., W. HARRE, H. KREUZER, E.-R. Look & B. MATTIAT. 1975. K-Ar-Ages of Eocene to Oli- gocene glauconitic sands from Helmstedt and Lehrte (Northwestern Germany). Newslett. Stra- DALRYMPLE & D. A. CLAGUE rthward movement of the mpero . GRONLIE, G. . Tertiary paleogeography of the Norw rer an-Greenland Sea. Norsk Polarinst. Skr. 170: 49-61. Grow, J. A. & T. ATWATER. 1970. Mid To à tonic iini in the Aleutian arc. Bull. Geol r. 81: 3715-3722. GUPTA, H. K. AF M. DzLANY, editors. 1981. Zagros, Hindu Kush, ome Geod Te Geodynamics Series 3. Amer. Geophys. re Washington, D.C., and Geol. Soc. Amer., Boulet, olorado HAGGERTY, i O. SCHLANGER & I. P. = 9c 1982. Late » Cretaceous and Eocene vule kan HAMILTON, W. 197 078: l- . Geol. Surv. 345. HAMMOND, S. R., D. Epp & F. THEYER. Ee gene relative motion between the Pacific ue mi mantle, and the Earth's spin axis. Nature HARRIS, A. W. & W. R. WARD. a stion of plat- i. 10: aiy bodies. Annual Rev. Earth Planet. Sci 61-108. Polar HARRISON, CG & ies LINDH. 1950 A Wande North Ane RE Mesozoic and Ades Jour. Geoph 1903-1920. & 1982b. Compariso hot spot and geomagnetic field re Nature 300: 251-252. mA "d "ES M 83: 26-38. l May 19 Hie UE “& W. C. PITMAN, Ill. 1970, MAE lineations i in the North Pacific. Pp. RAE D. Hays (editor), Geological Investigat North Pacific. Mem. Geol. Soc. Am ^ yd HEATWOLE, H. & R. LEVINS. 1972. n (et the Puerto Rican Bank: flot sam HE restrial animals. Ecology 53: 112-1 B. Grass & H. W W. MENARD: HEEZEN, B. C., i Manihiki Plateau. Geol. Soc. Amer. Abstracts grams, iine City Meeting, P- 76. cH. 1971. HEINTZ, E. & M graphique des s sous-continent in Pout ER, J.-L. 1983. LaGrandeCoupure 0 ) 1983] rasie occidentale pour les faunes continentales " Miocéne nep Compt. Ren ebd. Séance 01- HEISSIG, K. 1979. "Die hypothetische Rolle ndr nun Oligoziin. Neues Jahrb. Geol. Paliontol. MERE 96. 1981. gradual, not catastrophic, change $ the end of the Cretaceous. Nature 292: 529-5 HILGENBERG, O. C. 1933. Vom ule i Erdball. Giessmann and Bartsch, Berlin. (Not seen. See Carey, 1958: 300.) HoFFMAN, P. F. 1979. Comment [on Badham, 1978]. Geology 7: 226 2: 83-9 HICKEY, L, J. Land plant evidence compatible . B. & M. C. BOULTER e- construction of Palaeogene climate os palyno- logical evidence. Nature 301: 147-150. Moon D. W. 1979. Earth's cratering rate. Nature YH. Huxıey, T. H. 1872. A Manual of the Anatomy of Vertebrated Animals. D. Appleton, New Yor IRVING, E. 19 Paleopoles and paleolatitudes ‘of ——, J. W. H. Moncer & R. W. YoLe. 1980. New eomagnetic evidence for displaced terranes in British Columbia. Pp. 441-456 in D. W. Strang- way eee The Continental Crust and Its Min- eral s. Geol. Assoc. Canad. Special Pap. JACKSON, E. D., E. A. SitvER & G. B. DALRYMPLE. 1972. Haw sapie chain and its derum to Cenozoic ien tectonics. Bull. Geo Soc. Ame x -618. Jacops, ^ = jian 1981. x Moear To- America. Tectonophysics AN, P. Coney & M. BECK. growth ay f western North America. Sci. HONOR OWENS Es 1982. The Amer. 66. ieweg, Mee 1971. The Bpéndiné Earth, Some Conse- Monogr. Nat. Philos. 37. Pergam reer New York, Toronto, Sydney, Braun- Juroy, D. M. 1981. True polar wander. Tectono- 6. 2 1982. The flies fan n A. T. OH ^ Pp. 54-58 in C. Simon (editor), Hawaii: howcase of tiras Nat. Hist. 91(12). McKENNA—HOLARCTIC LANDMASS REARRANGEMENT 485 Kaura, W.M. 1978. An expanding Earth on the pásis of sea-floor spreading and subduction rates: Com ment. Geology June, 1978: 377 KEIGwIN, L. D. JR. 1980. Palaeoceanographic change in the Pacific at the Eocene-Oligocene boundary. Nature 287: 722-725. ., S. D'HoNpr & T. L. Sacred 1983. A ocen rine m Science 221: 150 KENNETT, J. P. & N. ^ SHACKLETON. 1976. Oxygen isotope MEM for the development of the psy- Krem, F.W. 1 Earthquakes at Loihi Submarine Volcano and the Hawaiian hot spot. Jour. Geo- Be Res. 87: es 726. eg RM, 1980. The ex- the Tibetan sedimen sui series, Thakkhola region, Nepal Himalaya. Earth Planet. Sci. Lett. 51: 381— 405 J. W. PEIRCE. 1979. India's and Australia’s pole path since the late Mesozoic and the India- Asia collision. Nature 282: 605-607 KowaLEvskv, V. O. 1873. Sur l'Anchitherium au- relianense Cuv. et sur noue ew des €— Mém. A p. Sci nt Péters- 7, 20(5): i-iv i 73 : 1965. Tertiary fresh-water molluscs from Pacific islands. Malacologia 2: 189-197. LAM, H. J. 1934. Materials towards a study of the flora of the island of New Guinea. Blumea 1: 115- 1 > LARSON, R. r & W. c. PITMAN, III. 1972. World- atian and its sig ati Bull. Geol. Soc. Amer. 83: 3645-3662. LEINDERS, J. J. M. & J. E. MEULENKAMP. 1978. A Microstonyx tooth from veer oo paleogeo- graphical M of Creta mal associa Kon. Ad, Akad. We- tensch., Ser. D useniol 81: 416-424. D. MACGINITIE. 1972. Devel- Mountains. Pp. n A. Graham (editor), denies and Palcoiorstics of Asia and Eastern ica. Elsevier Publ. Co., Amsterdam, todoa. New Yor. LEPICHON, X. & R. HEIRTZLER. 1968. Magnetic LK anomalies in the Indian Ocean and sea-floor spreading. Jour. Geophys. Res. 73: 2101-2117. Li, C.-K. & S.-v. TING. 1983. e Paleogene mam- mals of China. Bull. Carnegie ven Nat. Hist. 21: 1-98. LoweLL J. D. Soc. Amer. 83: 3091-3102. MACARTHUR, R. H. & E. O. WiLSON. 1967. iW 1972. Spitsbergen Tertiary orogenic Spitst F Z Bull. Geol. Princeton, New Jerse MALAHOFF, A., G. M. ey J. C. WILTSHIRE & H.-W 1982. Geology and chemistry of hy- ts cano Loihi, Hawaii. Nature 298: MALUSKI, H., F. Proust & X. C. Xi4o. yr PAr/ 486 ANNALS OF THE MISSOURI BOTANICAL GARDEN *°Ar dating of the trans-Himalayan calc-alkaline magmatism of southern Tibet. Nature 298: 152- 154. MaRLow, M. S. & A. K. Cooper. 1983. Wandering terranes in southern Alaska: the Aleutia Micro- plate and nissim tions for the Bering Sea. Jour. Geophys. Res. € 3439-344 MARSHALL, HECHT. 1978. Mammalian faunal dyuimies of the Great — Inter- aleobiologv 4: 203-206. , S. D. WEB B, J. J. SEPKOSKI, JR. & D. M. RAUP. i Potassium-argon ages from lavas of the Hawaiian Islands. Bull. Geol. Soc. Amer. 75: 107-128. CELHINNY, M. W. 1973. Palaeomagnetism and plate tectonics. Cambridge Earth Science Series, Cam- bridge. , B. J. J. EMBLETON, X. H. MA & Z. K. ZHANG. 1981. Fragmentation of Asia in the Permian. Na- ture 293: 212-216. , S. R. TAYLOR & D. J. M 1978. Lim- its to the expansion of Earth, n, Mars, and ercury and to changes in the ode con- stant. Nature 271: 316-321. McKenna, M. C. 73. Sweepstakes, filters, corri- dors, "Noah's arks, and beached Viking funeral ships in palaeogeography. Vol. 1, pp. 295-308 in D. H. Tarling & S. K. Runcorn n (editors) emer e of Continental Drift to h Sciences. Aca- demic Press, London. ——. 1975. Fossil mammals and early Eocene North Atlantic land continuity. Ann. Missouri Bot. Gard. 62: 335-353. as Eocene Paleolatitude, Climate, and Mammals of Ellesmere Island. Palae eogeogr. Pa- ‘chainlink Palaeoecol. 30: 34 49-362. ———, Ob R. oc A. B. Chiarelli Aha iit aoo Biology of the New World Monke eys ental Drift. Plenum Press, New Y ork, 35-338 in G. Nel- son & D. E. Rosen i aa Vicariance Biogeog- raphy, A Critique. Columbia Univ. Press, New x 83. Cenozoic paleogeography of North At- lantic ser bridges. Pp. 351-399 in M. HPE Ð ov, M. Talwani & J. Thiede (editors), Struc- ture sue Development of the Greenland-Scotland = ew Concepts and Methods. Plenum, New McVrrrIE, G. a stant. Pp. 19-28 in Se and New You: MCWILLIAMS, M. O. & D. G. Howrrr 1982. Exotic terranes of western California. Nature 297: 2156 17. MELVILLE, R. 1981. itors), V. "e An A Critique. Columbia Univ. Press, Yor New MrsERvEv, R. 1969. Topological inconsistency of continental drift on the present-sized Earth. Sci- ence 166: 609-611. . Late Cretaceous and Paleogent sedimentation and tectonics in the dian Arc- tic Islands. Pp. 221-272 in A. D. Miall (editor) Sedimentation and Tectonics in Alluvial Basins. Geol. Assoc. Canad. Special Pap. 23. — K. G. & W . B. CURRY. of Biscay. Nature 296: 347-350. 507 19 Phanerozoic plate bound- eenei with Programs, San Francisc 41. RS ater R. & R. L. Larson. 1975. i ne etic anomalies, oceanic plateaus, a and "s 957 in R. L. Larson, R. Moberly et abili Ber f the Deep Sea Drilling Project. Reports e Deep - Was ur DC Mam S, H SIGURDSSON & R . GOODW i celand. Earth Pl : 197-205. MoRGAN, W. J. 1972. Dei mantle convection = and plate motion. Bull. Amer. Assoc. Petro 56: 203-213. NAFZIGER, R. H. & F. DACHILLE. 1965. oí of the probability of Earth-comet oe snare Abstracts with Programs, Evaluation s of 1979. A theory S. V. M. CLUBE. 82: 455 450 & D. E. Rosen (editors), bese Bi E A oe ace ra Univ. — —— & N. PLATNICK. de: Th rphy: wides sade tisgeoprapiiy Sys Zool. 27: inane” & 1981. ches rests with coniferous forests o * 2 AA 4 1 gions. Later on these dry regi t eastward up to the west part of Gansu (Kansu). Finally, the Talimu Basin became desert and some parts of Qinghai and Gansu gradually changed to semidesert (Fig. 3). By the Quaternary, the topography of China was almost identical to that of the present but with slight differences. The Gulf of Pohai ap- peared in North China during the interglacial periods (Fig. 4). Hainan Island was once con- nected with the continent, but later a strait formed between them due to local subsidence of land. In northeastern China, the Taixinanling and the Xiaoxinanling Mountains were elevated, and in North China Taihangshan and the Inner Mon- golian and the Loess Plateaus were raised. By that time the Himalayas and the Xizang Plateau were rapidly uplifted. The Xizang Plateau was only about 3,000 m in elevation in the early Pleistocene, but it reached 4,000 m in the late Pleisto din the Hol went up to 4,500 o 5,500 m (Xu, 1981). At the same time, the vou 494 ANNALS OF THE MISSOURI BOTANICAL GARDEN a j ^ 3 of i FIGURE 3. Neogene vegetation of China: 1. Temperate forests and grasslands to semidesert-desert mU | northwestern China; 2. Temperate to subtropical deciduous forests and grasslands of northeastern an la 14 4 £ rs 1htranica China: 3 tral and East China; 4. Subtropical evergreen forests m crm mangrove vegetation of South China; 5. Subtropical deciduous and evergreen forests of Yunnan and Xizang. mountains of North China were further uplifted. Inner Mongolia, Shanxi, Shănxi, and a part of Gansu became table-lands. Loess blanketed much Europe. According to the estimate of Antevs (1928), by the time the ice ; ick- sheets reached their maximum extent gis z made from the present depths of shore a coral reefs, etc. give a minimum figure pe 60 m (Brooks, 1950). In this way many pear. seas around the world would have dE So it is reasonable to assume that all the seas along the Pacific coast of China m pi at least several times during Pleistocene., Bis dition, some of the southern J apanese ! the were connected with each other and W! main continent of Asia. Taiwan and H connected with the mainland of China. in wer eastern Asia and the Malaysian Penins connected with Borneo, the Philippine se Celebes, Sumatra, and Java. Moreover, uc also some land bridges between the ur i islands, the Moluccas, New Guinea, an lia. At present, the western Aleutians are v on both the north and the south by deeP Island | J | | 1983] HSU—CRETACEOUS AND CENOZOIC VEGETATION t z 6 FIGURE 4, tem due to oscillation of temperature; 5. Vegetational changes in Xizang; 6. Cert y g 7. Migration of plants via temporary land connections. but the Bering Sea and the north side of the Aleu- hans and the Alaska Peninsula are bordered only by a shallow sea 27 to 95 m deep. So this portion of the Bering region would have become a very wide land mass favorable for plant migration between eastern Siberia and Alaska. However, Qurino th retreated and a large mass of the glaciers melted, * level would have returned to the level of the Interglacial periods, more or less like that of the present, Plant migration would have been tem- Porarily blocked by sea between them. CLIMATE It is generally accepted that Cretaceous tem- ue was higher than that of the present. Ac- ing to the estimate of Schwarbach (1963), average annual temperature of the Late Ju- lE in the world ranged between 25°C and 28°C. iat Early Cretaceous, the maximum temper- 96 wes 24°C, but in the Cenomanian (about to 103 million years ago), sea temperature Changes in Quaternary vegetation of Chin perate deciduous forests; 3. Expansion of dry climate a: 1. Appearance of taiga; 2. Migration southward of flora in northwestern China; 4. Changes in vegetation Ce * 1 . 4 1 ta was reduced to 16°C. In the Coniacian-Santonian (about 80 to 90 million years ago), sea temper- ature rose to 22°C, but by the end of the late Cretaceous it declined again to 20°C. Saito and van Donk (1974) determined the isotopic tem- perature of bottom-water at the same sites around 14 to 15°C for Campanian (about 70 to 78 mil- lion years ago) and early Maastrichtian (about 68 to 70 million years ago), about 12°C higher than that of the present. It suggests a mild tem- perate climate in polar regions with surface-water around 14 to 15°C. Sea surface temperature de- clined over 2°C from the Middle to the Late Maastrichtian (about 66 to 68 million years ago), and further declined by 1.5°C from the late Maas- trichtian to the early Paleocene (about 64 million years ago). Gordon (1973) suggested that “Cretaceous Visi flowing through the Tethys and across the northwestern Pacific Ocean in a circumglobal band of warm water with its own characteristic fauna." Western winds probably existed at high 496 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Voi 70 latitudes in the Northern Hemisphere, and cur- wind system existed during the Mesozoic and Tertiary. During the Late Cretaceous, China mainly lay in the tropical to subtropical region of about 5 to 40°N, but th rtl t part pr obably was in the warm temperate zone. Due to plate tec- tonic movement of the Pacific plate and the In- dian plate against the Eurasian plate, China would have drifted some 10 to 13? northward, as men- tioned above, from the Late Cretaceous to the Quaternary. Accompanying the rapid shifting of geographical position and uplift and the appear- ance of a series of great mountain ranges in the western part of China, the climate changed grad- ually from warm to cooler; so the basic devel- opment of the climate in China may be divided into three stages: Late Cretaceous to Eocene, Oli- gocene to Pliocene, and Quaternary. The annual average temperature from the Late Cretaceous to the present has decreased about 12°C. During the Late Cretaceous, the warm Pacific current bathed the eastern shore of Asia, steadily drifting poleward with the warm Indian Ocean drifting in a northwesterly direction around the northern shore of the Tethys. Accompanied by warm winds these currents produced the warm climate in China and caused abundant rainfall in the northeastern part and less in the southern and western parts. There still was a great expan- sion of dry belt from southwestern, Central, South to East China. Summarizing, in China the cli- mate of the Late Cretaceous was basically sub- tropical. By Eocene time, the warm Pacific and Indian ocean currents were still present. Frakes and Kemp (1972) determined from oxygen isotopes that the mean annual temperature of the warm Pacific Ocean current along the eastern coast was 33*C and that of the warm Indian Ocean current 37°C. The mean annual temperature around Beijing was calculated to be slightly more than 20°C. At the same time, the climatic conditions in other areas were similar to those of the Late Cretaceous. Some large lakes appeared south of Beijing and southern Liaoning, so the climate of these parts was probably wetter than earlier. In the Oligocene, the mean annual tempera- ture of the Pacific Ocean current along the coast of southern China declined to 18°C, although the direction of the current was the same as before. The temperature of the Indian Ocean current then dropped to 34°C. The mean annual tem- perature of Beijing has been calculated to be about 15 to 20°C. The mean annual temperature is now about 15°C. By this time the climate of China south of the Qinling Mts. turned warm and wet and that north of the Qinling Mts. became slight- ly drier and less warm. During the late Tertiary, the average annual Himalayas—which were at least about 2,500 to 3,000 m in elevation, the Kunlun, the Altyn Ta and the Hengduan Shan, became barriers, block- ing the monsoonal winds from the Indian Ocean passing into Tibet. The climate of this part of China, such as the Talimu Basin, the Quidam Basin, Inner Mongolia, and the western part of Gansu, became drier and warmer. As a result of rapid uplift of the Himalayas, the climate of Xi- zang and Qinghai became drier and colder. By contrast, the eastern part of China was subjected to the moderating influences of the Indian and the Pacific Oceans and became warm and pe Due to the barrier of the Qilian Shan and Qinling Mts. across the middle part o running from west to east, cold air masses Siberia were blocked and the climate of — China became warm and wet. According t0 d results of recent investigations on the -— China during the Pleistocene, there is no re evidence of glaciation on the mountains m eastern part of China, except the T à the Qinling Ranges. But the tempt E northeastern China became cooler and drier than before and was temperate 1n Quaternary. In comparing it with that 0 Cretaceous, the average annual tem from would have declined 14 to 24°C. Td iN the plant fossils found at Weinan (092 5 : in Shánsi Province and Panxian (104.7. : in Guizhou Province, the difference in te od and ature between the Wisconsin Glacial Per! p the present in the eastern part of China's 8°C lower. à : As the Quaternary is the most pe period of the upheaval of the Himalayas and during Plateau, the climate of Xizang Plateat dle Pleis interglacial periods of early and mi tocene was relatively moderate. late Pleistocene and Holocene the f ester? came dry and very cold. Other parts © c China also became dry and cold. pit of the temperature and amount of precip. = ——— nÁÀ 1983] southernmost part of southwestern China and the coastal region of South China, Guangdong, and Guangxi, the Hainan Island, the southern part of Taiwan and the islands of South China Sea were tropical. LATE CRETACEOUS VEGETATION IN CHINA Psi the Late Cretaceous and Cenozoic, the d into two major los zones, the northern and me southern. as clearl d by a chain of itibuntain r rop running from west to east, that is, the Kunlun, the Altyn Tag, the Qilian Shan, the Qinling, and the Dabie Shan. The vege- tation as a whole was subtropical. According to Berry (1937) ied Axelrod (1952), by the middle Cretaceous the floras of the world were already differentiated into Arcto-Creta- ceous, Tropical Cretaceous, and Antarcto-Cre- pie 5 Provinces. The Late Cretaceous was fur- and Tethyan CTukhtejan: 1969)— caver ee to the Arcto- and Tropical Cretaceous Flora The Boreal-Cretaceous flora was temperate " nature, cen chiefly of mesophyllous decid- uous trees and shrubs together with ginkgoes, conifers, dt ferns. This flora was widely dis- tributed in the northern part of North America, the Arctic, Greenland, North, Central and north- ‘astern Europe, Kazakhstan, Siberia, the Far East ofthe U.S.S.R. , Korea, and Japan. The Tethyan- Cretaceous flora was subtropical in Baur. con- it some | iticu: comparatively narrow-leaved forms, In this flora, Lauraceae, hone ar evergreen Fagaceae, and palms were charact istic, as well as Myrtaceae and Sapotaceae. It was widely distributed in the southern part of North erica, southern England, southern Europe, the Baltic, the Caucasus, m southern part of the pian Sea, Turkman e main part of E. and seed Hohe In the south- sia of this region, a broad, seasonally dry Ke extended from Spain and north Africa, ime. ugh West and Central Asia to eastern China i northern Indo-China, judged from litho- es data. South of the subtropical Tethyan- i taceous flora lay the Paleotropical Cretaceous ora. Fossils known from Nigeria, Egypt, Syria, the and Iran indicate the existence of a tropical m = the Malaysian type. Remains of man- (roc ud. -— of the Tethys (Takhtajan, 1969). HSÜ — CRETACEOUS AND CENOZOIC VEGETATION 497 LATE CRETACEOUS FLORA OF NORTHERN CHINA Late Cretaceous fossils of northeastern China have been recorded from the Hunchun Group i in hun Basin of Jilin and tl of iia of ‘Heilongjiang (Fig. 1: M of dir plants, though Erden a assem- blages included ferns, gymnosperms, and her- baceous plants. Flora of the lower part of the Hunchun Group and Hunchun Basin (130°E, 42.8°N) of Jilin, of Turonian to Senonian age (95 to 67 million years ago) represents a mesophyllous deciduous forest characteristic of a warm-temperate, humid cli- mate. In this flora the conifers G/yptostrobus europaeus (Brongn.) Heer and Metasequoia cu- neata (Newb.) Chaney are very abundant. An- giosperms include Populites cf. litigiosus (Heer) Lesq., Juglandites poliophyllus G. et L., Trocho- dendroides vasilenkoi Iljin et Rom., Protophyl- lum multinervis Lesq., P. haydenii Lesq., P. spp., and Leguminosites (Guo & Li, 1929). A similar flora of the Late Cretaceous from Wuyun (129.8*E, 49.11°N) in Heilongjiang flour- ished in a warm temperate to subtropical humid climate and is under investigation by Tao of our laboratory. In this flora Osmunda greenlandica yea Brown and the conifers Metasequoia and Se with a few sn of Thuja cretacea (Heer) Newberry. The angiosperms include Protophyllum cf. mi- crophyllum G. et L., Pseudoprotophyllum cf. dentatum Hollick, Betula prisca Ett., Alnus, Pop- ulus carneosa (Newberry) Bell, Ziziphus phos- phoria Krysht., Mahonia cf. furnaria Brown, Menispermites borealis Heer, M. obtusiloba Lesq., M. kuliensis Tanai, Ampelopsis acerifolia (New- berry) Brown, Debeya tikhonovichii (Krysh.) Krassilov, Trochodendroides arctica (Heer) Ber- ry, Tetracentron, Sorbaria, Tiliaephyllum cf. tsagajanicum Kapacnorb., Viburnum cupa- nioides (Newberry) Brown, V. antiquum (New- er Hollick, V. asperum Newberry, Bauhinia, TY. 102.1 , Cissus nonna (Lesq.) Brown, nw cf. turcomanica (Krysht.) Kor., and C ycloca These fossils lived on the eem of ancient Songhua Lake. They contributed to a meso- phyllous deciduous forest, with a few evergreen plants, including Mahonia in the forests. At pres- ent, Bauhinia, Pterospermum, Hat and Cissus occur in tropical-subtropical r Judged from the fossils, the pa Cretaceous 498 flora of northeastern China belonged to the Bo- real Cretaceous Flora. It contained a great many temperate taxa, such as A/nus, Betula, Corylus, Populus, Salix, Tilia, and Zizyphus, and a few warm temperate to subtropical ones, notably Mahonia, Glyptostrobus, Metasequoia, Sequoia, Platanus, Tetracentron, Cyclocarya, Myrica, Nymphaea, and Trochodendroides. The warm temperate to subtropical elements, such as Am- pelopsis, Viburnum, Sorbaria, Rhus, Cercidi- phyllum, and Schisandra, are quite abundant. Moreover, tropical elements, such as Bauhinia, Cissus, Pterospermites, Dombeya, Dombeyopsis, and Grewiopsis, were associated with the ancient taxa Protophyllum, Pseudophyllum, and Dry- ophyllum. Altogether about 40 species have so far been recorded. Twelve of them, Osmunda greenlan- dica, Glyptostrobus europaeus, Thuja cretacea, Trochodendroides arctica, T. vasilenkoi, Meni- spermites borealis, Betula prisca, Populites cf. li- tigiosus, Debeya tikhonovichii, Ziziphus phos- phoria, Tiliaephyllum cf. tsagajanicum, and Rhus cf. turcomanisa seem similar to or nearly iden- tical with European, Central Asia, and Siberian Cretaceous plants. Eighteen of them, Osmunda greenlandica, Thuja cretacea, Metasequoia cu- neata, Trochodendroides arctica, Cercidiphyl- lum arcticum, Protophyllum multinervis, P. hay- denii, Populites cf. litigiosus, Menispermites obtusiloba, Ampelopsis acerifolia, Pterosper- mites auriculaecordatus, Pseudoprotophyllum cf. dentatum, Viburnum antiquum, V. cupanioides, V. asperum, Cissus marginata, Populus carneo- sa, and Mahonia cf. furnaria seem similar to the North American Cretaceous plants. LATE CRETACEOUS FLORA OF SOUTHERN CHINA Turning to southern China, fossils of Brachy- phyllum rhombiomaniforum Kuo, Cinnamo- mum hesperium Knowlton, C. newberryi Berry, Nectandra prolifica Berry and N. guangxiensis o have been recorded from the Boli Forma- tion of Cenomanian age in the Yongning of Guanxi (Guo, 1979) (Fig. 1:2). Palynological data show that the flora of the late Early Cretaceous to Cenomanian in Jiangsu and Zheijiang was dominated by Classopollis (pollen of Brachyphyllum and Pagiophyllum), Ephedra, Quercus, Schizaeaceae, and Ulmaceae. During this period pinaceous pollen was less common (Wang et al., 1979: Song et al., 1981). ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo.. 70 These plants indicate that a dry-climate flora fl ished in thi t ical i Ofthe species noted above, similar fossils have been recorded | in North America, such as Cinnamomum hes- — perium, C. newberryi, and Nectandra prolifica. | Clearly, the Late Cretaceous flora of southem | China was part of the Tethyan-Cretaceous flora. | PALEOGENE VEGETATION IN CHINA | The Chinese Tertiary flora can also be basically divided into two zones, demarcated by the same | mountain ranges as those of the Late Cretaceous. The northern Paleogene floras were further dif | ferentiated into two floras and those in the south | into three. These were controlled by geographical | and climatic factors. At that time, China was in a lower latitude, under a subtropical climate. e northern Paleogene flora was comp. | mainly of deciduous plants, such as Carpinus, Alnus, associated with some evergreen species | such as Quercus, Dryophyllum, and conifers Se- | quoia, Metasequoia, Glyptostrobus, Taxodium, and Torreya. | The southern Paleogene flora, which was char- | consisted baceous ground cover was chiefly of drop | and Schizaea.™ | momum eve e e | as Quercus (including Cyclobalanopsis), | nopsis, Lithocarpus, and palms. A DRY CLIMATE FLORA OF NORTHWESTERN - This flora developed in Xinjiang, Qingh4^ - western Gansu at the beginning 0 grassland vegetation occupied the pr beginning but gradually became sem! 2p A 4: x: 1 1 1 records vy gocene some Abies and Picea forests with - | were flourishing in the high hil rus fot- side of the Zhungaer Basin, as Were rests wer l ests in the lower latitudes. Deciduous ie Car | composed of Betula, Tilia, Acer, Qué rocatyt | pinus, Castanea, Juglans, Alnus, Some ele” Rhus, Carya, Ginkgo, and Magnolia. the foot ments of Elaeagnaceae were scattered a plum- hills. The undergrowth included herbs% | 1983] HSU—CRETACEOUS AND CENOZOIC VEGETATION baginaceae, Primulaceae, Labiatae, Compositae, and others. The drought-resistant plants, Ephed- ra and chenopods, were widely distributed on the plain (Hsu, 1956). Some forests of Abies, Picea, Pinus, Cedrus, Betula, and Quercus lived on the hills around the Quidam Basin. Trees of Magnoliaceae and Pro- teaceae grew scattered in the valleys. The un- dergrowth included herbs of Compositae, Cru- ciferae, Gramineae, and others (Hsü et al., 1958). In western Gansu, pine forests were better de- veloped in the hills around the Jiuquan Basin. Magnolia and Ginkgo grew on hilly tracts. Drought-resistant Ephedra was widely distrib- uted on the plain (Song, 1958). WARM-TEMPERATE TO SUBTROPICAL DECIDUOUS AND CONIFER FORESTS IN NORTHEASTERN CHINA AND NORTH CHINA This flora was the successor of the Boreal Late Cretaceous mesophytic flora of northeastern China, the eastern part of Inner Mongolia, and North China. Alth gh Dal g d soils have so far been recorded in this area, a sporo- pollen assemblage in Fushun (123.9°E, 41.8°N) indicates that some mesophyllous forests flour- ished, in which Sphagnum, Lycopodium, Se- i Quercus, Ulmus, Salix, Buxus, Liquidambar, Cornus, and Magnolia. The gen- eral aspect of this flora was very similar to those «t the same age found in East Siberian and Sa- T Island (Song & Liu, 1976; Sun, Du & Sun, = Celtis, Nelumbo, Trochodendron, Cercidi- Phyllum, Schisandra, Cinnamomum, Lindera, , Ulmaceae, Rosaceae, and Rham- Naceae were most abundant. Most of them are 499 deciduous, except Dryophyllum, Cinnamomum, and Sabalites. Most genera are subtropical to warm temperate, but Osmunda, Ginkgo, Meta- sequoia, Sequoia, Glyptostrobus, Torreya, Tax- odium, Keteleeria, Dryophyllum, Meliosma, and Firmi. l t subtropical, and Lygodium. Cinnamomum, Acacia, and Meliosma are chief- ly tropical. In this flora, one-fourth of the species, such as Osmunda lignitum, Alnus corylina, Betula po- puloides, Celtis preacuminata, Dryophyllum saf- fordii, Lindera antique, Hamamelites inaequalis, Rosa hilliae, Acacia aquilonia, Mimosites var- iabilis, Ampelopsis acerifolia, Fraxinus rupina- iburnum speciosum, Sparganium anti- S eus, an from the illustrations, seem similar to the Paleo- cene-Eocene plants of North America. Similarly, more than one-fourth of the species appear to be similar to European, Central Asian, and Siberian plants. These include Osmunda lignitum, Ly- godium kaulfussii, Ginkgo adiantoides, Glyptos- isti Taxo- ungeri, Cercidiphyllum arcticum, Lindera anti- qua, Phellodendron grandifolium, Paliurus col- ombii, Rhamnus duensis, Fraxinus juglandina, F. rupinarum, and Viburnum nordenskiöldii. Taxodium, Sequoia, Comptonia, Fothergilla, and Sabal are extinct in China, but still live in North America. By this time Metasequoia, Alan- gium, Trochodendron, Pterocarya, Platycarya, Zelkova, Cercidiphyllum, Cinnamomum, Koel- reuteria, and Paliurus existed in both northeast- ern China and North America. They now occur only in eastern Asia. Middle Eocene microfossils in the coastal re- gion of Bohai also show that the climate of North China was warm temperate, characterized by a predominance of Taxodiaceae, Betulaceae, Ju- glandaceae, Ulmaceae, Sapindaceae, and also Alangium, Platycarya, Lygodium and Osmunda. By late Eocene-early Oligocene, Ephedra and Schizaea were abundant there. During the mid- dle Oligocene, Quercus, Ulmus, Alnus, Salix, and Melia became dominant, and, in the late Oli- gocene, Ul d Jugland bec pre dominant. Asa whole, the Paleogene flora of North China was composed of Abies, Larix, Picea, Cedrus, Keteleeria, Pinus, Tsuga, Carya, Engelhardia, 500 Platycarya, Pterocarya, Alnus, Betula, Carpinus, Fagus, Quercus, Ulmus, Celtis, Trochodendron, Nelumbo, Magnolia, Liquidambar, Melia, Elaeagnus, Tilia, Rhamnus, Nyssa, Cornus, Aralia, Symplocos, Fraxinus, and Lonicera, as- sociated with some elements of Liliaceae, Chen- opodiaceae, and Compositae. DRY CLIMATE FLORA WITH SCATTERED FORESTS IN THE SUBTROPICAL ZONE OF CENTRAL CHINA Judged from a Paleocene sporo-pollen assem- blage found in Jiangxi (Sun & He, 1980), the flora was composed chiefly of Schizaeaceae, Pterida- noniaceae, Olacaceae, and Palmae. The genera Schizaea, Pteris, Ephedra, Ulmus, Hemiptelea, and Quercus were most abundant, followed by Corylus, Sorbaria, Lespedeza, Rosa, Chenopo- diaceae, Cruciferae, Leguminosae, Polygalaceae, and Lythraceae. Up to the Eocene, Quercus, Xylosma (Fla- courtiaceae), Anacardiaceae, Hamamelidaceae, Myrtaceae, Symplocaceae, Nyssaceae, Santala- ceae, Sapotaceae, and Olacaceae were very abun- dant. The temperate elements Betula and Ju- glans were less abundant. The sporo-pollen assemblages of Hunan and Hubei of the same ages were similar to those of Jiangxi as described above. By the late Eocene, megafossils of Cinnamo- mum cf. lanceolatum Heer and Comptonia an- dersonii Florin were recorded in Hunan (Sze & Lee, 1954). Some megafossils of the form genus Palibinia have been found in Shanxi, Henan, and Hunan (Tao, 1965; Li, 1965; Liu & Kong, 1978) (Fig. 2:3). The na i cori ture of the leaves indicate drier sites. Associated with these in Henan are leaves of Ulmus, Fraxinus, Sophora, Crataegus, Campylotropis, Lespedeza, Cercidiphyllum, Zelkova, Dalbergia, Zanthoxy- lum, and pollen of Ephedra, Carpinus, Myrica- ceae, Taxodiaceae, Cruciferae, and Leguminosae (Liu & Kong, 1973). Palibinia has been found in Bembridge, Isle of Wight, England, and Turk- menia of the same age. Evidently this flora was closely connected with the subtropical Tethyan Tertiary region of southwestern Asia and south Europe. ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 DECIDUOUS AND EVERGREEN FORESTS WITH CONIFEROUS FORESTS IN THE SUBTROPICAL ZONE OF COASTAL REGION OF EAST CHINA This flora covered the coastal region of Jiang- | su, Zhejiang, and Fujian provinces. As men- d e , = d z re, We z 1 dai ence parallel to the coast line. By Paleocene-Eocent 1 loni] 4 £ 4c nhia time ly of Ulmaceae (Ulmus, Zelkova, Aphananthe, Celtus, Trema, and Ostrya), Betulaceae (Carpi- nus, Alnus, Corylus), Juglandaceae (Carya Pterocarya, Juglans), and Fagaceae ( Fagus, Quercus) associated with Taxodiaceae and Pi- , naceae (Pinus, Keteleeria, Cedrus, Picea, Abie) — forest flourished in these mountains. Some trop- ical elements of Abelia, Rhus, Proteaceae, Mag- from palynological records (Wang et al. Song et al., 1981; Petroleum Geological Explo- ration Team of Xijiang, 1979) (Fig. 2:4). Up to the Oligocene, forests of Ulmaceae, Fi gaceae, and Meliaceae were abundant. In flora there were some elements of Magnoliaceae, Juglandaceae, Myrtaceae, Rubiaceae, Moraceae, Araliaceae, Betulaceae, Oleaceae, Eu aceae, and Rosaceae associated with Tilia, Rhus, and others. By this time, herbaceous plants, a as Convolvulvus, Gentiana, Hedyotis, and Co ositae, were present. Among the gymnos Taxodiaceae became better developed than pinaceous plants. Ephedra an been recorded, though not abundant. dicates that this flora was more humid of Central China (Wang et al., 1979). DECIDUOUS AND EVERGREEN FORESTS MIXED WITH CONIFEROUS FORESTS IN SUBTROPICAL ZONE AND MANGROVE VEG ai ALONG THE SHORE OF THE SOUTH CHINA This area covered most parts of South China. including Guangxi and Guangong andi the South China Sea. By the Paleocene, for- was composed of deciduous and evergreen a ests mixed with some conifers. These forests mainly composed of Ulmaceae, Loran became abudi Imaceae ( This than that Other predominant plants were U “Co mus, Celtis), Rutaceae, tanea), Juglandaceae (Carya, Juglans, e : ya), Liquidambar, Alnus, Salix, A^ Rhus Taxodiaceae (Sun, 1982). Megafossil phorbi- d Schizaea hav’ | | THE ETATION | | slandso 1983] (Schenk, 1883), Osmunda, Lygodium, Palibinia, Cinnamomum, Nelumbo, Trapa, Nordenskioel- dia, Goeppertia, Eucommia, Cyclocarya, Citrus, Ocotea, Dryophyllum, Nectandra, and Sabalites (Guo, 1965) were recorded (Fig. 2:5). Ofthe 18 species, one-third of them have been recorded from the Paleogene of North America, such as Lygodium kaulfussii Heer, Osmunda lig- nitum (Giebel) Stur, Nelumbo protospeciosa Sa- porta, Nordenskioeldia borealis Heer, and Dry- ophyllum puryerensis Berry. Others have been found in Europe, Kazakhstan, and Japan, such as Palibinia laxifolia Korovin, Lygodium kaul- fussii, Cinnamomum naitoanum H. & T., C. lar- titii Watelct, Nelumbo protospeciosa, and Nor- denskioeldia borealis. Up to the Oligocene, judged from palynolog- and the subtropical Liquidambar. The tiópidil Caesalpinia (Leguminosae) and the water fern Ceratopteris were well represented. Mangrove vegetation appeared along the Pacific coast. This orà was closely related to the Paleogene flora of Borneo. cond types of pollen are identical in each area, especially Florschuetzia (possibly pollen of cian, SUBTROPICAL HIGHLAND FLORAS OF XIZANG, YUNNAN, AND GUIZHOU Some fossils were recorded from the Xigaze Group of the Yarlung Zangbo valley (Guo, 1975) and the Qiuwu and the Menshi Series ofthe west- ‘most corner of Xizang (Geng & Tao, 1982) (Fig. 2:6), poindisg Salicaceae (Salix, Populus), oraceae us), Cercidiphyllaceae (Cercidi- in Arca (Aralia), Fagaceae (Quer- (Sp C ntaceae (PES Liliaceae (Dianella), EM (Cyperacites), and Typhaceae (Ty- s The age of these fossils was formerly as- mig by the authors to the Late Cretaceous and te Cretaceous to Eocene, respectively. After Pronti investigation by field geologists, both have ai shown to belong to the Eocene (Yin, per- live 7| Most of tnn genera now Ph n tropical regions. Ficus is pantropical — is distributed in ibe et and Pe ig Africa. Dianella is distributed in tropical ia, Australia, New Zealand, and Polynesia. HSU—CRETACEOUS AND CENOZOIC VEGETATION 501 Of the 23 species, half of them, as judged from illustrations, seem similar to Paleogene plants of North America, such as Populus latior Al. Braun, Salix meeki Newberry, Juglandites sinnatus Les- quereux, Ficus daphnogenoides (Heer) Berry, F. myrtifolia Berry, F. stephensoni Berry, Cercidi- phyllum ellipticum Brown, Rhamnites eminens minalis L. and Spiraea alpina Tureq., resemble those of Europe and East Asia A tropical-subtropical Pisces flora occurs in Jinggu, southern Yunnan. It is mainly com- posed of evergreen plants, Lauraceae (Phoebe, Machilus and Nothophoebe), and Fagaceae (Lithocarpus, Quercus and Dryophyllum). Oth- ers were of Annonaceae (Annona), Araliaceae (Oreopanax), Combretaceae (Terminalia), Mo- us ceae gymnosperms (Cephalota Shrubs of Rhus, Rosa, Jasminum, and Sorbus were part of the understory. Similar specimens, such as Zelkova ungeri, Myrica banksiaefolia, and Carya cordioides, have been found in Europe and Central Asia, and Oreopanax oxfordensis and Quercus simulata have been recorded from North America. NEOGENE VEGETATION IN CHINA By Neogene time China was ata slightly higher latitude. Neogene -a were more gods the nd her- bagni | plants were e better developed. TEMPERATE-WARM TEMPERATE FORESTS AND GRASSLANDS AND DESERT-SEMIDESERT FLORA OF NORTHWESTERN CHINA This flora was developed in Xinjiang, Qinghai, and Gansu, but later extended to eastern Inner Mongolia. 4 re 1 di + Dy E 3 Larix and Picea forests began to develop in the Altai Mts., he alpine Sabina fi d mountains of northern Xinjiang. Deciduous for- est composed of Betula, Ulmus, Salix, Populus, Juglans, Alnus, Corylus, and Tilia occurred in some foothills. Grasslands with Artemisia, Ephedra l Most parts of the Talimu Basin were covered by grass- land. Some trees of Taxodiaceae, Palmae, Myr- taceae, Lauraceae, Magnoliaceae, Anacardi- 502 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 aceae, Ulmaceae, Quercus, and Juglans were scattered in the foothills north of the Talimu Basin (Hsii & Li, 1980). Subalpine coniferous forests composed mainly of Pinus, Cedrus, Abies Tsuga, and Juniperus were flourishing on the northern slope of the Kunlun Mts. (Hsii & Li, 1980). Deciduous forests of Quercus, Castanea, and A/nus occurred in the valleys. Grasslands of chenopods, Polygonaceae, and Cyperaceae, as- sociated with Ephedra, covered the plains of the Quidam Basin and Gansu (Hsii et al., 1958). Af- ter the withdrawal of the Obé Sea from Central Asia, the Mediterranean elements Tamaricaceae and Zygophyllaceae (Nitraria and others), mi- grated through Central Asia to the Quidam Basin and Gansu (Li, 1960). In certain places salt marsh vegetation appeared (Fig. 3:1). Active orogeny of the Himalayas and other mountains in Xizang caused the climate of northwestern China to become drier than before. The alpine forests of Larix and Picea, in addition to Abies, were better developed in the Altai Mts. Some pine forest mixed with Betula, Ulmus, Quercus, Corylus, and Tilia appeared in the foot- hills. But Picea forests were better developed in the Tianshan, western Kunlun, and Qilian Shan, with an understory of some Rosa and Spiraea shrubs (Hsü & Li, 1980). On the plains occurred - Fraxinus, and Celtis were scattered in some val- leys (Hsü & Li, 1980). WARM TEMPERATE TO SUBTROPICAL DECIDUOUS FORESTS AND GRASSLANDS OF NORTHEASTERN AND NORTH CHINA Palynological investigation shows that at the beginning of the Neogene forests of Picea, Pinus, Carya, Fagus, Quercus, Alnus, and Celtis were l Juglandaceae, and Rosaceae. Among them some thermophilous elements such as-Carya, Liquid- ambar, and Corylopsis were present (Li, 1982). The late Miocene flora of Shandong (Fig. 3:2) is well represented by the Shanwang flora (Hu & | Chaney, 1940; Inst. Bot. and Inst. Geol. and Pa- leont., 1978). It was mainly composed of decid- uous trees and shrubs associated with some ev- ergreen elements. The flora includes Fagaceae (Castanea, Castanopsis, Quercus), Betulaceae (Alnus, Betula, Carpinus, Corylus), Juglandaceae (Carya, Platycarya, Juglans), Ulmaceae (Ulmus, Celtis, Zelkova), Salicaceae (Populus, Salix), Moraceae (Broussonetia), Hamamelidaceae (Fothergilla, Hamamelis, Liquidambar), Rosa- ceae (Crataegus, Eriobotrya, Malus, Prunus, Rosa, Spiraea), Aceraceae (Acer), Anacardiaceae (Rhus, Pistacia), Sapindaceae (Aesculus, Koel- i ea), Legumi- onymus), Rhamna Paliurus, Ziziphus), Tiliaceae (Tilia), Cornaceat (Cornus), Ebenaceae (Diospyros), Bignoniaceat (Catalpa), Vitaceae (Ampelopsis), and Simari- baceae (Ailanthus), plus some subtropical plants of Araliaceae (Kalopanax), Magnoliaceae (Mag nolia), Lauraceae (Lindera, Litsea, Cin mum), Rutaceae (Evodia), Euphorbiaceae (M T lotus), Ulmaceae (Aphananthe), and trop! plants of Sterculiaceae (Commersonia), Mora- ceae (Ficus), and Vitaceae (Tetrastigma ochningense, Acer subpictum, Viburnum nordenskióldii, Corylus MAM < jiy Juglans acuminata, Populus glandulifera, en argusta, and Astronium truncatum, se "m mon to those of Europe, Central Asia, -— beria. It is interesting to note that Fot sequoia, Sequoia, Glyptostrobus, lived there (Song et al., 1964), though palynological investigations, 1 of Abies, Picea, and Pinus existed 1n th part of North China. Mixed forests us, Qu Cryptomeria, Pinus, Betula, Salix, a: cus, Ulmus, Zelkova, Fraxinus, an no lot" : re ‘flourishing; but many tropical trees we shansi ger present. Megafossils found in Taigu. | HSU—CRETACEOUS AND CENOZOIC VEGETATION 1983] were composed mainly of Quercus, Ulmus, Acer, Amelanchier, Ribes, and Leguminosae, all of which are native plants now. No taxa are known to be identical with those of North America. SUBTROPICAL EVERGREEN AND DECIDUOUS FORESTS OF CENTRAL AND EAST CHINA During the Miocene, the vegetation of Central and East China turned into coniferous and ev- ergreen and deciduous forests in sharp contrast ly composed of Ulmaceae (Ulmus, Celtis, Zel- kova), Juglandaceae (Carya, Juglans, Pterocarya), Betulaceae (Betula, Alnus, Corylus), Fagaceae uercus, Fagus, Castanea), Hamamelidaceae (Liquidambar, Fothergilla, Corylopsis), and Myricaceae (Myrica), mixed with some ever- green trees like Magnolia, Quercus, Symplocos, etc. Carya, Pterocarya, Sapindus, Ilex, Symplo- cos, Corylopsis, Pittosporum, Melia, Liquidam- bar and some elements of Rutaceae, Euphorbi- aceae, Myrtaceae, and Olacaceae existed there, (Fig. During the Pliocene, herbaceous plants be- et al. die So AM 1981; Zheng et al., 1981) 3:3). ie me dominant and were compo chiefly of Ulmus, Celtis, and Quercus. It is worth noting that the tropical Caesalpinia, Nyssa, Bux- us, orchids, and some Annonaceae and Euphor- d * are occasionally recorded. Salt marsh *Betation existed in the coastal region (Zheng etal., 1981). By this Eo the flora of western Sichuan was a mainly of Castanopsis, Quercus, an : nodaphne, associated with some Acer and orbus (Guo, 1978). 2 503 SUBTROPICAL EVERGREEN FORESTS AND TROPICAL MANGROVE VEGETATION IN SOUTH CHINA During the Miocene, forests of this region were composed mainly of Castanopsis, Quercus, Lithocarpus, Juglans, Castanea, Carya, Ptero- neratia appeared along the coast (Sun et al., 1981) (Fig. 3:4). SUBTROPICAL EVERGREEN AND DECIDUOUS RESTS ON THE YUNNAN AND XIZANG PLATEAU By the Miocene, the vegetation of southern Yunnan was composed of Cupressaceae (Calo- cedrus), Aceraceae (Acer), Alangiaceae (Alan- gium), Lauraceae (Laurus, Cinnamomum), Le- guminosae (Albizzia, Cassia, Dalbergia, Desmodium, Sophora, and Pithecelobium), Ju- glandaceae (Pterocarya ), Faga aceae (Castanea, m in Vietnam (Inst. Bot. and Inst. Geol. and Pa- laeont., 1978). The Miocene flora of northern Yunnan in- cluded Pinus, Cupressus, Fagaceae (Dryophyl- lum, Quercus), Lauraceae (Sassafras, Phoebe, Cinnamomum), Ulmaceae (Zelkova), Anacar- diaceae ally Pistacia), and Rhamnacaee (Pal- iurus) (Fig. By the Pli f th onifers appeared. The chief elements were fu sclerophyllous Quercus semecarpifolia, spathu- lata, pannosa, monimothicha, and gilliana. Oth- ers were Acer paxii, Celtis bungeana, Viburnum A EAS Betula, Populus, Ulmus, Michelia, d Rhododendron, associated with some Ro- saceae and Leguminosae. Confers were domi- nant, including Pinus yunnanensis, Cedrus deo- dara, Picea, Larix, and Tsuga. Pinus yunnanensis was very abundant there, but Cedrus deodara had migrated to the southwestern part of the Hi- malayas (Tao & Kong, 1973). During this time, the forests of northwestern copes were composed mainly of Fagntaee, Lauraceae, Leguminosae, Theaceae, and niaceae. ce half of the plants are identical to the recent ones. Most of them belong to Casta- nopsis, Ulmus, Eurya, and Strychnos. In addi- 504 tion, many species of Rhamnaceae, Fagaceae, Anacardiaceae, Leguminosae, Rosaceae, Alan- giaceae, Lauraceae, and Myricaceae also occu- pied the southern slope of the Himalayas. This suggests that during the Neogene there was an extensive migration of plants between Xizang and Yunnan and the northern part of the South Asian subcontinent (Tao & Du, 1982) (Fig. 3:5). By the Miocene, vegetation of central Xizang was composed mainly of Juglans, Sophora, Pop- ulus, Betula, Carpinus, Ulmus, Ribes, Crataegus, Quercus, Thermopsis, Rhododendron, Salix, Cedrus, Tsuga, Sabina, and Podocarpus. Similar vegetation was flourishing in the western Xizang (Hsii, 1981) (Fig. 3:5). Passing to the early Pliocene, forests of Abies, Picea, Tsuga, Cedrus, Lithocarpus, Quercus, Fa- gus, Salix, Carya, Melia, and Ilex were wide- spread in northern Xizang. It is interesting to note that pollen of some tropical elements, such as Podocarpus and Araliaceae and spores of the epiphytic fern Drynaria, were found at some sites. Up to the late Pliocene some subalpine forests were widespread on the northern slope of the Himalayas. Cedrus deodara and some sclero- phyllous evergreen oaks, Quercus semecarpifo- lia, pannosa, and senescens, were predominant. The general aspect of the flora and the floral com- position were much like those of the Pliocene flora of northern Yunnan (Xu, 1981). QUATERNARY VEGETATION IN CHINA The Quaternary vegetation in China did not differ very much from that of the late Pliocene. However, there were some evident changes, since the evolution of plants was rapidly progressing. The change of vegetation during this period was oscillating temperature, lowering and rising of sea level, and variance in topography, especially the uplift of the Qinghai-Xizang Plateau, the Hi- malayas, and other mountains. APPEARANCE OF A COLD TEMPERATE FOREST ZONE IN THE NORTHERNMOST CHINA From the late Pliocene up to the late Pleisto- cene, some Siberian, cold-temperate taxa pene- trated southward 4° to 5° to form taiga in north- ern Xinjiang and northern Heilungjiang. Forests of Larix gmelini migrated southward to-and-fro several times from the Altai Mts. This forest passed along the Beitashan Range, was widely distributed in the Dzongarian Basin, and intrud- ed the east end of the Tianshan Range (Chang, ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo.. 70 1959). Forests of Larix gmelini also migrated southward several times as far as the Sanjiang plain (Kong & Du, 1982). By the time of the Late-Glacial, Abies- Picea forests associated with Betula were developed on the plain, leaving the taiga forests flourishing in the Taixinanling and the Xianxianling Mts. (Fig. 4:1). SHIFTING OF THE SOUTHERN BOUNDARY OF THE WARM-TEMPERATE, DECIDUOUS-FOREST ZONE The southern boundary of the warm-temper- ate zone shifted southward 3? to 5? during the Quaternary Period. Although the Qinling Range is an old mountain range running from west t0 east, across China, it was not high enough to separate climates in eastern China in the early Pleistocene. The palynological record shows that the subtropical plants Ginkgo, Podocarpus, Ce drus, Tsuga, Pseudolarix, Cunninghamia, Pter- ocarya, Platycarya, Carya, and Magnolia still resided in Shănxi (=Shenxi), north of the Qinling Mts. (Inst. Bot. and Inst. Geol., 1966). Only after the middle Pleistocene, were the Qinling Mt uplifted to the present altitude, thus forming à barrier to prevent seasonal winds from the south passing over to North China. Since then, the cli- mate of North China has become drier and coo" er, and the Qinling Mts. have begun to act oh barrier between the subtropical and the Wa temperate zones of Eastern China. Acco : the previously mentioned subtropical plants n longer lived in North China (Fig. 4:2). OF EXTENSION OF THE DRY-CLIMATE s NORTHWESTERN CHINA TO NORTH CH : : acid zone was By the early Pleistocene the semiarid zone ™” K ins In limited to northwestern China. Some basin the middle Pleistocene, the dry-cl tended westward to North China. tion of Qinhai. A/nus and yn : components in the deciduous fores s. ;ccatiol- late Pleistocene, due to the effect of — the size of forests and grasslands vini (Fig duced. Some grasslands became sem! 4:3). ‘tors and In North China some forests of conifers broad-leaved trees still flourished p ped int? Pleistocene. Grasslands were well deve ~~ 1983] the middle Pleistocene. During the late Pleisto- z cene, gr were mucn reduced, and their composition became simpler. The forests were composed mainly of Ulmus, Celtis, Morus, Quercus, and Populus. By this time herbaceous plants wer C ; (Inst. of Botany and Inst. of Geology and Pa- laeontology, 1978). CHANGE OF VEGETATION DUE TO OSCILLATION OF TEMPERATURE During the Pleistocene there may have been as many as 17 glaciations (Heller & Liu, 1982). Due to oscillation of temperature during that time, Abies-Picea forests were better developed in the eastern part of China, but during intergla- cial periods the forests decreased in area. Faga- ceae was the most common group in the sub- tropical and moist zone, while Ulmaceae was the dominant one in the warm-temperate and drier regions, During the last glacial (Wisconsian) period Abies-Picea forests were better developed on the northern slope of the Qinling Range. Picea wil- sonii forest once extended down to the hilly lands of 490 m in Weinan (109.5°E, 34.5°N) of Shanxi (Fig. 4:4), but is widely distributed in the hills of Hebei, Shanxi, Shanxi and northern Sichuan at an altitude of 1,600 to 2,500 m. Abies and Picea forests once flourished in the foothills around Beijing 430 to 450 m (Kong, 1976), but now these forests occur only above 1,600 to 2,000 m in this area. Fossils of Picea have been found in Panxian (104.7*E, 25.8*N) of western Guizhou at an al- titude of 1,000 to 2,000 m, but this spruce now lives at an altitude of 2,200 to 2,400 m in north- em Guizhou. By this time Abies-Picea forests Were also widely distributed in East China. Now only a few trees of Abies survive as relicts in the a of Zeijiang and northern Guangxi (Fig. 4: VEGETATIONAL CHANGES DURING THE QUATERNARY IN THE XIZANG PLATEAU During the early Pleistocene, the Xizang Pla- "idus about 3,000 m in elevation. Subalpine ps ts of Cedrus, Pinus, Betula, Quercus, Car- Hi us, and Alnus were widespread in the central A eis, while subalpine forests of Larix, ies, Picea, Pinus, and Sabina flourished in HSÜ — CRETACEOUS AND CENOZOIC VEGETATION 505 marix and Nitraria had already migrated from the Qinghai Plateau to northern Xizang (Fig. 4: During the middle Pleistocene, the Xizang Pla- teau was uplifted to 3,500 m in altitude. Subal- pine forests of Picea, Cedrus, Sabina and scle- rophyllous evergreen Quercus semecarpifolia, pannosa, and senescens were widely distributed in the central Himalayas. But subalpine forests of Pinus, Betula, Quercus, Carpinus, and Alnus still occurred there. Due to decreased tempera- ture and rainfall only few forests could survive in northern Xizang, so xerophytic plants Ephed- ra, Tamarix, Nitraria, Artemisia, and chenopods were better developed there. As the result of uplift, as the Xizang Plateau rose to 4,000 m, few forests could survive in northern Xizang. Passing to the Post Glacial, the altitude of the Xizang Plateau was 4,500 m in elevation, and the vegetation of most parts of Xizang became alpine steppes. ACCELERATION OF EVOLUTION DUE TO TOPOGRAPHIC AND CLIMATIC CHANGES The speed of evolution of the angiosperms of Xizang was accelerated by the uplift of the Xi- zang Plateau. Plants in Xizang adjusted to the influences of topographic and climatic changes by evolving some new genera and species. More than 32 new genera of Umbelliferae, Primula- ceae, Solanaceae, Compositae, and Gramineae have recently been discovered in Xizang (Wu et al., 1981). Most of the species of Gramineae are polyploids (Liu, personal communication). At the same time, a new regi I (Wu et al., 1981) (Fig. 4:6). MIGRATION OF PLANTS BETWEEN MAINLAND OF CHINA AND THE ADJACENT ISLANDS Lowering of sea levels caused by the with- drawal of ocean water to form the Pleistocene ice sheets caused the Pacific coast to extend east- ward to the Taiwan-Ryukyu Area-Kyushu dur- ing the glacial periods. This enabled some plants from Fujian to migrate directly to Taiwan and from Guangdong to Hainan, and vice versa (Fig. 4:7). THE RELATIONSHIP OF CHINESE PAST FLORAS WITH THOSE OF NORTH AMERICA Paleobotanical records previously mentioned tell us that most of the Late Cretaceous and Pa- leocene-Eocene plants were widely distributed in 506 Laurasia. During the Late Cretaceous less than one-third appear to be similar to European, Cen- species. By the time of Paleocene-Eocene, one- third to one-fourth of the species seem similar to those of Europe, Central Asia, and Siberia, and one-fourth to one-half seem similar to those of North America. This reflects an extensive mi- gration between the Chinese and North Ameri- can floras via Central and West Asia and Europe before the middle Eocene, as there were no ef- fective barriers such as ocean or high mountains in these regions this time there was also no land bridge available in the Bering region for migration of plants between Asia and North erica. By the Oligocene most species in China were of eastern Asiatic origin. One-tenth were com- mon with the plants of Europe and Central Asia and less than one-tenth appear to be similar to those of North America. Until the Miocene about one-tenth of the plants recorded in China were common with the European, Central Asian, and Siberian species, and only one-twenty-fifth seem similar to the North American species. Up to the Pliocene, most of the Chinese plants were of east- ern Asiatic origin, only a few were common with European, Central Asian, and Siberian species, and very few appear to be similar to North Amer- ican species. This means that after the Eocene Chinese vegetation was rapidly modernized by a loss of the American elements. Some Chinese Oligocene and Miocene species common with those of North America are remnants of meso- phytic plants once T distributed over all the northern hemisphere I agree with the view of Li (1971) that most of the present isolated and disjunct genera of eastern Asia, especially of China, and eastern inction, except the local areas with pentes pr favorable climatic condi- tions” as claimed by Schuster (1976 At present, eastern North America and eastern Asia (especially the southern part of China) are two important relict temperate centers in the northern hemisphere, because in both regions there were no extensive Pleisticide glaciations, and they both have more favorable climatic con- itions. However, of these two regions, eastern ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 Asia (especially southern China) has more relict genera than eastern North America LITERATURE CITED 1928. The last glaciation. NOR York, r. Geogr. Soc., Research Series pE D. I. 1952. A theory of Ra evo- lution. Evolution 4: 29 Berry, E. W. 1937. Tertiary floras of Eastern North America. Bot. Rev. (Lancaster) 3: 31-46. Brooks, C. E. P. ep CM Through the Ages. ANTEVS, i Ernest Be: CHANG, pi 1959. micans hical distribution of the eastern Tianshan forest. Papers on the Sym- e Natural Factors in uic Sinjiang AE sium of th i e Press, tonomous Weiwoer Region. Scie 4956. m Drewry, G. E., A. T. S. RAMSAY & A. G. SMITH. l Climatically controlled sediments, the geomag- netic field, and trade wind belts in Phanerozoic time. J. Geol. 82: 531-553. FLORIN, RUDOLF. 1958. On Jurassic taxads and Co nifers from northwestern Europe and castei nland. Acta Horti Be J: 257-405 d . 1963. The distribution "or conifers an ss and space. Acta Horti 1972. Influence of Frakes, L. A. & E. M. KEMP. "fime ms continental position on a Tertiary with 110-125. Science d Beijing. (In n Chinese English summ : Goop " . 1947. The Geography of the Floweriné Plants. Longmans, Green & ; Gorpon, W. A. 1973. Marine li 269-294 currents in the Cretaceous. J. Geol 81 Re Gray, A. 1846. Analogy between Me flora d Ars and god ot d United States. Am d IIQ): 1 a. 1979. GUAN, E wie XIUMEI & SUN XINHU On sporo ed pa : phy of the Neogene of Bohai. Select the First Symposium of the Palynolosica. Societ of China. Pp. 64-70. Science Press, Chinese.) T ossil Guo, Suenos, 1965. On the dod 9 rti orma $- E RAM e Pelsebnt ol. Sin. 134: 9? . 1975. The plant fossils at the he rm ports n gma Press, Beijing. (In Chinese.) Acta ——. 1978. Pliocene floras of western Sichuan. —— ma m—À nnam om | 1983] Palaeontol. m E 343-350. (In Chinese with JONES sh sum 979. 1 lan coi and el. Tertiary flo- ras from the southe m Guangdon uangxi E Cenozoic red beds of South tme a 223-231. Science inset end (In Chin —— — & Li Haomin. 1979: gg p Le flora from Huncian of Jilin. Acta Palaeontol. Sin. 18(6): 547-560. (In Chinese with English summary.) HELLER, F. & T.-S. Liv. 1982. Magnetostratigraph- ical dating of loess deposits in China. Nature 300 431-433. Hsu, JEN. 1956. Sporopollen assemblage from the l ificance. News yanma nese Palaeontological Society 1. (In 1978. On the paleobotanical evidence for continental drift and Himalayan uplift. Palaeo- botanist 25: as 142. E HEN & CHow Hor. 1958. Sporo- pollen assemblages from the Tertiary prio of the Tsaidam Basin and their geologi gnifi- cance. Acta Poterie 2 E 429-440. X Oe des glish s ——— __ Chinese wi — —, ZHU JIANAN, CHEN Yi Yr, rede SHUYING, Hv YUFAN & Une WEIQING. 1974. New genera and species of the late Triassic plants from Yungjen Yunnan, I. Acta Bot. Sin. 16(3): 266-278. (In ese with English summ — — & LiSivin. 1980. Development and evolution of the Chinese Cenozoi 65-76 in e Miocene flora from Shantung Province, China. Paleontologia ; Sinica, EIN Ser. A (1). MUR , ACADEMIA SINICA & CENOZO Researches i in Ce- nozoic c palaeobotany of Nantian, Shánxi Proceed- of| Hoiian, ‘Shanxi. Pp. 157-182. Science Press, Beijing. (In Chinese.) Eid OF BOTANY & INSTITUTE OF GEOLOGY AND ALEONTOLOGY, ACADEMIA SINICA. 1978. Ce- nozoic plants from China. Fo ae: plants of China Es - Science Press, Beijing. cade nese. G, ZHAO-CHEN, Du NAIQ erus ene & Tao JINRONG, 1976. iris and climatic changes uring the past one hundred million years of Bei- jing. Aca Phytotax. Sin. 14(1): 82-89. (In Chinese ao English sum ary. & Du Nargru. 1982. The climate in canes 000 years (abs IN aternary Woes: v Environment of China. OMIN, 1965 peine aa ostatki kion (rri gor chashan po khenyan. cta “aye Sin. 13(3): 540-547. (In Chinese HSÜ — CRETACEOUS AND CENOZOIC VEGETATION 507 e Morris Arboretum Monograph reprinted in 971.) 1972. Eastern Asia-eastern North America species-pairs in wide-ranging genera. Pp. 65-78 in A. Graham (editor), dae ae and Paleofloristics of Asia and Eastern dem Elsevier Co., mste iios pelea aba ork. Li on 1960. Don eo: of the vege- tion on the northern side of the Kunlun Ranges. Its formation in relation to the arid region. Acta Liu, YoNG-AN & K eue of late Eocene from Wucheng, Henan and thei cance in botany and palaeoclimatology. pn Bot sot. Sin. 20: 59-65. (In s Chinese with English sum- mary. McEruiNNv, M. W. 1970. Formation of the Indian an. Nature 228: 977-979. Pis. GEOLOGICAL EXPLORATION TEAM OF DEPARTME TONGJI UNIVERSITY & THE LABORATORY OF THE THIRD MARINE GEOLOGICAL SURVEY. 1979. Sporopol- len assemblage of or seen pud of Xijiang. ogy Report of Tongji Uni- versity. (In grains from the co astal region of Bohai. Sc Press, Beijing. (In Chine with hehe su ummary 3 D. I. AXELROD. SAITO, T. & J. VAN Donk. 1974. _Oxygen and carbon Tertiary foraminifera. Micropaleontology 20: 152- 177: SCHENK, A. 1883. Pflanzenreste aus dem Tertiár des südlichen China. In Richtofen's China. Bd. 4. S. 268 SCHUSTER, R. M. 1976. einige tectonics and bearing and dispersa ri sperms. Pp. 48-136 in "Charis B. Beck (editor), Origin pu hee Evolution of Angiosperms. Co- lumbia Uni s. SCHWARZBACH, M. 1963. Climates y^ the Past—An edited by R. O. Muir.) D. van Nostrand Co. SONG, pore and pollen — from the red beds. of Chiu-Chuan , Kan- su, an Acta Palaeontol. Sin. re 159-167. (In Chinese (In Chinese pem English ——— &Li rupes spen Tertiary spo- 508 £ €x t Mem. Inst. Geol. Palaeontol. Acad. Sin. 3: 179-190. (In Chinese.) Tertiary palynological assemblages from Jiangsu. Geological Publishing House, Beijing. (In Chinese.) SUN, XIANGJUN. 1982. Palynoloflora of Liushegang Formation Ras rid Oligocene) in the north- ern part of Sea. Acta Phytotax. Sin. 20(1): Pass dn Chinese with English summary.) ——, Du N seen & SUN MENGRONG. 1980. Paly- nological investigation on the Fushun Group (Pa- leogene) ot northeas In Hong You- i the Fushun Coal Press, WE (In & HE YUEMING. 1980. Pild sporopollen assemblages of Jiangi. Science Press, Beijing. (In Chinese.) , KONG ZHAOCHEN & Li MINGXING. 1980. leogene new pollen genera and species of ta outh a Sea. Acta Bot. Sin. 22(2): 191-199. (In Chinese with English summary. , Li PUN & Li MiNGxiNG. 1981. Oli- gocene palynoflora i in the northern part of South China Sea. Acta Sete ka 19(2): 186-194. (In Chinese with English mary.) SZE, m C. & LEE Homer. 1954. PAE Tertiary flora m Hunan. Acta Sci. Sin. 3(2): 205-2 218. inburgh. (E a TAO, JUNRONG. 1965. A ub Eocene florula from = district Weinan of central Shensi. Acta Bot. 13(3): 272-278. (In Chinese with English sum- G ZHAOCHEN. 1973. The fossil flora - Acta Bot. Sin. 15(1): 120-126. (In Chinese with English summary.) ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 —— & Du Naraiu. 1982. Neogene flora of Teng- chong Basin in western Yunnan, China. Acta Bot. Sin. 24(3): 273-281. (In Chinese with English summary. VAN DER LINDEN, W. J. M. 1975. Mesozoic and Cai- nozoic opening of the Labrador Sea, the North Atlantic and the Bay of Biscay. Nature 253: 320- _ 324. WANG, XIANZENG, ZHOU SHANFU & XU SHUJUAN. 1979. Discussions on the palaeogeography and palaeoclimatology of the late Eocene epoch and igocene epoch, north Jiangsu. Acta Bot. Sin. 21(2): 149-156. (In Chinese with English abstract) Wu, ZHENGYI, TANG YANCHENG, Li XIWEN, WU SUGONG 1981. Dissertations "upon nalization of Xizang flo ie ee cid bii Proceeding Xizang (Tibet) Plateau (Beijing, c: Géoló gical pem sconce studies of Qinghai-Xi zang Plateau, Vol. II. Pp. 1219-1244. Science Press, Beijing. ; NGYI) On the significance of Xu Ren (HsÜ JEN). roceedings posiu ang (Tibet) Plateau i jing, China). Pie d ecologia E i i Plateau. Vol -Xizang Pla Liu X1aNncat, WANG G XIAN ‘emia Sinica jiang Inst. Geol. Ages Academia Si 1981 (3): 29-90. 15 pls. (in Chinese wi summary.) is —— VEGETATION OF CHINA WITH REFERENCE TO ITS GEOGRAPHICAL DISTRIBUTION Hou, HsioH-Yu (Hou XUE-Yv)! THE PHYSICAL/GEOGRAPHICAL BACKGROUND F CHINA'S VEGETATION General references for this section include: Editorial Board of China's Vegetation (1980); Hou (1956, 1982). China is situated in the eastern part of Asia, on the western coast of the Pacific Ocean. The distance from east to west measures over 5,000 km, and from north to south, over 5,500 km. The topography of China is a three-step west- east staircase (Fig. 1). It begins with the Ching- hai-Tibet Plateau, 4,000 m above sea level. Crossing Kunlun and Chilien ranges on the pla- teau’s northern edge and the Hengtuan Moun- tains on its eastern edge, the land slopes away to highlands and basins mostly from 2,000 to 1,000 m above sea level; it then descends further east- ward to hilly regions and plains below 1,000 m. The summer monsoons greatly dominate the climate of the country. The southeast monsoon arising over the Pacific Ocean mainly influences the eastern half of China. The southwestern mon- soon arising above the Bay of Bengal and the Indian Ocean mainly influences China's south- em and southwestern regions. This is why coastal eastern China, especially the southeastern part, has comparatively heavy precipitation that is s concentrated within a few summer months. _ annual rainfall ranges from 600 to 1,000 mm In the northeast, and is mostly between 1,000 to 2.200 mm in the southeast. Atmospheric mois- ture decreases as one goes westward. Due to the Parts of China have a low rainfall, mostly re- ceiving an average of less than 100-500 mm an- "ally. The land lying between the two regions ae above is semi-arid, having a mean ie rainfall of 300-400 mm. Thus, three aera regions, namely forests, steppes, and *Serts, correspond to the humid, semi-arid, and ‘nid climates (Fig. 2). use the northernmost point of China is at bun. iN Bs, Beijing, People’s Republic of China. ANN. Missouri Bor. GARD. 70: 509-548. 1983. about latitude 53°N, and the southernmost one cold-temperate, temperate, subtropics, and trop- ics (Fig. 3). Accordingly, there occur four vege- tational regions, namely, needle-leaved decidu- ous forests, broad-leaved deciduous forests, broad-l d orests, andt ical i Ulva evergreen seasonal and rain forests. Because both altitude and terrain also strongly influence the climate, climatic regions are very complicated in China (Fig. 4). The greater part of the Chinghai-Tibet plateau ranging from 4,000 alpine meadows, alpine steppes, and alpine des- erts. There is a close relationship between the char- acter of the soil and the nature of the underlying rock. However, climate, topography, and vege- tation are just as important as rock in determin- ing the ultimate nature of the soil. Because of the variation in soil formation factors mentioned above, the soil varies greatly from place to place in China (Fig. 5). In the same climatic zone, dif- ferent vegetation types can occur on different soils. In a country like China where numerous vari- ations occur in physical features, climatic, and soil conditions, it is not surprising that many types of vegetation are to be found. CONCEPTS AND CLASSIFICATION OF VEGETATION General references for this section include: Braun- Blanquet (1932); Chien et al. (1956); Küchler (1964); Tansley (1953); UNESCO (1973); Walter (1973); Warming (1909); Weaver and Clements (1938); Whit- aker (1962). Different schools of geobotanists have devel- oped various concepts concerning plant com- munities (the vegetation). I maintain that a plant Laboratory of Plant Ecology and Geobotany, Institute of Botany, Academia Sinica, 141 Hsi Chih Men Wai 510 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vou. 70 | AX ^ pid yr 2, 3 Chienling LA Second-step ig rer FiGuRE 1. Map showing topography of China. community, occupying a certain space, is an ag- gregate of plants occurring together in a common habitat. The groups of associated plants are to a certain extent similar in structure (including pop- ulation, stratum, synusium), floristic composi- tion, external appearance of the dominant species, and intimate functional interrelationships among each other, and in their relationship to environ- mental factors. From this viewpoint, a plant community can be considered to be a system that includes not only the plants of which it is com- posed, but also the natural geographic environ- ment of habitat that forms a recognizable self- containing entity. It also means that either the climax or serial stages, and either natural or man- made ones are all considered to be part of the plant community (Hou, 1960, 1979, 1982). Based on the plant community concept just given, the following four criteria should be thor- oughly considered when classifying vegetation: (a) life or growth forms of dominant plants; (b) strata and synusia; (c) dominant floristic com- position with the help of character (indicator or diagnostic) species; (d) habitat or environment. Only two of the higher vegetational categories i ighest one ! in China are used in this paper. The high ! s i » which oome | can be called “vegetation type, pec the plant community in which th done n i i er have a commo | plants in the dominant lay sodioa forests | S mis | deserts and xerophytic shrubs, —_ ceous vannas, meadows, and swampy vegetation. , “hed i The vegetation category mainly d" vhid this paper can be called “formation ¢ formation is to a great extent similar to the d referst0 group” (UNESCO, 1973). This — gee flo- communities that have a or i get- ristic composition, including fa zo forts era, in addition to the same life d m^ of their dominants. The — tially t0? : essentitt ORT vertical certain latitudinal, longitudinal, - referring ® nation. Thus, supplementary n dded 10 the climate, soil, and landforms are i give? names to help in the identificati category. 1983] HOU — VEGETATION OF CHINA LEGEND Humid Region = Ze eq a Mer m if A SIE, k it ——— imue. ace N p Eoo dan FIGURE 2. Map showing regions of atmospheric moisture of China: 1) Humid region with non-distinct dry season; 2) Bum region with disti 6) Extremely arid region VEGETATION TYPES OF CHINA AND THEIR GEOGRAPHICAL DISTRIBUTION NEEDLE-LEAVED FORESTS l. Needle-leaved deciduous forests (light taiga) of the cold-temperate zone or on the mountains ofthe temperate zone. Reference: Chang (1955). Larch forests, being light-demanding, can endure Passam degree of dryness than spruce and fir Pio, They can be found on either dry, sunny pes or in moist valleys and lowlands. Larix $melini forests occur chiefly on the Greater gan Mountains in extreme northern China, Picus the climate is continental boreal. These rests have an undergrowth of dwarf shrubs, the rier with Vaccinium vitis-idaea and the wetter with Ledum palustre. In the Altai, in north- ‘esternmost China, the predominant tree is re- y Larix sibirica, often growing with Pi- foe in the shrub layer of this forest (Fig. 1 Needle-leaved evergreen forests (dark taiga) nct dry season; 3) Semi-humid region; 4) Semi-arid region; 5) Arid region; on mountains of the temperate zone. References: Chen et al. (1964); Chow and Li (1964); Integrated Survey Team of Sinkiang and Institute of Botany, Academia Sinica (1978); Teng (1947). The shady coniferous forests dominated by Picea obovata, Abies sibirica, and Pinus si- birica are confined to the northwesternmost cor- ner of the Altai Mountains, whereas those of Picea jezoensis, P. koraiensis, Abies nephrolepis, A. holophylla, and Pinus sylvestris var. mongo- lica are extensively distributed in the northeast- ern mountains of China. The climate of the two above-mentioned regions is much more humid than that of places where only spruce forests oc- cur. On the mountains of Tianshan, Chilienshan, and Honan, which are situated in the central part of the arid desert region, only spruce forests, dominated by Picea shrenkiana and P. crassi- folia, but no fir forests, are found (Figs. 7, 8). 3. Needle-leaved evergreen woodlands on semi- E sandy soil of the temperate zone. Referen Zhao (1958). The precipitation of the Lin steppe, which ranges from 300 to 400 mm yearly, 512 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 === ES 2 c K ES 6 E. RE 7 Plateau SSSS 3 perate; 3) FiGURE 3. Map showing regions of atmospheric warmth of China: 1) Cold-temperate; 2) Tem i Wi arm-temperat mountain clima is much less than in humid areas. However, the sandy soil can retain larger amounts of rainwater than the loamy soil on which grasses grow. More or less open and scattered woodlands dominated by Pinus sylvestris var. mongolica grow on the sandy soils. This type of woodland has an un- rae —— T ud . 3 nm . (Fig. 9). 4. Needle-leaved evergreen forests of the tem- perate zone. These forests are distributed on the hills and mountains of the North China Plain that are under the influence of the summer mon- soon. One type of forest, consisting of Pinus ta- bulaeformis inland and P. densiflora along the coast, occurs on acid brown soils. Another group, dominated by Platycladus orientalis, is seen on neutral or calcareous Korichnevie soils. This type of forest is usually associated with deciduous broad-leaved trees and shrubs (Chen et al., 1954, 1965; Chow, 1963) (Fig. 10). 5. Needle-leaved evergreen forests of subtropi- cal and tropical zones. References: Chow e; 4) Transitional subtropics; 5) Subtropics; 6) Transitional tropics; 7) Tropics; te. 8) Cold high- : 8y Wu (1957); Ho and Chen (1963); Kiang (193 phe f (1955). These forests are composed o. | Taxodiaceae, and Cupressaceae, and er p accompanied by some evergreen broa! ve” trees and shrubs. Forests growing on ec P. include Pinus massoniana, P. yunnane in the armandii, and Cunninghamia lanceolata subtropics, and Pinus khasya and P. Pini ta niana in the tropics. The montane arg wanensis forests and mixed Cathar aa phylla forests are also found on a id "a subtropics. Another group, including on lime | funebris and C. duclouxiana, thrives stone soils (Figs. 11, 12, 13). n mou 6. Needle-leaved evergreen forests es nd tr tains (subalpine conifers) of nie 3) gdito ical zones. References: Chiang Cogo; Tet ecreases wit rial Board of Szechuan’s Veg (1948). Because air temperature " increasing altitude, on mountain — ical 2,500 to 4,000 m above sea level 1n 1983] HOU— VEGETATION OF CHINA 513 A". Es = ES d (lj Se : N NN : NS X RV BN NS VI Tine: A K LEGEND E LY u « VI [^2] m EE va Ew ASSY Vi — M ———M rr IGURE 4. Map showing climatic regions of China: I) COLD-TEMPERATE ZONE—Humid region; II) TEMPERATE ZONE—II1. Humid region with distinct dry season, II2. Semi-humid region, II3. Semi-arid region, II4. Arid region with evenly-distributed rainfall, 115. Semi-arid and humid montane region; III) WARM- TEMPERATE ZONE~III1. Humid region with distinct dry season, III2. Semi-humid region, III3. Semi-arid region, III4. Arid region, III5. Extremely arid region, III6. Montane semi-arid region, III7. Montane extremely arid ok IV) TRANSITIONAL SUBTROPICS—Humid region with non-distinct dry season; V) SUB- season; VII) TROPICS—VII1. Humid region with non-distinct dry season; VII2 and tropical zones, climati diti found fers have the following characteristics: (a) There t, in some ways, resemble those of the cold- are many species of fir and spruce, namely, Abies lemperate zone, Because of this, evergreen co- — faxoniana, A. georgei, A. forrestii, A. faberi, A. le fargesii, A. cheniensis, A. squamata, A. delavayi, ferous forests consisting of fir, spruce, and pine uam can Survive on the mountains at low latitudes. A. spectabilis, A. kawakamii, Picea asperata, P. Although there are obvious similarities between purpurea, P. likiangensis, P. brachytyla, P. mor- the mountain climate of the subtropics and trop- risonicola, P. spinulosa; (b) The tree layer is as- ics and that of the cold-temperate zone, there are sociated with evergreen, broad-leaved Quercus also many differences. On southern mountains and Rhododendron species, as well as Tsuga chi- wo day temperatures in spring and autumn are nensis and T. dumosa, which are not found in higher and the light and heat intensity at midday the northern evergreen conifer forests; (c) In the are much greater throughout the year. In addi- shrubby layer there are no arctic plants, such as he the precipitation and humidity are much Vaccinium vitis-idaea and Ledum palustre, but '£her than those of the temperate zone. It is rather higher evergreen shrubs such as Rhodo- small wonder that the southern subalpine coni- dendron and Quercus species, as well as bamboo ANNALS OF THE MISSOURI BOTANICAL GARDEN ; 1 5 : 3. FiGURE 5. Map showing soil regions of China. FOREST SOIL REGIONS: 1) Brown conifer soil region; 2) Dark brown soil region; 3) Brown soil, Korichnevie soil, alluvial soil region; 4) Yellow ME Korichnevie soil region; 5) Yellow and Red Earth Subalpine conifer soil region; 8) Lateriti RASSLAND soil region; 14) Zierozem soil region; 15) i : 4 d -mountain mea: di steppe soil region. DESERT SOIL REGIONS: 17) Gray-brown desert soil region; 18) Brown desert 19) Cold high-mountain desert soil region. shrubs such as Sinarundinaria and Pleioblastus ig. 14). (E BROAD-LEAVED FORESTS AND WOODLANDS evergreen conifer and deciduous broad-leaved forests. They grow on acid dark-brown soils on the northeastern mountains of China. The forests : . Juglandaceae Oleaceae, Salicaceae, Araliaceae, etc. The follow- ing trees can be noted as representative: Ulmus soil, Yellow ; 6) Red earth region: ) 0) Phosphorus lime soil reg dow soil region; propinqua, Maackia amurensis, ed mandschurica, Tilia amurensis, T: o rica, Acer mono, Juglans mandschurica, jn davidiana, P. koreana, Betula platyphylle ben us mongolica, Carpinus cordata, Phell "i amurense. e evergreen conifers ges by dominated by Pinus koraiensis, feque?" axus Cus; t Abies holophylla, and rarely by (Fig. 15 2 Broad-leaved deciduous forests of the pel perate and subtropical zones. Referen - (1958); Chen et al. (1965); Numata (19 p broad-leaved deciduous forests (known pose! i m ulacea climate in North nd i these forests differ from those 1n region; } -—9 1983] fe > z y "E nISRBAMAIZMN, ' t r © "M". P 6 HOU — VEGETATION OF CHINA e . :x i dg one 515 a he. T Y VM = x hs ds PH: MS 2 N -J A i & yer "17 p: A41 T ee ees ee FD TREYT tons Wa. a the Greater Khingan Mountains is the typical FIGURE6. Larix li : g gr *getation of cold-temperate regions. UR United States in the absence of beeches Mid Spp.). These humid-loving trees are con- oi o the humid subtropics in China. This type Mum on can be classified into two groups, ks : deciduous oak forests and mixed hard- sive rests. The oak forests are found in rel- Y Open and light habitats. The sequence of horizontal distribution of the dominant oaks in Q. variabilis, Q. acutissima, Q. glandulife second group of forests has no dominant species, but i d of mixed hard ds. The mixed E" ANNALS OF THE MISSOURI BOTANICAL GARDEN Uy ale {a amm 4 A Beer ( aT à) yd, re SA- æ "4 "EU edet Y TET vo e mom à vof "L] oome dus "2 £ dicii i d Ficure 7. (Upper) At altitudes between 1,100 and 1,800 m on the Chanapa Mountains, perle Beni evergreen forest dominated by Sae) nephrolepis and Picea jezoensis associated with its secondary gov E 8. we forest i is ay on shady slopes and e * a Wi „igni hrenkið altitude in northern pog of the desert region, Picea $ ain steppe on sunny s HOU — VEGETATION OF CHINA D d e z 7 s | 3E Te ee moi RA ay FIGURE 9, (Upper) In the mee region sandy soil can retain larger amounts of rainwater than the loamy soil on which grasses grow, so that on the sandy soil grows scat ttered ea 2 is dominated by Pinus uen is var. mongolica wi": an unde — wth an B f ound cover of steppe vegeta Ficu E10. (Lower) e warm-t rate r n, deciduous oak El Mu beech growing on slightly a sit 2 T However, Pinus rebua mid. Lam with an undergrowth of Cotinus coggygria var. cinerea r$ her ANNALS OF THE MISSOURI BOTANICAL GARDEN N unning ipa: 11. (Upper) After rie of the broad-leaved evergreen forest i in am the subtropical region, C eco aoe ine oppe r slope) and Phy. llostachys pubescens (lower slope) take over ye seconda! j : wer) Pinus m tea ac foren t unde rowt growth ot de eon daara "ERE with Rhododendron simsii Sh an rer HOU — VEGETATION OF CHINA FIGURE 13. Upper) Cupressus funebris, a calcium-loving tree, is the secondary growth of the mixed deciduous and evergreen broad-leaved forest on limestone soi soil. es GURE 14. (Lower) Picea fabri forest mixed with Tsuga chinensis is seen on the western mountains in the i Szechuan Basin from 2,000 to 3,000 m elevation. 520 forests, growing on acid soils in shady and moist slopes or valleys, consist of Acer and Tilia species, Fraxinus mandshurica, Ulmus propinqua, Bet- ula platyphylla, etc., whereas those growing on al ; phora japonica, Broussonetia papyrifera, Tilia mongolica, Dalbergia hupeana, and Ailanthus altissima (Fig. 16). 3. Montane microphyllous deciduous forests of temperate and subtropical zones. Dominant trees of these forests include species of Betula and Populus. They are usually secondary, re- placing cold-temperate or subalpine conifers af- ter felling or fires. Betula platyphylla forests are distributed in the cold temperate zone and on mountains of the temperate zone, whereas the forests dominated by Betula albo-sinensis and B. platyphylla var. szechuanica frequently occur on the mountains in the subtropics. However, Bet- ula ermanii often forms pure, virgin stands lo- cated at the upper limit of montane evergreen conifer forests in the temperate z and its low- er limit can reach 1,000 to 1,800 m in the north- eastern mountains of China. Populus davidiana can accompany Betula platyphylla; however, it can also be found in areas where the climate is warmer (Fig. 17). 4. Microphyllous deciduous woodlands of the temperate zone. References: Chen and Chow (1957); Ching (1959). There are two groups of this type of woodland. One type is distribute along the rivers of the arid region and is domi- nated by Populus euphratica (P. diversifolia). This species is widespread on the flood plain of the Darim River, which is fed by water from the glaciers and snows of the high surrounding mountains. It also is frequently encountered in a limited area on river banks in all desert regions. The soil in which P. euphratica grows is slightly saline with a water table of about 1 to 3m Another species is Populus pruinosa, found on soils with a lower salt content. Elaeagnus an- gustifolia woodland also is found on saline soil. Ulmus pumila woodland, which occurs more fre- quently on sandy soil in the semiarid steppe re- gion, sometimes also is found in desert regions that have a water supply (Fig. 18). 5. Mixed broad-leaved deciduous and ever- green forests on limestone soils of the subtropical zone. References: Chow (1957); Chu and Wen (1953); Wang (1956). Because limestone has ex- ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. cessive drainage through sinkholes, and because 10 rock surface temperature fluctuates greatly be- tween day and night in the subtropics, the local | | | soil climate is comparatively dry. For this reason, 4 C UP LAS fold ) the mixed fi loving, xerophilous, deciduous trees. Most be- long to the Ulmaceae, Cornaceae, Juglandaceae, Moraceae, Leguminosae, Sapindaceae, Anacar diaceae, Rhamnaceae, etc. The most common evergreen broad-leaved trees are Cyclobalanop- sis (Fagaceae), Cinnamomum (Lauraceae), 0s- manthus (Oleaceae), Eriobotrya, and Photinia (Rosaceae) (Fig. 19). Mixed broad-leaved deciduous and ewr green montane forests on acid yellow-brown soil of the subtropical zone. References: Edi Board of Szechuan’s Vegetation, 1980; Jian & al, 1975; Kiang, 1958; Wang, 1964. This type of mixed montane forest is widely distributed on mountain slopes at 1,000 to 1,900 m level in the humid subtropics. Because of lower tempera- tures on mountains, the warm-loving evergreen species of broad-leaved evergreen forests M placed by the cold-resistant ones, mixed “a deciduous broad-leaved trees. The trees kn mixed forests belong mainly to the Fagaceae, eraceae, Betulaceae, Hamamelidaceae, icd ceae, Araliaceae, Magnoliaceae, Lauracea®, Theaceae, Ericaceae, Symplocaceae, s aceae, Rosaceae, Oleaceae, and Tiliaceae. forests are charactized by the presence humid-loving trees, including Fagus ° ngl - F. longipetiolata, and F. lucida, species tha 2 of China. It not seen in be noted that these forests embrace many chaf acteristic species such as Cercidiphyllum JO" icum var. sinense, Tetracentron sinense, schneidera sinensis, Daviaia 1177. Aes: Camptotheca acuminata, Nyssa sine chi- culus chinensis, A. wilsonii, Liriodenna nense, and Emmenopterys henryi (Fi: he sib 7. Broad-leaved evergreen for es oft of References: Edito (1973 Lin et al. (1965); Liu (1939); Shan and Liu a sified as belonging mainly to tw group contains evergreen o orests, is i 515, mainly of Cyclobalanopsis, Casta"? p involucrals HOU — VEGETATION OF CHINA — asa D. were deciduous we ieget leaved evergreen forest is found on the 100 m elevatio us koraiensis is the typical = whereas the via Fraxinus, Phellodendron Carpinus, and Juglar o 500 m elevation Bon mongolica ANNALS OF THE MISSOURI BOTANICAL GARDEN — rm "a ame i - b Jah tee PD Pen BE »* uL n ee n ad 2005 URE 17. e date Betula ermanii occurs at the upper limit of forest lying between 1,800 an f the temperate re don meado* woodland Md Mii Populus diversifolia is frequently distribute soil with ground water at a depth of several meter HOU — VEGETATION OF CHINA FIGURE 19, (U 1 d broad-leaved deciduous and evergreen on on limestone hills, mixed broa TE. en is es rers me qtii, = i es belong mostly to the Ulmaceae, Anacardiaceae, a aved e are Cyclobalanopsis (Fagaceae), mon ev e), DA “Eriobotrya (Rosace FIGURE 20. ( eii Mixed icons d in n forest iod with beech is found at about Basin. l 800 m elevation in the mountains of the western “abet 524 Lithocarpus, which occur in the northern sub- tropics. The dominant members of this group in the eastern region include Cyclobalanopsis glau- ca, Castanopsis eyrei, C. sclerophylla, and Litho- carpus glaber, and are often intermingled with coides, C. delavayi, Castanopsis delavayi, C. or- thacantha, Lithocarpus dealbatus, but without species of Fagus. The second group consists of mixed evergreen broad-leaved forests, dominat- ed by warm-loving species of Castanopsis, as well as those of Lauraceae and Theaceae. In the east- ern region, the dominant trees are Castanopsis hysterix, C. carlesii, C. faberi, C. armata, C. chi- nensis, C. concinna, C. fissa, Schima superba, Adinandra bockiana, Polyspora axillaris (Gor- donia axillaris), Ternstroemia gymnanthera, Hartia sinensis, Eurya loquaiana, E. pseudo- polyneura, Cinnamomum parthenoxylum, C. japonicum, C. chingii, and Phoebe bournei. In addition, there are some species of Hamameli- daceae and Elaeocarpacaeae, which are frequent- ly mixed with species of Fagus. In the western subtropical region, the forests consist mainly of Castanopsis orthacantha, C. hystrix, Machilus yunnanensis var. duclouxii, M. kurzii, Actino- daphne reticulata, Schima wallichii, S. argentea, Manglietia insignis, Magnolia championii, and Illicium yunnanense (Fig. 21) ergreen sclerophyllous woodlands is restricted mainly to ANNALS OF THE MISSOURI BOTANICAL GARDEN the dry, sunny slopes on the mountains of the eastern Chinghai-Tibet Plateau at elevation: ranging from 1,000 to 4,200 m. The climate is warm with a distinct dry season, and the annual precipitation is less than 1,000 mm. Under these climatic conditions, the dominant evergreen oaks have small, thickish, leathery, spiny leaves that prevent excessive transpiration. Thus, they form hard-leaved woodlands similar to those around the Mediterranean Sea. The oaks are Quercus semecarpifolia, Q. aquifolioides, Q. spinosa, Q. longispica, and Q. senescens. They can form for- ests, woodlands, or scrub in various habitats on the mountains; however, dwarf woodlands art commonly seen (Fig. 22). 10. fc mhnn MOO hrania tropical zones. Reference: Editorial Board of Szechuan’s Vegetation (1980). These bamboo forests are extensively distributed in subtropical and tropical China. There are many Species of bamboos. Some bamboo forests are natural, but many are man-made. Different species adapt t0 different habitats. Species of Sinarundinaria are usually found naturally at high elevations in the have many edible, i5 whereas cens, the young shoots of which are more commonly planted on acid soils, Sinocalmus affinis is grown on calcareous i n marshy soils near river banks, Phyllostac congesta is frequently encountered. b 11. Tropical broad-leaved semi-ewrgree jor ests (tropical semi-evergreen seasona and e. References: Chang et al. (1955); pes Chu (1955); Kwangtung Institute of pon (1976); Li (1956); Li et al. (1964). These a0 are very similar in structure and appearan be tropical rain forests, but as the dry anoibt comes more distinct they merge into type. Because the dominant trees are ergreen and partly deciduous, these i known as *semi-evergreen forests. — we differs from the tropical rain forests 10 x characteristics. The trees of the upper p but lower and very few of the larger trees“ — tressed. Furthermore, lianas and pipi pret. noticeably scarcer than in the tropical rain s" These forests can be classified into bu í One group growing on limestone soils is ' »i spread throughout southwestern K posed of southern Yunnan provinces. It is com joli calcium-loving trees, such as U/mus ian uit Gironniera nitida, Celtis austro-sinensis i HOU — VEGETATION OF CHINA FIGURE 21 (U d on the acid soil in the subtropical region is dom- . pper) Broad-leaved evergreen forest found on th ep I inated by evergreen oaks, consisting of Castanopsis, Lithocarpus, Cyclobalanopsis, and of Schima superba (with White-colored flowers). IGURE 22 wer) On the lee side of wet wind in the mountains of western Szechuan Basin, sclerophyllous d Wood ii. land dominated by Quercus aquifolioides is found. 526 ceae), Colona sinica, Burretiodendron hsienmu (Tiliaceae), Pometia tomentosa, Pseudonephe- lium confine, Delavaya yunnanensis (Sapinda- ceae), Muricococcum sinense, Drypetes perreti- culata, D. confertiflorum (Euphorbiaceae), Garcinia tinctoria, G. paucinervis (Guttiferae), Caesalpinia sepiaria (Leguminosae), etc. The second group is found on acid soils, mainly on western Hainan Island. It consists of trees such as Kleinhovia hospita, Pterospermum hetero- Hainania trechosperma, and Microcos panicu- lata (Tiliaceae). 12. Tropical broad-leaved evergreen rain for- ests. Reference: Kwangtung Institute of Botany (1976). Tropical rain forests are found locally on the eastern sides of Hainan and Taiwan Islands anni in eee Yunnan Province where the lo- throughout the year. The ERN are E ai by nu- merous species of evergreen trees. Giant trees, usually clothed with ferns, mosses, and epiphytes onging to Orchidaceae and Araceae, reach a height of more than 30 m or more and often exhibit plank-buttresses and cauliflory. Stran- glers are also characteristic These tropical rain PR are remarkable in raceae, Myrtaceae, aceae, Apocynaceae, Sterculiaceae, ui rog Palmae, Rubiaceae, Myrsinaceae, Leguminosae, Dipterocarpaceae, Meliaceae, Sapindaceae, Aquifoliaceae, Thea- ceae, Rutaceae, Fagaceae, Proteaceae, and Sa- mydaceae (Flacourtia aceae). Some represenative species found in tropical rain forests in different localities are: Vatica as- trotricha, Amesiodendron chinense, and Tarrie- tia panin in southern Hainan; Diptero nee yunnanens s, Crypteronia paniculata, Pomet unnan; and Myristica cagayanensis, Pterospermum niveum, Sideroxylon duclitan, and Artocarpus lanceolata, in southeastern Taiwan (Fig. 23). SCRUB AND COPPICEWOODS 1. Broad-leaved deciduous scrub of the tem- perate zone. Reference: Chen (1958). Leaf- shedding by shrubs of the temperate zone in p" Ww ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 northern and northwestern China is an adapta- tion to a cold season. The broad-leaved decid- uous scrub is actually composed of two groups. One consists of primary vegetation that occurs on semi-arid sand dunes, with species of Cara- gana, Salix, and Artemisia, and on arid saline meadow soil mainly with species of T: amarix The second, consisting of secondary vegetation formed after the felling of temperate forests, 1- cludes northern and southern scrubs. The north- ern temperate shrubs are dominated by Corylus heterophylla, Lespedeza bicolor, and coppice- wood of Quercus mongolica, whereas the south- ern temperate shrubs are dominated by Vitex negundo, Zizyphus jujuba, Cotinus coggygria var cinerea, Coriaria sinica, Dalbergia hupeana, and wheat strobilacea (Fig. 24). 2. Mixed broad-leaved evergreen and decidu- ous aus on acid soils of subtropical and tropical ical and tropical forests after felling or fires, an is characterized by sunny and acid-loving plan's In the subtropical zone, scrub vegetation is dom- inated by shrubs belonging to the Ericaceat, Theaceae, Hamamelidaceae, and Fagaceae. Rho- dodendron simsii, R. ecorum, Vaccinium o long to the Myrtaceae, Euphorbiaceae, e caceae, Sterculiaceae, Rubiaceae, Theaceae Re cies 4 myrtus pn eh candidum, M. normale, Apo i nensis, A. yunnanensis, Schima villosa, in lichii, Phyllanthus embrica, and Psychotri ra. 3. Mixed broad-leaved evergreen and deci ia af ous thorny scrub on limestone a of subtrop and tropical zones. Reference: Ho 1 ( scrub forms successional comm stone regions in subtropical and tropi Calcareous limestone soil is excessively due to sinkholes, and therefore has low holding capacity. The shrubs on these $9 hibit greater defoliation and a mark rosa chi- al- wate! ils e* and high lime content of the soil. In the subtropical zone, the plants the Rosaceae, Rutaceae, Caprifoliaceae. X o Ó((a— MÀ M I Á— A g, — — * n HOU — VEGETATION OF CHINA * 7 å $ d p QURE 23. During the dry season in Yunnan Province, because of the presence of clouds that compensate là à shortage of rainfall in the winter, tropical rain forest is found locally in the valley. The tree is Antiaris Xicaria (Moraceae) ts Anacardiaceae, Rhamnaceae, Legumi- reticulatus, Prinsepia utilis, and Platycarya stro- Ju €, Euphorbiaceae, Ulmaceae, Meliaceae, bilacea. In addition to the above-mentioned glandaceae, etc, Typical genera are Rosa, Ru- families, shrubs occurring in the tropical zone yl also consist of plants belonging to the Moraceae, Urticaceae, Sapotaceae, Sterculiaceae, Apocy- naceae, Flacourtiaceae, Sapindaceae, etc. The 528 representative species include coppices of Ficus spp., Alchornea trewioides, Boehmeria nivea, Clausena excavata, Sapium rotundifolium, Ster- culia lanceolata, Desmos cochinchinensis, Pleo- mele nepalensis, Pterospermum heterophyllum, Terminalia hainanensis, and Wrightia haina- nensis. . Broad evergreen sclerophyllous forests or scrub (mangrove) of the littoral subtropical and tropical zones. Reference: Kwangtung Institute of Botany (1976). Along tropical and subtropical coasts of southern China, mangrove forests or scrub are distributed in muddy tidal marshes of estuaries. The best-developed mangrove forests, which can reach 10 to 15 m and have climbers, are found in the eastern part of Hainan Island at about 19°N. With increasing latitude, the num- ber of species decreases until one species, Avi- cennia marina, remains. It occurs as 1 to 2 m- high shrubs at 27°N in North Fukien Province. The evergreen leaves of mangroves possess a marked xerophilous structure and succulent character adapted to soil salinity. The chief man- grove genera are Rhizophora, Bruguiera, and Ceriops. They have viviparous seedlings and stilt roots or pneumatophores. On the northern coast of the transitional trop- ics occur only five species, namely Avicennia ma- rina, Acanthus ilicifolius, Kandelia candel, Bru- guiera conjugata, Aegiceras corniculatum, which form shrubs less than 3 m tall. On the southern coast of Hainan Island more species of mangrove are present, namely, Rhizophora apiculata, R. era racemosa, Excoecaria agallocha, Hibiscus ti- liaceus, Sonneratia caseolaris, Xylocarpus gra- natum, Scyphiphora hydrophyllacea, an Heritiera littoralis. These mangrove species can reach 10 to 15 m, but are usually 5 m tall, and can include some climbers, namely Dischidia chinensis, Derris trifoliata, Caesalpinia nuga, and Stenochlaena ilicifolius (Fig. 25). 5. Broad-leaved evergreen succulent scrub on coral islands of the tropical zone. Reference: Chang (1974). In the South Sea of China are numerous coral islands where evaporation is greater than precipitation because of strong winds spiration. The vegetation is poor in floristic com- position. Plants with thick succulent leaves are ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 namely Scaevola frutescens, S. hainanensis, dis is found in the center of the coral islands. . Broad-leaved evergreen semi-sclerophyllous thickets (shrublands) on the mountains of tem- perate and subtropical zones. Reference: Zhuo (1975). Areas dominated by semi-sclerophyllous beyond the limits of arborescent growth in the eastern Chinghai-Tibet plateau in alpine regions from 3,200 to 5,000 m in elevation. Typically, these communities are dense and are composed of dwarf, evergreen shrubs 2 to 15 dm tall. The shrubs are usually creeping, and their leaves are often leathery, thus reducing transpiration. Rho- dodendron species mixed with Sinarundinaria are frequently seen in the subalpine belt - they are always secondary after the felling nm alpine conift L i : such as species of Salix, Potentilla, and Cara gana, are also associated. : 7. Broad-leaved deciduous thickets po shrublands) on the high mountains of dien and subtropical zones. Reference: amicos Wang (1966). Alpine broad-leaved deci¢ a shrublands are distributed on the pp the eastern Chinghai-Tibet Plateau ce s northwestern high mountains. Tum gei in the high m pend Pe shady slopes in t e Banc non : a t the alpine meadows are seen on sunny slopes i on same elevation. The thickets are gp posed of Salix oritrepha, S. cupularis, 4 jubata, and Potentilla fruticosa. i 8. Tundras with evergreen Wer high erences: mosses (mountain dwarf-shrub tundra, ; mountains in the temperate zones. Re a), The Chen et al. (1964); Chow and Li (1969) mountain dwarf-shrub tundra is poorly oul oped on the summit of the temperate b seat tains of the Changbaishan, Altai, an dims Khingan ranges. The soil is acid, and the ah. is characterized by high eum strong a short growing season, an ev- The vegetation consists of dense £r ows UE ergreen dwarf shrubs, mainly belong!” th small Ericaceae, about 10 to 20 cm high and wit" and tough 1 R. confertissimum, R. Fei a doce caerulea, Arctous japonicus, » aa as biricum, and deciduous dwarf shru | HOU — VEGETATION OF CHINA (Upper) Broad-leaved deciduous scrub is the secondary growth of the warm-temperate forest. . rea ady slopes, whereas Ziz zyphus ju ujuba is found on sunny slopes. (Lower) The best-developed mangrove, with 16 or 17 species, is encountered in the eastern part Island. The mangrove trees or shrubs have stilt roots. 530 Arctous ruber, Salix rotundifolia, S. polyadenia, Betula rotundifolia, and mosses and lichens (Fig. 26 9. Alpine deciduous cushion sub-shrubs mixed with herbs of temperate and subtropical zones. Reference: Chang and Wang (1966). The cushion sub-shrubs occur principally on the summits of mountains on the northwestern Chinghai-Tibet Plateau from 4,800 to 5,200 m in elevation. The plants assume the form of creeping dwarf sub- shrubs 2 to 5 cm high with richly developed, branched, horizontal axes that spread over the ground. The branches of these cushion plants are so compacted together that they touch each other on all sides. The climatic conditions under which this vegetation type grows resembles those of the mountain dwarf-shrub tundra in coldness and strong wind. However, they differ from those in that the humidity and precipitation are much lower, and the soil is alkaline. Because of this, acid-loving, ericaceous plants are absent, and the floristic composition of this vegetation type is entirely different. The plants belong mostly to the Caryophyllaceae (Arenaria musiformis, A. diapensoides), Primulaceae (Androsace tapete), Leguminosae (Caragana saja, Oxytropis spp.), Compositae (Saussurea involucrata), Rosaceae (Potentilla spp.), as well as species of Cyperaceae and Gramineae. DESERTS AND XEROPHYTIC SHRUBS 1. Temperate deciduous dwarf semi-shrubby deserts. Reference: Integrated Survey Team of Sinkiang and Institute of Botany, Academia Si- nica (1978). These deserts are stony, hilly, or sandy-pebbly with an extremely arid climate, and occur in western China. Only a few low-growing xero-halophytic plants occur infrequently in the cracks of rocks and in erosion gullys. However, a vast area of hills and the gravelly piedmont is devoid of vegetation. The soil contains a certain amount of gypsum. Dominant plants are Salsola passerina, Reaumuria soongarica, Sympegma regelii, Iljinia regelii, Anabasis salsa, A. brevi- folia, and Nanophyton erinaceum. All these species are xerophytic halophytes, and the first two are gypsum plants that contain 1-4% of sul- fur (Fig. 27). 2. Temperate dwarf hairy sub-shrub deserts with ephemeral forbs. References: Integrated Survey Team of Sinkiang and Institute of Bot- any, Academia Sinica (1978); Li (1961). This type of desert is found on loess-like, salt-free ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 soils. It occurs on the lower slopes of mountains in the Dsungaria Basin where, although the an- nual precipitation is not more than 200 mm, it is distributed evenly throughout the year. The vegetation is predominantly composed of hairy dwarf-shrubs, namely, Artemesia kaschgarica, A. borotalensis, and A. terrae-albae, as well as some ephemerals that develop in spring, when the soil is moist, but fade quickly with the coming of summer. The perennial ephemerals commonly seen are Poa bulbosa var. vivipara and Carex physodes. Annuals are Trigonella arcuata, Me- niocus linifolius, Tetracme quadricornis, and Le pidium perfoliatum, as well as Ferula spp., which are occasionally encountered. 3. Temperate succulent halophytic dwarf sub- References: Integrated Survey 30% NaCl or Na,SO,. Halophytes, such et A lidium foliatum, K. cuspidatum, K. gracie . ne- caspicum, Halostachys belangeriana, meon soil. à Temperate shrub and sub-shrub de References: Hwang et al. (1962); any, Academia Sinica (1978). This type is widespread on the sand dunes and p! plains in the arid region. In the eastern ye sd Gobi plains with about 200 mm annual ead itation, the predominant shrubs are Can =. tibetica, Tetraena mongolica, Potaninia E golica, Ceratoides latens, and oe. : mopiptanthus mongolicus. On the san which can retain larger amounts of water, aciphylla, Caragana microphylla, shinskii occur. In the extremely arie western region, Ephedra przewalskii, lum xanthozylon, Z. kaschgaricum. = sphaerocarpa, and Calligonium roboro scarce on the piedmont plains along thec the suf- of temporary streams that are fed by | 1983] HOU — VEGETATION OF CHINA 531 FIGURE. 26, (Upper) Mountain dwarf-shrub-tundra, dominated by Vaccinium vitis-idaea, Rhododendron as reum, and ohare species of nd is seen above 2,000 m on the upper mountinas of Changpaishan in . tern China, above the for sts. n IGURE 27. (Lower) illy d lacking all vegetation, in the Turfan Basin weg annual precipitation usually is less than 20 mm. However, in the oasis, grapes grow very well under irrigatio 532 rounding snow-covered, high mountains. The sandy Takla-Makan desert is essentially devoid of vegetation, but Tamarix spp. are sparsely dis- tributed in the dune valleys (Figs. 28, 29, 30). 5. Temperate semi-arboreal deserts (Saksaoul communities). References: Hu (1963); Inte- grated Survey Team of Sinkiang and Institute of Botany, Academia Sinica (1978). Haloxylon am- modendron and H. persicum are characteristic plants growing on the sandy deserts. Originally, H. ammodendron was widely distributed in the arid region of China, but it is now scarce in the Darim Basin. This species is a halophyte and is always associated with numerous saline plants. It usually grows on the bottom of sand dunes with moist and slightly saline soil or on salty soil with ground water at a depth of 5 to 8 m, which can be reached by the roots of this plant. This low tree can attain a height of 5 to 7 m and form small woods. Haloxylon persicum, a non-saline plant, is confined to the slopes or ridges of sta- bilized or semi-stabilized sand dunes in the Dsu ia Basin, where annual precipitation is distributed evenly throughout the year. Its growth is dependent upon rainfall, but not ground water. It often grows together with Calligonum leuco- cladum, Artemisia santolina, and A. terrae-al- bae, as well as some spri h Is (Figs. 31, 32) 1 6. Cold high-mountain deserts with creeping or matted dwarf sub-shrubs of the temperate zone (cold high-plateau deserts). Reference: Inte- grated Survey team of Sinkiang and Institute of ; » Oxytropis poncinsii (Leguminosae), Ajania tibetica, A. scharn and Saussuria glandulifera (Compositae). They are all mostly 4 to 15 cm tall (Fig. 33). 7. Tropical and subtropical succulent thorny ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 shrubs. Reference: Editorial Board of Sze- chuan’s Vegetation (1980). This type of vegeta- tion, which always forms a complex with savan- i bt icalandt ical China. is ewhat allied to the southern Asiatic deserts. Because the potential evaporation is much greater than the precipitation in places where this type of vegetation occurs, the vegetation consists of xe rophytes and succulents that require a certain minimum amount of water uptake even in pe riods of drought. Therefore, under warm and climatic conditions, plants with succulent stems such as Opuntia ficus-indica and Euphorbia roy- leana, and those with succulent leaves such 4s Aloe vera var. chinensis, Kalanchoe pinnata, etc. predominate. Additional plants are Zizyp mauritiana, Z. montana, Acacia farnesiand, Campylotropis yunnanensis, C. mue leana, Pandus tectorius, and Phoenix hanceana, which are shrubs that have small, hard leaves or are clothed with hairs or spinules that offer some protection from excessive transpiration. STEPPES AND SAVANNAS 1. Temperate forb-grass steppes phytic steppes or meadow steppes). Li (1962); Zhu (1958). This type of ste" found in the temperate zone under a semi-h io climate with an annual precipitation of 35 500 mm. It is composed mainly of mee ora l aem i z D E me umer- subordinate tussock grasses mixed with n : as ous mesophytic herbs. It also 1$ eque panulaceae, etc. Represen : are Aneurolepidium chinense, Stipa Filifolium sibiricum, etc., us shrubs including Prunus sibirica, ja licus, and Ulmus pumila in the no 4 south are Bothriochloa ischaemum ame meda triandra var. japonica, which are hus sati ly associated with shrubs such as Zizyp -—À var. spinosus, Sophora viciifolia, Ha m rhamnoides, Rosa hugonis, and Prinse? flora (Fig. 34). erences 2 Pb needlegrass steppes. Re teppe is Li (1962); Wang (1963). This type er inset distributed in the semi-arid region of nest Mongolia Plateau and Loess Plateau wi baicalensis mixed with jx mo T: 4 | — HOU — VEGETATION OF CHINA URE28. (Upper) Zygophyllum, Calligonum, seque are sparsely distributed on the piedmont plains in rary te. channels of tempo runoff in the southern regi 29 (Lowen Ephedra przewalskii occurs on ja 0 piedmont plains in the desert region. 534 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo.. 70 FiGURE 30. (Upper) The wanderin marix spp. can be found very infrequ m above the ground. FiGURE 31. (Lower) In the Soongaria Basin of the to the slopes or ridges of semi-stabilized or stabilize 8 sandy desert Takla-Makan is essentially devoid of equation IR a ently in the dune hollows. This air photograph was taken fined desert ten, the growth of Haloxylon persicum nee d sand dun 1983] HOU — VEGETATION OF CHINA 535 had a ground FIGURE E Haarala saline on Precipitation is 250 to 350 mm. The vege- "s I5 principally composed of xerophytic tus- £s Sses, such as Stipa grandis, S. krylovii, reviflora, and Cleistogenes squarrosa, that Possess very extensive, finely branched root sys- water py at a depth of 5 to 8 m, which cs can 1 be reached by roots a the E et It peni iy is pud on moist and tems and narrow leaves that regulate their tran- spiration by inrolling. On sandy soil, sub-shrubs, including Artemisia frigida, onn prostrata, and Thymus serphyllum, and spiny shrubs, consist- ing of Caragana microphylla, € stenophylla, etc., ANNALS OF THE MISSOURI BOTANICAL GARDEN n Da Hi DN FIGURE 33. (Upper) The high -cold desert o orthwesternmost th e wrong coran Plateau has v ccurring at 5,000 m elevation in the n acta wi ibet 00 ery few species of plants. The dominant plants are Ceratoides comp . e easter ower)Mesoxerophytic or meadow — is distributed on the North East Plain and th pepo” Mongolia Plateau, areas d n The dominan Aneurolepidium chinense, and for bs are nume ates 2°) vo AM aes Oe ae 1983] are frequently found. In mountains of the desert region, steppes dominated by Stipa capillata, S. krylovii, and Festuca sulcata, also occupy a vast area (Fig. 35). 3. Temperate dwarf-shrub or sub-shrub nee- or semi-des- the annual precipitation varies from 150 to 250 mm, and where the winters are dry and cold. The vegetation consists mainly of xerophytic sub- shrubs and dwarf grasses about 10 to 20 cm high. The extremely xerophytic grasses most frequent- ly encountered are Stipa gobica, S. glareosa, S. sida, A. xerophytica, Ajania achilleoides, A. tri- colia, A. fruticulosa, Ceratoides latens, Salsola herbs of this type of steppe (Fig. 36). This type of vegetation also is found on mountains of the region. 4. Subtropical and tropical savannas. Refer- ences: Editorial Board of Szechuan's Vegetation (1980); Kwangtung Institute of Botany (1976). Savanna-like vegetation in China is confined to à limited part of the subtropics and tropics. Sa- m can be defined as homogeneous vegeta- on, consisting mainly of warm-loving and me- ‘oxerophytic tussock grasses with a more or less M scattering of tall and small, evergreen and *ciduous trees and thorny scrubs. The domi- ving grasses are Heteropogon con- » Cymbopogon distans, etc. Savannas are eres in the subtropical zone in deep valleys iod Hungtuan Mountains with locally dry and Zi climate. There, the grasses are associated with PR. ce mauritiana, Acacia farnesiana, and ted inia racemosa. Another group is distrib- Tai on the sandy, tropical coasts of Hainan and viti P Islands where the grasses occur mixed bizzi acourtia indica, Pandanus tectorius, Al- d Procera, and Phoenix hanceana (Fig. 37). x b: High-mountain steppes of temperate and á Topical zones (cold high-mountain *Ppes). References: Chang (1963); Wang and HOU —VEGETATION OF CHINA 337 Li (1981). The major portion of western Ching- hai and east central Chang Tang from 4,200 and 5,300 m above sea level in the Chinghai-Tibet Plateau is characterized by a continental type of climate with an annual precipitation of 200 to 300 mm. The annual mean temperature is below 0°C, and the wind is strong throughout the year. Low temperature is overcome during sunny weather by the int insolation; thus daily tem- perature fluctuation is great. Under this dry and high-cold climate, the vegetation is composed mainly of cold-xerophytic, dense tussock grasses less than 20 cm tall and dwarf sub-shrubs with small, thick, hairy leaves. The dominants are Stipa purpurea, S. subsessiliflora var. basiplu- mosa, Festuca olgae, Carex moorcroftii, and Ar- temisia wellbyi, mixed with alpine elements such as Kobresia spp., Thalictrum alpinum, Leonto- podium alpinum, and Thermopsis alpina (Fig. 38) MEADOW AND SWAMP VEGETATION 1. Temperate meadows. References: Chen et al. (1954); Hou et al. (1953); Hou (1954); Wang and Li (1981). Most temperate meadows are not genuinely virginal, but, except for some saline meadows, are communities established under the disruptive influence of human activity and do- mestic animals. The plants of this vegetation type are mostly mesophytes, but the floristic com- position varies with soils. On neutral or calcar- eous soils occur Phragmites australis, Calama- grostic epigejos, Hemarthria compressa var. japonica, Deyeuxia langsdorffii, D. hirsuta, and D. angustifolia, along with many species of forbs belonging to different families. However, there are numerous species of forbs, such as Sangui- mountain slopes after the cutting of forests. The floristic composition of halophytic mead- ows is entirely different from those mentioned above. This type of meadow is widespread in humid, semi-arid, and arid regions of the tem- perate zone in tidal estuaries along the coasts and inland salt marshes where many species of halo- phytic grasses and herbs are the same. They are Aeluropus littoralis, Aneurolepidium dasysta- chys, Puccinellia distans, Scorzonera mongolica var. putjae, Xanthium sibiricum, Suaeda salsa, S. glauca, Atriplex littoralis, A. sibirica, Salicor- nia europaea, Limonium aureum, Polygonum passerina, etc. Allium indere is a mm. FiGURE 35. about 200 m ANNALS OF THE MISSOURI BOTANICAL GARDEN (Upper) The t Aon Steppe of Stipa grandis occurs in Central Inner Mong obica, $. Tt steppe has dwarf Hire about 10 to 20 cm tall, nhu ` Stipa g Salsola tf semi-shrubs a € Artemisia frigida, Ajania spp., Ceratoides AN rainfall i$ characteristic herb of this type of steppe. The annual m HOU — VEGETATION OF CHINA eu 37. (Upper) On the lower slopes on the lee side of the Hengtuan Mountains in western Szechuan unnan pr ment © mets ical savannas are found. The dominant grasses are Heteropogon contortus, nus yunnanensis, Zizyphus mauritiana and Acacia farnesiana. is r^ RE 38. andis Cold high-mountain steppe, occurring in the seii part of the Chinghai-Tibet Plateau, ominaied by Stipa purpurea. and S. iste Sn var. plumosa 540 sibiricum, and Triglochin maritimum. Ach- nantherum splendens, Karelinia caspica, and Al- hagi pseudalhagi are confined to the arid region. However, the associated halophytic shrubs and «11h-«h SUD gl Win vary- ing moisture conditions. 2. Cold high-mountain meadows of temperate Tibet Plateau beyond the limit of tree growth at elevations from 4,000 to 5,200 m and with an annual precipitation of 350 to 500 mm. They are mainly composed of numerous species of Ko- bresia, such as K. pygmaea, K. tibetica, K. roy- leana, and K. humilis and y alpine fort I as Polygonum viviparum, Thalictrum alpinum, Pedicularis spp., and some alpine species of grasses. Successional subalpine meadows rich in forbs and grasses, with a few species of Kobresia, follow deforestation of subalpine meadows (Fig. 39). 3. Subtropical and tropical meadows. Refer- ence: Kwangtung Institute of Botany (1 976). The meadow vegetation or mesophytic grassland of humid subtropical and tropical zones is, in gen- eral, not a climatic climax, but rather a seral stage of succession formed after the felling of the for- ests. However, it can gradually merge into forests with the i i farb f . Thegrassy or herbaceous taxa are sensitive to the combined influence of soil acidity and air temperature. The taxa of meadows found on subtropical acid soils are typified by numerous graminoids, such as ever, on tropical, deep, i on open roadsides. The tropical acid meadows are usually composed of shrubs belonging to the Euphorbiaceae, Melastomataceae, etc. n the calcium or limestone soils the grami- noids are represented by Sporobolus elongatus, Bothriochloa intermedia, Neyraudia reynandi- ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 ana, Heteropogon contortus, Pogonatherum crinitum, Capillipedium parviflorum, and Ar- 1hrnvnn hicenirhice Q 1 +t} A i tad by pure stands of Cyclosorus acuminatus, with Pteris vittata in the cracks of limestone rocks. Nevertheless, Chrysopogon aciculatus, Hetero- pogon contortus, Apluda mutica, Kyllinga bre- vifolia, and Scleria levis can be considered rep- resentative species on tropical limestone soils. On tropical, coastal, sandy saline soils, often affected by sea spray, the meadow vegetation consists of graminoids, including Spinifex litto- reus, Sporobolus virginicus, Perotis indica, Cy- perus malaccensis var. brevifolius, together with forbs such as Ipomaea pes-caprae, Wedelia pros- trata, Limonium sinense, Suaeda australis, and shrubby Vitex trifolia var. unifoliata (Fig. 40). 4. Swamp The term “swamp der similar climatic conditions, they are 9!" entiated by edaphic factors. VEGETATIONAL REGIONS OF CHINA The longitudinal zonation of vegetation - temperate zone, which is closely related ie: is east-to-west variation in annual precipitation, o 3» China, the different forest regions correspo Fie the latitudinal zonation from north to south 43). In the mountainous areas of the country zonation of the vegetation is vertical. Based on the principles of vegetatio d tion, the country can be divided into eight 1960, tational regions (Fig. 44). References: Hou( 1961, 1964, 1981a, 1981b, 1982). leaved 1. Cold-temperate deciduous needle- forest region. This region occupies the -g P north of China, embracing a very small : i Moun annual precipitation of 500 mm. nual temperature is below 0°C, and predom minimum is — 50°C. The forest consists lic sol inantly of Larix gmelini on acid podzo ori Many cultivated crops and fruit trees do? par- en in this region; however, potatoes, spenn ley, and cabbage grow well. 2. Temperate deciduous broad-leaved forts | HOU— VEGETATION OF CHINA NS TUSA 4 itu - « NP eut NN » - 59999 ws AN di "SURE 39. (Upper) Cold high-mountain steppe meadow is found at about 5,000 m elevation in the recon y i Plateau. This type of vegetation consists of numerous species of Kobresia such as K. pygmae . tib wer) hh ia cadi de is frequently encountered in the ground layer of Pinus masso- a "y Woodland. | n places, fern sS pere tands. It, along with various associated plants, indicates strongly 1d soil in the humid subtropics e tropi cvs isana S Feds tp — N3GPIVO 'IVOINV.LOS I3010SSIN AHL AO STVNNV een ar — — - " - — - - 000 q^ A Bue eC „50-50 Rainfall 60-250 (mm) 100-506 z ipee Humid WEST } -- - Arid Region - - — - Extremely Arid Region, — - - - Arid Region-- —,- - - - Semi-arid Region- — 4-- Region —- -Region--] EAST (Semi-desert, Desert) (Desert) (Semi-desert, Desert) (Steppe) "Np SR (Broad-leaved PR Tienshan Dsungaria eciduous Forest) gm Ys Basin Alasha Plateau Inner Mongolia Plateau Northeast | Changpei Mountains 3,000 m in 1,500 m = A) Doe cH Plai m 4 ' E.L. Gray-desert Soil Bro Soil | Brown-desert Soil i Chernozem x Dark Brown y | wn-desert ! i Chestnut Soil Dark Chestnut Soil Soil | 80 90° 100° 110° 120° 130° FiGURE 42. Schematic rainfall and soils in relation to vegetational regions varying from west to east in the temperate zone of China (about 42°N). VNIHO JO NOLLV.LH3O3A — NOH [£861 evs {400-2008 0-6 600-900 900 gem PE 5,000-5* 500: 500-100-99? 5,200- e 15 nr, > i e -32 -14 m. gerunt 2t re. ee -——— T Vor sum 270 - ap eR oe 25:9 Dg io - 29 130 - 240 | o EA 100 - 180 Transiti onal Warm-temperate Broadleaved —— apg oasian peara K „— Tropical Deciduous Forest tropic ergreen Fores Needle-leaved Mixed Broad-lea Deciduous Deciduous & N Needle-teaved I Evergreen Fores Changpei Mountains North China Plain r U L i i l l ! D [ L D | Chingling : h t : isse Taiposhan South Yangtze Hills niea cio roma E. i i l NORTH | i 1 I i l Red z gr ded Earth, [ Rend SOUTH N.L. Brown Conifer : ı Brown Earth, Korichnevie | Soil Dark Brown Soil x Alluv ial Soil i 53° 50° 45 35 I Dire Soil \ | Laterite Soil 18 le N V , to LLO°E FIGURE 43. Schematic climate and soils in relation to vegetational regions varying from north to south in the eastern section of China (about 120° ). brs Nadav) 'IVOINV.LOH I310SSIN AHL AO STVNNV 0L noA] eS oO 1983] HOU— VEGETATION OF CHINA 545 LEGEND D E—TLTÓ——A———J — — —1. Eu RU, TS it Ds T MIT, | sm jd Taipei ON, fox : eet 6 I raed s , ES Sey ete iton uf SX ea". s ROY I KARR Oe 4 6 Sy i nd QS EUR QS * EXD F RY I Oi p 2) Temperate deciduous broad-leaved forest r gion; 3) Subtropical evergreen broad-leaved forest region; -5)T d regi, A P d In this region the annual precipitation e a from 600 to 900 mm. The northern part, € mean annual temperature varying from a dominated by Betula platyphylla; Bet- has a forest; and the dwarf-shrub tundra. Wheat, m eys and alluvial plains, soybeans, spring ed din aize, kaoliang, millet, and rice are plant- à year «s ge Southern part of Region 2 principally cov- needle-leaved forest is dominated by Pi- nus tabulaeformis or P. densiflora and grows on acid, brown forest soil. Mixed deciduous broad- leaved forest and the Platycladus orientalis forest thrive on the neutral or calcareous Korichnevie soil; Sali i widely distributed along the sea coast. The land is almost entirely unde cultivation and usually is planted three times in two years with winter wheat, cotton, peanuts, sweet potatoes, maize, kaoliang, and millet. Rice is locally planted once a year. Region 2 is the main production area in China of deciduous fruits, namely, apples, pears, grapes, persimmons, chestnuts, walnuts, and Chinese tes. 3. Subtropical evergreen broad-leaved forest region. Occupying a vast area of the south- weste d heastern part of th try, this region is characterized by different types of sub- tropical forests. In the eastern subregion, the cli- mate is moist and warm. The annual precipita- tion is 1,000 to 2,000 mm and the mean annual temperature varies from 16° to 20°C. The ever- 546 green broad-leaved forest, dominated by Cy- clobalanopsis glauca, Castanopsis eyrei, Litho- us massoniana an Cunninghamia lanceolata, and the Phyllosta- chys pubescens forest grow on acid, yellow or red, olic soils. On the limestone soils occur mixed deciduous and evergreen broad-leaved forests characterized by trees belonging to the Ulmaceae and some other calcium-loving families. There are also forests dominated by Cupressus funebris, Sinocalmus affinis, etc. The vertical distribution of the vegetation can be represented by the east- ern slope of Erlanshan where evergreen broad- leaved forest, mixed evergreen and deciduous broad-leaved forest, Abies fabri forest, and Sin- arundinaria scrub appear successively from the base to the 3,300 m top of the mountain. In the western subregion where the climate is warm and rather dry with an annual precipita- tion less than 1,000 mm, sclerophyllous ever- green oak forest, dominated by Quercus aqui- do ny slopes, and succulent thorn scrub dominated by Opuntia fi- cus-indica and Euphorbia royleana occurs in the deep, dry valleys. There the evergreen broad- leaved forest consists of Cyclobalanopsis glau- coides, Castanopsis delavayi, and Lithocarpus dealbatus. Pinus yunnanensis, P. armandii, and E dp? deae (C 1unre«s«oius "T: p the dominant trees of the evergreen needle-leaved forest. In the eastern subregion two crops consisting of summer rice and winter wheat or rapeseed are usually planted each year. Double-cropping of rice, followed by a planting of a cool-loving crop, also can occur. Cotton, ramie, peanuts, sweet tatoes, maize, and sugar cane are the common crops. Subtropical fruits such as Citrus spp., M y- rica rubra, and other cultivated trees such as tea, tea oil, loquat, tung oil, and mulberry are widely grown. In the western subregion, however, sum- mer rice or maize and winter wheat or rapeseed are planted annually, whereas the fruit trees are mainly deciduous. 4. Tropical seasonal rain forest region. This nually. In well-sheltered valleys one can find thick, broad-leaved evergreen rain forest with a wide variety of trees of different heights, some up to 40 m. The massive trunks are supported by but- tresses tailer than a man and entwined with thick vines. Lovely ferns and brilliant orchids flourish ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 on some of the trees. Tropical broad-leaved semi- evergreen forest (tropical seasonal forest) occurs on limestone hills or on drier sites. Along the sea coast, mangroves containing Kandelia candel, Rhizophora mucronata, R. apiculata, and Bru- guiera sexangula are confined to saline, muddy, swampy soils. Some species produce a tangle of arching prop roots from their base, which enables them to stand up to strong wind and waves. Oth- er species have air-roots with tiny holes that pro- ject above the mud and supply the submerged roots with air. A number of them produce vr viparous seeds that germinate while still attached to the tree. When these sprouted seeds drop, they readily take root in the shifting mud. In this region rice frequently yields two crops a year. Sweet potatoes, maize, and peanuts can be raised in the winter. Also cultivated are ev- ergreen orchards of tropical fruits, bananas, pine- apples, lichees, longans, mangoes, and betelnuts, as well as plantations of industrial crops : rubber, coconut, coffee, pepper, vanilla, and a palm in the southernmost part. There are fields of sugar cane, manioc, and sisal hemp: i 5. Temperate steppe region. The plain Inner Mongolia Plateau is a vast grassland with with a mesoxerophytic steppe and apei dominated by Aneurolepidium chinense e lifolium sibiricum that grow on loamy che ik em or dark chestnut soils. Sparse mesos tic woodland dominated by Pinus sylvestris "5" mongolica or Ulmus pumila is found on soils. Typical steppe occurs in the wes s of the Inner Mongolian and Loess plateaus, ail the xerophytic perennial grasses, Stipa gr ud S. krylovii, S. bungeana, S. brevifolia, S. 8 pi S. glareosa, and Cleistogenes squarrosa are pi inant. The deciduous scrub, including - : Caragana spp., and Artemisia spp. £' Owe sandy soil or sand dunes of this region. e yii meadows and the annual succulent Md high vegetation frequently occur on soils WI". ground-water tables. Grasses and forage plants in this area are highl making the region China's main livestoc area. The main domestic animals here cattle, and horses. : number of places in the eastern part pn region have been turned into agricult ing sod Their main crops are soy beans, sorghum sugar beet M — L—Á————— 1983] 6. Temperate desert region. This region in- cludes the western parts of Inner Mongolia, Kan- su, the Tsaidam Basin in Chinghai Province, and the whole of Sinkiang. The characteristic feature of the climate is its dryness; the annual mean precipitation is less than 50 to 150 mm in most places. The sandy deserts and areas of gravelly gobi contain a poor variety of plants. Growth is sparse, and there are stretches that are devoid of vegetation. Kalidium spp., Nitraria sibirica, Ha- lostachys belangeriana, and Halocnemum stro- l aceum is confined to solonchak soil. Sympegma regelii grows very sparsely in the rocky desert. Most of the above-mentioned plants make ex- cellent camel fodder. The leaves of some of these plants are vestigial, their function having been taken over by green branches to reduce transpi- ration. i the temperate desert region Populus eu- poo E d, sometimes with Elaeagnus an i la, is found on river banks. The leaves ke : us make an excellent fodder for camels ss he oa This tree is the main source of timber Boe inek Tamarix scrub occurs on saline Soils with a high water table. The vast grasslands on the slopes of the high Mountains in the ili i Ch T d:iffus ilien, ienshan, Kunlung, are timber forests consi tin dt 7 sate On the upper «i sisting mostly of Picea spp. ice and s Slopes. The melting of the perpetual the ELS on the peaks irrigates the oases in lis regio S below. The low-elevation basins in n are centers for growing fruit, such as able f, e à sd Concentrating sugar, the fruits are ex- y sweet. In addition, China's best long- -cold meadow and steppe region. The lui-Tiber pan’ southern portions of the Ching- and hay ateau are 4,000 m above sea level * no summer. The melting glaciers pro- Da marshes, lakes, and meadows with Pygmaea, K. tibetica, etc. The high- HOU — VEGETATION OF CHINA 547 mountain alpine steppe occurs in the central part and is dominated by Stipa purpurea and S. sub- sessiliflora var. basiplumosa. Yaks, Tibetan goats, and sheep are the main domestic animals there. In the valleys of the plateau grow such cold- resistant crops as rapeseed, spring barley, tur- nips, etc. They are locally seen at elevations be- low 3,200 m in the north and 4,200 m in the south. 8. High-cold semi-desert and desert re- gion. The northwestern corner of Tibet, at 5,000 m above sea level, has even sparser vegetation and fewer varieties of plants. Only short, small sub-shrubs are found here. Some of these are eepers, ion plants, forms that are adapted to resisting wind and cold and preserving moisture. Ceratoides compacta and Ajania tibetica are frequently found on the sandy gravelly desert. There is no agriculture in this region. haran th h O = LITERATURE CITED BRAUN-BLANQUET, J. 1932. Plant Sociology: The Study of Plant Communities. McGraw-Hill Book o. Inc. CHANG, H. T. 1974. Vegetation of the Sisha Islands. Acta Botanica Sinica 16(3). Science Press. (In et al. 1955. The plant communities of Ting- hushan, Kaouai county, Kwangtung Province. Chungshan Univ. Bull. No. 3. (In Chinese.) Science Press. (In Chinese.) . 1966. Vegetation of the Cen- tral Tibet. Science Press. (In Chinese.) CHANG, Y. L. 1955. The plant communities of the Khingan Mountains. Contribution to Plant Ecology and Geobotany, No. 1. Science Press. (In Chinese.) CHEN, C. C. 1955. Observation on ecologically mor- phological characteristics of certain alpine plants i Chienlin Mountains, Kansu Pr 4(1). Science Press. (In Chinese.) Y.Chow. 1957. Vegetation ofthe Sulian River valley, Kansu Province ribution to Plant Ecology and Geobotany, No. 15. Science Press. (In Chinese.) CHEN, C. T. 1958. Vegetation of Tsiwuling forest , Kansu Province in relation to soil and water conservation. Contribution to Plant Ecology an Geobotany, No. 2. Science Press. (In Chinese.) , C. T. WANG & H. M. TunG. 1954. The plan com iti f hill d h in [I ienyunkan, Kiangsu Province. Acta Geographica Sinica 20: 285-311. Science Press. (In Chinese.) on the northern slope of the Changpai Mountains, 548 Province. Acta Phytoecologica et Geobo- tanica Sinica 2(2). Science Press. (In Chine: e mental characteristics P tain vegetation in relation to deve opment of forestry and side-lines in Huairou ounty, Peking. Phyt logica et Geobo- tanica Sinica 3(1). Science Press. (In Chinese.) CHIANG . 1963. On problems of phytocoeno- w and ated by Populus euphratica and P. pruinosa. Contribution. to the physical notitia ons of Sin- egion Chinese.) incipal vegetation types and their charateristics of the northeastern i nica Sinica 2: 349: 370. Science Press. (In Chines EDITORIAL BOARD OF CHIN s VEGETATION, China’ i Vega Science Press. (In ms EDITORIAL BOARD OF SZECHUAN’S VEGETAT Szechuan Province’s Vegetation People’ s Press of Szechuan Province. (In Chin Ho, S. Y. & S. P. CHEN. TA Ve egetational region- alization of Kwangtun vince. Acta Phytoe- cologica et Gasketing Sinica 1 1(1-2). Science in 1951. The ed com- munities of southern Kweichow Province. Acta 54. The Indicator Plants of Acid, Calcar- cous, and Saline Soils in China. Science Press. (In Chinese.) 1956. The vegetation of China with special reference to the main soil types. Report for the 6th may oils 964. ry. On the principles, criteria, and units DAE a pee alntgiobalination. t Geobotanica Sinica 2: 2 ————— (Editor- and parera 1979. Vege- ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 tation Map of China (1: 4 000 000). Cartographic Publishing House, Peking. (In Chinese with En- glish legend.) —. 1981a. A further discussion on the principle and scheme of vegetational regionalization. of China. Acta P. 5: 290-300. Science Press. (In Chinese with En- glish summary. . 1981b. The geographical distribution of vegetation of China related to the horiz se vertical zonation. Act Bot. Bor.-Occ. Sinica 1(l- 2). (In Chinese with English summary m 1982. The Vegetation Geography of China with Reference to the Chemical Composition of Dominant Plants. Science Press. (In ese. etal. 1953. A aperi study on je ka sociations of Peitaiho, ovince. ape Press. (In Chinese with Engish summary.) Cuu. 1955. The vegetation of (e n Visi Kwangtung Province. Contri- : bution to Pl ant Ecology and Geobotany, No. Science Press. (In Chinese.) 1963 Saksaoul deserts of the north weg Hu, S. T. China. Acta P. 1(1-2). Science Press. (In Chinese J f dy desert vegetation 0 Hwang, Y. S. etal. 1962. Sandy ioratioā tion. Science Press. of Sinkiang and its Utilizati Fangs JIAN, Z. P. et al. 1975. Beech forests of Fi nica 2(1). Science Press. (In Chinese.) cn 1964. Mute natural vp of the conterminous United States. Amer. Soc. Special Publ., No. KWANGTUNG INSTITUTE OF BOTANY. Province's vegetation. Science 1976. Kwang: Press. (I ese.) Li, C. C, et. al. h 1964. Discussion on the crit c aracteristics for region onalizing trop: teria ics and SU wpoint Acta Phy 5. Sc toecologica et Geobotanica Sinica 9: 253-25 ; region. part of Lungchow county, Kwangs! No. & a to Plant Ecology and Geobotany, Scien deserts 1961. The fundamental clan o( A in the northern Sinkiang regi tgp A 1983] Sinica 9(3—4). Science Press. (In Chinese with English summary. LiN, Y.etal. 1965. Main vegetation types of Chiangsi. Univ. Bull. No. 3. Liu, F. S. 1964. Discussion on the characteristics of three vegetational zones. Acta Phytoecologica et Geobotanica Sinica 2: 256-261. Science Press. (In inese. Liu, T. N. 1939. Phytogeography of southern and southwestern China. Chinese Journal of Biology 1: 21-27. (In Chinese.) . Icones of the Woody Plants of North- eastern China. Science Press. Num, M. 1974. The Flora and Vegetation of a — 1955 pan. Kodansha Ltd., Tokyo Suan, J. H. & F. S. Liu. 1963. The main fnm type of Chekiang Province. Acta Phytoecologica et SMehotanica Sinica 1(1—2). Science EO (In 55. The British pus "d their eir ecological aspects. Bot. Bull. Acad. Sin. 1: 187-200. COR 1948. dA geography ofthe East-Tibet Pla- teau. Bot. B cad. Sin. 2: 62-67. UNESCO. 197 n iulersational Classification and Wa Torn of vegetation. Unesco, Paris. Vegetation of the earth in relation to E: ate and the eco-physiological conditions. Springer-Verlag, New York Wana, C. T. & B. S. Li. . Them d characteristics of high cold-steppes in the Chang- i Acta Phytoecologica et ern Sinica 5: 1-13. Science Press. (In c, HS S. 1964. The distribution of halophytic Plant communities of Sinkiang Region in relation HOU — VEGETATION OF CHINA 549 to soils and groundwater. Acta d pose din et n Chinese.) vicinity of Yenshan, Linkwei county, Kwangsi Re- Contribution to Plant Ecology and Geobot- any, No. 7. Science Press. (In Chinese 4. A preliminary investigation on the phytocoenosium-charateri stics of med Cathaya forests in Kwangsi. Acta oda et Geo- botanica Sinica 2: 103-110. Scene jn (In Chinese. WANG, Y. F. 1963. The fundamental characteristics of steppes in the Tienshan Mountains. Acta Phy- toecologica et G Sinica 1(1—2). Science Press. (In Chinese. WanMiNG, F. 1909. Oecology of Plants. Oxford Univ. Pre WEAVER, J. E. & F. E. CLEMENTS. ipa JB Ecol- ogy. McGraw-Hill Co., Inc., New WHITAKER, R. : Cheatin, of natural communitie Rev. (Lancaster) 28(1). Wu, C. L. 1955. Preliminary investigation on the Na ae Sinica 21(3). Science Press. (In ZHAO, s. L “1958. Preliminary report of Pinus peor vestris var. mongolica woodland growing on arid sandy soil in Inner Mongolia. COS tibution e Plant Ecology and Geobotany, No. 2. Science Press. (In Chinese.) ZHU, p A 1958. General comments on the steppes northeastern China. Northeast Teacher Univ. Bull. NG - a Chinese Znuo, Z. D. S. The fundamental characteristics and uses j^ alpine evergreen shrubs of Kansu Province. Langchow Univ. Bull. (Natural Science) No. 2. (In Chinese.) QUATERNARY HISTORY OF DECIDUOUS FORESTS OF EASTERN NORTH AMERICA AND EUROPE! MARGARET B. Davis? ABSTRACT The temperate deciduous forest of North America is more diverse than the deciduous forest of western Europe. his difference has traditionally been explained by greater survival in North erica . . . t of deciduous species during the Quaternary. More recent investigations have shown, however, tha late-Tertiary a minor component. During the Quaternary, LFA onifer s, with deciduous angiosperms 1 e +h LDuranean 2 ^ angi d inental flora, leaving a few species and genera of angiosperms as the dominant trees. Cold, dry, Nene Ses climate during the glaciations caused the extinction of conifers; deciduous trees apparen ies th IM ti diti : c ard re CORO MAN CD * s E 4. VOU E ag - . L . T, t were North America, in contrast, temperate deciduous forests are quite similar to the ties pe k present in the late Tertiary. During th e : e ais. ula- deciduous angiosperms were displaced from the Appalachian mountains, arbe in n ed tions in the lower Mississippi valley or on the southern coastal plain. Coniferous fores by spru Quaternary climatic ern species distri- th i i ; i ternary, butions can no longer be considered relicts of Tertiary distributions. Throughout the Quate: Temperate deciduous forest grows over a wide area of eastern North America. The forest is rich in numbers of species, especially the mixed me- sophytic forest communities of the southern Ap- 2 ANN. Missouri Bor. GARD. 70:550—563. 1983. nosperms such as Sequoia dominant - A regions, and evergreen angiosperms p di where. He argued that a number of maj ud matic changes occurred during the Tertiary, led pecially during the Oligocene; these changes : o ponents of the Tertiary flora. Thus changes * had also occurred in Europe, where +’ f ras contained many genera and wee minani fers (Wolfe, 1979). Sequoia was the tree in some regions (Traverse, 1982). E Mixed coniferous-deciduous fore d during the Pliocene is a new interpre atrast V forest history that stands in marked ud the traditional view. The traditional M. pe int that deciduous forest persisted 1n. d = the early Quaternary Period, when 1n ' This work was supported by the National Science Foundation. : is. Minnesota 5545$ Department of Ecology and Behavioral Biology, University of Minnesota, Minneapolis, 1 1983] verity of climate caused extinctions of many an- giosperm trees (Tralau, 1973; Campbell, 1982). In contrast, Wolfe emphasized that coniferous species and genera were the important losses from the European flora during the Quaternary. The Taxodiaceae, for example, once so important in the Black Sea region, were eliminated entirely (Traverse, 1982). The angiospermous genera that remain today represent differential survival of one component of what had been mixed conif- trous-deciduous forest (Wolfe, 1979). Wolfe pointed out that Europe today has the type of climatic regime that elsewhere in the world sup- ports mixed coniferous forest; the dominance of deciduous species in the region today is therefore anomalous. Eastern North America, in contrast, has a climatic regime typical of deciduous forest regions. Today it supports forests dominated by deciduous angiosperms, just as it did during the late Tertiary (Wolfe, 1979). The Quaternary pollen record adds a useful perspective to these differing views of the origin i the European deciduous forest and the rela- 'onship of deciduous forests in eastern North an ig and Europe. First, the Quaternary rec- ila ws that dramatic changes in the geograph- ranges of forest species occurred during the Pana We can no longer speak of Tertiary remaining” in a region throughout the Fa, Bographical region where they now occur pean on, modern geograph nges ot be used to identify the locations of “relict” 45 units extinctions that occurred during the Quaternary, and PA the differential extinctions of conifers Rien. angiosperms document that the ed forests voc ity of Quaternary climate affect- ica. These EU ently in Europe and North Amer- Up of the Tee factors contributed to the make- Merica modern temperate forest floras of North in and Europe. M rin how Quaternary climate two Cotes floras into modern floras, the Mates and Per Quaternary climate, glacial cli- The two Spin climates, are impor! ent e ffects € appear to have had quite differ- een Slacial phases, i.e., times when ice "S present Com- P > DAVIS—QUATERNARY HISTORY 551 prised about 90 percent of the time during the Quaternary period. During these long, cold in- tervals, temperate species survived in small pop- ulations that were susceptible to extinction. The severity of the climate, both in terms of average temperat tinentality, and droug ht; the ex- tent of geograp p cies; the sizes of populations; and the community composition of forests in refuge areas; all had an effect on the probability of extinction for indi- vidual species. Interglacial intervals comprised a much small- er proportion (about 10 percent) of the Quater- nary period. They were characterized by climates similar to those of today, which in Europe and eastern North America seem to bear a general resemblance to late-Tertiary climate. Each in- terglacial was short, lasting only 10,000 to 15,000 years, and began and ended with a sudden, major climatic change (Emiliani, 1972; Broecker & Van Donk, 1970). The interglacials, although favor- able for survival and population expansion of temperate fi t trees both in Europe and m North America, were times of vegetational in- stability. During interglacials the geographical distributions of temperate species shifted many hundreds of kilometers, and the composition of forest comunities changed rapidly. 2x 13 rs tz GEOLOGICAL EVIDENCE OF EVENTS DURING THE QUATERNARY PERIOD The exploration of the deep sea by geologists in the last twenty years has led to a new under- standing of Quaternary events, revolutionizing our thinking about the time scale of glaciation. Many marine cores include sediment extending through the entire Quaternary Period. Previ- ously, four major glaciations were recognized within the Quaternary. We now know that there were 18 or 20 glaciati during the last 2 million years, the time interval now assigned to the Qua- ternary. Each of these glacial cycles lasted about 100,000 years (Hays et al., 1969). Figure 1 shows oxygen-isotope paleoclimati ds for the last 800,000 years. The climatic events are well dat- ed: the last interglacial (the earliest part of stage 5) started 125,000 years ago, lasted about 15,000 years, and ended with a sharp decline in tem- perature that initiated the last glaciation. Warm conditions returned, followed by a cold period, then a short warm interval. Seventy thousand years ago a long cold interval (stages 2—4) began, which culminated in the glacial maximum 18,000 to 20,000 years ago (Broecker & Van Donk, 1970). ANNALS OF THE MISSOURI BOTANICAL GARDEN all as today (isotope curves above dashed lines) for only short periods dur- ing the Quaternary. The three cores record nine glacial- interglacial cycles during the last 800,000 years (mod- ified from Johnson, 1982). A number of aspects of this record are im- portant to vegetation history. The first is that changes occurred much more rapidly than pre- viously supposed. We had imagined four rather leisurely glacial-interglacial cycles, each lasting several hundred thousand years, with long in- terglacials equal to, or longer than, glacial pe- riods. The newer data indicate that interglacial periods were brief, 10,000 to 15,000 years in length, while the glacial intervals were much longer, lasting about 100,000 years. During gla- cial intervals there were fluctuations of climate culminating in a severe, but relatively brief gla- cial maximum and then a rapid, large temper- ature rise into the next interglacial (Broecker & Van Donk, 1970). The present interglacial, the Holocene, is typical in length for an interglacial, meaning that events recorded in Holocene sed- iments can be used as an analog for events during previous interglacials. FULL-GLACIAL VEGETATION OF NORTH AMERICA egions. More than 1,800 surface samples are available for comparison from different parts of eastern United States and Canada (Overpeck & Webb, 1983). [VoL. 70 These can be compared with quantitative sam- ples of modern forest to show how the percent- ages of pollen relate to the percentages of growth stock volume in the modern forest (Delcourt et al., 1983). In addition, pollen concentrations in sediment can be corrected for the rate at which the sediment matrix has accumulated, and thus the actual numbers of pollen grains deposited every year on the sediment surface can be - termined (Davis & Deevey, 1964). Comparisons of modern pollen deposition rates and forest rec- ords show that pollen deposition is proportional to the density of source trees within a 15 km radius of the deposition site (Davis et al., 1973; Webb et al., 1981). Although imprecisions arise because pollen grains are produced in different amounts by different species and are transported different distances, we nevertheless can use pol- len to obtain not only a list of the flora, but also roughly quantitative information about the abundances of different kinds of trees. Macro fossil studies are used increasingly in combine tion with pollen studies to provide species " tifications and definitive proof of the l presence of particular plant species (e... Watts, 1979; Davis et al., 1980). WS Eastern United States is shown in Figure 4 with black dots indicating sites where S are securely dated by radiocarbon at 18, dd 20,000 years before present, when the 10¢ p at its maximum. Most of these sites have dern within the last 15 years. 0 "m new p : à the vegetation during the last glacial by very different from modern vegetation ME 1980, 1984). Furthermore, most of the P ; communities were dissimilar from amy communities that are widespread : esi trary to previous thinking, the latitudin tion of modern forest communities was merely displaced southward in frere ^ placement of climatic zones (Martin, 1 ; 1 1 + of the developed, without modern analogs. South ward at high elevations al (Watts, 1979; Maxwell & Davis, was at about 1,000 m elevation in wes land (Watts, 1979). oo Along the Atlantic coastal plain uet w niferous forest, with Pinus, Picea, nape picti rix (Watts, 1984; Whitehead, 1973, 198). 1972). Tredl* tern Mà? siue deni dst dadas ct iid 1983] was common in the north, but farther south, it became quite rare and Pinus was dominant. The pollen assemblages from South Carolina leave it ambiguous whether there was dense forest or semi-open vegetation (Whitehead, 1973; Watts, 1983), but it is clear that deciduous trees were rare or absent. On the southern coastal plain and piedmont as far south as northern Florida, Pinus was the dominant tree (Watts, 1970, 1983; Watts & Stuiver, 1980). There were only trace quan- lites of deciduous tree pollen at sites north of northern Florida (Watts, 1970, 1980; Watts & Stuiver, 1980; Whitehead, 1973, 1981). On the western flank of the Appalachians, Pi- nus pollen dominated the fossil assemblage (Del- court, 1979). But in the Mississippi valley, Picea pollen was abundant, and Pinus pollen was ab- sent (Delcourt et al., 1980). Farther north and West on the Great Plains, Picea and Larix were the dominant trees. Pinus was present before Mis years ago, but it became extinct just be- ore the last glacial maximum. Unlike the mod- a boreal forest, which contains both Pinus and Finis especially in the central part of Canada, 4 dominated the full glacial forest in the cen- ee Part of the continent, and Pinus was absent. ere Is Some controversy about the presence of linn E Quercus pollen is present (1596) at an um site at stratigraphic levels that appear to Sea emp (Griiger, 1972). Some authors have ted that oak grew together with spruce Ree prc regime in which seasonality was is LE" (Wright, 1982; Solomon, 1982), td ers suggest. that the pollen assemblage * iraran a tundra in which oak pollen carried tit rom a distance stands out in a pollen Ter ge from unproductive tundra (King & Y, 1980; Maxwell & Davis, 1972). Oak in a with spruce today on sandy soils ign rn Manitoba, producing similar pollen Pes E" (West, 1961) and macrofossils in- Quies oak occurred in Iowa 12,000 years 1980) Spruce was still present (Baker et al., mun’ deciduous forest today reaches its maxi- : meat in the Appalachian moun- wth maximum size OMA k The high diversity of this forest jag d to speculation that the south- DAVIS—QUATERNARY HISTORY 444 PICEA (and DECIDUOUS TREES) Gulf o? Mexico mmm —cum —nHmum o km 500 L l fi ci dd S 90 80 FIGURE2. Outline map of North America, showing the extent of ice and the locations of major kinds of vegetation 18,000 years ago. Black dots indicate lo- with certainty at to 18,500 years. d proves this idea, showing instead that deciduous forest was completely displaced from the Appa- lachian mountains. Full-glacial pollen deposits are dominated by coniferous pollen; they do not show a predom- inance of deciduous tree pollen, such as occurs oday in d li t from all regions inated by deciduous forest, especially regions of mixed mesophytic forest (Delcourt et al., 1983). Deposits dated at 18,500 years before present in northern Florida contain 80 percent pollen from southern pine species. Farther south there are no records, although sand dunes and other geo- morphic evidence suggest that the region was too dry for extensive decid forest (Watts, 1983). Itis unclear where all the decid f i Anm [d survi g Fossil pollen, seeds and fruits from temperate deciduous trees suggest that some species grew together with spruce near Memphis in the Mississippi Valley (Delcourt et al., 1980). In northern Florida, de- ciduous tree pollen (mainly Quercus and Carya) makes up 15 to 20 percent of the pollen assem- blage (Watts & Stuiver, 1980). Deciduous tree pollen increases in abundance rapidly at many southern sites during the late-glacial period, sug- 554 gesting that at least small populations of tem- perate trees such as Fagus grew nearby in very small, scattered refuges at the time of the glacial maximum (Watts, 1984; Solomon et al., 1981). During the coldest part of the last glacial pe- riod, small populations of deciduous tree species may have grown in unusual habitats where local conditions compensated for the regionally cold, dry climate. These refuges were small in area and scattered throughout the region, apparently com- prising only a small fraction of the total land- scape area. In a regional sense pine was more abundant than deciduous trees: it dominated the pollen assemblages. There was no large area 18,000 years ago that could have been mapped as a region of predominantly deciduous forest. In their recent, detailed reconstructions of full- glacial vegetation, Delcourt and Delcourt (1981) mapped the coastal plain as ““Oak-hickory south- ern pine" evergreen forest. The existence of this forest type is not well documented, however, as evidence has not been published from sites on the coastal plain except northern Florida, where it is certain that sedimentation occurred contin- uously during the full-glacial (Watts, 1984). Al- though Delcourt and Delcourt implied a higher frequency for angiospermous trees than I have suggested, they were in agreement that the full- glacial vegetation of this region was different from the modern mixed mesophytic forest of the Ap- palachians. HOLOCENE FORESTS Additional information about the history of deciduous tree species comes from records of vegetation after the ice began to retreat. In this case we have many sites, not only those in Figure 2, but also many north of the glacial boundary, where lakes and bogs are plentiful. These local- ities are indicated by the small numbers in sub- sequent figures; references are given in Davis (1981). Pollen records were used to determine the | pollen deposition rates in lakes show a relationship to geographical range limits. Accumulation rates for all species studied are higher within the range limit than outside of ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 it, and higher still where the species is abundant. These relationships were different for the differ- ent species that were studied, depending appar- ently on the production rates and dispersal ten- dencies of the pollen. Generally, pollen deposition rates are at least 5-fold, and in some cases 10- fold higher within the range than outside (Davis et al., 1973). On this basis I have deduced that the earliest steep rise in the rate of pollen de- position in fossil deposits, that is a tend m a ern sediments within the range of the source species, can be used as evidence of local arrival. In a few instances where data were available only as relative counts, I used a steep rise in pollen percentages in radiocarbon-dated pollen per centage diagrams. In fitting isochrones, however, more weight was given to data points based on deposition rates. Recognizing that pollen percentages can bem low as one percent near the geographical limit for a species, I have avoided using a pori | pollen percentage value to indicate presence. p | percentage values are difficult to deter | curately because of statistical errors. Mei à | many cases (Davis et al., 1980). There are pe - where macrofossils have been found at je i tion expansion occurred sites where populati p per W- show later arrivals are farther north. ee : ing northward range extensions : seve pec ared from p fn ure ee UIC GULIN. (Figs. 3, 4, 5). Arrival times (rounded to 1,000- ee onan vals) ee indicated on a m pue age. TH EE) ey H "erred earliest that show year jnter | i nes were dra were closest to refuge areas, and sites : later arrivals are farther north. Mare nnrthivarrgad s feavera ites ished po" est trees have been prepared from. pep above len evidence using the criteria descr! (Figs. 3; 4, S); r pices- Figure 3, for example, shows data fo gov The stippled is th i where P pe “gf T} i o rs on the map $ today. The small numbe puse nt rival times" in thousands of years be a at each site. Isochrones have been pa "T necting points of similar age, tracing th scatter o edge of the expanding population. The j points indicates the uncertainty of ue nations of arrival dates. Pollen percen | 1983] DAVIS—QUATERNARY HISTORY 555 3 £11 e T oen er ay "e ; We s TOM — Tm "d ES [PE NUN oT A, E dz 4 eal o o >14 o ^ 23 12 (E ac Larix /arícina . f Larch b Q 400 km ^ ims ML LO l a io T SM dune Ree n a 122 Om BLM" E 12 n muon " 13 1 T y 14 9 e [9] 19 21 i ^ d Abiss balsamea : BalsamFir à Q400 km ^ T enini. SE the retreat of the ice. Small numbers indicate 5 ge s pe sites over a wide area have been plot- Webs oie series of maps by Bernabo and Pis 7), Webb (1981), and Huntley and the se dna Pollen percentages are affected by ound of pollen produced by other individual sites, as indicated by fossil pollen (see text). species, as well as by the amount of pollen pro- duced by the population of interest. Nevertheless a map of percentages shows the region where the species was ir constant pollen background) and it shows the 556 c= L, =B 7 iot - 87 Sh AS. s 2 ——À oH O : E 2 pian A o 10°10 7 A » 3 eMo? jo 10. (+ D B . Ow o 213 $2 on «Sy [e] o . Pinus Strobus | White Pine ANNALS OF THE MISSOURI BOTANICAL GARDEN FIGURE 4. Range extension of two conife [i Quercus and Ulmus. Both coniferous s somewhere in the vicinity. Oak and located in the southern Mississippi retreating edge of the migrating population (Ber- nabo & Webb, 1977). Data showing later arrivals at northern sites than at southern sites can be used to calculate rates of range extension in me- ters per year. Most tree species were able to ex- tend their ranges northward 200 to 300 meters oe Ne | iravude i Ulmus SPP Elm Q 400 km and uferous tree species that commonly grow in deciduous foro | pecies expanded from the eastern coastal elm expanded in a northeasterly direction from refuges that were se valley region (from Davis, 1983). $ 1 re plain, suggesting 2 parently | per year, a rapid rate for trees with long er | ation times. These estimates are SupPo E data from Europe, which suggest similar P pii expansion there (Firbas, 1949; Huntley & | 1983). | m the In northeastern United States, far ss | a = 1983] DAVIS—QUATERNARY HISTORY 557 RR c VN cer Spp. Hh Maple Q. . 400 km 5 à was ex — where trees grew during the glacial gradual E forest vegetation gained species through» through time, increasing in diversity iive oe the Holocene. Certain forest species necticut atd late— Castanea arrived in Con- manh S recently as 2,000 years ago. Sites in uthern Appalachians, such as Anderson Fig ; bes 2 Range extension of four deciduous tree species or ge ds Slowest species to extend its range northward. Beech moved nort ding westward across the lower Great Lakes region (from Davis, 1983) n nera following the retreat of the ice. Chestnut hward east of the Appalachians, Pond in Tennessee (Delcourt, 1979), were closer to refuges and much less influenced by migration histories (Solomon et al., 1981). The replacement of Picea and other boreal species by deciduous forest or by mixed decid- uous forest was time-transgressive, occurring 16,000 years ago in the south and 10,000 years 558 ago farther north. The maps (Fig. 3) show that northward expansions of boreal species occurred at different rates. Picea and Abies, which have overlapping geographical ranges at the present time, did not expand northward together. In some regions, such as northern New Hampshire, Abies arrived after Picea had declined in abundance. These results refute the idea that two trees that co-occur today in a recognizable plant commu- nity necessarily expanded together in the past in response to climatic change. The rate at which each species could expand into available habitat depended not only on suitability of climate, but also on the dispersal of seeds and the ease with which new individuals could become estab- lished. Many deciduous tree species moved into tanea (Figs. 4, 5). Ulmus (Fig. 4) ranges far north into Canada at the present time (stippled area on Fig. 4); its physiology appears adapted to the cooler climates that may have prevailed 10,000 to 11,000 years ago. Acer (mainly A. saccharum) also arrived very early at some midwestern sites, 12,000 to 11,000 years ago, expanding rapidly into the northeast, where it arrived about 9,000 years ago. The rapid and early expansion of Quercus is surprising, as its distribution today is temperate rather than boreal. Furthermore, it has heavy seeds that depended on animals for dis- persal. Birds can be very effective dispersal agents, however, because birds such as blue jays and crows not only feed on acorns but fly consider- able distances with them, caching them in wooded areas (Van der Pijl, 1969). It can be argued that animal dispersal is more effective than wind dis- tablishment may also have affected the success of Quercus. Its range extension was rapid, more rapid than any of the deciduous trees that today grow in mature forests. Also moving from the west, but expanding more slowly, was Carya spp., which arrived early (10,000 B.P.) in the Midwest, even at sites close to its present northern bound- ary. But it was slow to cross the Appalachian mountains and arrived late in New England, only 5,000 years before present. Castanea pollen is found in quantity near Memphis, Tennessee, as early as 15,000 years ago (Delcourt et al., 1980). It then expanded northeastward up the Appalachian mountain ANNALS OF THE MISSOURI BOTANICAL GARDEN chain (Delcourt, 1979). The expansion was rel- atively slow, averaging about 100 meters per year. Castanea is dispersed by birds; however, it is completely self-sterile, which may mean that the statistical chances for the establishment of pop- ulations are much lower than for self-fertile species. It was successful once it became estab- lished, becoming abundant throughout the Ap- palachians, and dominant in many areas (Braun, 1950). The direction of expansion ofall the above named species suggests that they survived the last glacial period somewhere west of the Ap- palachians, perhaps in the southern Mississipp! Valley (Delcourt & Delcourt, 1981). Fagus grandifolia (Fig. 5) moved northward along the Atlantic coast east of the Appalachians. Beech then expanded westward through the Mo- hawk Valley into the Great Lakes region, spread- ing southward into Ohio. The movement of beech into Wisconsin and Michigan is being studied in my laboratory by means of closely spaced sites, our results suggest that beech has been dis several times across geographical barriers 30 » 100 km wide. Detailed studies of migration his tories indicate the mechanisms of dispersal for plants, and give us some idea of the role of geo- hical barriers in preventing or slowing 8° graphical spread (Woods & Davis, 1982; Webb, 1982) (Fig. 5). species, Pinus strobus p Tsuga canadensis, expanded from a refuge tinental shelf, which has long been suggested T a refuge area (Fernald, 1925). At least their PO appeared first at sites in the mid-Atlantic regi (Fig. 4). Both species then expanded wes Tsu- and northward. Although Pinus strobet with : n. mixed € ga canadensis grow today in eas they were ot growing together with many of the decim trees during the last glacial period. In other p : the communities in the refuge areas Were ent from modern communities, comp? tant point, because trees were expo petition from species co-occurring in refug ca for about 50,000 years. The makeup of the® is- munities is important when the evolution tory of speci pone | 3 : hetween are considered. In contrast, deciduous re we have been growing in the Appalac Study them today for fewer than : and for less than half that time in the 000 yeas 12,000 ME —————— — I —n,. 1983] Appalachians. In the northeast, deciduous forest communities have gained new species through- out the Holocene, increasing in diversity as ad- ditional species extended their ranges to this re- gion. Many communities that appear similar today have had different histories even during the last few thousand years. For example, Genan Cas- tanea forests in the A contained Castanea for different seis of time, 8,000 years in the southern Appalachians and only 2,000 years in the north. Thus, present sim- ilarities in community composition do not nec- essarily mean that forests have had a long, sim- ilar history. Range limits change constantly in response to climate. For example, Pinus strobus moved north of its present limit about 5,000 years ago (Teras- mae & Anderson, 1970), when the climate was warmer. As the climate cooled in the last few enin, the te: has MR tUd southward; at Dear the western limit of P. sapis Mp it lo expand in Minnesota (Jacobson, 1979). Species ranges are dynamic, expanding and contracting in a sensitive way in response to minor changes E We have a demonstrated — adouch P ntains have S 5 ange westward a 01 - n within the last i 000 years, presumably with the "Little Ice Age" (Woods re Davis, 1982). any of them may represent geo- Voce distributions that have been attained Mrs and that will change again as the anges The Holocene i is similar in length and com- uu in climate to previous interglacials. We conclude, therefore, that rapid northward ue of deciduous trees occurred many Previously during the Quaternary. DECIDUOUS FORESTS OF EUROPE A detailed knowledge of Pleistocene biotic his- tory is available from Europe, where interglacial its have been studied in detail, making it Possible to observe the timing of changes in the a composition of temperate forest com- xg Recent work has indicated that ex- *ly severe climatic conditions prevailed in Pe during the glaciations. Reconstructed “aN Surface temperatures 18,000 years ago show DAVIS—QUATERNARY HISTORY 559 o o GLACIA GETATION POLAR DESERT xtent of Weichselian ice Sheet ARTEMISIA o i -STEPPE STEPPE-TUNDRA ? 9e » * 40 40 " ARTEMISIA - STEPPE . A ? G o 500 — —CB—À km i o 10 20 30 FIGURE 6. i jm map of Europe, showing the dis- tribution of ice sh and major vegetation types dur- ing the full-glacial, L 18, 000 to 20,000 years ago. that sea ice covered the ocean as far south as Iceland and polar water flowed south to the Bay of Biscay. Surface water temperature fell 10°C. The warm Gulf Stream, instead of flowing north- eastward across the Atlantic and are the west- ern coast of Europe, as it doe , flow eastward across the mid- lu inn abena past the Iberian peninsula (Climap, 1976) by lowered sea levels, providing a large land area in Northern Europe and a relatively continental climate. The low countries and Denmark were “polar desert.” Farther south, the vegetation was a tundra with abundant Artemisia (Peterson et al., 1979). Artemisia steppe also characterized the Mediterranean region. Similar vegetation grew in Turkey, and east to the Iranian Plateau (Van der Hammen et al., 1971) (Fig. 6). Cold and dry climate is indica ted. Pollen diagrams from interglacial deposits in Europe are similar to Holocene pollen diagrams; they show the successive arrivals of different species as trees. pundet. their ranges northward ern n Europe, interglacial pollen diagrams begin with a boreal flora of Pinus and Betula, and end with Pinus, Betula, and Picea; in the intervening temperate zone the diversity of the flora gradu- ally increases as more and more deciduous tree Palynological Sequence of Grande Pile Peat Bog Est.Age X10") Depth Holocene 10 - 500 18 + 700 - 900 Glacial Interval WALL 70 F 1300 90 -1500 110 Interglacial - 1700 r25 -1900 [e] 50 100 (cm) Percent Arboreal Pollen FIGURE z. Percentages of tree pollen in the Grand Pile pollen diagram (from Woillard, 1978). species appear. The younger interglacials and the Holocene are marked by the very late arrival of both Carpinus and Picea; these genera appeared earlier in the older interglacials. he last aximum ANNALS OF THE MISSOURI BOTANICAL GARDEN forest returned, including all tree species. [VoL. 70 percentages of tree pollen. Tree pollen was abun- dant during the last interglacial, which extended from 125,000 to 110,000 years ago. The inter- glacial terminated with a change in climate 110,000 years ago that is also recorded in deep sea sediments. Forest disappeared from the vi- cinity of Grand Pile and was replaced by an Ar- temisia steppe. One hundred thousand years ago, the climate became warmer again, and the forest returned. A forest succession similar to the Ho- locene is indicated by the pollen diagram, as all the forest species, including Picea, returned. Pic ea had had a late arrival during the preceding interglacial interval, but it must have been grow- ing nearby during the subsequent cold stadial, because it was available for rapid recolonization when the climate warmed 100,000 years ago. A 5 eon p i g ahont 90,000 years ago. Again, the forest was replaced by Artemisia steppe, and again, about 10,000 years ua : the climate became cold again 70,000 years ag? and a long cold period ensued, culminating 1 the glacial maximum 18,000 years ago. During this long period, Artemisia steppe prevailed lo- cally and the deciduous forest and Picea appa! ently retracted to distant refuges. In her classic paper, Reid (1935) emp the importance of east-west trending mo in the extinction of deciduous forest spec! i Europe. She believed that the glaciers were ve companied by cold climate, which caused = rywhere north 0 this mountain barrier. Only those species thal hasized untain. es in data imply, however, that although foi oe appeared locally in response to cold o : the onset of glaciation, all species Were Wer to return as soon as the climate becam? ihe again. A more important extinction time he i very long and very severe cold period e 000 cluded the glacial maximum (70,000 to ps d years ago), rather than the first on devel 100,000 years ago that accompanied the during opment of continental ice sheets. It was ore, long cold periods, lasting 50,000 years OF a that trees like Picea became locally extinct viving only in refuge areas far to the east cel the next interglacial, the Holocene, began, fact it was slow to return to western Euro ert yi never did return during the Holocene to cinity of Grand Pile. panded jn Distributions of deciduous trees €X —— — ———— 1983] Europe during the Holocene just as they did in North America. It seems likely that somewhere on the southern slopes of the Alps and/or in the Balkans there were habitats during the full-gla- cial that were suitable for deciduous trees (Van der Hammen et al., 1971); probably small pop- ulations were scattered about in pockets of fa- vorable environment. Maps of Holocene pollen frequencies in Europe show that many deciduous tree species expanded first in southeastern Eu- Pie and then moved westward (Huntley & Birks, 3). The gradient in ocean across the Mediterranean (Climap, 1976) may have been correlated with more favorable tem- peratures on land as well The succession of interglacial deposits in Eu- rope record the progressive extinctions of tree species during the Quaternary (West, 1970; Wolfe, 1979). Different interglacials witnessed the de- velopment of different communities of decidu- ous trees, depending on which species arrived in the British Isles. Floristic differences among the British interglacials were also caused by evolu- tionary changes. West (1980) suggested that Pic- ea, for example, which appeared in Britain dur- ing early interglacials, must have been migrating 7 the British Isles from refuge : areas in western t a ual loss of biotypes that ` were able to survive in Teluges in western Europe. Picea appeared later and later in successive interglacials, and during the mo it failed entirely to expand into Great Brita. i Additional evidence for evolutionary changes me vided by Corylus pollen n percentages in the A Ssive British interglacials (West, 1980). In *early interglacials of East Anglia, Corylus pol- i pa are low. Later in the Quaternary sn € more abundant, and finally, in the Ho- it P ovod early and built up a Iam E d. Wes before e oth 1 — that Corylus may have adapted to in- Si EM conditions. Whereas Picea was less and haus uL and Tsuga became extinct alto- Orylus, a deciduous angiosperm, be- * more and more abundant. DISCUSSION E Xtinction rates in Europe during the Quater- te Were higher oe coniferous trees than for MPerate deciduo The m in Quaternary deposits led to the devel- op : : ment of the predominantly deciduous forest DAVIS—QUATERNARY HISTORY 561 that we see soday "^ xcci AS Barone. TN factors may h g the Quaternary: 1) they might not have been able to expand into forested landscape as ogi as the deciduous trees, thus losing out in c e mi- —€— of Picea is influenced by soil changes, (EE: 1964). If so, the expansion of Picea l not explain Picea's history during the early Qua- ternary, however, when it appeared during the early phases of interglacial sequences in Britain. 2) Alternatively, conditions in the refuge areas ea mid-way through the Quaternary (West, 1980) may be related to the very severe conditions that existed during the glacials in the western Medi- terranean Tégion. A very -— and probably con- would be a mistake to study the present climate and try to use it to explain all aspects of the distribution of the modern flora (Wolfe, 1979). Instead, the longer intervals of time during which the climate was very different must be taken into account, i.e., the glacial intervals, which com- prise nine-tenths of the Quaternary Period. Dur- ing these longer intervals, extinctions of conifers occurred, although pony, temperate angiosperms were able to surviv The Quaternary fossil record has further im- plications for the geoflora concept. Geofloras are forest communities that were believed to persist, unchanged in composition, over many millions of years. In contrast, Holocene forest commu- nities are ephemeral, composed of newly im- ag species, and easily invaded by addi- onal species. The deciduous trees growing in our forests that have persisted from Tertiary time are aggressive species that have been able to move pidt onto deglaciated territory. Any species that has not been able to disperse its seeds easily, to extend its range, and to penetrate forest com- munities, has become very rare or extinct. Wolfe (1979) compares our tree flora to secondary species in the forests of eastern Asia. The nature of modern communities calls into question whether the concept of closely co-adapted com- 562 munities such as “geofloras” describe Tertiary forests. The deciduous forest communities of eastern North America and western Europe remain a grab-bag of species with characteristics that en- abled them to survive during the Quaternary in the very unstable environment of interglacials, when rapid dispersal and colonization abilities were selected for. During the glacials, dry and continental climate exerted a different kind of selective pressure. Campbell (1982) has com- pared the tree floras of Europe and eastern North America, invoking these factors to explain some of the differences. The modern deciduous forests is appropriate to the forests of eastern North America bear a sim- ilarity to Tertiary forests of the same region, the marked changes that have occurred in Europe caution us against using the present forest as a model for the forests of the Tertiary. LITERATURE CITED ANDERSEN, S. T. 1964. Interglacial plant successions n the light of environmental changes. Report of A Congress 2: 359-368. Baker, R. G., K. L. VAN ZANT & J. J. DULIAN. 1980. Three late- glacial pollen and plant macrofossil as- semblages from Iowa. Palynology 4: 197-203. BERNABO, J. C. & T. WEBB, MI. 1977. Changing pat- North America: a mapped summa: ua Research 8: 69-96. B ocium BRAUN, E. L. 1947. LL ot of the itas forest of eastern North America. Ecol. Mon 17: 211-219. . 1950, reprinted 1964. Deciduous Forests of Eastern North America. Hafner Publishing Co; New York BROECKER, W. S. & J. VAN EN 1970. Insolation changes, ice volumes and the O'8 record in dee eep sea cores. Review of Geophysics and Space Phys- ics 8: 169-198. CAMPBELL, J. JN: " - of California . Carnegie Institute of W. Publication 553. i cat — — Members. 1976. The surface of i th earth. Science 191: 1131-1136. à ing M. B. 1976. Pleistocene aa eroeraphy of tem- d Man 13: 13-26. . 1981. Quaternary history and the stability of 177 in H. H. Shu deciduous forests. Pp. 132-1 gart D. C. West & D. B. Botkin (editors) F : sion. Springer-Verlag, New NT Sones 83. Holocene vegeta io 5 eastern U States. Pp. 166-181 in H. E. Wright & S. PM ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 (editors), The Late Quaternary of the United States, Vol. II. University of Minnesota Press, Minne- li AKER & T. WEBB, III. 1973. Cal- & Plant Ecology. Blackwell, London & E. S. Deevey. 1964. Pollen accumulatio rates: estimates from late-glacial sediment of Rot ers Lake. Science 145: 1293-1295. R. EAR & L. C. K. SHANE. 1980. Ho- locene climate of New England. Quaternary Re- search 14: 240-250. ,R.B. & T. Wese, III. 1975. The contemporary istributi tion a pollen in eastern North America: m with the vegetation. Quaternary Re- search 5: 395-434. DELCOURT, H. R. 1979. Late Quaternary vegetation history of the eastern highland rim and adjacent Cumberland Plateau of Tennessee. Ecol. Monogr. 49: 255-280. DELCOURT, P. A. & H. R. DELCOURT. 1981. ear B.P. I (editog, Geobotany I. Plenum Publi Co., New ——, R. C. Bris R & L. E. Lackey. 1980. Quaternary - vegetation pesete: of Mens C Embayment. Quaternary Rese Massil . DAVIDSON. A 1983. and calibration of modern pollen-vesetanen e lationships i in the southeastern United Sta E^ Paleobotany and Palynology 39: I view o ehe EMiLIANI, C. 1972. Quaternary hypsithermals: esearch 2: 270-273. ae FERNALD, M. L. 1925. Persistence of plants pi glaciated areas of Boreal Am ca. Mem. Acad. Arts 15: 238-342. FiRBAs, F. 1949. Spät- und aisi — geschichte Mitteleuropas nordlich pen, 1. Gustav Fischer, mes: pp of s of Nir consinan vegetation in det U.S.A. Bull sash 2715 Pac eomagnetic, biostra climatic record. Bull. Geol. Soc. Am 14 Honey, B. & H. J. B. Birks. 1983. pe Past and Present Pollen for Europe 0-1 An Atlas of 000 Years re Cambridge University Press, i ut Jacomsow,G.L. Jm. 1979. The paleco oloy oe 691- pine (Pinus strobus) i in Minneso 7 Jonnson, R. G. - EA versal dated at 790,000 year B. P. by —— icon a tions. Quaternary 17: 135-147. tmn Kine, J. E. & E. H. Linpsay. 1980. Tar tas biotic records from spring — ? coil souri. Pp. 63-78 in W. R. Wood & R. B.M Poio; Prehistoric Man and His En c Press, New hie T ju & M. B. Dvd 1972. Pollen ^ dence of Pleistocene and Holocene vegeta | | 1983] the Allegheny Plateau, Maryland. Quaternary Re- search 2: 506—530. Moe, D. 1970. The post-glacial immigration of Picea ies into Fennoscandia. Bot. Not. 123: 61 Ohare. T T 2 WEBB, III. 1983. Calibration of nu merical diss milarity measures for matchi modern and fo ssil spectra. Bull. Ecol. Soc. (e 64(2):155. PETERSON, G. M., T. WEBB, III, zs thai pip, T. VAN DER HAMMEN, T . A. WUM & F. A. STRUT 1979. The continental record of environmental nitial investi- 782, 1935. British floras diitecedenit to the at I e. In Discussions on the origin and oerni pé the British flora. Proc. Roy. Soc London, Ser. B Biol. Sci. 118: 197-241. SOLOMON, A. M. 1982. Plant community respons to decreased seasonality during full glacial tae AMQUA deis 1982: 82. , D. J. A. SoLoMON. 1981. The role ofi climate Point and species immigration in for- est succession. Pp. 154-177 i West & D. B. Botkin (editors) ay Succes- sion. Er CA New TAM AE, J. & T. W. ANDERSON. 1970. Hypsi- — al range uo of white pine (Pinus stro- of bus L.) in Quebec, Canada. Canadian Journal 13, 3. Some ary plants. In A. tlam (editor), Atlas of reaktante El- k sevier, Amsterdam. RAVERSE, A. 1982. Res esponse of world vegetation to Neogene tectonic and climatic events. Alcheringa ; 6: 197-209. ^N DER PIL, L. 1969. Principles of Dispersal in Higher Plants. op a New York. DER HAMMEN, T., T. A. W RA & W. H. ZAG- ( . The floral iic of a? late Ceno- zoic of Europe. Pp. 391—424 in K. K. Turekian Da , The Late Cenozoic Glacial Ages. Yale niversity je New Haven. VAN Di Whi Late Qua te ernary vegetation history at Perd Pond on the inner coastal plain of South lina. Quaternary Research 13: 187-199. : 1983. Vegetation history of eastern United DAVIS—QUATERNARY HISTORY 563 States 25,000 to 10,000 years ago. Jn H. E. Wright & S. Porter (editors), The Late Quaternary of the United States, ies s I, the late Pleistocene. Uni- versity of Minnesota Press, Minneapolis. & M. ST TUIVER. "1980. Late Wisconsin climate ofn ern Fl he origin of species-rich : 325-327. $ Long- distance dispersal and the n into Wisconsin. AMQUA Afin. 1982: WEBB, ae III. 1981. The past 11,000 years of vege- tatio nal change in eastern North America. Bio- science 31: 501—506. Howe, R. . BRADSHAW & K. HEIDE. Estimating plant abundances from pollen iew , S. 1981. of Paleobotany and Palynology 34: 269- 00. West, R. G. 1961. Late and postglacial vegetational history in Wisconsin, particularly changes asso- cita with the Valders readvance. Amer. J 259: 766-7 ————. 1970. Pleistocene history of the British flora. Pp. 1-11 in D. Walker & R. G. West (editors), Studies in the Vegetational History of the British Isles. Cambridge University Press, Cambridge. . 1980. Pleistocene forest history of East An- glia. New P 85: 571 WHITEHEAD, D. R. 1973. Late-Wisconsin ean | North ica. Quaternary Research 3: 621-631. . 1981. Late-Pleistocene vegetational changes in northeastern North Carolina. Ecol. Monogr. 51: 451-471. WOILLARD, G. M. 1978. Grand Pile peat bog: a con- iras pollen record = the last 140,000 years. rnary Research 9: 1- 978. A paleobotanical interpretation hern hemisphere. 79. Temperature parameters of humid to of other regions of the northern hemisphere and Australasia. U.S. Geological Survey Professional Paper 1106. Woops, K. D. & M. B. Davis. 1982. Sensitivity of 1982. ety of natural sys- WRIGHT, H. E., JR Po di JA Abstracts 1982 12-16. THE TIBETAN PLATEAU IN RELATION TO THE VEGETATION OF CHINA DAvID H. S. CHANG! Vegetation zones of the eastern coast of China represent a series of forests from the tropical to southeastern coast to the northwestern interior (11). These patterns are quite different from those in western Eurasia where the vegetation pattern of the western part from south to north is: trop- and savanna, Mediter- and coniferous forest. This series shows a pat- tern, drier in the south and moister in the north, that basically reflects Hadley's classic diagram of atmospheric circulations (10, 25, 26, 28). There- fore, the differences between the eastern and western parts of Eurasia reflect an *asymmetric" distribution of vegetation (Fig. 1). The geograph- ic distribution of vegetation in China-Eastern Asia should be interpreted primarily in terms of pat- terns of atmospheric circulation in Asia because it reflects the “projection” of atmospheric cir- culation onto the land surface of the earth. Recent research in atmospheric physics indi- cates that the Tibetan Plateau in eastern Asia is a huge “Hot Island" that hinders atmospheric circulation (5). The powerful thermal and oro- graphic effects of the Plateau cause great changes in the atmospheric circulation of the northern hemisphere, especially in Asia, and have a direct effect on patterns of vegetation in that part of the orld rid. Hereafter I will discuss the uplift of the Plateau in relation to atmospheric circulation and the distribution of vegetation in China. The most important events in this regard are the formation of the Tibetan High, the establishment of the nnana ti 1 and +h 254 f VAL UL the Eastern Asian Monsoon. FORMATION OF THE TIBETAN HIGH AND THE PLATEAU ZONATION OF VEGETATION IN TiBET The Tibetan High, which arises from the pow- erful thermal effect of the Plateau in the summer and the cooling effect ofthe Plateau in the winter, ! Cornell of China. ANN. MISSOURI Bor. Garp. 70: 564-570, 1983. is a vast system of the atmospheric circulation in the upper atmosphere (5) (Figs. 2 and 3). Its center is located in Ali, the western part of the Plateau. This region has very little precipita- tion—only about 50 mm annually—and is there- fore a desert or semidesert climate. Figure 4 shows the extent of the resulting vegetational region. The vegetation is suffi t desert and steppe desert types that are composed mainly of Cer atoides latens. In the northwestern part of the Plateau, the climate is very dry and cold by re son of high altitude and its more northern lati- tude. A sparse high-cold desert of low suffrutes- cent and cushionlike Ceratoides compacta has developed there. On the vast flat of the central Plateau, the annual precipitation 1ncrease about 200 mm, and a high-cold steppe vegetatio" of Stipa purpurea prevails. This is the major vegetation type ofthe Plateau. In the eastern part of the Plateau there is a cold, low-pronups P E : 3 f der this cold and wet climate, a special kind o east to northwest, grading from der ow and scrub through high-cold steppe and - to high-cold desert. This sere of high-cold aad tation was formed and developed in the Qu@ nary. ' The Tibetan High influences not only the "T teau itself, but is also the center of a great in climate and vegetation: the southe the Plateau is the wet, rainy, and ern Himalayas, the annual rainfall exceeds p the mm, making this one of the rainiest m have world. Luxurious tropical mountain pec developed there. On the opposite. north of the Plateau are the Kunlun Mountains: University, Ithaca, New York. Current address: Agricultural College, Xinjiang, People's R este uoa mmm emma ~ 1983] CHANG—TIBETAN PLATEAU WESTERN FRINGE CONTINENTAL INLAND EASTERN FRINGE ATLANTIC COAST OF W. EUROPE & N. AFRICA MID EUROPE - ARABIAN PENINSULA WESTERN SIBERIA -CENTRAL ASIA- GANGES PLAIN PACIFIC COAST OF EASTERN ASIA TUNDRA BOREAL CONIFEROUS FOREST (TAIGA) TEMPERATE STEPPE SUBTROPICAL DESERT TEMPERATE ERA z DECIDUOUS FOREST TEMPERATE Y TEMPERATE DESERT DECIDUOUS FOREST SESE a ee B MEDIT. | : TE oni pea“ to eae TIBETAN PLATEAU SUBTROPICAL EVER- E HIGH - COLD GREEN BROAD-LEAF s VEGETATION FOREST SAVANNA TROPICAL SEASONAL & RAINFOREST ARABIAN SEA DESERT & | TROPICAL SEA- SAVANNA SONAL FOREST TROPICAL MONSOON RAINFOREST INDIAN OCEAN SOUTH CHINA SEA FIGURE 1. Asymmetric distribution of horizontal vegetational zones in Eurasia and North Africa. - TROPICAL EASTERLY JET m Á Zu. — aÁ— É — — —— — -— F 'GURE2. The summer circulation above East Asia (22). 565 PACIFIC OCEAN 566 ANNALS OF THE MISSOURI BOTANICAL GARDEN POLAR FRONT JET — — —M _ N ARCTIC FRO 5 ter + TROPICAL SEASONAL FOREST & RAINFOREST REGION SENSU FIGURE 4. The map of vegetational regions of East Asia. [Vo.. 70 — —À ~ — — — -= 1983] are the driest mountains in the world. Desert vegetation covers the mountain slopes from the foothills up to the snowline. At the base of the Kunlun Mountains there is a vast expanse of sand desert, the Taklamakan—the arid core of Asia, where annual precipitation is less than 10 mm. East of the Plateau, across the Traverse Mountain Range and reaching to the shore of the Pacific Ocean, is an extensive subtropical ever- green broad-leaved and temperate deciduous broad-leaved forest. To the west of the Plateau, through the arid, hot, and desert valley of Kash- mir, extending to the Middle East and the west coast of North Africa, is the greatest area of sub- World. The different vegetation zones on the Pla- teau itself reflect the close connection of the cli- mates, floras, and vegetation types of the Plateau to the surrounding areas. THE NORTHWARD MOVEMENT OF THE WESTERLIES, ESTABLISHMENT OF THE MONGOLIAN-SIBERIAN COLD HIGH, AND FORMATION OF THE DESERT VEGETATION IN CENTRAL ASIA tropical and subtropical latitudes. The Tibetan Plateau, which is situated between 30° and 40°N latitude, is a hugh orographic dynamic and ther- mal hindrance to the Westerlies. The Westerlies divide into southern and northern branches at = Western end of the Plateau. In the summer, i orani branch shows an anticyclone curve pe orthern Sinkiang, situated on the north side the Plateau. A high pressure zone caused by e compensational descending effect of the Ti- tan High occurs above Gansu and Sinkiang. A — anticyclonic system—the Mongolian- venti an Cold High—caused by the Plateau pre- thes thermal exchange between Siberia and teas as. Ocean, and causing an accumulation Air Over the continent, is formed in win- : This Mongolian Cold High is the controlling m of the eastern Asian weather during the winter, It makes the winter climate cold, dry, CHANG—TIBETAN PLATEAU 567 and lacking in precipitation in this area, and has caused a vast temperate desert zone to develop in the plain of Central Asia. It ranges from Iran to Sinkiang, Gansu, Qinhai, Inner Mongolia of China, and Western Mongolia. Its northern limit reaches nearly to 50°N latitude; far beyond the of Central Asia would have a moist climate that would support forest or grassland vegetation and the wet, hot, and fertile plain of India would be arid and hot climate in central China, but it was much warmer and moister in Dzungaria, which had a forest-steppe climat d vegetati t that time. After the uplift of the Plateau and the west- ward movement of the Tethys Sea, the climate in central China became moist, and in northern China the climate changed from warm and moist to cold and dry. The northwestern part of China became a temperate desert (7). The formation of deserts proceeded with uplift of the Plateau and moved gradually northward (4). The Plateau prevents the effects of marine monsoons reaching the Central Asian desert re- gion. Only a remnant of the Southwestern Mon- soon reaches the Alasan and Hexi Corridor in the summertime, and then causes a concentra- tion of precipitation. Therefore, there have de- veloped some annual summer ombrophiles in the flora that are characteristic of the eastern Central Asian desert. Dzungaria is located in the western part of Central Asia and receives some of the Mediterranean subtropical air masses. This causes more spring rainfall and has led to de- velopment ofa vegetation containing spring-om- i i I teristi the western Central Asian desert. The Beishan Gobi, East Sinkiang, and the eastern part of South Sinkiang are located in the most arid core of the desert, and are under the influence of the center of high pressure and powerful anticyclone. There is very little precipitation, usually not more than 10 mm. The desert vegetation is sparse, and there often appear vast bare areas of the “Gobi,” drift sand, and “‘Yardang.” 568 EXPANSION OF THE STEPPE VEGETATION AND COMPRESSION OF THE DECIDUOUS BROAD-LEAF FOREST BY THE EFFECT OF THE MONGOLIAN COLD HIGH There is no significant orographic hindrance to atmospheric circulation in the eastern part of Asia. Therefore, the cold Mongolian-Siberian anticyclone caused by the uplift of the Tibetan Plateau invades southward in Eastern Asia. This has permitted steppe vegetation, which is adapt- ed to arid and cold tinental climate, to expand southeastward. This has restricted the area of the mesic deciduous broad-leaf forest, which re- quires a temperate, moist climate. Several thou- sand years of agricultural activity have destroyed the primeval forest vegetation in the central plains of China (the Yellow River Basin), one of the original centers of ancient culture, and acceler- ated the development of the steppe. Generally, the area of broad-leaved decid f in east- ern Asia is more limited and smaller than that in Western Europe and North America. It dis- appears abruptly in the central part of eastern Asia. The spring drougl ] tł ld winter that are caused by the Mongolian High are the most important limiting conditions for development a d decid ft i as hroad.leave viveu tern Asia The temperate steppe zone, which crosses the center of Eurasia, bends southward in eastern Asia and reaches it southern limit at 34°N lati- tude, a large southward deviation. The eastern As 1 d PE 11 P J cold climate controlled by the Mongolian-Sibe- rian High and the moist climate under the influ- ence of the eastern Asian Monsoon Low. The steppe zone can be divided into three transitional subzones: the desert-steppe, the typical steppe, and the forest-steppe. Communities in these sub- zones are dominated by various steppe grasses. The sub pi ly reflect the transition from arid to semi-moist climate, and from desert to forest vegetation. ALEUTIAN Low AND PRESENCE OF THE NORTHEASTERN MIXED NEEDLE-BROAD-LEAF FOREST A temperate, mixed, needle-broad-leaved for- ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo.. 70 passes by the coast of northeastern Asia, creating a moist and rainy climate (27). In terms of at- s acid i Tad east A add aioi der the control of the Aleutian Low in the winter, and transitional between the hot Tibetan Plateau and the cold Okhotsk Sea in the summer. The northern limit of the effect of the warm-humid Eastern Monsoon, which produces abundant precipitation, is in this region. Therefore, eastem China is more moist than northern China, and serious droughts are rare. According to the research of Kasahara and Asakura (21), if the Tibetan Plateau were not present, the Aleutian Low and the Okhotsk High might not form or might be unstable. It might be, then, that the Plateau affects the distribution of the northeastern mixed needle-broad-leaved orest. Not more than 500 km from the coast, the maritime climate is replaced by a continental climate controlled by the Mongolian High, an the mesic mixed forest disappears. Beyond be taiga of the Great Xingan Mountains, the vege tation rapidly becomes steppe. PREVALENCE OF THE EASTERN MONSOONS AND DEVELOPMENT OF THE BROAD-LEAVED EVERGREEN FOREST 1 ond SOUS The subtropical climate of lot eastern Asia is controlled chiefly by the ee pressure system of the Southeastern and 50 tes western Monsoons in the summer. In the yet it is influenced strongly by the Mongolia® High. In this region monsoons are more we à; developed than elsewhere and are mone ffects ofthe presence ofthe Tibetan Plateau. Thee of the Plateau are: n The importance of the Southwestern M pue (21). There is a branch of the humi€ >. eastern Monsoon that comes from Ocean, passes over the mountains O Yunnan and Burma, and blows into ith the Eastern Asia in summer. It converges and northwestern current at 30-36°N pore si for the southern and southeastern i tropical broad-leaved evergreen fores increases and maintains the stabili sity of the monsoons in Eastern Re ks theri mits the monsoons to reach to bo -——— ———À—s———— r] j 1983] and southerly extremes (15, 16). The reinforced uc 1 za 4 * fy whi h © , c is a major factor for the development of the southern and southeastern Asian broad-leaved evergreen forest. The tropical cyclone. Typhoons also have an important effect of increasing the water supply forthe eastern Asian broad-leaved evergreen for- est. The Tibetan High is a strong influence in turning typhoons northward. The southwestern low vortex. This powerful vortex forms on the plateau and moves eastward in the summer. It is an important weather sys- jee producing summer rainstorms in eastern sia. Therefore, the Tibetan Plateau is an orograph- ic and thermal factor that maintains, reinforces, and even forms the eastern Asian Monsoons and other weather systems. Its effects are indispen- sable for the development of the southern and southeastern Asian subtropical broad-leaved ev- ergreen forest. The savanna or thorn scrubs of subtropical desert do not develop in a monsoon climate. Only a few relicts of these types of vege- tion are present in rain shadows of the dry and hot valleys of the Traverse Mountain Range, where there is a hot low-pressure system formed by the descent and convergence of the Westerlies. ESTABLISHMENT OF THE TROPICAL EASTERLIES AND DISTRIBUTION OF THE TROPICAL RAINFOREST AND MONSOON FOREST IN SOUTHEASTERN ASIA ne Powerful Easterlies, which occur south of latitude in southeastern Asia, India, and hei. A frica in the summer, are one of the high- " Ion systems that control the climate by ese tropical areas (19). They are supported ~) the Tibetan High and Saharan High (23) and icd prevalent at the south edge of the teau 24 : . i h : The Easterlies bring abundant rainfall into thern Asia and are the major supplier of water Bu. in winter, the dry season of such tropical man » the tropical deciduous forest enters dor- ^Y. In North Africa and the Middle East, dis: à powerful descending air-flow on the of tro ™ side of the Easterlies, causing a vast area Pical desert. — Southwestern Monsoon, caused by the of the Plateau, has had an important effect CHANG—TIBETAN PLATEAU 569 on the development of the tropical rainforest in the southwestern part of tropical China, India, and the Peninsula of Middle South. The warm, humid air of the Monsoon comes from low al- titudes over the Indian Ocean and blows into Bengala and the eastern Himalayas. It is trans- formed into a cyclone curve by the horseshoe- shape of the mountain barrier, and produces abundant rainfall in the mountains. The annual precipitation of 5,000 to 10,000 m maintains luxuriant montane tropical rainforests and sea- sonal semi-evergreen forest. Because of the oro- graphic and humid-hot monsoon effects, the most limit here at 29°N latitude. The tropical seasonal forests of Xishuangbanna in Yunnan are also supported mainly by the Southwestern Monsoon. However, as the cyclonic winds of the eastern Himalayas turn westward and become an east wind along the southern foothills of the Himalayas into Punjab, the rainfall decreases sig- nificantly westward and the dry season becomes more severe. Thus, there are no rainforests on the southern foothills of the central Himalaya Mountains. Instead, there are monsoon forests that are mainly composed of the deciduous Sho- rea robusta. Additionally, the Tibetan Plateau becomes a hindrance in the western part of Eastern Asia in the invasion of the Siberian cold current in the winter. This is an important reason for the north- ward deviation of the limit of tropical vegetation there. Because of this obstruction, the Yunnan Plateau, the eastern Himalayas, and the Indian Peninsula are rarely invaded by cold currents. The temperature in the winter there is higher than that further east. Because the eastern plain of China lacks the protection of high mountains, the cold current from the Mongolian Cold Hi can reach to the far south, producing cold waves 1n uic + P leading to cold injury of rubber trees, bananas, é : 1 A NEA al 4. : 1 and other trop Tops p g limit southward to the Tropic of Cancer. There- fore, the northern limit of tropical vegetation in China reaches higher latitudes in the west than in the east. CONCLUSIONS 570 northward extension of the subtropical desert zone and formation of temperate desert in Cen- tral Asia: the Victims ot the toipre steppe fbroad- leaved deciduous forests: the preservation of the mixed needle-broad-leaved forest in northeast- ern Asia; the reinforcement of the eastern Asian Monsoon region causing an expanded develop- ment of the south and southeast Asian subtrop- ical broad-leaved evergreen forest, instead of nere savanna and desert there; and the odifi rthern limit of the tropical forest zone in parts of E ei Asia. These features of vege- tation zonation, all the result of the uplift of the ibetan Plateau, are unique to Eastern Asia. LITERATURE CITED (1) THE METEROLOG IC BUREAU OF CHINA. 1975. The Level Climate of China. Science Press. (2) SIXTH Group, First ION OF RESEARCH I T East he Meteorologic d Js the Qinghai Xizang Plateau (1975-1976). Pp. 7 (3) THE ses idea OF dis U ATMOSPHERIC NZHOU. 1976. The Movie of st Be Qinghai-Xizang sissy Academia Sinica (4) THE pore GROUP OF CLIMATIC CHANGES OF T GIC ERE PLATEAU. 1977. The Cind Regime Before and After the Uplift of the Qinghai-Xizang parser The Institute of ography, Academia Sin (5) Yie DU-ZHENG & ZHANG JiE-GiAN. EA The heat- Hn of the Qinghai- -Xizang Plateau to the east- SULI r. Zhongguokexue 3: 301-320. Jordin vai PHY, 1977. The na. Sect. 10. The Peking rmal Uni Mu Pg. 64. 46—53. nce Press. (8) gere ens (CHANG HSINSHIH). 1978. The ateau zonation of vegetation in Si Bot. Sin. 20: 140-149, EN (9) THE RESEARCH GROUP OF THE Low BARIC SYSTEM OF THE QINGHAI- XIZANG PLATEAU. low of the Qinghai- -Xizang Plateau mid-su ummer. Site 9: 4- (10) Ha XüÜEJ-YÜ, n the en and the- ore aree of es vegetation regionalism Bot. Sin. 9: 275- "e im idee (11) tea kür- -YÜ & ZHANG XIN-sHIH. 1980. The phic Distributed Pattern of the Vege- ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 pom of _ The Nee of China. Pp. —738. Science P (12) 2 Fu-j1, Lu Lonc- HUA & JIANG FENG-YING. southern rology 7: 9-11. (13) Sie enti xi. 1962. The Problem of Monsoon. em e Problems of the Eastern Asiatic Mon- JEF n. Pp. 2-11. Science Press. (14) e You-xt. 1977. The Influence of m bes Plateau to the Climate of China. ni Ms terologic Collectanea of pi rit: nghai-Xizang Plateau (1975-1976). Pp. 3 (15) o, You-x1, XO Suu-vmo, Guo QIi-YUN & ANG MIN-LI. The Mut Regions dd Regional Climates of China. Some Prob- lems kA the Eastern Asiatic Monsoon. Pp. 49- nce Press. Asiatic Monsoon Climate. ner isa a. Asiatic Monsoon. Pp. 12-27. ce Pre (17) Tue: RESEARCH ns OF THE PLATEA ATLAS. 1977. The Characters ofhei er + Field upon the Qu ang bre Adjacent Regions. The Meteorologic d ea of the Qinghai-Xizang Plateau ( -10. 976). e (18) THE RESEARCH uerus OF THE PLATEAU CLIMA ix ATLAS. 1977. The Characteristics of the Th tribution of Aleme of the Plateau. Meteorologic Collectanea of the Q Qinghai- ang Plateau (1975-1976). Pp. 22-33. per mperate cyclone 21) iai Tapasui. 1974. The T iben M, d Climate of he World. The AERE: | the World. (22) FLoHN, H. 1968. puo to a meteorol- ogy of the Tibetan Hi ands. Aso ence Paper 130. Colorado p^ . The (23) REITER, E. The Tropic Eastern Jet tream. Free Atmospheric Cli j ii (24) RAGHAVEN, K. 1 The Tibetan we jor tropic eastern jet stream and Appl Geophysics 110. (25) Tron e 1968. The Cordilleras of the Tropica Am s. Aspects of the Climatic, F the Tropic Soil and Climate, D. 5-56. (26) Waa ace V. R. 1953. A. Zones 9% Azerbaizanskoi SSR ae (27) MakEev, P. S. 1956. The Physi Landscape. Geographgiz. Sov- (28) SCHENNIKOV, A.P. 1950. Plant Ecology: and Cenoz ets (29) He ian zb. Late Cretaceous con in China, emphasizing bea rth America. Ann. — ———— ——— wn A COMPARATIVE STUDY OF THE VEGETATION IN HUBEI PROVINCE, CHINA, AND IN THE CAROLINAS OF THE UNITED STATES ZHONG CHENG! The study of the floristic relationship between A involve extensive research of widespread genera with disjunct distributions as well as intensive investigations of paired species occurring in lim- ited areas. My report rey ts the result of field observations and herbarium and library inves- tigations of two comparable areas: Hubei Prov- ince in Central China and the Carolinas in Fast- ern U.S.A. It covers the geographic background and general aspects of the floras of the two re- gions, and compares the similarities and differ- ences of the vegetations they support. GEOGRAPHICAL BACKGROUND AND GENERAL PECTS OF THE FLORAS Hubei, a medium-sized province in central China, has an area of 187,000 square kilometers andis located in the middle portion ofthe Yangtze Valley, between 29°05’ and 33°20’ north latitude. lt is surrounded on three sides by mountains, ie the higher ones in the northwest reaching 000 m or more in elevation. Eastern Hubei is hilly and the central-southern portions of the Province form the J ianghan Plain. ci € flora of Hubei is one of the richest in cae especially in the western Hubei moun- Ous region where remnants of virgin forests um According to preliminary inves- iai em the province has about 3,816 species 1,165 2 varieties of vascular plants belonging to >: 93 genera and 207 families. This flora is com- eng tropical, subtropical, and temperate ele- E : i vergreen coniferous forests occur in some of the higher 1 : a. SES in western H $ NAM ry: Wa ubei. Small deciduous forests occur at some intermediate elevations, but usually the decid- uous and evergreen coniferous species intermin- gle. At lower elevations, pe by broadleaf evergreen forest. Limestone areas are sometimes occupied by Cupressus funebris, especially where the rock is near the surface, as in the Changjiang (Yangtze) Gorges. The predominant species of deciduous trees in Hubei belong to the following genera: Fagus (F. longipetiolata, F. lucida, F. engleriana), Casta- nea (C. seguinii, C. mollissima), Quercus (Q. acutissima, Q. acutidentata, Q. aliena, Q. var- iabilis, Q. fabri, Q. glandulifera var. glanduli- fera, Q. glandulifera var. brevipetiolata), Betula (B. luminifera, B. insignis, B. utilis—all in west- ern Hubei mountains), Carpinus (C. fargesiana, C. fargesii), Juglans (J. regia, J. cathayensis), Pterocarya (P. stenoptera, P. hupehensis), Platy- carya (P. strobilacea), Acer (A. davidii, A. mono, A. amplum, A. wilsonii—most in western Hubei mountains), Populus (P. adenopoda, P. davidi- ana, P. wilsonii, P. lasiocarpa—most in the west- ern part of the province), Sorbus (S. alnifolia, S. wilsoniana, S. hupehensis, S. folgneri—most in western Hubei mountains), and Tilia (T. tuan, T. oliveri—in western Hubei mountains). The major species of the broad-leaved evergreen trees belong to these genera: Quercus (Q. glauca, Q. fargesii, Q. myrsinaefolia, Q. engleriana—most in western Hubei mountains), Castanopsis (C. tibetana, C. sclerophylla), Lithocarpus (L. gla- ber, L. henryi), Phoebe (P. neurantha, P. chinen- sis), and Cinnamomum (C. camphora, C. wil- sonii, C. bodinieri—most in western Hubei mountains). The major evergreen conifers are Pinus massoniana and Cunninghamia lanceo- lata, which usually occur in areas below 1,200 m in elevation. Pinus armandii occurs between 1,200 m and 2,500 m in northwestern Hubei, and Abies fargesii and Abies chensiensis occur at the higher elevations in the northwestern portion of the province. We visited and collected specimens in the UVVUpIVA ‘Wuhan Institute of Botany, Academia Sinica, Wuhan, Hubei, People's Republic of China. ANN. Missouri Bor. GARD. 70: 571-575. 1983. 572 A 1 southern Apr hi tai April to late June and September, 1982. Although our visits were short (three months), the field observations and a study of the American botanical infor- mation obtained from literature have convinced us that Hubei Province and the Carolinas, even though separated by a great ocean and thousands of miles of land, are very similar, not only in their climates, but also in the vegetation. North Carolina and South Carolina are bor- dered by the Southern Appalachians on the west and by the Atlantic Ocean on the east. This area has a total area of 217,000 square kilometers, and lies between 32? and 36?40' north latitude, a little further north than Hubei Province. The Carolinas are naturally divided into three phys- iographic regions: the Mountains in the west, the Piedmont, and the Coastal Plain in the east. In general, the mountains of the Carolinas are similar to the western Hubei mountainous re- gion. They vary in height from about 350 to 2,025 meters: Mt. Michell at ca. 2,025 m is the highest mountain in North America east of the Mississippi River. In western Hubei, the moun- tains are somewhat higher, with an average el- evation of 1,000 meters. The highest peak, Non- name Peak, in the Shennongjia Forest District, reaches 3,105 meters. On some of the high summits of Carolinian mountains occur fir forest of Abies fraseri and spruce forest of Picea rubens. The forests of most areas below the fir and spruce forests are com- posed of mixed deciduous species of Quercus (Q. alba, Q. rubra, Q. coccinea), Betula (B. lutea, B. There are two species of Liriodendron, one from China, and the other from the United States. In western Hubei, small areas of pure Liriod- endron forest are rare; usually it is mixed with other species and is not abundant. The American species, Liriodendron tulipifera, is common in the Carolinas, where there are large areas of pure yellow poplar forests. Aesculus, distributed in temperate zone, consists of about 25 species. In Hubei Province there is only one species, Aes- culus wilsonii, and its distribution is limited in western Hubei mountains. There are three species of Aesculus in the Carolinas, only one of which is a large tree, Aesculus octandra ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo.. 70 Amelanchier arborea and its two varieties (var. laevis and var. austromontana) are widely dis- tributed in the mountainous region. In Apriland _ May, their white to pinkish flowers are conspic- | uous along the mountainsides and in the valleys. In Hubei, there is a single species of Amelanchier, | A. sinica, which is a shrub and is distributed at _ lower elevations. ; The predominant conifers in the mountains of the Carolinas are Pinus echinata, P. strobus, P rigida, and P. virginiana, Picea rubens, and Abies raseri. Generally, the Piedmont region of the Caro- — linas is comparable to the hilly areas of eastern Hubei, despite the fact that the elevation of the Piedmont is lower than its counterpart in China. In the Piedmont, altitudes vary between 100 and 350 meters, whereas the Dabieshan Mountains in eastern Hubei are about 1,700 meters. The vegetation of both regions is primarily mix deciduous forest. Important elements of the Piedmont region are Quercus (Q. alba, Q. p cinea, Q. stellata, Q. michauxii), Fagus (F. ds i difolia), Carpinus (C. caroliniana), and RT (B. nigra). The major deciduous trees 1n jc : Hubei are: Quercus (Q. variabilis, Q. acut : tata, Q. aliena, Q. dentata), Castanea - a lissima, C. seguinii), and Alnus (A. es at ui i i i nus , Of particular interest is the ge ^ nde one urring . 1 ra) comparable to the Jianghan Plain bio bei Province. Both constitute the larges : i ions lowest elevation of the respective reg! The watt lai e ilandica © uercus laevis and Quercus mariland s mon in the Coastal Plain region. portant components in the lowland Coastal Plain of the Carolinas large red by six species of 1 ove y pines. diet (o long-leaf pine (Pinus palustris), loblolly —————————— = — RE — CH- 1983] ZHONG-HUBEI PROVINCE TABLE 1. A summary of the floras of Hubei and the Carolinas. 573 Spermatophytes Angiosperms Pteridophytes [Gymnosperms Monocots Dicots Total i Caro- Caro- Caro- Caro- Caro- Categories Hubei linas Hubei linas Hubei linas Hubei linas Hubei linas Families 35 16 7 3 23 31 142 130 207 180 Genera 67 30 23 8 169 226 906 687 1,165 951 Species 182 89 37 21 629 940 2,968 2,310 3,816 3,360 Infraspeci- c 0 0 5 0 0 42 237 132 242 174 nus taeda), pond pine (Pinus serotina), short-leaf pine (Pinus echinata), slash pine (Pinus elliottii), and spruce pine (Pinus glabra). A COMPARISON OF THE FLORAS The floristic relationship between Hubei and the Carolinas are very close, with comparable number of species, genera, and families. The number of vascular plants of both regions, taken from the Flora Hupehensis (Wuhan Institute of Botany, 1976, 1979) and Radford et al. (1968) are shown in Table 1. ; Families that contain more than 20 species of ems, 10 species of gymnosperms, 25 species of cotyledons, and 50 species of monocotyledons are given in Table 2. Pics floras of the two regions have numerous ues and genera. Of 207 families of ens plants in Hubei and 180 in the Caro- '5, approximately 75% are represented in both ‘gions. Many of the genera are common to both “gions. For example, in Urticaceae, there are 15 senera in Hubei; 5 of these are represented in the Adee Le, Urtica, Laportea, Boehmeria, ia and Pilea. In the Betulaceae, Betula, ioi orylus, Carpinus, and Ostrya are genera es "wm to both floras. In Saxifragaceae, there 10 genera in Hubei, and 14 in the Carolinas, of which are shared. thas Od and American botanists have shown Ene nd genera the species that occur in M E sia and North America are closely re- dé bi or instance, in Buckleya, Phryma, Cor- fe iphylleia, and Podophyllum there are paired €s in the floras of Hubei and the Carolinas. Teig are, however, some obvious differences een the two floras. First, at the family level, are 44 families in Hubei not represented in the Carolinas, and 29 families in the Carolinas not occurring in Hubei. Families in the Hubei flora not in the Carolinas are as follows: Ferns — Angiopteridaceae, Plagio- gyriaceae, Gleicheniaceae, Dicksoniaceae, Mon- achosoraceae, Lindsaeaceae, Sinopteridaceae, G ogrammaceae, Parkeriaceae, Loxogram- maceae; Gymnosperms — Ginkgoaceae, Taxa- ceae, Podocarpaceae, Cephalotaxaceae; Dicots — ceae, oracea etracentraceae, Eupte- eaceae, Cercidiphyllaceae, Sargentodoxaceae, Illiciaceae, Pittosporaceae, Euco ceae, giaceae, Myrtaceae, Trapaceae Myrsinaceae, Pedaliaceae, Gesneriaceae; Mono- cots — Stemonaceae, Zingiberaceae. Families in the Carolina flora not in Hubei flora: Ferns—Isoetaceae, Grammitidaceae; Di- cots— A Bataceae, Empet Cac- taceae, Diapensiaceae, Cyrillaceae, Sapotaceae, Cabombaceae, Martyniaceae, Elatinaceae, Po- dostemaceae, Sarraceniaceae, Plumbaginaceae, Dionaeaceae, Turneraceae; Monocots— Brome- liaceae, Arecaceae, Zannichelliaceae, Zostera- ceae, Ruppiaceae, Juncaginaceae, Xyridaceae, Mayacaceae, Cannaceae, Marantaceae, Haemo- doraceae, Burmanniaceae. Secondly, in the flora of Hubei the woody ele- ments are more diversified than those of the Car- olinas. Based on an incomplete survey, there are about 650 species and varieties, belonging to 200 genera and 73 families, of woody plants native to Hubei. The number of species and varieties of woody plant in the Carolinas flora is 217, belonging to 82 genera and 46 families. The gymnosperms are much better represent- 574 TABLE 2. Major families in Hubei and the Carolinas. ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 Hubei Carolinas | Species Species | Categories Families Genera & v Genera & vars. Ferns Aspidiaceae 1 (1) 7 2 thyriaceae 6 20 (placed in Aspidiaceae) Gymnosperms Pinaceae 8 18 4 14 Dicots Salicaceae 2 31 2 (13) Fagacea 5 47 3 43 Polygonacea 6 72 7 45 Caryophyllaceae 15 36 17 53 Ranunculacea 21 94 16 67 Berberi 8 48 7 (8) Lauraceae 9 54 4 (5) Brassicaceae 22 45 26 73 ifragaceae 18 60 14 34 Rosaceae 31 215 23 117 baceae 55 131 50 184 Euphorbiaceae 21 55 11 45 racea 2 30 1 (10) Rhamnaceae 8 39 J (5) Apiaceae 34 75 36 84 Ericaceae 7 41 19 10 Primulaceae 4 47 6 (18) Lamiaceae 42 105 34 90 Scrophulariaceae 25 79 30 85 Rubiaceae 23 47 11 37 Caprifoliaceae 10 61 6 28 Asteraceae 70 306 85 449 Monocots oacea 70 150 93 ee: Cyperaceae 14 120 17 196 Liliaceae 36 140 35 106 Orchidaceae 44 104 os neo db NM | ed in Hubei than in the Carolinas. Metasequoia glyptostroboides is one of the most outstanding and significant species in the Hubei flora. This relict species was found in western Hubei, where there are about 5,000 very large met ia trees growing in a small valley in Lichuan Xian. Thirdly, each flora has its own endemic and unique genera and species. In Hubei they are: aruma (endemic genus) (Aristolochiaceae), Tetracentron (Tetracentraceae), Sinofranchetia (Hamamelidaceae), Bretschneidera (Bretschnei- deraceae), Fortunearia (Hamamelidaceae), Ho- 1 linas: Dionaea (endemic genus) (Dionaeaceae) Hexastylis (Aristolochiaceae), Houstonia (Ru- biaceae), Kalmia, Leiophyllum, Menziesia (Eri- caceae), Galax (Diapensiaceae), Gaylussacia (Ericaceae), Sarracenia (Sarraceniaceae), am thorhiza unculaceae), Ceratiola en ceae), Spigelia (Loganiaceae), and Conop (Orobanchaceae). endemic to the perennial The monotypic genus Dionaea, Carolinas, is a rare and fascinating known only from the Coastal Plain. Fourthly, there are also some dil vegetation types. In western Hubel, above 2,000 m, large areas of arrow pc (Sinarundinaria nitida) forests are comm oet the lowland of the Jianghan Plain of ape 100 are large numbers of aquatic plants, W! or more species and varieties. The unique vegetation ndv a aive ) of the Piedmont, there are crops, such as 40-Acre Rock, in penes which support many lithophytes, SUC ZHONG— HUBEI PROVINCE pusillum (Crassulaceae). In the Coastal Plain, ex- tensive sandy areas such as that of Cheraw in South Carolina, support pine forests with Pinus palustris as the most common species, and usu- ally not mixed with shrubs. Herbs in sandy areas there are Amsonia ), Selagi- nella arenicola (Selaginellaceae), and “Opuntia compressa (Cactaceae). To a visitor from China, the grasslands of the Carolinas are a distinct fea- ture of a landscape not found in Hubei Province. LITERATURE CITED BouFFoRD, D. E. 1982. Plants of the Appalachian Mountai iwu Zazhi (Plant Magazine) 1982(3): 42-43, back cover. (In Chinese.) BRAUN, E. L. 1974. Deciduous "Fo rests of Eastern North E ae of the edition of 1950.] Free Press, New Duncan, W. H. S. An scular Flora of Georgia: xd ^n: ai “University of Georgia SS 315 NORTH CAROLINA DEPT. OF AGRICULTURE. 1896. h Carolina and Its Resources. M. I. & J. C Stewart, Winston. RADFORD, A. E; H. E. Aures & C. R. BELL. 1973. Manual of the Vascular Flora of the Carolinas. University of North Carolina Press, Chapel Hill. SARGENT, C. S. (editor). 1913-1917. Plantae Wilso- nianae. 3 Vols. University Press, Cambridge Wu CHENG-YIH. 1979. The regionalization of Chinese flora. Acta Bot. Yunn. 1(1): 1-22. (In Chinese, English summary.) WUHAN giis or BorA ACADEMIA SINICA. 1976. Flora Hu serum "Vol. 1. Hubei People's Press, Wuhan. (In Chin 79. Flora Hisebensie M 2. Hubei Peo- ple’ s Press, Wuhan. (In Chine: Flora Hupehensis, Vols. zu Forthcoming. (In Chinese.) 1980. Shennongjia Plants. Hubei People's Press, hires (In Chine YiNG TSUN-SHE ONG & CHANG 97 . Ob tions of the flora and vegetation of Mt. Shennongjia in Western Hupeh, China. Acta Phytotax. Sin. 17(3) 41-60. (In Chinese, English summary.) T . Volume 70, No. 1, pp. 1-210 of the ANNALS OF THE Missourt BOTANICAL GARDEN, was published on | Moveniber | 1983. : 4 Volume 70, No. 2, pp. 211-420 of the ANNALS OF THE MISSOURI BOTANICAL GARDEN, was published on April 1984. Vegetation Map of China 350 $R iü E E ATTE 2S A fe) seen asip Aw bâ L SAR tr HK. $ (1) SEE eA LL b Rep TT san acne ZRH b MAHE ia = agi bw Xo c -— KM druide oes one. " aE a, vole eek oe Pk HAHA, £LESR AE . Á (8) ilk 3l ( xp Sod) HER PALE PR BR LH MUN llc RES a AHERE Wl ee (Due ser Bà bz (5)8ERAE , ME RETO aK II rci uE NE EL IE CE Ex c 90 95 100 l5 — e 115 T20 BARBARA a BAR (13) JE PE E ERAT (16) tct sn SRT EPR z | pua ne TII 1 tk gabe miis Eu orai / E (12) SE AGE , AVE LL HEBER b REI, Edd “RSH 1 acs EM o NM UM EL ane a cam ck iB AB bpt., BA Cc] WER FPE (18) Sivit IRE EE AS Eg an ama aceeoret DURS] Mai Gio avex [2] aennaneen. mE LE HA. Bt, WRIT ee ESC LETT WER BH (15) 3E ve ae aa baa E BER AE. BORRAR IU RNM BU Pe y ‘loca arr FEST om tap La seis pup wd "Men sen meroa AEDA | dein ^» | Bep 0 QOGMERARAGE 0 m _* BEDA anti, dH ERRARE (BO) ERI REA RR : LE Lead RERER AJA P ) err. K — (83) ili fr AE KR RC p d age pu e wg DESET] je RAT E] hien d ba cu os — RHF 54 | SHENA [55] ase. aea aremania X . REBAR a 29E bz MF JSTOR . ARERR 80 m 7 nnals of the Missouri Botanical Garden 70(3). 1983. 499 FbEXE. RARI Az RY a SEMRAHERS , FFER dd | GONERA , ME A AAC ES A HE Sen b ARCA. BRE (4 BE ARME *X. Fd. cenensawas a Ced ERE bem — (3 (39) iB f Ji EE ACH PE tu xrar te, E * ow piho oh E a ` , AMF r “ My e i 2 z i um i d ; 145 : Re OR I. — SURE , SESE He AB EX BR, MR WH ae a my. ntm mE UR BB AAS RE. BSB, WR I. FAARF , OT RE Meer KE EX HS, HE. IS S EE CEN sie, EX, OF. SR HE, BE E 4,8,4 OR d BS. KE BE J 2()he EX, BR, RE, WS ME AR. III. zk&—4F URIE), WER. OT mæ a. : E CS] me RARE SR LR NAR BO ea ee ee a eo GAR, BA HE GRA m Z HS AR, AX, EE HE, EI , HA. hI. R., pa V. REBAR, MRR AmE Wpbk REREESHR MEEK, HR AB. BR, BK AR. BK HE E36, 224 UR. SF. HR, BÈ. BF. ME SR » Js A; 14 000 000 Ht 0 150 450 75028 Legend to *Vegetation Map of China" (1: 14 000 000) NATURAL VEGETATION |. Needle-leaved Forests (1) Needle-leaved deciduous Mene T ies cold-temper- ate 22 és or i e mountai he w zone. T: x fore a. L. pen. ju L: ibid (2) Made a eigen forests on Kom of the t te zo a. Picea-Abies. b. Picea. opine abe Pinus rest c. Pinus. (3) Needle-leaved evergreen cds on semi-arid sandy soil of the temperate zo 3. Pinus sylvestris var. mongo qe odland. (4) Needle-leaved evergreen forests of the temperate 4. orest. a. P. tabulaeformis. b. P. densiflora. (5) Naedindeaved evergreen forests of subtropical or tan nd E Pm forest. Cunninghamia B cealate fore (6) a Heal evergreen ‘inate, d mountains of sub-tropical and tropical zones 9. Abies-Picea forest with Tsuga. Broad- leaved For xed no dua deciduous and needle-leaved eaf Pin (8) Broad-leaved E Lorene of the ee and sub-tropical zo 11. Deciduous ins forest. a. Quercus mongolica. b. Q. liaotungensis. c. Q. variabilis, Q. pe econ Q. glandulifera. 12. Mixed forest containing Acer, Tilia, Fraxinus, Ulmus, and Betula. 13. Mixed forest containing ans ea and trees belonging to Ulmaceae on limestone soil. (9) Montane microphyllous deciduous ean of ene and a hea zones. 4. Be orest. (10) i OA ie Iioa woodlands of the tem- per 15. Ulmus pumila woodland on semi-arid sandy soil. 16. Populus diversifolia woodland on arid saline meadow soil. (11) Mixed broad-leaved deciduous and evergreen forests on limestone soils of the subtropical zone. 17. Mixed forest containing Platycarya strobilacea, M quare eie trees belonging to Ulmaceae, (12) Mixed broad-leaved deciduous and evergreen ontane forests on acid yellow-brown soils of the saeia zone. 18. mea forest containing Cyclobalanopsis, nou sis, Fagus 19. Mixed forest containing poh acacia iroad: leaf trees, evergree s, and Tsuga. a. Cyclobalan- opsis glauca, ralis longipetioloata, Tsuga, etc. b. Quercus aquifolioides, Acer, Betula, Tsuga. - (13) m leaved evergreen forests of the subtropical 20. Mixed forest UM Cyclobalanopsis, Casta- nopsis, Lithocar, EA be on INE Castanopsis, Laura- e, Theacea (14) Bread ERE. evergreen forests of the transitional ical zone. a forest containing Castanopsis, Laura- * ib: with some trees belonging to ropie (15) Broad leaved evergreen eq Aes. woodlands or a d subtropical z 22 ne cus aquifolioides forest or scrub. (16) Tropical broad- leaved semi- giat Meg forests ests). 24. Semi- -evergreen i forest on Toe e soil. 25. Semi-evergreen forest on acid pues He (17) T 26. Tropical rain forest. (18) Tropical broad: leayed evergreen goleraphyllous ee deo 1. Artemisia loamy desert with (27) xad ate succulent halophytic dwarf Mile p ER des 42. Kalidium s saline desert. (28) "T, hriuih desert 43. Ephe edra przewalskii, pride, xanthoxylum Nitraria sphaerocarpa, Calligonum spp. gravelly 44. Caragana tibetica, Ceratoides latens sandy grav- elly de 45. Polanini ‘mongol ica, Ammopitanthus mongoli- cus, Tetraena mongolica sandy gravelly desert. > 6. Artemisia sphaerocephalla, A. ordosica sandy desert. 47. Calligonum sandy des 48. Sandy desert with ssa Tam (29) alee semi-arboreal deserts (dep commun- ne ose persicum, H. ammodendron sandy . 50. Haloxylon Heide Reaumuria soongarica : a dese 27. Manoni Scrub and Coppicelands (19) Se leaved deciduous scrub of the tempera: 8. Corylu us heterophylla, Lespedeza bicolor, ES mongolica scru 29. IE davidiana, Spiraea pubescens mon- tan 30. Vitex negundo var. ee Zizyphus jujuba, Cot coyggyria var. ? b. (20) fee road: leaved setae ud scrub acid soils of subtropical pns tropical zo 31. Rhododendron, oe m scrub. ma, Aporosa scru (21) Broad: auod ever dee succulent scrub on coral lands of the tropical zon Hee evala frutescens, a hainenensis, Pistonia (22) Broad-leaved evergreen semi-sclerophyllous thickets (shr ET on the mountains of temperate and sub- tropical zo Facbdendran’s rub. (23) Broad. leaved decla oas thickets (alpine shrublands) high mountains of temperate and subtropical zones. 39. po spp., Dasiphora fruticosa, Caragana jubata scru (24) Tundras with evergreen dwarf- Rid is mosses on A ied cde in the temperate zo 36. Montane tundra RUD. Salix roundifolia, vios vitisidaea, (25) Alpine deciduous od sub-shrubs mixed with herbs, combined with codd os vegetation of pe quic and subtropical zo 3 musiformis, peer Un tapete cush- ion-like vegetation. one erts is pe Shr nperate deciduous bac sub- — deserts. ma eee oda dese avelly des imuri a sceneries revel desert. mpegn TS Poets (30) Cold init arain DU with oss ds or matted dwa rf sub-shrubs of the temperate ratoides compacta, Ajania cd sandy grav- Steppes and Savann (31) Temperate forb- grass steppes mesoxerophytic step- es or meadow 53. ean e chinense steppe. 54. Filifolium sibiricum big tals 95. EE ischaemum, Themeda triandra var. steppe with ca EPEE of forbs. (32) Temperate redeas s. 56. Stipa grandis, S. krylovii steppe. 5r Stipa bungeana, S. breviflora steppe. 58. Te (33) Temperate dwarf-shrub or sub-shrub needlegrass 59. Stipa gobica steppe. a. with Artemisia ete, b. with Stipa glareosa, Artemisia E ylla 60. Stipa breviflora, Ajania fruticulosa steppe. 61. pie Apis dwarf-shrub or dwarf Sub shrub needle- grass steppe. (34) 2 pup. steppes of temperate and subtropical 62. ee Poa, Festuca st ee 63. Stipa velie rea steppe. with Stipa subsessili- flora var. basiplumosa. b. ‘with Festuca pseudoo- vina. 64. ne aoe cc var. basiplumosa, Ceratoi- pe. (35) subtropical d tropical sa ropogon nsa tus. goce den distans sa- ee ws with thorny scrub. Meadow and Swamp Vegetation (36) Temperate meadows. b meadow 7. Halophytic grass abd forb meadow. (37) Laine and subalpine meadows, locally combined with scrub in temperate and subtropical zones 68. Kobresia mm K. deed meadow. od Grass, forb, Car 0. Halophytic grass, yer abrasa meadow. (38) Temperat pan d agm pul ebrii Carex rhynchophysa wamp. (39) High- usce allele swamps of temperate and subtropical z 72. Carex, Kobresia swamp. CULTURAL VEGETATION |l. Onecr op annually, cold-resistant economic cro 73. p ng wheat, soybean, corn, millet, Rap beet, 74. i ie wheat, millet, potatoes, sugar beet, flux They nent barley, spring wheat, pea, potatoes, rape- eed. |. Two crops annually or three crops in two years (with rice locally). 76. R wheat, soybean, opc Mein Sweet pota- es, tobacco, apple, pear TE. Winter wheat, kaoliang, Ei millet, three crops e, pear, grape, persimmon, chestnut, walnut. 78. Winter (spring) wheat, Mee millet, three crops in wo years or two c crapa. annually; cotton, grape, vue melon, pear, apric Two crops containing upland and rice annually; subtropi- cal evergreen and deciduous economic forest and or- ch 79. panne rice, C a (rapeseed) two crops y; cotton, peanut, tea, pomegranate, at. 80. Summer (corn), winter wheat (rapeseed), two crops annually; potatoes, — tea, lacquer, red bayberry, walnut, apple, pea Double-cropping rice followed by a cool-loving crop annually or three uplan sited annually, and subtropical evergreen economic forest and or 81. Double-cropping rice followed by winter wheat (rapeseed) or green manure; cotton, ramie, mul- < 82. Single or oats Bogen rice or by winter eet eat (rapeseed) a 2 potatoes, peanut grains, five one aon years; n ange, tung o oil, A ha a 83. Since or double -cropping rice followed b winter manure annually, or , soybean, crops annually, ramie, jute, tea, tea oil, red bay- berry, orange, loquat. Double-cropping rice followed by warm-loving crops, and tropical evergreen economic f rd. 84. Double-c eet potatoes, rcane, manioc, litchi, longan, banana, pineap le. 85. Double-cropping rice, winter deed sugarcane, vanilla, sisal, rubber, coconut, coffe Made by Hsioh-Yu Hou (Hou Xue-Yu), Institute of Botany, Aca- demia Sinica, Peking Systematics Symposia Published - > h 1953 the Missouri Botanical Garden began holding a series of maai systematics symposia each fall. Since 1969, with the sixteenth symposium, the Garden has published the proceedings of most in its quarterly series, ANNALS OF THE MISSOURI BOTAN- ICAL GARDEN. Those published symposia which are available are listed below, and the price of all except the twenty-fourth i is $10.00. The twenty-fourth, which was reprinted separate ly from the ANNALS; is available for $6.00. Many of these symposium issues of the ANNALS contain other papers, but each issue listed below contains the complete - | proceedings of the particular symposium. The proceedings of the 1981 symposium will be published during the: second half of 1982, and if you wish to purchase a copy of that symposium, you may subscribe at this time. Upon publication, your copy will be sent. Sixteenth Annual Sagres 1969— TROPICAL, ISLAND BIOGEOGRAPHY. 56(3). .. 127 pp. $10. Seventeenth * ope Symposium, 1970— HYBRIDIZATION, EVOLUTION, AND SYSTEMATICS. 59(3). 142 pp. $10.00. Twentieth Annuak Symposium, den eie ANIMAL COEVOLUTION. 61). jo ... pp. $10.00. ; | Twenty-first Annual Symposium, 1974 BIOGEOGRAPHY. 620). 1 43 pp. $10.00. Twen enty-second Annual Symposom I HE POPULATION ~ LEVEL. 63(2). 100 pp. $10.00. a = Twenty-third Annual Symposium, 1976—CHEMOSYS 'EMATICS. : 4(2). 98r pp. . $10. 0 e Twenty-fourth Annual Srpen, 1977 -SYSTEMATI 3 'UDIES IN AFRICA. 2 $6.00. 4 pp. $10.00. * sixth Annual Symposium, 1979—EVOL . . GRAMINEAE. 68(1). 104 pp. $10.00. — Twenty- -seventh Annual Symposium, 1980— REPRODUCTIVE STRATEGIES IN : a AND ANIMALS. 68(2). 146 pp. $10.00. nty-eighth Annual Symposium, 1981— BIOLOGICAL. STUDIES IN CENTRAL . AMERICA. 69(3). 304 pp. $15.00. Twenty-ninth Annual. Symposium, ET MPERATE EAST 1982— -BIOGEOGRAPHICAL. RELATIONSHIPS IRN AS IA AND 1 MSATER p NORTH I ould t be 5 peopl: a $1.00 are made until payment is supinus d to c ee n invok nts. ay ved. Mail fara with your check or money rer, payable to Misco Bowsicst Garden, OF THE MISSOURI BOTANICAL GARDI ANNALS, published quarterly, contains papers in systematic botan plant biogeography, reproductive biology and other bi of i matic bosses d odes from symposia on a wide) val a both closed please find $. — — — to cover my introductory indi- B vidual Subscription to the AN- NALS at $30.00. or $35.00. . = (check one). Please begin my a 2n subscription with the 198331.) S. or eu ( ) volume es one). by A. Gray in his fi ilarities of temperate Asian and North Ame Vot ornus ier M ANNALS SSOURI BOTANICAL GARDEN 'OLUME 70 1983 NUMBER 4 | CORNUS KOUSA CONTENTS BIOGEOGRAPHICAL RELATIONSHIPS BETWEEN TEMPERATE EASTERN ASIA AND TEMPERATE EASTERN NORTH AMER- ICA: The Twenty-Ninth Annual Systematics Symposium On the e of Pacific Intercontinental Discontinuity Wu Zhengyi .... = 577 Temperate cue The North Pacific "Connection AK: ke "mE — 591 The Floristic Relationships of the Temperate Parent Repons of China and the United States Tsün-shen Ying .. pence 597 North American Trees with Relationships in Eastern OP - Elbert E 9$ 2 Little, Jr. . Distribution Paen among poems of tl he c North Temperate > De- ciduous Forest Biota Robert T. Allen .. Es 6 Biogeography of Oaks in the Arcto-Tertiary Province Axelrod . Genetic Divergence within a the Geos s Liriodendron ( (Magnoliaceae) Clifford R. Parks, Norton G. Miller, Jonathan F. W 'endel & Karen M. McDougal . — ea 6 Biogeographic, Taxonomic, and Cladistic ; Relationships between East | Asiatic and North American Crataegus J. B. Phipps — 667 Dadia L Contents continued on back cover VOLUME 70 1983 NUMBER 4 ANNALS OF THE MISSOURI BOTANICAL GARDEN The ANNALS contains papers, primarily in systematic botany, contributed from the Missouri Botanical Garden. Papers originating outside the Garden will also be accepted. Authors should write the Editor for information concerning arrangements for publishing in the ANNALS. EDITORIAL COMMITTEE NANCY Morin, Editor Missouri Botanical Garden CHERYL R. BAUER, Editorial Assistant Missouri Botanical Garden MARSHALL R. CROSBY Missouri Botanical Garden GERRIT DAVIDSE Missouri Botanical Garden JOHN D. Dwyer Missouri Botanical Garden & St. Louis University R GOLDBLATT Missouri Botanical Garden Published four timesa year by the Missouri Botanical Garden, . Louis, Missouri 631 ISSN 0026-6493 he Busi Office ofthe Annals. OR. O. Box 368, 1041 New Hampshire, Lawrence, Kansas 66044. Subscription price is $60 per volume U.S., $65 Canada and Mexi $70 all other countries. Personal subscriptions are available at $30 and $35, “respective y Airmail delivery charge, $30 per volume. Four issues per volume. Sa E xt j Lawrence. Kansas 66044 F Op © Missouri Botanical Garden 1984 -—————— ————————————— "Áo" l—————————— Iu—Á—H— ANNALS MISSOURI BOTANICAL GARDEN VOLUME 70 1983 NUMBER 4 ON THE SIGNIFICANCE OF PACIFIC INTERCONTINENTAL DISCONTINUITY WU ZHENGYI! The abundant disjunctions in the ranges of Plant families and genera in Eastern Asia and temperate North America is one of the most fa- es E important patterns of discontinuity bes of biology. From the time Asa Gray (1859) ue attention to these discontinuities in iná ridi a century ago, they have been studied God Pe by many authors (e.g., Li, 1952; bec , : 4; Graham, 1972; Wood, 1972). For- a non was believed that there were only about me ra of plants with such ranges, but now it Own that there may actually be more than amili o r. Woody Table 5 that include trees or shrubs are listed in yr api are 45 genera distributed among Boiss of which four (Illiciaceae, Calycan- ene) iti a stricto, and Styraca- esas «i It the same sorts of disjunctions in tem—with f genera they include. The same pat- Die aces amilial and generic disjunctions ap- e Y parallel—is exhibited by families Sed of herbaceous plants (Nelumbona- To i, Saururaceae, Phrymaceae, Croomiaceae), "Idraceae, which exhibits a similar pattern of distribution. All of them are disjunct between eastern Asia and eastern North America, al- though Nelumbonaceae extends through the tropics to northern Australia, and Illiciaceae, Schisand and Ny tend to the West Indies and to southeast Asia. Styracaceae have a range similar to that of Hydrangeaceae sensu stricto, and might be derived from the latter through genera such as Deutzia. A single species of Styrax extends from eastern Asia to the Med- iterranean region. For Hydrangeaceae, Hydran- gea and Deutzia range through the mountains of Mexico and the Andes to Chile. In agreement with Good (1974), I believe that many of these genera could be survivors of an ancient flora that failed to persist in Europe and western Asia. I do not agree with Good, however, that these plants can appropriately be regarded as “circumboreal” in distribution. Rather, I be- lieve that they might be relicts of a warm tem- perate to subtropical tane flora derived from a Tertiary Paleotropical flora that was present on the mountains in warm-temperate to subtropical climates in the Old World in early Tertiary or perhaps late Cretaceous time. The persistence of such ancient patterns of distributions is suggest- ed, among other lines of evidence, by the exten- sion of some of these genera far southward into the tropics or even to the lands of the Southern Hemisphere; even in these cases, however, the centers of distribution are definitely in North America and temperate Eastern Asia. l Kunmi : : : unming Institute of Botany, Academia Sinica, Kunming, Yunnan, People's Republic of China. ANN : Missouri Bor, GARD. 70: 577-590. 1983. 578 TABLE l. Families with trees or shrubs. Number of Genera Taxaceae (Torreya) 1 Pinaceae (Pseudotsuga, Tsuga) 2 Cupressaceae (Calocedrus, Chamaecyparis, Thuja) 3 Magnoliaceae (Liriodendron, agnolia 2 Illiciaceae (I/licium) 1 uraceae (Sassafras) 1 Berberidaceae (Mahonia) 1 Calycanthaceae (Calycanthus) 1 Theaceae (Gordonia, Stewartia) 2 Iteaceae (Itea 1 Hydrangeaceae (Decumaria, Hydrangea) 2 Rosaceae (Photinia, Physo- carpus, Sorbaria, and herbaceous genera) 3(-5) Caesalpiniaceae (Gleditsia, Gym 2 nocladus Papilionaceae (Lespedeza and 4 Hamamelidaceae (Hamamelis, Liquidambar) 2 Santalaceae (Buckleya, Pyrularia) 2 Rhamnaceae (Berchemia) 1 Anacardiaceae (Toxicodendron) 1 e (Carya) 1 Juglandacea Cornaceae (Bothrocaryum = Cornus t. Thelycrania) 1 Nyssaceae (Nyss 1 Araliaceae (Aralia, Oplopanax, and herbaceous Panax) Ericaceae (Leucothoe, Lyonia, Pieris, Therorhodion, and Hugeria) 5 Styracaceae (Halesia) 1 Oleaceae (Chionanthus s.s., Osmanthus) 2 Rubiaceae (Cephalanthus and 2 herbaceous genera) 1(-3) Caprifoliaceae (Symphoricarpos d 1 herbaceous genus) 1(-2) TOTAL 28 families 45 genera Among the genera listed in Appendix 1, some tha m + s i: 1 1: /"1 . } of yp ( g Glehnia of Um- belliferae, Phryma of Phrymataceae, Symplo- carpus of Araceae, Diarrhena and Schizachne of Graminae) belong to two distinct categories of distribution. Phryma, for example, appears to be one ofthe relictual representatives ofthe Tertiary flora previously mentioned. It has a typical pat- ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 tern of disjunction between eastern Asia and eastern North America. In contrast, Glehnia, Symplocarpus, Diarrhena, and Schizachne (for a discussion of Schizachne and Diarrhena see Koyama & Kawano, 1963) might actually have a Beringian distribution and might therefore be direct decendents of the flora that was common through Beringia was still quite might have been distributed throughout Arctic Europe by way of northeastern Siberia and Ja- pan, temperate North America, and even the mountainous regions ofthe western United States Genera of this group appear to be “younger m their distribution than Phryma, and their path- way of distribution can be regarded as very ob- vious. Other oligotypic genera mentioned in Appen- dix 1 can be analyzed as follows. First, families and six genera of conifers are common in eastern Asia and North America. Among them, a, and Chi am al- forests in these regions like the oligotypic gener Torreya, Thuja, and Calocedrus. oa cedrus with three species in New Caled al two species in New Zealand; Papuacedrus di New Guinea to the Moluccas (Li, 1953)]. "i gests that the range disjunction of Cal might have originated as a result of a Go nan distribution that extended throug Southern Hemisphere. ndwa- h the E 1n many oligoty P E the been fully discussed by Li (1952). pens e (Fumariaceae), Decumaria (Hydran (Bignoni- Chionanthus (Oleaceae), and Camp v ce richlY aceae). Genera with three sp /———— —— —— HR a guum nsn are Uem "npememehe 1983] WU -INTER( d ` FIGURE 1. The distribution of Torreya. M FIGURE 2. The distribution of Pachysandra. FIGURE 3, The distribution of Carya. 579 580 represented as those with two species. Among them are many that have two species in eastern Asia and one in North America, like Calocedrus; similar patterns occur in Diphylleia (Podophyl- aceae), Boschniakia (Orol I ), and Tulo- tis (Orchidaceae), among others. The corre- sponding pattern with two species in North America and one in Asia occurs, for example, in Zizania (Gramineae). Of these oligotypic genera it can be said that the species are obviously closely related in every case and that in general they represent either primitive families or relatively primitive lines within phylogenetically advanced families. It is clear that they constitute relicts in every sense of the word. Magnoliaceae, one of the more ancient families of dicotyledonous plants, has in general a distribution pattern that can be rep ted by that of the genus Magnolia. In general, the family is disjunct between eastern Asia and North America but extends south in the mountains into tropical southeast Asia and to South America. Liriodendron, on the other hand, is strictly an eastern Asian-eastern North American genus confined to temperate and sub- tropical zones. The western limit of Liriodendron in China extends to southeastern and northeast- ern Yunnan, where many other relict genera such as Rhoiptelea, Diplopanax, Annamocarya, and Davidia are also known. Liriodendron also ex- tends to the northern part of Viet Nam where the most primitive species of Platanus (P. kerrii, with unlobed | d pi e tion) exists, as well as to adjacent mountains of southeastern Yunnan. In view of these distribution patterns it can be hypothesized that both Magnoliaceae and Platanaceae originated in the subtropical mountains of southeastern Asia. Ithough the theories that formerly placed Nelumbonaceae as an ancestor of the monocots have now largely been discarded, this family cer- tainly represents a very ancient lineage that has been in existence for a very long time both in Asia and in America; it might represent a sur- vivor of the floras that existed in both areas in nunculaceae, Paeoniaceae, Berberidaceae, and Podophyllaceae, there are at least four bispecific or trispecific genera that are disjunct between eastern Asia and eastern North America. Among them. Jef»r« Ist no : all KISEL XN Se VA LMI ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 dophyllum and Dysosma, which have three to nine species in China but a single species referred to Podophyllum sensu stricto in North America. y 65), ob- viously is another relict genus ultimately derived from the same stock. Another very interesting small family is the Saururaceae; it has a strictly eastern Asian-east- ern North American disjunct distribution. Both paired genera (Houttuynia and Anemopsis) and paired species (Saururus chinensis and S. cer- nud) are represented among the four genera and six species of the family. This family certainly originated in the subtropical region during the Late Cretaceous and early Tertiary. Among the members of this family, Houttuynia occurs from the Himalayas to Japan; Anemopsis in Mexico and the southwestern United States; Saururus from the Philippines and Indochina through the Yangtze Valley; and Gymnotheca with two close- ly allied species of scattered distribution 1n southeastern Yunnan, Guangxi, izhou, ichuan. Saururus is morphologically T and geographically quite similar to Zippelia ( Cir- caeocarpus") of the Piperaceae, a large panum ical family that apr to be ultimately of (ont wanan distribution. The possible relationship between the Saururus and Zippelia should investigated further. E Menispermum, which has many perianth "m ments that are irregularly spirally arranged an more than four whorls of stamens, is one i most primitive members of the mainly trop! : and subtropical Menispermaceae. It is the only member of this family that has a disjunctio ange eastern Asia and temperate NO : America, which suggests that AMenisperm" could be an ancient relict derived from jener ical mountains in Laurasia. It might be close the ancestral form of this now primarily troP! amily. The genus Penthorum (Penthoraceae) e regarded as intermediate between the Crass w ceae and Saxifragaceae. Its disjunct dinit pattern coincides with its primitive morP o d and critical phylogenetic position. Similar sta i ments could be made for Tiarella (Saxifragac ; See -— - FIGURE 4. The distributions ef Camptotheca (horizontal lines) and Nyssa (vertical lines). .FIGURES, The distributions of Buckleya (dashed outli lines), and Symphoricarpos (shaded). 6. The distributions of Illicia Outline) and S i e: Kadsura (dashed bis ceae: Illicium (vertical lines); and Schisandracea chisandra (horizontal lines). 582 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 D. FiGURE 7. The distributions of Calycanthaceae: Calycanthus (vertical lines) and Chimonanthus (shaded). Ad 1 FIGURE 8. The distribution of Nelumbo. - FIGURE 9. The distributions of Magnoliaceae (shaded): Liriodendron (solid) and Magnolia (vertical lines) ——X 9 — am ——— —— m ŘŘ—_ M oe a Y 583 1983] wU Asian-eastern North American pattern of distri- bution. It seems clear that they, too, are derived from the ancient flora that is being discussed here. There are a number of phylogenetically more advanced families that exhibit similar disjunc- tion in range. These include Rubiaceae, Cam- panulaceae, Scrophulariaceae, Bignoniaceae, Phrymaceae, and Labiatae of the Sympetalae, as well as Cyperaceae, Gramineae, and Orchi- daceae among the monocots. The fact that such advanced families share a similar disjunction in range might be taken as evidence that some an- glosperm families are older than currently be- lieved. Following this review of a few of the oligotypic genera that are represented both in Asia and in orth America, we now pass on to consider some of the genera represented by more species in one or both areas. Among the with a disjunction in range between eastern Asia a E cuu America. Abelia (Caprifoli- dur or example, has 10 to 14 species widely Istributed in the forested half of China. In the Macon Hemisphere it is represented by two ca that occur only in Mexico. Cleyera (Tern- CU in contrast, has a distribution that West Ex Mexico, extends to Panama and the Whose D but is represented in Asia by a single Valle at ranges from south of the Yangtze Pending Japan. Deutzia (Hydrangeaceae) and Sis \ lid ) are similar in their es but in these cases differentiation of have pe ern and Western Hemisphere elements k Cip so far that they must be regarded loni, : genera: Deutzia-Neodeutzia; Disty- iuba, Inadendron. Paired genera that collec- North Present disjunct ranges between Asia and prising merica are listed in Appendix 2. A sur- "à i "^n relationship between the floras of "Wee Mexico has been detected recently by usekom (1971), who regards Meliosma x common to these very widely separated Hes e farther south, I would like to point out ranges “a also is a pattern of disjunction in the ica, Abo ween tropical Asia and tropical Amer- distant ini genera are common to these very early o oai and they could represent a very families ern of disjunction in range. Among the represented primarily by trees and shrubs are genera such as Talauma (Magnoli- aceae), Anaxagorea (Annonaceae), Litsea, No- thaphoebe, Persea, and Phoebe (Lauraceae), Hedyosmum (Chloranthaceae), Eurya (Tern- stroemiaceae), Saurauia (Saurauiaceae), Sloanea (Elaeocarpaceae), Microtropis (Celastraceae), Pristimera (= Reissantia) (Hippocrateaceae), Sageretia (Rhamnaceae), Picrasma (Burseraceae), Sapindus (Sapindaceae), Meliosma (Meliosma- ek Donini anyi dug " 1 ceae), Turpinia (Staphy ) Z Spondias (Anacardiaceae), and Gaultheria (Eri- caceae). Many of these genera are common or even dominant in the tropical and subtropical evergreen forests that extend from southeastern Tibet to Taiwan in China. They can also be re- garded as descendents of the sort of vegetation that seems to have predominated in the sub- tropical mountains of the Northern Hemisphere in the Late Cretaceous and the early Tertiary. Juglandaceae presents an interesting example of the differentiation of these warm temperate to tropical northern floras (Lu, 1980). Annamocar- ya, which extends from tropical Indochina into southwestern China, is very likely to be near the common ancestor of Carya, which is disjunct in distribution in eastern Asia and eastern North America. It is distributed in eastern Asia from northern Viet Nam to the southern part of the Yangtze Valley. Pterocarya and Platycarya, in contrast, are mainly eastern Asian, with Platy- carya extending west of the Caucasus; Cyclocar- ya is endemic to China, but known from the fossil record in the Paleogene of North America China, is a genus paired with Oreomunnea, represented by three Mexico and Central America. the Juglandaceae, is the Chloranthaceae. The sent distribution of this small family can be characterized as primarily tropical and subtrop- ical, but when its distribution is examined in detail, it appears that it can best be regarded as being of Gondwanic origin (see also Raven & 584 ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 | FiGURE 10. tributi fS 1 iopsis (Shaded), Gymnotheca (solid), Houttuynia (vertical | lines), and Saururus (horizontal lines). | FIGURE 11. The distributions of Juglandaceae: Annamocarya (solid), Carya (horizontal eut rocary (dashed outline), Juglans (shaded), Oreomunnea (diagonal lines), Platycarya (heavy outline), an (vertical lines). uos ngelhardtia | FIGURE 12. The distributions of Castanopsis (vertical lines) and Lithocarpus (horizontal lines). | | | | | en a a a EE ee WU —INTER«( Y 585 1983] Axelrod, 1974). The relatively close relationship between Ascarina (with eight species from Ma- laysia and Polynesia, including New Zealand) and Ascarinopsis (a monotypic genus of Madagascar) suggests that these genera differentiated when migration between Australasia and Africa was much easier than it is at present. Hedyosmum, on the other hand, is very well developed with about 40 species in tropical America and the West Indies, and one in the Old World, where it occurs from Sumatra and Borneo northward to Hainan in southern China. The most primitive member of Chl t iving at the pres- ent time, however, is certainly Sarcandra, which that genus from tropica Asia north- Ward through eastern Asia. Chloranthus extends further north to the mixed forest in northeastern and is clearly of Late Cretaceous origin. D oM by two species that have a dis- fiis sui distribution in the Himalayas and MEL a Asia. One of these Asian species Secs; epal and the other ranges from Pegu- d. Eee Yunnan and the upper he ag alley between Yunnan and Sichuan. demic pe is together with the monotypic and en- of Asia ius Nouelia that occurs in the same part Maias a Small use belonging to the trie ily cun perhaps the most primitive of the We ibis mpositae. In the tribe Mutisieae, there itive in 2 genera that appear to be prim- and South x characteristics and occur in Central Well repr merica and in Africa, where they are in the IM both in the tropical regions and tisieae cosi more temperate portions. Mu- igin. inly appear to be of Gondwanic or- The two m : embers of Mi naceae re- cently disc trastemona Mainland E in Yunnan and Fujian on the tastemon ig ina are extremely interesting. Mi- wi, 196 Ee by one species in Japan : ), and one or two species in Taiwan (Liy : jian $ Lai, 1976) (one of which extends to Fu- ), . Pan Species from Indochina and south- Species į unnan, one species in Borneo, and one in Sumatra, with an additional two species completely isolated from their Asian relatives in Mexico (Matuda, 1947) and Central America. In southeastern and temperate eastern Asia, every species of this genus is, without exception, a root parasite of Castanopsis. Castanopsis is a domi- nant genus that forms extensive forests in the subtropical to tropical mountains from which so many of the genera discussed in this paper are derived. If Castanopsis ever occurred in the New World, it is now extinct, but it has been replaced during the course of evolution by the related ge- nus Chrysolepis, which consists of two species in California and Oregon. Despite the absence of Castanopsis and related genera in Mexico and Central America, Mitrastemonaceae, which are parasites of this genus in Asia, have persisted in a more depleted vegetation than has survived in Asia. I would now like to summarize with a few general conclusions derived from the preceding discussion: (1) The discontinuity in plant range that spans the Pacific Ocean is a very important one for the study of floristic evolution. (2) The eastern Asian-eastern temperate North American, and the tropical Asian-tropical Amer- ican disjunctions both appear to have been de- rived imilar ancestral floras and are closely related to the tropical southeastern (Indo-Ma- layan) and temperate eastern (including Sino- Japanese and Sino-Himalayan) Asian elements. (3) All of these disjunctions in range seem to have been derived from the same upper Creta- ceous and lower Tertiary paleotropical flora that occurred on the mountains within the tropical to subtropical zones (Axelrod, 1952, 1960). (4) The pattern of these disjunctions is con- sistent with an origin of some of the constituent oups of angiosperms in the Late Cretaceous or even earlier (Axelrod, 1961). LITERATURE CITED AXELROD, D. I. 1952. A theory of angiosperm evo- lution. Evolution 4: 29 0. . 1960. The evolution of flowering plants. Pp. 227-307 in S. Tax (editor), Evolution after Dar- win. Vol. 1, The Evolution of Life. Univ. Chicago Press, Chicago. . 1961. How old are the angiosperms? Amer. i. 259: 447-459. Be ets C. F. van. 1971. Revision of Meliosma (Sabiaceae) section Lorenzanea excepted, living and fossil, geography and phylogeny. Blumea 19: -529. dia 1974. The Geography of Flowering Plants, 4th ed. Longman, London. 586 GRAHAM, A. (editor). 1972. Floristics and Paleoflo- merica. Elsev- m rvations upon the relationship of the Japanese fo to that of North America and of other parts of the northern hemisphere. Mem Amer. Acad. Art 2 Koyama, T. n S. KAWAN 1963. grasses with North pec and eastern Asiatic distribution. Canad. J. Bot. 42: 859-884. Li,H.L. 1952. Floristic relationships between eastern Asia and eastern North America. Trans. Amer. Philos. Soc. 42: 371—429. Reprinted 1971, with a forward, pp. i ii, and without pagination as Morris Arboretum Mono . 1953. Present distribution and habitats ofthe conifers and taxads. Evolution 7: 245-261. Liu, T. S. & M. J. Lar. 1976. Rafflesiaceae. In H. L. Li, T. S. Liu, T. C. Huang, T. Koyama & C. E. DeVol (editors), Flora of Taiwan 2: 582-584. ANNALS OF THE MISSOURI BOTANICAL GARDEN Critical taxa of [VoL. 70 Eu; Z M. 1980. On the geographical distribution of Phytotax in . 20: 257- [o e geographic significance. Amer. J. Bot. 69: 275- MATUDA, E. 1947. On the pus Mitrastemon. Bull. Torrey Bot. Club 74: 133-14 Onwi, J. 1965. Flora of Land (English edition.) Smithsonian Institution, Washin Raven, P. H. & D. I. AXELROD 1974. ind iosperm n and past continental move Ann. Missouri Bot. Gard. 61: 5 : SMITH, A.C. 1947. The families Ill and Schi- sandraceae. Sargentia 7: 1-224. Woop, C. E. 1972. Morphology and phytogeography: the classical approach to the study of disjunctions. Ann. Missouri Bot. Gard. 59: 107-124. 1983] WU-—INTER(¢ APPENDIX 1. Eastern Asia and eastern North America intercontinental discontinuities. Number of Bed. Distribution in Family Genus N. Am E. Asia N. Am. Cupressaceae Calocedrus (Heyderia) 2/1 S to N Burma, NE Pacific N. Am. Siam Cupressaceae Chamaecyparis 3/3 S to Taiwan Cupressacea Thuja 3/2 Pinaceae Pseudotsuga 5/2 W N. Am. Pinaceae Tsuga $=10/2 Taxaceae Torreya 5/2 Florida and California Anacardiaceae Toxicodendron 18+/8 S to S. Apocynaceae Amsonia 1/17 N. Am. Apocynaceae Trachelospermum 10/1 SE United States Araliaceae Aralia 30+/4+ S to Indo-Malaya Araliaceae Oplopanax 1/1 Araliaceae Panax 7/2 Mahonia 40+/30— S to Sumatra S. Am. ; Bignoniaceae ampsis 1/1 S to W. Indies Bignoniaceae Catalpa 7/4 S to W. Indies ceae Pachysandra 3/1 E United States oo Gleditsia 10/2 Tropical Africa S to S. Am piniaceae Gymnocladus 3/1 S to Assam, Burma : ^ Cillycanthaceae Calycanthus 1/1 SW and E United tates oe Heterocodon 1/1 W N. Am. Ca epi Symphoricarpos 1/15 | Pt oliaceae Triosteum —6/4-5 C positae Brachyactis 4/1 N Asia N. Am OMpositae Cacalia +60/4 N. Am. + and S to S. Am C E Poma Crossostephium 1743/1 California me Bothrocaryum (Cornus 1/1 NE Am sect. dos crania) : | Penthoraceae Penthoru z 1/1 S to Indochina Atlantic N. Am. -Diapensiaceae ANUS 531 SE United States | paese Leucothoe 4/3(-40) S to S. Am. E waceae Lyonia 9/21 S to Himalayas W. Indies Ericaceae Pieris 6/2 : NW Am Ericaceae Therorhodion 1/2 NE Asia . u, - 2/1 : i * T 12/1 S to W. Malaysia Atlantic N. Am. | pees R (incl. 60-100/2 S to tropical Asia W United States | Pew Chrysolepis) | Án. | F mariaceae Adlumia NE Asia EN. Am à Dicentra 24 1/1 mae fom i ee Hydra idaceae—Liquidambar 2+1/3 sw aa aen Dv ^ Hydr Ceae Decumaria 1 S to Indoc i Uii um Hydrangea 45/35 S to Uoc Java Atlantic N. Am., Chile Juglandacea lliciu 214/ S to W. Malays — W. Indies I paresis E N. Am. Labiatae e Carya 2/23 S to In E io Mexico Labiatae Agastache 1/20 "m E United States Loganiacea Meehania 5/1 sp. comp N Mexico " Gelsemium 1/2 S to N Borneo, Suma- ee tra 10 000 10 cuim 588 ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 APPENDIX 1. (Continued). Number of 2 ea in Distribution in Family Genus N. Am E. Asia N. Am. Magnoliaceae Liriodendron 1/1 S to N Indochina Magnoliaceae Magnolia 30+/5+ S to Java Venezuela Menispermaceae Menispermum 1/2 Atlantic N. Am., S to Mexico Monotropaceae Hypopitys 1/1 S to Himalaya Colombia Nelumbonaceae Nelumbo 1/1 NE Australia S to Colombia Nyssaceae Nyssa 6/4 S to W Malaysia E N. Am Oleaceae Chionanthus s.s. 1/2 EN. Am. Oleaceae Osmanthus 15/3 S to SE Asia Orobanchaceae Boschniakia 2/1 NW N. Am. Papaver Stylophorum 2/1 Atlantic N. Am. Papilionaceae Amphicarpaea 2-3+1/1 S. Africa 2S to tropical Am. Papilionaceae Apios 6/4 Papilionacea Cladrastis 4/1 EN. Am. Papilionaceae Lespedeza 65+/15-30 Sto Australia Papilionaceae Thermopsis 7/23 E United States Papilionaceae Wisteria 7/3 Am. ryma Phryma 1/1 Podophyllaceae Achlys 1/1 Pacific N. Am. Podophyllaceae Caulophyllum 1/1 Podophyllacea Diphilleia 2/1 Atlantic N. Am. Podophyllaceae Jeffersonia 1/1 Podophyllaceae Podophyllum IA E N. Am. Polygonaceae enor 2-3/1 S to Philippines Ranunculaceae Enemion (Isopyrum) 1/1-3 Rhamnace Berchemia 134+/1-2 S to tropical Africa Atlantic N. Am. Rosaceae Acomastylis 1/12 Rosaceae Photinia 40/20 S to Sumatra Rosaceae Physocarpus 1/9-10 NE Asia Rosaceae Sieversi 1/1 NE Asia Aleutian Isl. Rosacea Sorbaria 5+/1 ?W to central Asia Rubiaceae Cephalanthus 1+1/5 Africa to warm Am. Rubiaceae Kelloggia 1/1 Yunnan-Tibet SW United States Rubiaceae Mitchella 1/1 NE Asia Atlantic N. Am. Santalaceae Buckleya 2/1 S United States Santalaceae Pyrularia 2/2 S to Himalaya SE United States Saururaceae uru 1/1 S to Philippines E United States cr Astilbe 15/2 xifragaceae Mitella 2/12 i Saxifragaceae Tiarella 1/6 Pacific and Atlantic 2x Am Schisandraceae Schisandra 17/1 S to tropical Asia Scrophulariaceae Castilleja 1/30-200 N. Am. to S. Am- Scrophulariaceae Orthocarpus 1/25-30 W Am. Am. Platina Veronicastrum 1/1 NE Asia temperate NES Solan toucher c a. 4/1 sa m) Styracaceae Taea 1/3 SE United States n Gordoni, 8-32/1 S to Indo-Malaya SE United Suet Stewartia 7-10/2 incl. Hartia, S to Indo- E United States Umbelliferae Glehnia — | NE Asia Pme M AD —— — | 1983] WU —INTER( Y 589 APPENDIX 1. (Continued). Number of ar a Distribution in Family Genus N. Am. E. Asia N. Am. Umbelliferae Osmorrhiza 1+/10-14 W to Caucasus S to Andes Vitaceae Ampelopsis 15/2 Vitaceae Parthenocissus 9/2 Alliaceae Nothoscordum 1/2 NE Asia, 4 African N. Am., 28 S. Am. Araceae Symplocarpus 2/1 NE Asia Atlantic N. Am. Croomiaceae mia 2/1 E United States Taceae Trichophorum 1/3 Gramineae Diarrhen 1/1 Gramineae Hystrix 2/3 S to New Zealand Gramineae Muehlenbergia 6/94-110 S to Andes Gramineae Schizachne 1 S to Mt. SW United States Gramineae Zizania 1/1-2 S to Burma Liliaceae Aletris 13/5 S to Himalaya Liliaceae Clintonia 2/4 Liliaceae ispo. 8/12 S to Indochina Liliaceae Smilacina 14/11 S to Himalaya Central Am. Liliaceae igadenus 1/20 S to Central Am. Orchidaceae Erythrodes 3-:-/97— S to Indo-Malaya, Po- Argentina lynesia Orchidaceae Pogonia 2/14 (40 to tropical S. Am.?), S to W. In- dies, and tropical S. Am. ie fam Tipularia 2/1 N to E Asia 3+/ca. 40 Kamchatka to W Him- alaya : steraceae Phyllospadix 2/5 Pacific coast of N. Am. 590 APPENDIX 2. Paired genera of Eastern and Western hemispheres. ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 Family Name Genus in Eastern Hemisphere (Species number, distribution) Genus in Western Hemisphere (Species number, distribution) Cupressaceae Taxodiaceae Berberidaceae Podophyllaceae Saururaceae Papaveraceae Hydrangeaceae Rosaceae Hamamelidaceae Fagaceae Cornaceae Caprifoliaceae Campanulaceae Calocedrus (2) Glyptostrobus (1) Epimedium (21, W to Mediterra- nean) Plagiorhegma (1) Houttuynia (1) Macleaya (2) Deutzia (50) Rhodotypos (1) Distylium (15, to Malaysia, Java) Castanopsis (120, to tropics) Dendrobenthamia (12) Weigela (12) Homocodon (1, Yunnan) Heyderia (1) Taxodium (3, to Mexico) Vancouveria (3, Pacific N. Am.) Jeffersonia sensu stricto (1) Anemopsis (1) Bocconia (1) Neodeutzia (1-3, Mexico) Neviusia (1, SE United States) Molinadendron (3, to Mexico, Central Am.) Chrysolepis (2, W United States) Benthamidia (Cynoxylon) (3, to Central Am.) Dierville (3) Heterocodon (1, W N. Am.) VETE NOE NNNM Ee aaa TEMPERATE FLORAS: THE NORTH PACIFIC CONNECTION A. R. KRUCKEBERG! ABSTRACT id E em the floras of eastern Asia and eastern North America are widely recognized, ose affinities also exist between some elements of the floras of northeastern Asia (China, eastern pan) a indicated in taxonomic of the vicariants. By gro envi T om t ronment testing, it should be possible to determine the degree o ary study possible to test the degree Ov : P i e years—and at this present sympo- doe tanists and biogeographers have em- the connections between the floras of E E 8 E d y -J = nz o E * o a Q S S 8 iin. (1972) remind us that the North the E. Japan and mainland China to Wmber of flog acific Northwest provides a large World E "o links between the oriental Old It has dm € western New World. the N n remarked (Mizushima, 1972) that ion has been over- What e ^ : affinities remain, though, are substantial A —Á e and suggest that some taxa survived the erasures of much of that flora of the late Tertiary. Or, perhaps, the Pacific Northwest in the late Ter- . E] 1 3 * » 1 4 1 4 EN. uat y uv p F east to west—up the coast of western North America and down the east coast of Asia. Could, for example, vicariants in Lysichitum and Rubus have arrived in eastern Asia in a counter-clock- wise migration? Further, it is probable that the number of vicariants and disjunctions in the North Pacific floras might have diminished fur- ther during the Pleistocene. So, we have a set of ZA iog trib proposes ways t vergence between vicariant taxa. EXAMPLES OF EASTERN ASIA-WESTERN NORTH AMERICA FLORISTIC AFFINITIES Vascular plants with amphi-Pacific affinities can be categorized in three modes: taxonomic, geographic, and ecological. Furthermore, these i Department of Botany, University of Washington, Seattle, Washington 98195. ANN, Missouri Bot. GARD. 70: 591-596. 1983. TABLE 1. Some taxa of eastern Asian—western North American affinities. Eastern Asia Western North America Range Continuity Habitat Taxonomic Status Woodwardia orientalis Sw. Cryptogramma crispa (L.) R. Br. Pinus thunbergii Parl.-P. densiflora Sieb. & Zucc. Calamagrostis purpurascens R. Br. Deschampsia atropurpurea (Wahl.) Scheele Lysichitum camtschatcense Schott Fritillaria camschatcensis Ker Maianthemum dilatatum Nels. & MacBr Platanthera chorisiana Rchb. f. Alnus crispa Ait. Coptis japonica Mak. Achlys ap DC. var. japonica (Max.) Fukuda (Small Geum li. Smith var. nipponi- cum Hara Rubus pedatus J. E. Smith R. vernus Focke € bronchialis L. ssp. funstonii Hult. W. fimbriata J. E. Smith C. crispa P. contorta Loud. C. purpurascens D. atropurpurea L. americanum Hultén & St. John F. camschatcensis M. dilatatum P. chorisiana A. crispa var. sinuata Regel C. aspleniifolia Salisb. A. triphylla var. triphylla S. bronchialis ssp. austromontana Pi G. calthifolium R. pedatus R. spectabilis Pursh Japan & Washington Circumboreal Circumboreal Eastern Asia + western North America Cosmopolitan Japan to California Japan to Washington Japan to western North America Japan to Washington Eastern o North America dig 8 Gall British Columbia Japan to California Siberia—North America (circumbor.) Japan to British Columbia Japan to southern Oregon Japan to NW California Discontin. Discontin.? Discontin. Contin.? Contin. Discontin. Contin.? Contin.? Contin. Contin. Discontin.? Discontin. Discontin. Contin.? Discontin.? Discontin.? Mesic coastal to montane Mesic-xeric, coast- al/montane Mesic-xeric, coast- al/montane Mesic-xeric, coast- al/montane Mesic, montane Mesic Mesic, coastal Mesic Mesic Mesic, montane Mesic Mesic Montane Mesic, coastal Mesic-montane Mesic Vicar. spp. Conspecific Vicar. spp. Conspecific Conspecific Vicar. spp. Conspecific Conspecific Conspecific Vicar. vars. Vicar. spp. Vicar. vars. Vicar. vars. Vicar. vars. Conspecific Vicar. spp. Nadav) TVOINV.LOG INNOSSIN JHL JO STVNNV 76S 0L 10A] TABLE 1. (Continued). Taxonomic Eastern Asia Western North America Range Continuity Habitat Status Spiraea betulifolia Pallas ssp. aemi- S. betulifolia ssp. lucida Circumboreal? Contin.? Mesic-xeric Vicar. spp. liana Hara (Dougl.) Hitch. Oplopanax horridum (Smith) Miq. var. O. horridum Eastern Asia-western & Discontin. Mesic Vicar. vars. japonicum Ha central North America Acer japonicum Thunb. A. circinatum Pursh Japan-Northern California Discontin. Mesic, coastal Vicar. spp. Cassiope stelleriana (Pall.) DC. C. stelleriana Eastern Asia to Washing- Contin.? Mesic-montane Conspecific Menziesia pentandra Maxim. M. ferruginea Smith Japan-Oregon Discontin. Mesic-coastal Vicar. spp. Rhododendron kamtschaticum Pall. R. kamtschaticum Japan to Alaska Contin.? Polar/boreal Conspecific Fauria crista-galli Makino ssp. japoni- F. crista-galli Japan to Washington Contin.? Mesic-coastal Vicar. vars. cum Gillett Boschniakia rossica Fedtsch. B. rossica Eastern Asia-Alaska Contin.? Mesic-coastal Conspecific B. rossica B. hookeri Walpers Eastern Asia- Washington — Contin.? Mesic-coastal Vicar. spp. Galium kamtschaticum Stell. var. acu- G. kamtschaticum var. ore- Eastern Asia-Oregon Contin.? Mesic-coastal Vicar. vars. tifolium Hara ganum Piper (— G. ore- ganum) Valeriana sambucifolia Mikan var. V. sitchensis Bong. Eastern Asia-California Discontin. Mesic-montane Vicar. spp. fauriei Hara Campanula lasiocarpa Cham. C. lasiocarpa Japan-Washington Contin.? Montane Conspecific Artemisia arctica Less. ssp. arctica A. arctica ssp. arctica Japan-western Contin.? Boreal/montane Conspecific North America Achillea sibirica Ledeb. A. sibirica Japan-British Contin.? Montane? Conspecific Columbia Erigeron thunbergii A. Gray E. glaucus Ker Japan-California Discontin. Coastal Vicar. spp. [£861 | NOLLO3NNOO OIJIOVd HLYON — OW3H34O(13N £6¢ 594 ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 categories are best viewed as elements of over- lapping, multidimensional matrices. For exam- ple, some are strictly amphi-Pacific plants [AcA- lys triphylla (Smith) DC.] that are conspecific throughout their North Pacific distributions; others are circumboreal [Cryptogramma crispa (L.) R. Br.] or circumpolar [Phippsia algida (Sol.) R. Br.]. Some taxa are differentiated into vari- ties/sul i ies along the North Pacific Ly v E 3 hl rim. Others are i ting t ) Some north Pacific taxa reach the full extent of the distributional arc, from Japan or adjacent mainland Asia, more or less continuously to the southern limit of the Pacific Northwest in north- ern California; others go only part of the way— for example, from Japan to the Aleutians or to laska. These are but some of the permutations of the amphi-Pacific floristic theme. Table 1 samples some of the various types of amphi- Pacific continuity. From Table 1 we can glean some inferences about amphi-Pacific distributions: most exam- ples are herbaceous perennials with coastal, me- sic ecologies; bog species are common; and, apart from circumboreal and circumpolar examples, the strictly circum-Pacific cases can range from eastern Asia to western North America more or less continuously while others reach North America discontinuously (only the Aleutians, Alaska, British Columbia, or Washington). No one can be sure what these variant distri- bution patterns are trying to tell us. I suspect they are the combined result of obvious phe- nomena: (1) the capriciousness of dispersal/dis- tribution, (2) specific physiological tolerance ranges, and (3) the influence of past glacial and orogenic events. SOME GEOLOGICAL HISTORY IN THE NORTH PACIFIC AREA The time from the Paleocene to the present (ca. 75 million years) has spanned major geolog- ical and climatic changes. The major phenomena include change in coastlines (including the flux Of epiric seas); sea floor spreading; episodes of orogeny, especially vulcanism; and Pleistocene glaciation (and interglacial/post-glacial amelio- rations). Through this span of turbulent time, floras waxed and waned, but remnants of floras had to persist to give their present amphi-Pacific distribution. The early Tertiary seemed to have provided the most continuous stretch of habitat for tem- perate taxa. The mixed mesophyte forests of those times did pervade the North Pacific land masses (Wolfe, 1969, 1972). And, by the Miocene, most of the major floristic elements of the North Pa- cific rim were in place. Climatic change and geo- logic upheaval intervened to disrupt the contin- uities in varying degrees— from extinctions to major disjunctions. The latest perturbations, the Pleistocene ice encroachments, were the most recent to have affected this fluid, transitory biota. It is instructive to note that Li (1952) inferred a difference in age among the various vicariants between the New and Old worlds. He (Li, 1952: 373) pointed out that the western North Amer- ican-eastern Asian group of plants is a different and more recent one than the eastern Asian- eastern North American group. Li bases this as- sertion on (1) the relatively continuous, rather than discontinuous distribution of the amphi- Pacific plants; (2) their coastal, cool temperate to arctic distribution; and (3) their predominant herbaceous life-form. Li’s arguments remind us that the intercontinental migrations and conflu- ences of taxa (and of floras?) have taken place repeatedly throughout the past, have taken dif- ferent paths, and today are expressed in à range of discontinuities. GENETIC AFFINITIES OF VICARIANTS AND DISJUNCTS It is assumed that the plants that show caste Asia-western North America similarities 2 closely monophyletic (conspecific to vic taxa). It then follows that each example i: show some degree of genetic affinity. ee logical resemblance is, of course, the initial y for inferring genetic similarity. Further testing the inferences of relationship from mor must rely on genetic methods. Attempts 1o of vicariants is a necessary next step. crossability and interfertility can be- afinity measures of affinity. Other tests of genetic | dit could include isozyme comparisons v14 ge MU trophoresis and could eventually Rs gs parisons of DNAs of paired taxa. For no i will consider the evidence from crossing t€? L Several alternative hypotheses regardıng netic affinity of amphi-Pacific taxa are : e and (1) vicariants or disjuncts are conservativ per have not diverged in any significant way e : valuable o are reproductively isolated. It is to ene ont that each particular pair of taxa will v 1983) or the other of these hypotheses. Besides genetic/ reproductive isolation, another component of potential genetic difference is surely to be eco- typic differentiation of vicariants or disjuncts. From other genecological studies (e.g., Clausen & Hiesey, 1958; Clausen et al., 1940, 1948), one would predict that even conspecific taxa in Asia and western North America will possess detect- able ecophysiological differences that are bound to be genetically fixed. n what follows, I review some ofthe evidence from crossability studies as it bears on the issue of genetic Bhnity Bebyoen. bicontinental dis- juncts, Not all th the western North Co ed Asian theme ofthis pa- Per; rather, they illustrate the more general no- tion of degrees of genetic relationship between Congeneric taxa in different continents. Thus, — volving pines; and certain hardwoods (e.g., P via the north antic Between die Old World and the New, rather than to the North Pacific connection. Few or no deliberate crosses have been per- s -North American taxa. However, me bits of evidence come to light from hor- tic ic hk oe forestry, and agronomic literature, as oo Breeding tests by Wright (1955) with ^t E fom from North America and Asia range a crossability (P. jezoensis X P. ensis, to unsuccessful (P. asperata x P. sitch- i e from Pinus (Mirov, 1967) include h Ybridize n case of P. lambertiana. It does not i iie naturally with any of its western North With two n congeners, yet it is successfully crossed Pn E Asiatic pines, P. koraiensis and affinity. 1 li, also of the soft pine (Haploxylon) ye 2. ntentional crosses among some Diplox- America n, from eastern Asia and western North 1952) ave been reported as failures (Duffield, "€ e leaved trees. The Palmatum group of North sanh ) is largely Asiatic. The one western cross with rican species, A. circinatum, fails to both of : either 4. palmatum or A. japo nicum, Washi apan (fide J. A. Witt, University of p ington Arboretum). z arks et al. (1983), at this symposiu Pife E crosses between eb iis Sidi. ( oc. North America) way L chinensis China) yield viable F, see and KRUCKEBERG — NORTH PACIFIC CONNECTION 595 flavonoid comparisons, coupled with the cross- ing data, indicate that the two taxa “ ave not diverged — far from a nanmi common ancestor . Other PLUS plants. Crosses within and be- tween series (= sections, more or less) in Rho- dodendron are often successful, yielding progeny of varying fertilities. The well-kn azalea hy- ri f the Knaphi n Asiatic me crossed successfully with our western North American, R. macrophyllum Herbaceous monocots. Iris hybridizers have included some eastern Asia-western Nort America crosses in their many attempts to im- prove Jris for horticulture. Four different types setosa (North Am Vins (the so-called ‘cal. sib” hybrids), /. douglasiana x I. lactea (central Asia to Korea), and J. longipetala (northern California) x J. si- = rica. Herbaceous dicots. Very few cases of inten- tional crossings between eastern Asian and west- ern North American e found. Giv- en the very rich but diffuse Misión on plant breeding, more cases will surely come to light. Taylor's (1967) study of intentional crosses in aie includes the hybrid A. flabellata (Ja- an) x A. flavescens (western North America); the F, had 61% stainable pollen and normal mei- sis. The scanty record of intentional hybridization between eastern Asian and western North Amer- ican taxa makes it fruitless to draw any conclusions. Only when a program of deliberate crossing is mounted, involving a whole range of amphi-Pacific plants, can we test the genetic sp es of the “North Pacific Connection.” Many the plants of Table 1 would be ideal subjects iot such genetic testing. Indeed, the fuller lists of Mizushima (1972) and Hara (1952, 1956) pro- vide ample taxa for pioneering a genetic analysis of amphi-Pacific biogeographic connections. LITERATURE CITED CLAUSEN, J. & W. M. Hies 958. IV. Genetic structure of of ecological made ce Inst. Wash. Pub 596 , D. D. Keck & W. M. Hiesey. 1940. Exper- imental Studies on the Nature of Species. I. Effect of varied environments on western North Amer- ican plants. Carnegie Inst. Wash. Publ. 520. —— & ——. 19 II. Environmental responses of climatic races of Achillea. Carnegie Inst. Wash. Publ. 581. DUFFIELD, J. W. 1952. — M and species hy- tion in the genus . Forstgenet. tpflanzenzücht. 1: 93-100. 72. Outline of the origin and histor- cal recognition of floristic affinities eter Asia and eastern merica. Pp. 7-16 in A. Gra- beni editor), Floristics and Falealii of Asia rn North America. Elsevier Publishing : Contributions to the study of vari- ations in the Japanese plants closely related to those of Europe or North America. Part 1. J. Fac i ect. 3, Bot. 6: 29-96 Contributions to the study of varia- tions in the J of Penin or North America. Part LT Fac S Univ. Tokyo, Sect. 3, Bot. 6: 343-39 Bh poscis 1952. Floristic relationships between astern Asia and eastern a. Trans. jue Philos. Soc. 42: 371-429. 2. Eastern Asia-eastern North America species-pairs in wide-ranging genera. Pp. 83-91 in ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 A. Graham (editor), Floristics and Paleofloristics of Asia and Eastern No gs — Elsevier Pub- lishing Company, Amst Mirov, N. T. 1967. The dee "Pinus. Ronald Press, York. MizusHIMA, M. 1972. Taxonomic comparison of vascular plants found in western North America and Apa: . Pp. 83-91 in A. Graham | (editor), Flo- ristics a E Elsevier Publishing Company, Amster- dam PARKS, C. R., N. G. MILLER, J. F. WENDEL & K. M. McDoucaAL. 1983. Genetic divergence within the genus Liriodendron (Magnoliaceae). Ann. Mis- souri Bo : 6. TAYLOR, R. J. 1967. Interspecific hybridization and its evolutionary pe as ie in the genus Aquile- gia. Brittonia 19: 374— Wo re, J. A. 1969. Neogene floristic and vegetational history of the Pacific Northwest. Madroño 20: 81- 110. : 72. An interpretation of Alaskan Tertiary floras. da 201-233 in A. Graham (editor), -— ics and America. Elsevier Publishing Company, e am. WRIGHT, J. W. 55. Species crossability in spruce in hi dea to distribution and taxonomy. Forest . 1: 319-349. THE FLORISTIC RELATIONSHIPS OF THE TEMPERATE FOREST REGIONS OF CHINA AND THE UNITED STATES! TsUN-SHEN YING? ABSTRACT diagrammatic illustration is used to explain the meridional shift i pe constituents of the Chinese h dec A deciduous forest sts from one in Regi and to one in Region IV ns I and II wit iduous trees vergreen undershrubs canopy ee some evergreen undergrowth, with deciduous and evergreen trees mixed. Five samples from the temperate forests of Region III, IV, and hi adn are VER to explain the altitudinal variation of the Chinese deciduous forests Bom eastern to wi a, with their obvious limit in western Sichuan and the exact nature of the floristic relationship is explained. by NE Sin showing (1) western United disjunction of the species of Clintonia Raf. and the e apa astern United States, (2) bicentric pattern of edis w con Michaux in a tern Ais and eastern United States, (3) bicentric pae ap he = fru ia Torrey ex Bentham & Hooker Satine to China (Yunnan) and the western United SPECIAL FEATURES OF THE TEMPERATE FORESTS OF CHINA South China is under the influence of a mon- soon climate and its characteristic vegetation is One of mesophytic forests. Northwestern China i land occupied by steppes and deserts. ls +o meal the vegetation of China is recog- R as being composed of eight major types, m I-VIII (Fig. 1, Wu, 1980). Regions I to and the northern portion of Region IV sup- Port temperate forests. a on the composition, structure, and p * ysognomy of the communities, these forests tend d-leaved for- Pis coniferous forests. Generally speaking, pu ir of the areas supporting deciduous nda is hot and rainy in the summer, and col Pd Ty in the winter. The mean annual tem- ure oT um for Regions II-IV are le 1 Shown j tolit: an obvious meridional shift of the nts of the increased number of ever- Shae replacing the deciduous ones in the diferent ¢ forests of China. The composition of orests is modified in accordance to the a ofeach community. From north to south, i i hi evergreen shrubs enter the deciduous -eaved forests of Region II. As hydro- E — m (DEB-8119209) to Dr. Peter H. vided by the Arnold Arboretum xn in IM For this support I am deeply nstitute of Botany, Academia Sinica, Beijing, People’ ANN. MISSOURI Bor. GARD. 70: 597-604. 1983. thermal conditions g ncrease, evergreen trees enter the temperate Eon in the southern areas as the second layer of the communities. As the hydrothermal condition increases continu- es Sobsciduenc henad thward eve , ously sou gr leaved species become mixed in the forests. In still more southerly regions evergreen broad- leaved species replace the deciduous ones (Fig. 3° 1 + =r Another f the tem- perate Chinese forests | is the variation from east to west. For example, the deciduous forests of Huangshan in Anhui occur at an elevation of 1,000 to 1,600 m. In Shennongjia in western Hubei, the deciduous forests occur at 1,700 to 2,200 m. On the southern slopes of the Qinling Mountains the deciduous forests occur at 1,800 TaBLE 1. Mean annual temperature and precipi- tation in Vegetation Regions II-IV. Mean Annual Mean Annual Tem perature Precipitation Region E mm II 0.3-5.5 562.6-657.8 III 7.8-14.3 584.0-835.8 IVB 14.3-15.4 903.9-1,134.4 grant from the National Science Fundin ided by a g for field work in the eastern av eg pesa M nip qm idu Additional support and facilities rvard University, which made possible a year of study at that s Republic of China. 598 ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 HAR “FRAG SZ LJ 250 500 Li earn l. . The vegetation regions of China (simplified from Wu, 1980). I = bis; enpate coniferous forest n; II = rate deciduous leaved forest subregion; V = Tropical monsoon and rain forest region; VI = Temperate steppe region; vem Temperate desert region; VIII — Qinghai-Xizang plateau vegetation region. to 2,600 m (Chiu & Chin, 1957; Chow, 1965; recognized in the temperate forests of China (Wu. : Ying et al., 1979). The deciduous broad-leaved 1965). Owing to geographic transformations 4? g. thro geo They omplexity is another feature of the Chinese achieved different distributional patterns. " deciduous forests. Approximately 930 generaare have been classified into the following tyPe* kd = Evergreen species $ = Deciduous species FIGURE 2. The gradual meridional change of constituents of the deciduous forests in China. —— A n a ER ———3À Tashueshan Mt. Yulungshan Mt, . Slope of Qinling Mt. Shennunjia Mt. Hwangshan Mt, Central Japan. DOE Wty FIGURE 3. Altitudinal change of deciduous forests from east to west in eastern Asia. 1 = Succulent thorny scrub; 2 = Evergreen broad-leaved forest; 3 = Evergreen deciduous broad-leaved mixed forest; 4 = Temperate coniferous forest; 5 = Sclerophyllous evergreen broad-leaved forest; 6 = Eve n coniferous forest; 7 = Alpine meadow and bush-wood; 8 = Subniveal open vegetation; 9 = Pine forest; 10 = Deciduous broad-leaved forest; 11 = Deciduous coniferous forest; 12 = Subalpine scrub. [£861 SdIHSNOLLV' T33I OLLSISIO TH — ONIA 66S 600 ANNALS OF THE MISSOURI BOTANICAL GARDEN TABLE 2. Eastern Asian endemic families represented in the deciduous forests of China and adjacent floristic regions. Total Number of Genera/Species Family Genera/Species in Chi Range Ginkgoaceae 1/1 1/1 China-J apan? ephalotaxaceae 1/9 1/7 China-Japan Nandinaceae 1/1 1/1 China-Japan? Cercidiphyllaceae 1/1 1/1 China-Japan Trapellaceae 1/2 1/1 China-Japan Sargentodoxaceae 1/1 1/1 China onl Eucommiac 1/1 1/1 China only Davidiace 1/1 1/1 China only Tetracentraceae 1/1 1/1 Sino-Himalayan Toricelliaceae 1/3 1/2 Sino-Himalaya Eupteleaceae 1/2 1/1 Sino-Himalayan-Japan elwingiaceae 1/4 1/4 Sino-Himalayan-Japan North Temperate Zone genera (ca. 300); (2) east- ern Asian genera (298); (3) temperate Old World genera (157); (4) temperate Asian genera (63); (5) eastern Asian-North American genera (115). Of these, the North Temperate Zone genera, eastern Asian genera, and eastern Asian-North Ameri- can genera are more prominent in the compo- sition of the temperate forests of China. Genera such as Acer, Carpinus, Populus, Salix, Fagus, Castanea, Prunus, Sorbus, and Tilia are found in the temperate forests throughout the northern hemisphere. The greatest concentration of species in some of these genera is in China, especially in central China. For example, the genus Acer con- tains about 200 species worldwide. Almost three- fourths of the species occur in China, distributed south of the Qinling Range and east of south- eastern Tibet, and 42 species are concentrated in central China (eastern Sichuan and western Hu- i). Dipteronia, the only other genus of the Ac- eraceae, is almost totally restricted to this same region. Another genus with similar species con- centration is Carpinus, with 40 species world- wide, of which 25 are restricted to central and southwestern China. Antiquity is another feature ofthe Chinese de- fossil species C. miofargesiana (Tanai, 1972). Fossil species of Carpinus are also found in Tertiary deposits of North America and Europe. The Occurrence of extinct and extant species of a genus in the same area is an indication of the antiquity of the com- munity in which the living species occurs. High endemism is another prominent feature of the Chinese deciduous forests. At the family level there are many monotypic or oligotypic, endemic families. Table 2 shows that there are 12 families with distributions limited to China species, of which seven occur in China, in central China. Of the ten families of -— sperms, Eupteleaceae, Tetracentraceae, i liaceae, and Helwingiaceae (the latter two o i: placed in the Cornaceae) are of Sino- Hime or Sino-Japan-Himalayan distribution. Core diphyllaceae is represented by two varieties ^ monotypic genus occurring disjunctly 1n pr China and Japan. Davidiaceae (also somet! E placed in the Cornaceae) and Sargentodom are endemic to China. Some of these fam i have fossil records in North America and Euro in Neogene or older strata. FLORISTIC RELATIONSHIP MANIFESTED THROUGH THE Decipvous FORESTS OF CHINA AND THE UNITED STATES ail in the arly well oun States, Europe, and t 1980)- (Daubenmire, 1978; Numata, 1974; pe The similarities of the geographical € [Vot. 70 | 1983] YING—FLORISTIC RELATIONSHIPS 601 TABLE 3. Comparison of the composition (trees) of the temperate deciduous forests in eastern North America and China. Not arranged according to dom- inance or closeness of relationship. (Derived in part from field work in China and the United States and from Fernald, 1950; Jepson, 1925; and Seymour, 1969.) China Eastern North America Acer buergereianum Acer rubrum A. mono A. saccharum Alnus japonica Alnus maritima A, nepalensis A. serrulata Aesculus chinensis Aesculus octandra Betula albo-sinensis Betula alleghaniensis B. chinensis B. lenta B. dahurica B. nigra B. platyphylla B. populifolia lifol arpinus turczaninowii Carpinus caroliniana Carya cordiformis C. glabra C. tomentosa Chosenia arbutifolia rnus chinensis Cornus macrophylla Fagus longipetiolata Fagus grandifolia F. lucida Pr mandshurica Fraxinus americana pet macregorii Halesia monticola ns mandshurica Juglans cinerea J. cathayensis J. nigra Liriodendron chinense Liriodendron tulipifera Cornus florida Magnolia officinalis Magnolia acuminata M. fraseri M. macrophylla M. tripetala Malus bacca ifoli ta Malus angustifolia M. sieversii "€ Q ad acutissima Quercus alba Q Pin z Q. coccinea : alena Q. falcata var. acutiserrata ata O. this . laevis Q. liaoty Q. laurifolia ngensis rin : mongolica Q. rubra moe Q. velutin Te iie Tilia americana Unc nns. T. heterophylla U S, ProPinqua Ulmus americana d U. ru ie conditions, and vegetation composition T c forests of China and the eastern and E lates are closer than those between China has a 9pe and the Caucasus. This phenomenon tracted the interest of numerous phyto- Phers. The dominant species of the decid- TABLE 4. Some common shrubs and herbs in Chinese and American temperate deciduous forests. Arrangement does not necessarily indicate close rela- tionship. (Derived in part from Fernald, 1950; Jepson, 1925; and Seymour, 1969.) China United States Adiantum pedatum Adiantum pedatum Clintonia udensis Clintonia umbellulata Convallaria keiskei Convallaria montana Hydrangea umbellata Hydrangea arborescens almia latifoli Maianthemum bifolium | Maianthemum canadense Pyrola rotundifolia Pyrola rotundifolia ar. americana Rhododendron Rhododendron maximum micranthum Ribes meyeri Ribes cynosbatii Tiarella polyphylla Tiarella cordifolia Trientalis europaea Trientalis borealis Vaccinium japonicum Vaccinium erythrocarpum . constablaei V. uliginosum V. vacillans uous forests of China and the eastern United States are often paired or closely related taxa of the same genera, as shown in Table 3. The relationship is not limited to the dominant tree species. The understory shrub layer and the herbaceous species of the deciduous forests also contain many genera common to the two regions, as shown in Table 4. In the temperate floras of China and the United States there are about 115 genera with disjunct distribution. Approximately 50 of these are oli- gotypic genera. In order to explain the floristic relationship between Asia and North America, the following genera restricted to China and/or Japan and eastern North America are selected: Cornus section Thelycrania (1:1); Chionanthus sensu stricto (1:1); Caulophyllum (1:1); Decu- maria (1:1); Diphylleia (2:1); Halesia (1:3); Ha- mamelis (2:4y; Jeffersonia (1:1); Liriodendron (1:1); Menispermum(1:1); Mitchella (1:1 ); Sassa- fras (1:1); Stylophorum (1:1); and Kelloggia (1:1). The figures in parentheses are the numbers of species in each area, eastern Asia to the left, east- ern North America to the right. The above species manifest some of the com- mon characteristics of the flora of eastern Asia and eastern North America. These can be divid- ed into three different distributional patterns: 1) eastern Asian-eastern North American (Fig. 4); 602 a 2) eastern Asian—western North American (Fig. 5); and 3) eastern Asian-eastern and western North American (Fig. 6). In addition, the monotypic or oligotypic fam- ilies such as Schisandraceae, Iteaceae, Saurura- ceae, and Illiciaceae are of special interest. In the deciduous genus Schisandra, the primitive sec- tion Pleio. Sphae- rostema both have a disjunct pattern of distri- bution from China southward to Sumatra and Java (Smith, 1947). Only one species, Schisan- dra glabra, belonging to section Schisandra is SN Kelloggia galioides MA FiGURE 5, ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 f * dl Diphylleia sinensis FIGURE 4. Bicentric distribution of Diphylleia in eastern Asia and eastern North America. known in eastern North America; two species of this section are in eastern Asia (Smith, 1947). Likewise, in the genus Jtea, most species occur in eastern Asia and thence southward to Java. Again, there is only one species in southeas Bicentric distribution of Kelloggia in eastern Asia and western North America. | lig y rge ny 1983] ) Clintonia Clintonia "m umbellulata andrewsiana e A YING— FLORISTIC RELATIONSHIPS IGURE 6. Tricentric distribution of Clintonia in eastern Asia and eastern and western America. southward to Indonesia. The genus Saururus oc- pein Vietnam. The genera restricted to North dni. and eastern Asia not only indicate a Es floristic relationship between the two re- $ they also provide reliable clues to their tropical origin. " ta all of the plants mentioned above grow Des ea deciduous forests in both regions. "n D e including morphology, coupled Provide ogy, plant sociology, and cytology, will tans fs Pa Clues that will be of great impor- in un erstanding the patterns of disjunc- in plant distribution. CONCLUSIONS cp tito forests of China represent a an flora. The dominant species belong evident E a Tertiary history. There is an lents fron, ual meridional shift of the constit- lo deciduo more evergreen species in the south cupied b P ones in the north. The areas oc- verted i4 €ciduous forests in China were con- MEL d a thousand or more years ago. tinous of the original forests in steep moun- t regions of Huanshan, Shennongjia, Qin- aliit rea, and Tashueshan indicate an ied ariation from east to west, accom- Mas by changes in composition. bution pam genera have a widespread distri- € temperate northern hemisphere. A tons f; mber of them are common to the decid- Fests of China and the United States. The shared genera include the dominant trees, shrubs and herbaceous plants. Many of these shared plants are not limited to large genera with wide- spread distribution but often belong to small, oligotypic genera such as Liriodendron and Sas- safras. The nature ofthe floristic relationship between the deciduous forests of China and the United States is exhibited by a tricentric distributional pattern in such genera as Clintonia with species in China, Japan, western United States, and east- ern United States, a bicentric distributional pat- tern as in Diphylleia with species in China, Ja- pan, and the southeastern United States, and another pattern of bicentric distribution in gen- era such as Kelloggia with species in China and western North America. LITERATURE CITED Cuiu, L. C. & C. S. CHiN. 1957. Observations of the plants of Mt. Yulungshan. J. Yunnan Univ. 4: 19- 130. (In Chinese.) Cuow, C. L. 1965. Plant communities of Huangshan. . 207-266 in P. C. Chen, P. C. Wu, P. H. Chiu, Hsu L. Chow (editors), Observationes i ce and Tech- art DAUBENMIRE, R. 1978. Plant Geography: with Spe- ial Refere to North America. A 1 ic P : ew Yo FERNALD, M. American Book Co., New York. Jepson, W. L. 1925. A Manual of the Flowering Plants of California. Univ. California Press, Berkeley. rk. L. 1950. Gray’s Manual of Botany. York 604 NUMATA, M. oe: 1974. The oe and Vegeta- pan. Kodansha Ltd., To SEYMOUR, F. c 1969. The Flora of NN England. C E. Tuttle Co., Inc., Tokyo. SMITH, A. C. 1947 . The families Illiciaceae and Schi- sandraceae. Sargentia : 1-224. TANAI, T. 1972. Tertiary history of vegetation in Ja- pan. Pp. m dod in A. Graham (editor), Floristics sia and Eastern North America. Elsevier Publ. Co., Amsterdam. ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 Wu, Y. C. 1965. On tropical eee x the flora of China. Scientia Sin. 1: 25-33. (In Chinese.) (editor). 1980. The Vegetation a China. Sci- ence Press, Beijing. (In Chinese.) Yin, T. S., C. G. MA & C. S. CHANG. 1979. Obser vations of the flora and vegetation of Shennungia in western Hubei, China. Acta Phytotax. Sin. 17(3) 41—60. (In Chinese, English summary.) | — 9 o À— 7 NORTH AMERICAN TREES WITH RELATIONSHIPS IN EASTERN ASIA ELBERT L. LITTLE, JR.! ABSTRACT Much information on distribution and taxonomy of North American trees with relationships in ono phs, is now available. East d richest sourc te seid ie hardwoods. Because of splitting of genera and species relationships are becomin: e are primarily generic. East sia is the geographic region or center with greatest net of remnants of the widespread Arcto-Tertiary flora, while Eastern North erica is second, with roughly one-half as many tree species. Europe and Western North America have smaller numbers, perhaps one-fourth. Many tree genera are widespread in all four regions. Most of the possible p Eastern North America and Eastern Asia have the most tree genera, about 16 in common. The relationships have been overemphas ized. They are readily explained by mountain-building and changes in climate, especially Pleistocene glaciation. À review of temperate North American trees With relatives in Eastern Asia is timely. Much information on distribution and taxonomy, in- cludine is now available and UN of 1 Fur in for hardwoods there is increased interest in foreign trees for in- rein in in forestry as well as in horticulture. ort America and: eastern Asia are becoming ob- Scured, bec. species » main conclusions are that pis relatian- Ips have been overemphasized and are readily Mid by earth changes accepted within re- nt yea Trees, though an artificial group oflarge woody a: it excellent examples for biogeograph- tive udies. They are large, conspicuous, rela- “y well known, of moderate numbers of Fia » and of economic importance. Thus, their ification has been studied in detail, and their aii distribution has been mapped. Also, va * Preserved as fossils and have a historical Fia though with many gaps. the icai of literature. Certain similarities in Oras of temperate Eastern North America tca oo Asia have long attracted the atten- lanists. These observations have been M to Carolus Linnaeus in 1750, Carl and Asa en; in 1784, Luigi Castiglioni in 1790, ut s Gray (Graham, 1972). Asa Gray (1840, 39, 1873, 1878) made important con- Some reprinted by Sargent (1889), (1972), and gun (1978). The notes tri tions, C: of Botany, U.S. National Museum of Natural History, Smithsonian Institution, bc 20560 Ann, Missouri Bor. GARD. 70: 605-615. 1983. made by Gray in 1840 were in a book review written when he was only thirty years old. As early as 1859, Gray cited the work of Darwin and Wallace on natural selection and he stressed the effect of glaciation on present distribution. Li (1952, 1972), who knew both regions, made a detailed analysis and reviewed the extensive literature. Hi g outline ranges iof. many tree genera. Atlases. Accurate distribution maps of most north temperate tree species have been pub- lished, replacing the general outlines by Schmucker (1942). No new maps are presented here. Worldwide generic maps have been com- piled, especially by Li (1952) and by Meusel et al. (1965-78). Atlas of United States Trees (Lit- tle, 1971-81) in six volumes m mately 684 tree species native in continental United States including Alaska. The Forest Ser- vice Checklist of United States Trees (Little, 1979) further compiles for each ofthe 216 native genera (Hosie, 1979) similarly describes, illustrates, and maps the species of that cons try. For Europe ps of vascular plants of the northwestern part (Hultén, 1950), central region (Meusel et al., 1965-78), and the paean (Jalas & Suominen, 1972-80, incom- 2. ete). Two volumes ofan atlas of the trees and shrubs of the Soviet Union have appeared (Bot. Inst. Washington, 606 Komarova, 1977). Maps for southwest Asia, or sia Minor, are being published with financial assistance from the United States Public Law 480 program (Browicz, 1978). For Japan there are similar atlases of forest trees (Japan Forest Tech- nical Association, 1964—71) and the flora (Hara, 1958-59; Horikawa, 1972-76). Monographs. Additional information on dis- tribution with species maps is available in cur- rent worldwide taxonomic monographs of north temperate tree genera. Examples are: Abies (Liu, 1971; supported by Public Law 480); Aesculus (Hardin, 1957, 1960); A/nus (Murai, 1964; Fur- low, 1979); Catalpa (Paclt, 1952); Cupressus (Sil- ba, 1981; Wolf & Wagener, 1948); Illicium (Smith, 1947); Lyonia (Judd, 1981); Nyssa (Eyde, 1966); Osmanthus (Green, 1958); Pinus (Mirov, 1967); Sassafras (Keng, 1953); and Stewartia (Spongberg, 1974). For Pinus, subdivisions as well as species have been charted (Critchfield & Little, 1966; Little & Critchfield, 1969). Other generic monographs cover geographic regions, such as North America. Forestry. Deciduous hardwoods (dicotyle- dons) are important in forestry as well as horti- culture. The Society of American Foresters at its annual meeting on Sept. 20-22, 1982 at Cincin- nati, Ohio, stressed the main subject or theme: "America's Hardwood Forests— Opportunities Unlimited." Seeds and other germplasm are de- sired for tree breeding or genetics research pro- grams. Eastern Asia is the world's largest and richest source of temperate deciduous hard- woods. Thus, many related species are available for introduction, selection, and hybridization. Also, Old World genera merit testing in new en- vironments, hopefully pest-free, in the New World. Foresters are actively exchanging seeds, information, and expertise. above) has resulted in the recognition of distinct species in the two geographical areas. TREE SPECIES IN BOTH TEMPERATE NORTH AMERICA AND EURASIA Many species pairs of morphologically similar and apparently closely related tree species have been listed from Eastern North America and Eastern Asia. Li (1972) included examples from these tree genera: Acer, Aralia, Cornus, Diospy- ros, Ilex, Rhus, Styrax, and Symplocos. At present only about five tree species of tem- ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 perate North America pted tive al in Eurasia. Trees are conspicuously absent in long lists of plants on both sides of the North Atlantic Ocean (Hultén, 1958), lists of circumpolar plants (Hultén, 1971), and a list of plants in both West- ern North America and Japan (Mizushima, 1972). (However, several tropical species, especially mangroves and other beach plants of southern Florida, are common both to New and Old Worlds. Juniperus communis, common juniper, per- haps the tree species with greatest range in the world, is circumboreal or circumpolar, extending northward beyond the limit of trees. However, in the New World it is usually a low mat-forming shrub, rarely a small tree. Two tree species of Alnus, alder, are widely distributed, though agreement on classification is lacking. A/nus incana of Eurasia has two New World varieties or closely related species: A, rú- gosa in northeastern North America and A. te- nuifolia in northwestern North America. Alnus sinuata, often included in the shrubby species 4 crispa, ranges from northwestern America into northeastern Asia. Also, two tree species of Salix, willow, S. alaxensis and S. bebbiana, extend from northwestern North America into northeaster North and. cidentalis, Fraxinus nigra, and Ilex mont i : closely few other paired species of Ilex are very dr. ^ and nia by two shrub varieties, var. californicus var. fulvescens and in western Mexi jaliscanus. PALEOBOTANICAL EVIDENCE i ; ust Of course, the study of plant relationships m' i h the plant kingdom can be reconst study of fossil plants found in the earth" (Taylor, 1981: 1). i ; diera i are doing a fine job pel mostly incomplete specimens and ur os scraps now available, though misde ad " have occurred. The work is slow, and ile, pet seem almost insurmountable. Meanwhile, rocks | a ——— Ee 1983] haps others may be permitted to consider and evaluate additional lines of evidence for rela- tionships and even to speculate. Unfortunately, some wild speculations have resulted, and the term “hysterical” plant geography has been pro- sed! Briefly, the fossil record reveals a widespread occurrence of many modern genera of woody plants, designated as the Arcto-Tertiary flora, more or less uniform across the temperate lands of the Northern Hemisphere, from the close of the Cretaceous period through the Cenozoic Era, roughly 65 million years in duration. Many gen- tra can be traced back that far, though their origins are obscure and the ancestries uncertain. Western North America has a fossil plant rec- ord known in much greater detail than has East- em North America. For example, Wolfe (1969) na vegetational history of the Pacific Northwest recorded from the early and middle Miocene flora these 17 tree genera now absent but persisting in a Liriodendron, Magnolia, Nyssa, Sassa- a and Tilia. Also, these genera now extinct in € New World: Ailanthus, Cercidiphyllum, and Zelkova, qu the middle of the Miocene epoch, some thr Ns pian the year. The fourth, Western North Ca, has other climatic patterns, such as dry One. ks me tree species. Europe and West- hans merica have smaller numbers, per- “mt as Many. lary pe and persistence ofthe Arcto-Ter- s and differences in present numbers of ii si in these regions are readily explained €nts of continental land masses un- LITTLE— TREES 607 and apparently shifted downward with advances of continental ice sheets and lower temperatures. In eastern North America similar movements of forest zones southward have been documented by pollen analyses or palynology (Davis, 1969). However, th g ti t tinuous southward but is present in the West Indies and on mountains of Mexico and Central America. In southern Europe, the Alps and east- west mountain chains, as well as the Mediter- ranean Sea, served as barriers to the southward migration before glacial advances. Obviously, extinction of many tree species followed. In Western North America the climates following mountain-building were drier and unfavorable du Ped 1 7. P ECIAM 4 Jd T trOpic al or sur p relicts (Acer, Alnus) in moist sites, such as val- PATTERNS OF DISTRIBUTION Four northern temperate forest regions can be designated by letter, from west to east (a fifth, the Colchic, from southeastern Europe to Iran, is not considered here): A, Western North Amer- ica; B, Eastern North America; C, Europe; and D, Eastern Asia. There are 15 mathematically possible combinations: widespread in all four (ABCD), in three regions (ABC, ABD, ACD, BCD), in two regions (AB, AC, AD, BC, BD, CD), and endemic to one region (A, B, C, D). Three combinations, CD, C, and D, are exclu- sively Old World and omitted here. Three others, AB, A, and B, are New World endemics to be treated separately. Many tree genera, some with numerous species common and dominant in temperate forests are widespread in all four regions or also beyond and have persisted in different climates (ABCD). Fa- miliar examples, not considered further here, with reference by number to published generic maps by Meusel et al. (1965-78), are: Abies (map 20a), Acer (map 276b), Alnus, Betula, Cornus, Cra- taegus, Fraxinus (map 340b), Juniperus (map 22d), Larix (map 21a), Picea (map 20c), Pinus (map 21c), Populus (map 112b), Prunus (map 226c), Quercus (map 121c), Salix (map 112b), and Sorbus. A few others, such as Aesculus, Cer- cis, Ostrya, Staphylea (map 275d), and Taxus 608 TABLE 1. Tree genera (7) confined to Eastern North America (United States or also southeastern Canada and Eastern Asia, with approximate numbers of species (BD). Maps by Li (1952) are cited by number. stern World North Eastern Genus Total America Asia Chionanthus* (map 42) 3 2 1 Cladrastis (map 26) 4 1 3 Gymnocladus (map 25) 4 1 3 Halesia (map 41) 4 3 1 Liriodendron (map 12) 2 1 1 Sassafras (map 16) 3 1 2 Stewartia (map 31) 10 TE A A Totals 30 11 19 * The related genus Linociera, if united, would add 80-100 species, mostly tropical and subtropical, and list. cause removal from the list (map 19d), have mostly restricted occurrence in the four regions with partly relict species. The eight remaining binati or both North American regions with one or both Old World regions. Tree genera under different patterns, with approximate numbers of species, are presented in Tables 1—5. The well-known pat- tern of Eastern North America and Eastern Asia (BD) is represented by the most tree genera, about 15 (Tables 1 and 2). This group can be divided into two, one extending beyond southward into subtropical regions, such as West Indies and mountains of Mexico and Central America and TABLE 2. Tree genera (9) confined to Eastern North America and also south to mountains of M f America, or West Indies and to Eastern Asia, with approximate numbers of species, including a tew (BD+). Maps by Li (1952) are cited by number. ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 in southeastern Asia to Indonesia. Pattern BCD has eight genera (Table 3), ABD has five (Table 4), and AD has four (Table 5). Patterns ACD and ingle t h, while C apparently ł gl g ABC and BC have none. Cupressus, cypress, apparently is the only tree genus absent from Eastern North America (ACD) Mediterranean region, and six in eastern Asia, mainly China. Arbutus, madrone, is limited to two warm tem- perate regions, Western North America and Eu- rope (AC; map 329c). There are eight or fewer species in Western North America south to Mex- ico (one to Nicaragua) and about ten others m the Mediterranean region of Southern Europe and Western Asia. Table 1 summarizes tree genera confined 10 Eastern North America (United States or also southeastern Canada) and eastern Asia ies IANA two) species in Eastern North America (on to Guatemala) and Glyptostrobus with one China; and Elliottia with one species very ar" and local in Georgia and Tripetaleia with two shrub species in Japan (map 36). in shrubs Bude Mexico, Eastern Central Nort America : Asia Genus World Total America (additional) West Indies _Eastem "5° Carya (map 2) 16 11 1 : Catalpa (map 46) 11 2 5 Gordonia (map 3 Sf 1 " Hamamelis (map 23) 6 2 1 : Illicium (map 14) 42 2 1 2 " Lyonia (map 32) 35 5 1 24 : Magnolia (map 13) 76 8 108 8 A Nyssa (map 5 3 ; nth 32 1 1 — Totals 254 m s 165 35 15 39 * To Ecuador in northern South America. 1983] LITTLE—TREES 609 TABLE 3. Tree genera (8) of Eastern North America and in both Europe and Eastern Asia but now absent from western North America, with approximate numbers of species (BCD). Maps by Meusel et al. (1965-78) are cited by num stern Genus World Total North America Europe Eastern Asia Carpinus (map 117d) 28 l 2 25 Castanea (map 121b) 12 4 l 7 Cotinus (map 273d) 4 l l 2 Fagus (maps 120c, 121a) 10 l 2 7 Liquidambar 3 1 (1) 1 Pistacia (map 274a) 10 2 5 Tilia (map 280c) 29 4 5 20 Ulmus _ 46 11* DTE 30 Totals 142 25 20 97 * Southwestern Asia. c Includes five species in Mexico, one also to Panam Table 2 has nine genera of mostly deciduous hardwoods of slightly broader distribution i in the seater, projecting southward into tropical mountains to Indone sia. Pie has eight genera mostly erue ti mai the deciduous forests of the Northern wate (BCD). They are now abse ws rom hens rn North America’ which pieron lacks i ona forest climate but were present there us locene. Some species extend slightly Ward, such as to Mexico or Panama or to ec 4. E genera (5) of both Western and East- les orth America and Eastern Asia but absent from Pe, with ritieni numbers of species (ABD). imc. 770 0 NY Western Eastern World North North Eastern — ms Total America America Asia ramaeeyparis 6 2 1 3 To. Ed 1 4 | 5 | l 4 Tideo 16 l 5 p ee 2. ee SE oe 0548.55. 20 259 E > 4 P Include« cheikh 4 from Mexico to Brazil. € in southern Texas and northeastern Mexico and one in Mexico and Guatemala. northern Africa or southwest Asia. Largest num- bers of species are in Asia. Buxus (Meusel et al., 1965-78, map 273b) might be added, with New World species in the West Indies, -> and Central America, though absent northward. Table 4 has five small genera visi from Eu- rope (ABD). The four genera of conifers were named from Eastern North America and after- wards found in the other regions and were pres- ent in Western Europe in the Miocene and Plio- cene. Table 5 has only four genera, which are con- fined to the arc around the North Pacific Ocean and which may not have been widely distributed in the past (AD). Three genera are mostly broad- leaved evergreens of subtropical and tropical TABLE 5. Tree genera (4) confined to Western North America and Eastern Asia with approximate numbers of species (AD). Maps by Meusel et al. (1965-78) are cited by number. estern North Eastern World Genus Total America Asia Castanopsis* (map 121b) 102 2 100 Lithocarpus (map 121b) 101 1 100 Photinia? 61 1 60 Pseudotsuga 6 2. 4 Totals 270 6 264 a Includes one tree and one shrub mae in North rysolepis * The one species in North America c placed in the segregate genus Heteromeles. 610 range in Eastern Asia with a single tree species disjunct in Western North America. Minor dif- ferences have been noted. Thus, the segregate genus Chrysolepis Hjelmqvist (1948) has been proposed for the two New World species of Cas- tanopsis. The segregate genus Heteromeles M. J. Roem. (1847) is now generally adopted for the single New World species also united under Pho- tinia. Calocedrus, if accepted as a segregate from Libocedrus (about ten species), would be added to Table 5, with one New World species and two Old World. Genera f decid t bsent ENDEMIC TREE GENERA About 23 additional tree genera (discussed be- low) are now endemic to temperate North Amer- ica (Little, 1979: 16-19). Some are old and relicts of wider fossil occurrence also in Eurasia, while others may be young. Several others of southern and southwestern United States not mentioned here are subtropical or tropical in origin. About eight genera that include at least one tree species have natural ranges confined to west- ern North America (A), most with only one or two species, and centering in California, part coastal and subtropical. They are: Adenostoma, two species, also in Baja California, Mexico; Cer- cocarpus, six species, about four additional in Mexico; Fremontodendron, two species, also in Baja California; Heteromeles, one species, often united under Photinia, also in Baja California; Lyonothamnus, one species, confined to four species, north to British Columbia; Sequoia, one species; and Sequoiadendron, one species, seg- regated from Sequoia in 1939. About 11 genera that include at least one tree species have natural ranges limited to temperate eastern North America (B; Little, 1980). Most are restricted to the United States and have only one species. They are: Asimina, eight species, one reme southern Can- ada; Cliftonia; Elliottia (the closely related genus Tripetaleia with two species in Japan); Frank- Pliocene), two species, one south to Guatemala. Leitneria with one relict species is in the dis- tinct family Leitneriaceae. Tertiary fossils from Mississippi, Oregon, and western Siberia have ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 been referred here. Maclura would not be en- demic, if the related genus Chlorophora of trop- ical America and Africa or an Asian genus were ited. Three tree genera have broad natural ranges orth America, both eastern and western (AB) and south into Mexico. Garrya, often placed in a separate family Garryaceae, has 14 species of shrubs and trees, mostly in western United States and Mexico, including one south to Panama, also one in the Greater Antilles (Dahling, 1978). This genus with a fossil record back to Miocene apparently originated in the Southwest. Prelea has three species, one in Mex- ico only (Bailey, 1962), and is recorded back to Eocene in western and eastern states and to Oli- gocene in Europe. Robinia has about ten species including shrubs, mostly eastern, one in Mexico only. Fossils have been referred to this genus in North America and Europe back to Paleocene and in Japan to Miocene. Young genera. Two endemic tree genera, one western and one eastern, apparently are young and lack a fossil record. Oemleria (Osmaronia) with one species may have originated from be nus, a genus with about 33 species of trees an shrubs native across North America and - 400 worldwide. Oemleria cerasiformis 1s à shru or rarely small tree occurring in mountain ei yons from California north to British o: Oemleria differs from Prunus in having us : five pistils instead of one, flowers unis well as bisexual, and leaf margins entire rà | than serrate. Nemopanthus may have evolved fro The latter has about 15 species in Eastern Be : America, is extinct in the western part, ane e widespread with 300—350 species mostly Ee nm of | ad 274). Nemopa (Meusel et al., 1965-78, map ) all réf m Ilex. Canada, along and mostly north of th Ilex in the glaciated region. The rel glaciated controversial, as are H ; s next new genus of North American tree B F insane aia lk iii SE o 1983] species of Abies local in the Santa Lucia Moun- tains of Central California merits watching. H. L. Mason in a paper at a scientific meeting in 1957, predicted that this species would be named à new genus. This distinct species at the border of the generic range is so different from all the other 38 species of Abies worldwide that it has been placed alone in a second subgenus, Pseu- dotorreya. Interestingly, it has been classed also as the most primitive species (Liu, 1971). SUBDIVISIONS OF GENERA Studies of relationships between tree genera of these four separate regions commonly do not ex- tend to the subdivisions, perhaps partly because of difficulties in classification of fossils. However, living or recent species show various patterns of distribution. A good example is the genus Pinus, pine, which has about 95 species of wide range in north temperate regions and southward in Topical mountains. It has a fossil record back to early Cretaceous 1967: 106) subgenera, Strobus or soft pines and Pinus or fils men are widespread, and their four sec- i: SO in both Old and New Worlds. How- Cr, the 15 subsections are more restricted, only three in both but four Old World and eight New c 5 ACD, I; BCD, 1; AB, 1; A, 6; B, 1; CD, 1; Ct D, 2. REEN io pa Pinus apparently originated in tem- iiaia tern Asia. The large number of species iiia des Mexico indicates a center of specia- (ACD) the Tertiary. Pinus subsect. Cembrae ; I$ represented by five species of high al- nd northern latitudes but is absent from "d oig America, which lacks high moun- albicans; e & Critchfield, 1969, map 4). Pinus iin. Whitebark pine, is a timberline tree of World m ern Rocky Mountains and the only New tole pi €mber of the group. Pinus cembra, Swiss and Ca ne, grows at high altitudes in the Alps The ag Mountains of southern Europe. Siberian r three species are Asian: P. sibirica, tral Sibe — pine, mostly in western and cen- inis cv P pumila, Japanese stone pine, from ‘ern Asia to Japan; and P. koraiensis, ad. Pine, from Korea to southeastern Siberia Pan. This subsection has been regarded as LITTLE— TREES 611 one of the most primitive in the genus and is characterized by a closed cone. Of similar distribution is a genus of birds, Nu- cifraga, related to jays and crows. These birds break open the cones and store and eat the large wingless seeds of sect. Cembrae, thus aiding dis- persal. The New World species is N. columbiana, Clark’s nutcracker or Clark’s crow (“camp-rob- ber”), and the Old World species is N. caryoca- tactes, European nutcracker, with ten named subspecies or races (Goodwin, 1976). Obviously, both pine and bird groups were present together in the ancestral Arcto-Tertiary forests and may h lved in symbiosis. Both bird species ł a somewhat varied diet including other seeds, berries, insects, and eggs and readily learn to eat foods of humans. The disjunct genus Magnolia has eight species in eastern United States, about 18 others south- ward in tropical America, and about 50 in East- ern Asia south to Java (Little, 1971). Seven of the 11 sections of the genus are entirely Asian, two are wholly American, and two occur in both regions. Nine sections are confined to a single region. The eight species of Magnolia in eastern United States are classified in four sections. Magnolia acuminata and a species of eastern China are in sect. Tulipastrum, sometimes treated as a genus. Two of the southern Appalachians, three of the Coastal Plain, and one of Mexico are placed with three of Asia in sect. Rytidospermum. Magnolia virginiana is sufficiently distinct to be alone in sect. Magnolia. Magnolia grandiflora of the post of sect. Theorhodon, composed of about 16 evergreen tropical species south to Puerto Rico, the ancient flat-topped peaks of southeastern Venezuela, and Ecuador. All species of the last section may have evolved from the same New World ancestor. Alnus, alder, is another illustration with wide- spread subgenera. A/nus maritima has an odd range in two areas more than 1,800 km distant: Coastal Plain in five counties of eastern Mary- land and one in southern Delaware and stream banks on granite bedrock in two counties of southern Oklahoma. This pioneer species in wet soil bordering water is regarded as a relict of former wider range with the Oklahoma area the older of the two (Furlow, 1979: 17). It is the only New World representative of A/nus subg. Clethropis, which ranges also from Himalayas to Japan and as fossil records in Western North 612 America. A/nus japonica of Japan has been treat- ed as a variety. TREES OF WESTERN AND EASTERN NORTH AMERICA Relatively few tree species confined to North America have broad east-west ranges and can be classed as transcontinental. Examples are mostly species of wide occurrence in the northern co- niferous forest, such as Picea glauca, Larix lari- cina, Populus tremuloides, and Betula papyri- fera. Southward, the few tree species that are f d the treel interi commonly grow in moist soil near water, for example, Sa/ix nigra and Acer ne, $ Every feld botanist traveling for the first time from Eastern to Western North America (or the reverse) has observed the striking resemblance of various trees and other plants to species pre- viously studied. These similar plants are known as paired species, parallel species, species pairs, east-west vicarious taxa, and east-west vicariads. More than two species may be involved. Long lists of these species pairs have been compiled (Wood, 1971). A few examples of tree species pairs may be noted, the western first: Pinus monticola and P. strobus, Juniperus scopulorum and J. virginiana, Quercus gambelii (also Q. lobata and Q. garry- ana) and Q. alba, Celtis reticulata and C. occi- dentalis, and Acer grandidentatum and A. sac- charum (also A. nigrum and A. barbatum). The explanation is obvious. The western and eastern forests became separated and isolated following western mountain uplifts and the ori- gin of drier interior plains with grassland vege- tation in late Miocene and early Pliocene more than five million years ago. One or both species of the pair changed during the interval though not into a distinct genus. RELICT SPECIES Some tree species of disjunct genera appear to be old and relict and declining, often rare. How- ever, Liriodendron tulipifera and Sassafras al- bidum of Eastern North America are widespread and common. They are also the northernmost members of their families, Magnoliaceae and Lauraceae. Many species of conifers are rare or local in North America. Thirty-five, mostly west- ern, have been the subject ofa special publication with maps (Little, 1975). Examples of relict tree species in Eastern North ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 America with published maps (Little, 1971-81, vol. 4) include: Cladrastis kentukea (map 35); otinus obovatus (map 46); Illicium parviflorum a endemic to islands of California, for example, Lyonothamnus floribundus (vol. 3, map 98) rep- resenting an endemic genus. General NM anni +h Ol UIC VVVL ogy of rare plants may apply also to many ex amples of disjunct genera and species. Griggs (1940) concluded that a species is rare because it cannot compete successfully with common plants, that most rare species have habitats in the early stages of the ecological succession, and that many rare plants have disrupted ranges and are slowly dying out. Examples of trees charac- teristic of dry rock outcrops in the upland succes- sion are Cotinus obovatus and Torreya taxifolia (Little, 1971-81, vol. 1, map 88-E). Trees of wel soil of stream banks in the lowland succession include Leitneria floridana (vol. 4, map 74) and Alnus maritima (vol. 4, map 8). Minor geographic variations may be o e "^ lcd dot fa ient careful studies of conifers have revealed addi- tional differences. In observing conifers for -— years, I have noted some local variations an have named a few. Opinions disagree whether the differences should be designated as -—— or species or left unnamed. An extreme -— is the genus Cupressus, Cypress, which has bee mentioned. As many as 15 New W 29-35). Th " United States have ranges so local and € that the isolated stations for each specie we i been counted. Clearly, these are relicts of form greater occurrence. DISCUSSION Several observations are offered here. My - terest in this subject goes back to 4 fasci pr? graduate course in floristics by an inspirit iy fessor, Henry Chandler Cowles, at the Univ Historical P geolog“. raphy has kept pace rogress 925 there When I studied historical geology a pe s of years: ————n—————— o EA LLL LLL AI LI TLE a ÁREA TT IO ÉD n. o PER o 1983] structural geology (1926), I was taught that con- tinents and ocean basins were fixed and that con- tinents moved up and down vertically (isostasy) but never horizontally. Alfred Wegener's hy- pothesis of continental drift proposed in 1915 did not come to the attention of plant geogra- phers until later. I recall an introduction to the subject by Wendell H. Camp at a field meeting of botanists in 1939. Some botanists readily grasped movement of continents as more plau- sible than land bridges. Others hesitated mo- mentarily, because physicists had no explanation or because the hypothesis seemed unnecessary. Now, continental movements provide an expla- nation for some problems of plant distribution! Recent acceptance of wandering poles will lead to additional explanations and may require re- vision of paleogeographic maps. The shift in pole Position and earth’s axis within the past 80 to 100 million years is estimated at about 20 de- x (Good, 1964; Harrison & Lindl, Time. One important concept in studies of Plant geography is time. A century ago James (1881: 67) observed “that the time necessary for the distribution of our plants has been suf- ficiently long.” He concluded also that the species of plants common to Europe and America had à common origin and migrated south. Time has been sufficient for many tree species to migrate lo the most remote part of the earth wherever nnns for establishment are suitable. It is Mie to conceive of continuous time longer an the few thousand years of written history. However, the time since the beginning of the erùary period, about 65 million years, is ade- beni lo explain present distribution of trees and Penis viia plants throughout the world wherever den suitable habitats. The time is ample for boom: by any means, gradual, in steps, or by confine in long distance dispersal. Of course, m movements speed the process. — inis. A key word in studying plant dis- idis. Is change, or dynamics. Just as all or- URN end in death, all species become extinct kih: into others by evolution. Likewise, the ing Istribution of a species is slowly chang- “ee of tree genera. It seems remark- at many extant or recent tree genera can back as much as 50 million years into C ne epoch or doubtfully farther into the us period. That is a very long time for a 1 e . H * H genus to maintain its identity and to continue LITTLE— TREES 613 to be common and widespread. In contrast, evo- lution of land vertebrates has been more rapid. Within the same time interval, Equus, the mod- ern horse, developed from the 4-toed ancestral dawn horse, H) therium (Eohippus). Eve ter separation of the land masses and forest re- gions in the Miocene roughly 15 million years ago, the tree genera in isolation maintained their identity. Obviously, these old tree genera are sta- ble and have a slow rate of evolution. Trees and herbs compared. Herbaceous plant species are considerably more numerous than trees in temperate regions though not as well preserved in the fossil record. Also, herbaceous species of wide distribution in temperate regions in both New and Old Worlds are much more numerous than trees, as noted above. However, many of the herbs are circumpolar across far northern regions above the tree line or climatic limit of trees. Thus, it appears that herbs have a more rapid rate of evolution than trees and that herbs migrate faster. Herbs have much shorter life cycles and more numerous generations and mostly small seeds adapted to long distance dis- persal. While one seedling in many years replaces a tree, an annual herb is followed by another The relationships of North American trees with Eastern Asia have been overemphasized and are readily explained by earth changes accepted within recent years. LITERATURE CITED BAILEY, VIRGINIA LONG. 1962. Revision of the genus Ptelea (Rutaceae). Brittonia 14: ; Bot. Inst. KoMAROVA. 1977. Arealy derev'ev i kus- tarnikov SSSR. Tom 1. [Atlas of Trees and Shrubs of Soviet Union. Vol. 1. Taxus through Morus.] Tissovya-Kirkazonovya, Leningrad. Bnowicz, KAZIMIMIERZ. 1978. Chorology of Trees and Shrubs in South-west Asia. Vol. 1. Polish Academy of Sciences, Institute of Dendrology, Kornik. Cain, STANLEY A. 1943. The Tertiary character of the cove hardwood forests of the Great Smoky Mountains National Park. Bull. Torrey Bot. Club 70: 213-235. CRITCHFIELD, WILLIAM B. & ELBERT L. LiTTLE, Jr. 1966. Geographic distribution of the m of the . 1978. Systematics and evolu- tion of Garrya. Contr. Gray Herb. 209. Sci. 57: 317-332. Eype, RICHARD H. 1966 614 the modern species and their fruits. J. Arnold Ar- bor. 44: 1-59. FuRLOw, JOHN J. 1979. The systematics of the Ame ican species of A/nus (Betulaceae). Rhodora 81: 121, 151-248. GooD, R. 1964. The Genera of aciei Plants. Ed. . John Wiley & Sons, Inc., New —— DEREK. 1976. Crows of the World. Cornell v. Press, Ithaca GRAHAM, | (editor). 1972. esi ges and h America Elsevier Publishing Co., Amster an GRAY, ASA. 1840. Dr. Siebold, Flora Japonica; sectio epics ccarini. Amer. J. Sci. 175-176. eed in Sargent, 1889, Graham, 1972, and Stuckey, 1978.) 18 nalogy ai the flora of Japan and that of i United States. Amer. J. Sci. Arts r. 2, 2: 135-136. (Reprinted in Graham, 1972, ru Due mad 1978.) — ——. 1859. Diagnostic characters of new species em. Am s n.s. 6: 377- 452. (Pp. 437-449 reprinted in Sargent, 1889, and Stuckey, 1978. 1873. Address of Professor Asa ab ay, ex- president of the CTS Proc. Am Advancem. Sci. 21: 1-31. (Reprinted in leat 1889, ais as 1978) ——. 1878. Forest ion. and archeology. Amer. J. Sci. ser. 3, 16: 85-94, 183-196. (Reprinted in Sargent, 1889, and Stuckey, 1978. "e 5 S. e A monographic revision of Os- a and America. Notes Roy. Bot. 542. F The ecology of rare plants. orrey e Club 67: 575-594. Hag Host 1958-59. Distribution Maps of Flowering Plants in Japan. 2 vols. To kyo HARDIN, James W. 1957. A revision of the Yom. n Hippocastanace ceae. Brittonia 9: 145-171, 173-195. 1960. Studies in the H 1ppocastanaceae, V. Species of the Old World. Brittonia 12: 26-38. Harrison, C. G. A. & T. LiNpH. 1982. dering curve for North America during the Me- o ma spaid J. Geophys. Res. 87(B3): HORIKAWA, YOsHIwo. 1972-76. Atlas of the Jap. : An Introduction to a Sido , Ed28, Yr Air ci Forestry n Governm ment Publishing Centre, Supply & Services nada. pee p 1950. Atlasof the Disuibonól of Vas- ar Plants in Northwestern Europe. Stockholm (Reprinted, ey revised, 1971 —— Amphi-Atlantic Plants and Th Phytogeograp hical Connections. K. Ta Sven 2 Vetenskapsakad. Handl. 7(1): 340 ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 . 1971. The Circumpolar Plants. II. Dicoty- ledons. Kongl. Svenska Vetenskapsakad Handl. 13(1): 463 pp. JALAS, JAAKKO & JUHA SUOMINEN (editors). 1972-80. Atlas Florae Europaeae: Distribution of Vascular Plants in Europe. 5 vols. Helsinki (incomplete). pee rie es F. 1881. On the geographical distri- on of the indigenous plants of Europe and the (editor). 1978. Essays on North Am Geography from the Nineteenth Century. Arno Press, New York.) JAPAN FOREST TECHNICAL ASSOCIATION. 1964-71. Il- of Japan. 3 vols. A taxonomic revision of Son J. Taiwan Mus. 315- 43 6. KENG, HsUAN. 1953. safras (Lauraceae). Quart. J Li, Hur LIN. Floristic hae between stern rn North Am . with new foreword, pp. i-ii, as Morris Arboret inc PEN 1972. Eastern Asia-Eastern North America aasien in wide-ranging genera. Pp. 65-78 in raham (editor), Floristics ee Paleofloris- tics et Asia and Eastern i c orth ca. Elsevier Publishing Co., Amste LITTLE, Eur Es JR. 1971. Endemic, disjunct and northern trees in the southern Appalachians. 4 249-290 in P. C. Holt (editor), The Distributio History ofthe Biota ofthe Southern Part II. Flora. vane Polytech. Res. yout Monogr. 2. poo 71-81. ne of United States trees. 1314, cate. Agric. Misc. Publ. 1146, 1293, 1342, 1361, 1410. 1975. Rare and eie gonia in mn Use States. U.S. Dept. Agric., rv. Res. Rep. rn Appalac Inst. State Univ. Mori eid States trees (native 1979. Checklist of United A ric. Han and se hep U.S. Dept. Agric., 37S p e and local trees in the ea stern de- 198 Fe ciduous legi biome. Pp. 89-103 in Peter P. & LS sions of the genus Pinus (pines Misc. Publ. 1144. Genus Liu, TANG-SHut. 1971. A Monograph of se Abies. Natl. Taiwan hes Taipei, Taiwan. ^ MEUSEL, Hiii t al. 5-78. Versi — ud Bee qiii t en Jen press Mirov, 5 1967. The Genus Pinus. Ronald ew MIZUSHIMA, MASAMI. scular omic compariso® mm 1972. Taxonon o A A | | | eT EE is 1983] LITTLE and Japan. Pp. 83-91 in Alan Graham (editor), Floristics and Paleofloristics of Asia and Eastern orth America. Elsevier Publishing Co., Amster- 1964. ete i SABURO. Phytotaxonomical and geo- I st ix sect. Japan Gov. Forest Exp. Stat. Bull. 107 PacLr, Jiri. 1952. Synopsis of the genus Catalpa eco = Candollea 13: 241-295. SARGENT, CHARLES SPRAGUE. 1889. Scientific Papers sa Gray. 2 E Houghton, Mifflin and Co., Boston and New 9 ist ous the Zone eand ceae). Phyt 49: 390-399. The rae a Ticiaceae and Schi- sandraceae. dS og fm tue J) —TREES 615 SPONGBERG, STEPHEN A. 1974. A review of decidu- ous-leaved species of Stewartia (Theaceae). J. Ar- bar. 55: nold Sruckey, bain | Be (edito r). 1978. Essays on North A n Plant Geography from the Nineteenth Rus. e o Press, New York. Tonm Tuomas N. 1981. a oiy; Introduc- on to Fossil Plant Biology. McGraw-Hill Book ce . New York. Wo LF, CARL B. & WILLIS W. WAGENER. New World cypresses. Aliso 1: 1-444 WOLFE, JACK A. 1969. Neogene floristic and vege- i lhi t. Madroño 1948. The 20: 83-110. Woop, CARROLL E., JR. 1971. Some floristic rela- tionships between the Southern Appalachians and ginia Polytech. Inst. State Univ. Res. Div. Mon- ogr. 2. DISTRIBUTION PATTERNS AMONG ARTHROPODS OF THE NORTH TEMPERATE DECIDUOUS FOREST BIOTA! ROBERT T. ALLEN?’ ABSTRACT The distribution patterns of seven different arthropod taxa representing eleven individual mono- phyletic lineages with representatives in eastern North America, Europe, and northeastern Asia were analyzed. Taxon cladog I ing relationship gtaxa nverted to area cladograms representing relationships among areas. Nine of the area cladograms were found to represent A congruent repetitive distribution pattern. The two remaining cladograms represented two additiona independent distribution patterns. The distribution patterns were correlated with geological events that have alternately divided and reunited the principal land areas during the past 180 million years. The largest and most diverse segments of the biota of the North Temperate Deciduous Forest presently occur in three disjunct areas: (1) eastern North America, (2) Europe, and (3) northeastern Asia. This has been known for almost one hundred and forty years. What is not known is the definitive relationships of these three related biotas t e another or tl lationships of these biotas to other, perhaps more distantly related biotas. To establish definitive relationships among biotas one must practice a science of com- parative systematic biology and employ an ob- jective methodology. The objective of this paper is two-fold. First, an objecti p ti thodol gy will brief- ly be outlined whereby biotas may be analyzed and definitive relationships established. Second, a number of monophyletic, arthropod taxa oc- curring in North America (NA), Europe (EU), and northeastern Asia (AS) will be analyzed to determine the cladistic and biogeographic pat- terns of these taxa. AN OBJECTIVE METHODOLOGY FOR COMPARATIVE SYSTEMATIC STUDIES Let us consider how we might compare the biotas of the three areas under consideration. We might begin by making lists of the similar en- demic taxa common to two or more areas. A number of such lists have been made. Asa Gray (1846) made such a list and recorded a number of plant species and higher taxa that were en- ! Published with the approval of the Director. ? Entomology De suggestions: Ronda Bitterly, James B demic to eastern North America and Japan. Some one hundred years later Li (1952) reanalyzed the floristic similarities between eastern North America and northeastern Asia in the light of modern systematic botanical data. These new data, presented by Li, showed that the species Gray had thought to be the same in the two areas were, in fact, different species. Li did, however, record 52 genera that were endemic to eastern North America and northeastern Asia. A num- ber of additional papers (Graham, 1972) since the work of Li have further documented the sim- ilarities of the floras in these two areas. Similar studies comparing the biotas of East rn North America and Europe have also be published. Lindroth (1957) presented a comp hensive list comparing the faunas of the pee boreal, and temperate regions of North Ameri with comparable areas in Europe. The 1983) S of [4] taxa with disjunct distributions 0 tinents. Studies comparing the biotas of Europe ps northern Asia or comparing the biotas o d northern Asia, and North America are e The work of Hara (1952, 1956) does pers of the plants of the three areas. But the m | the published literature compares only two : , Arkansas Agricultural Experiment Station. partment, University of Arkansas, Fayetteville, Arkansas 72701. ing aluable ivi itically reviewing the manuscript and offeri for typing vi the drafts and final copy of this paper ANN. Missouri Bor. GARD. 70: 616-628. 1983. 1983] areas and we may be able to form only two sets of comparative statements: (1) the biota of NA(e) (North America, east) is related to the biota of AS (Asia) (Fig. 1) and (2) the biota of NA(e) is related to the biota of EU (Europe) (Fig. 2). These i wo-entity com- risons and as such h w” informational content (Platnick & Nelson, 1978 Since two-entity statements are prevalent throughout systematic biological literature it is important to clarify the statement just made, i.e., that two-entity statements have “low” infor- mational content. In systematic biology two-en- lity statements may be of two types: (1) two- laxon statements— species A is related to species B; (2) two-area statements—the biota of area a is related to the biota of area b (Fig. 3). If one accepts the idea that extant taxa come from preexisting taxa through a series of intercon- nected, related ancient ancestors, i.e., evolution, thenall taxa and biotas are related to one another at some level. The question, then, is at what level the two taxa or biotas are related. To answer this question, a third entity must be considered so that a comparative statement may be formed. : Consider the proposition that A and B are more osely related to one another than either is re- ted to C (Fig. 4). This is a threc-entity com- cestor 2, than either did with C, and C's hy- na ancestor 1 (Fig. 4). We may also de- ea at A and B share one or more attributes If mmon that are not shared with C. for B. ussion just given seems too simple EM Ene student or practitioner of sys- Seque iology, consider the published litera- conse wi of information is replete with Me iiia er example of two- -entity statements. Pici © move toward an objective, analytical Systematic biology as suggested by Ian meen sac we must begin constructing com- etn oO composed of a minimum of e CoOnct three entity a in systematic biology begins with a to ls analysis of three or more taxa believed "isand p monophyletic relationship. The meth- "€ Procedures for executing a cladistic anal- ua" been discussed at length by a number Crisci rs including Hennig (1966), Ross (1974), and Stuessey (1980), and Wiley (1982). ALLEN—ARTHROPOD DISTRIBUTION Fig. 2 Fig. 3 Fig. 5 Pig. 6 D NA(e) G NA Ce) Fig. 7 Fig. 9 —2. Two area statements illustrat- between North America east A two tron g the relationships apie hy- A I n aiit ing in —4. three taxon statement illustrating the ata ahin E clado Ficures 1-10. ica, NA(e), Europe, EU, and Asia, AS. cladogram derived from Figure 5 illustrating the re- 7-8. Ta xon/area cladograms of two hypothetical monophyletic lineages.—9. An area cladogram m depict- ongruent distributional patterns arana in d8.— 10. An area cladogram illustrating ype of distributional pattern that ance exist between se North America, NA(e), Asia, AS, and Europe, EU. Here we need only note that there are four basic tenets necessary for a cladistic analysis of any taxon: (1) an analysis and comparison of char- acters and character states in the taxon being studied and in suspected related taxa; (2) the pos- tulation of apomorphic (derived) and plesio- morphic (ancestral) states for as many characters as possible from the analysis and comparison of character states in the study group and related groups; (3) the establishment of relationships (monophyletic lineages) based on the possession of shared apomorphic characters and the graphic illustration of these relationships in the form of a taxon cladogram; (4) an analysis of the bio- geographic relationships of the taxa and the areas in which they occur by converting n taxon cladograms to area cladograms. Let us consider a hypothetical example in which we cladistically and biogeographically analyze a monophyletic taxon with representatives in three disjunct areas. 618 Let us suppose that we discover a genus with representatives in each of the three disjunct north temperate deciduous forest areas: species A, NA(e); species B, EU; species C, AS. Let us also suppose that by applying the methodology for A at that species A and B are more closely related to one another than either is to species C. This in- formation concerning relationships may be ex- pressed in the form of a taxon cladogram (Fig. 5). This taxon cladogram is a statement that ful- fills the following three requirements for a sci- entific hypothesis (Ball, 1975): (1) the cladogram fits the known data about relationships; (2) the cladogram may be tested by a reanalysis of the known data and/or the discovery of new char- acter state data; (3) the cladogram predicts that the relative relationships of the three known taxa will remain the same no matter how many ad- ditional species may be discovered, i.e., species A and B will always sh t ancestor with one another than either will ever share with species C. Once the cladistic relationships of a mono- g d, other types of data, for example distributional data, may be plotted against the cladogram. Rosen (1975, 1978) and Nelson and Platnick (1981) have discussed the methodology as well as the scientific basis for converting taxon cladograms into area clado- grams by plotting distributional data on the tax- on cladogram. When this distributional data is plotted on the cladogram (Figs. 5, 6), we may see that we have a hypothesis relevant to the rela- tionships of the three disjunct areas in which species A, B, and C occur: eastern North America [NA(e), species A] and Europe [EU, species B] have shared a more recent common biota with one another than either area has shared with northern Asia [AS, species C] (Fig. 6). Using this methodology, it is necessary for a worker to make I fully realize that perhaps one-third of the plant species (as well as some ani ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo.. 70 The hypothesis about the relationships of bio- tas in NA(e), EU, and AS may be tested in at least two ways. It has already been suggested that the taxon cladogram may be tested by the further study of the characters on which the cladogram is based and the study of additional characters not previously used. In addition to this test on character states, a search may be made for ad- ditional monophyletic taxa with a distribution range similar to the first taxon. If such taxa are discovered, taxon cladograms may be deduced (Figs. 7, 8) and these taxon cladograms may be converted into area cladograms (Fig. 9). If the area relationships of the new, additional taxa are the same as the first taxa, the hypothesis is cor- roborated and may suggest a generalized distri bution pattern (Fig. 9). Ifthe new, additional taxa suggest a different set of area relationships (Fig. 10), then the original hypothesis may be incor rect or a second distribution pattern may exist In either case, corroboration of the first hypoth- esis may be objectively pursued. E Once corroborated and generalized distribu- tion patterns emerge, we may consider the phe- nomena that have contributed to the evolution of these patterns. We might ask, “Are the pat- terns due to the dispersal of organisms acros pre-existing barriers and to the subscinet ation and evolution of the organisms?" or pe the patterns due to historical factors that divi ancient cosmopolitan populations and stage for allopatric speciation to occur? pon attempting to answer these questions, let us sider the literature that is available anc a for taxa that are informative about biotic pe tionships between NA, EU, and AS, and pes how we can fit the available information wes objective methodology that has been 0u j USING PUBLISHED DATA IN A CLADISTIC-BIOGEOGRAPHIC ANAL * z * tterns. To objectively analyze distribution od two prerequisites are necessary: (1) 4 MES dnt letic taxon with distinct taxa, ME or data 2) a taxon cia ogram or more areas and (2) y be P be satisfied YSIS through original research or, in som by utilizing previously published inform The analysis of north temperate distri u terns presented in this paper relies he the work of other authors. These previo i lished data are not always in a form = directly : ladistic biogeograP. 4 EO ee ee Tea ee a ee ea MM OMM MU MM e o o Lh. 1983] and some rearrangement of the data is necessary. The following is a brief statement of how pre- viously published data used in this paper are pted in order to perform a cladistic-biogeo- graphic analysis. Three subject areas need to be considered: (1) the conversion of phylogenetic diagrams to cladograms; (2) the conversion of ta lad lad ; and (3) the accuracy of distribution records. iagrams representing the relationships of taxa to one another have been a part of systematic literature since the time of Haeckel (1866). These diagrams have been referred to by a variety of names, but the most common name appears to be "phylogenetic tree." More recently (1960s to the present), the diagrams have been referred to às cladograms. It has only been within the past thirty years that definitive and objective methods have been Proposed for the construction of cladograms. Central to this methodology is the establishment ofrelationships based on the possession of shared characters, In those instances where the meth- od of determining relationships is clear, one may baiona) T the use of the cladistic and distri- Fia: iniormation. Frequently it is difficult to lermine what method an author used to de- termine relationships. In those instances where method of determining relationships is un- or where relationships are based on some T other than the possession of shared apo- tphic characters, the published data is better to rest in pea ian tionship diagrams have also been pub- c mo r on™ relationship of four taxa A, B, C, "s ip 11) where A and B are more closely One another than either is to C or D, B, and C are more closely related than they are to D, the positions in some instances the location of taxa al nodes, in this case A and B, have er, 3 termin ALLEN —ARTHROPOD DISTRIBUTION 619 Gon wx» pO O »- m Fi. 12 Fig, 15 FicunES 11-15. 11-12. Taxon cladograms of a hy- pothetical group of species in which the positions of species A and B are reversed by rotating the cladogram at node 1. Such a rotation does not change the infor- mational content of the cladogram.— 13. An area cladogram in which the relationships of areas outside of North America, Europe, and Asia have been indi- cated by a dashed line.— 14-15. Area cladograms il- 1 : 1 P^ m £ amirrino in eactermm North America, NA (e), have been reduced to one area in Figure 15. been reversed (Fig. 12). Such a reversal has ab- solutely no effect on the interpretation of the data presented. The statement “A and B are more closely related to one another than either is to C” is the same as the statement “B and A are more closely related to one another than either is to C.” Put in another way, nodes in a clado- gram may be rotated on their axes. As you will see, this rotation procedure is often helpful in detecting congruent distribution patterns. 1 +1 mar this point it 1S WO the plants and animals we study are not obliged to fit precisely and neatly into either our classi- nv student of nature, even after a limited amount of experience, quickly recognizes the variation that is rampant in nature. This does not mean that biological patterns do not exist, for certainly they do. It does mean that while striving to detect patterns that are common to large groups oftaxa, we must also accurately record the variability we observe in nature. Patterns exist because indi- 1 1 4 A Å 1 > } + 1 A independently of one another. Individual species are constituent members of larger biotas. These biotas have shared common elements in their mind I have employed two methodological pro- cedures that have assisted in discovering com- mon distribution patterns between the biotas in North America, Europe, and northern Asia. The first methodological procedure that I have 620 used was necessary because many taxa occur in disjunct areas other than those under consider- ation. For example, Li (1952) noted a number of temperate deciduous forest taxa that occur not only in eastern North America and northeastern Asia but also in more southern tropical areas such as the Philippines. Other taxa such as some caddisflies (Insecta: Trichoptera) occur in eastern North America, Europe, northeastern Asia, In- U, and AS have been indicated by dashed lines (Fig. 13). It is felt that this method conveys all of the known data about the groups but allows one to perceive what the relationships are among the areas and taxa that occur in our three prin- cipal areas, NA, EU, and AS. The second methodological procedure used in detecting distribution patterns was first em- ployed by Rosen (1975, 1978). In this technique, identical areas that are juxtaposed to one another in a single monophyletic lineage are combined (Figs. 14, 15). When this combining method is used, the original data should always be pre- sented so that other workers will have access to all the facts and so that any conclusions may be judged objectively.5 The last topic to be considered here is the ac- curacy of distribution records. What is meant by accuracy is the tendency that some biogeogra- phers had in the past to list taxa as occurring in broad geographical ,ie., Asia, North Amer- ica. We now know that many of the geographical areas formally recognized as single units are ac- tually composed of two or more geographical subunits. These subunits have often had inde- pendent and therefore different histories. Even the relatively small island of Celebes is composed of at least two independent island masses that 3 At the first annual meeting of the Willi Hennig Society in 1980 Norman Platnick pointed out that Ro- s method of combining the same juxtaposed areas areas ofendemism (not disjunct), then one could argue a greement” in different cladograms might result from chance alone. Perhaps Platnick’s argumen is true. As long as the original, detailed data are (Allen, ms.). p ccurately we will be able to return to the ry. resented a problem Platnick Suggests exists if necessa ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 only fused together a few million years ago. When SUIIIIIIdl Bl OLIICI used in this paper, the distribution of each taxon and lineage has been verified and determined as accurately as possible. This has sometimes led to conclusions that differ from those expressed by the original authors. Using a cladistic methodology, including the methodological procedures just discussed, the distribution patterns of a number of north tem- perate, disjunct, monophyletic arthropod taxa have been analyzed. Let us consider these indi- vidual taxa, their taxon cladograms, and their area cladograms. TAXON AND AREA RELATIONSHIPS AMONG DISJUNCT NORTH TEMPERATE ARTHROPOD GROUPS A search of the literature revealed a number of arthropod taxa with distributions that might be informative about the relationships of tem- perate biotas and areas in North America, Eu- rope, and northeastern Asia. These groups occur in three orders of insects: the caddisflies (Tr choptera), the leafhoppers (Homoptera), and the beetles (Coleoptera), and in one group of spiders eae). (Aranea CADDISFLIES (INSECTA: TRICHOPTERA; GLOSSOSOMATIDAE, PHILOPOTAMIDAE, RHYACOPHILIDAE) Caddisflies are a moderately large order of aquatic insects. The systematics and Mee á these organisms have attracted the went ps a number of highly competent ento d many years. One of the leading authoriti caddisfly systematics was H. H. Ross. especially interested in the basic taxonomy biogeo hy of these insects. . Cuddisity ordi inhabit a wide range of w habitats, but many taxa are characteris mountain regions ire d made a special study on the syster" number f rint caddisfly taxa with ye wide distribution patterns and published s that sults in 1956. The phylogenetic — sl Ross presented for the mountain caddis (Fig. 16) have been changed to taxo grams (Fig. 17). These cladograms, NT been reduced to more simple cladog mi 18, 19) illustrating the area relationships Eurof North America (NA) (east and wen an (EU), and Asia (AS). Four mountain | | | | | Í PRODUCTA AS(n) AS(n) UNINACULATA NACw) NACw) VETINA Alw) EU BELONA ) NACw) LAEVIS EU ASCm VEPULSA NAW) NALS) PE NARVAE Stn) Her ig. 17 TRANSQUILLA AS(n) NACe) KISOENSIS — AS(n) NACw) YUKI I Stn) ASth) SIBIRICA AS(n) AGO MINORA NACe) MANISTEE NACe) CHER BLARINA NA(w) LITA NA(e) AMICIS NA (w) ATRATA Ale) VALUMA NA (w PELLISA NA (w RICKERI NA Cw) DEPRESSA AS(n) ABCHASICA AS(n) ISOR Alw) OPHRYS NA (w) VELORA NACw) EIA NAGO FicURES 16-19, Th e caddisfly genus Rhyacophila el Broup.— 16. E" phylogenetic chart pub- tc arta by Ross (1956) that have repre- hene €s in the north temperate areas are being Idered Er The four taxa and their asso- R mou i ea cladograms are as follows: (1) 6(E ono lonia group (Figs. 16-19), branch Hee 22); Glossosoma (Figs. 23-25); (3) 3). la (Figs. 26-27); (4) Agapetini (Figs. 28- When each of the reduced area cladograms ces ting the independent monophyletic lin- I$ considered (Figs. 31-36), two distribu- kis Rhyac Patterns seem to be identical: the pattern in dti Yacophila sibirica group (Figs. 31, 37) and in Glossosoma (Figs. 34, 40). We might lemperare Are the other patterns in these north th ays RE any taxa radically different from considered. When the area cladograms are M LL E 31-36) the area sequences do Other E ifferent on the one hand, yet on the M there also appears to be some degree D£ruence. Consider how the different pat- "d be arranged in linear sequences (Figs. 2. Now consider how these linear se- be arranged so that the same areas Tespond with one another (Fig. 43). This latter "gement (Fig. 43) of the linear sequences in now ask, *Are ALLEN —ARTHROPOD DISTRIBUTION LILLIPUNTA AS(n) MINUTA AS(n) ASCs) ARGENT IPUNCTELLA AS(s) AS Cn) DA AS(s) IN CASTAR AS(n) --- SCISSA IN EU ordi NACe) PEPI ge S n) NA (w) MARCI Sn) EU PHILOPOTAMO IDES EU NACe ae S EU NACw) TEDDY NA(e) NA (w) canoe TAA NA le) FENESTRA NAte) Fig. 21 LEDRA NACe) KIAMICHI NACe) SONOMA NACw) HYAL INATA NACw) VOCALA NA Cw) CLAVALIS EU(me) GLAREOSA EU MYCTA NACe) IGRIT. NACe) CARPENTERI NACe) PARANT NA(e) NKS NACe) V1BOX NACe INVAR NACe) KERNADA NA(w) COLORADENSIS NA(w) BIFILA NA(w) A ERI Alw) Fig. 20 FIGURES 20-22. The caddisfly genus re branch 6.—20. Taxon/area cladogram.—21- duced area cladograms. no way changes me original data but it does ac- as and it notes those areas in ninh members of a monophyletic lineage do not occur. PERDE we are peeing in eac the individual gment ofa la arger more inclusive pattern (Fig. 44). Let us zonsider additional taxa. LEAFHOPPERS (INSECTA: HOMOPTERA; CICADELLIDAE Because of his general interest in the evolution and biogeography of insects, H. H. Ross worked on a number of diverse groups. In addition to the caddisflies, leafhoppers, especially the genus Empoasca (Cicadellidae), were of interest to Ross. He published a number of papers dealing with PYRENACIUM EU PRIVATUM EU EU BEAUMONT! E ES TIMURE EU(me) NACw) ANALE S(n) AS CAP TACUM u t MALAYANUM AS(s) Fig. 24 CAUDATUM AS(n) HUM S(s) ASPERIGLOSSA NACw) MUROGLOSSA AS(s) CI POGLOSS S(n) LIPOGLOSSA AS(n) LIPOGLOSSA AS{s) PROTOGLOSSA AS(n) SINOGLOSSA AS(s) DIPLOGLOSSA EU SYNAFOPHORA AS(n) EJTI AS(n) LIVIDUM NACe) USSURICUM ASUn) VERDONA NA) AS ALTAICUM AS(n) NA(w) RIPAEGLOSSA — NA(w) EU Fig. 23 Fig. 25 E RES 23-25. The caddisfly genus Glossoso- a.— 23. — cladogram.—24-25. Reduced SUBT ERRANEA EU SUBNIGRA EU TRIANGUL MERA EU OCCIPITALIS EU VARIEGATA EU DIA PULLA EU CORV INA EU COP10S EU AEQUAL IS EU CONGENER EU MEDIA EU PAUL I ANA AF KYANA AF GRESSITTI AS(n) SARAWAKANA ASCS) MOHRI NACe) KISOENS!S AS(n) MONTANA AS(s) ULMER | AS(n) SPINOSA AS(n) YDA AS(n) ALTICOLA AS(s) RECTA IN CHINENSIS AS(n) RELICTA IN MOESTA NACe) GABRIELLA NA Cw) HAMATA NACe) parr NAC SHAWNEE ( L™ OCC IDEA (w) DORSATA NA(w.mx) CRUZENS!S NA(w) ANILLA NA(w) MPF I NA(w,mx) ARIZONENSIS NA Cw) PLANAE NA (w mx) ESPERONIS NA (w , mx) Fig. 26 OSTINA SA INSIGNIS SA FIGURES 26-27. The caddisfly genus Wormal- dia.—26. Taxon/area cladogram.— 27. Reduced area cladogram this genus. One of these papers is a taxonomic and biogeographic treatment of the empoascan subgenus Kybos (Ross, 1963). Discussing the distribution of the subgenus Kybos, Ross said, “Of the 62 species of Kybos available for study, 49 are known only from North America, 13 from Europe. It would seem certain that species of this subgenus occur in Asia also, but currently none are available from that area, either through the avenues of specimens or de- scriptions giving details of male genitalia. When these distributions are plotted on the family tree for the subgenus, chart 1 [Fig. 46], it is remark- able that all ofthe known European species occur NA COMATUS EU NA PUNJABICUS — IN EU COCANDICUS — EU(me) AS SINDIS IN EU AS KASHMIRENSIS IN ^U MEMBROSUS AS(n) AS BIDENS EU(me) AF KIRGISORUM — EU(me) u RUDIS IN Fig. 29 CHINEN AS(n) SIBIRICUS AS(n) CATARACTA ASCS) HAMATUS AS(n) TAGAPETUS AS(n) ULMER | G MONTICOLUS AU UNGULATUS AF DUBITANS EU Fig. 28 FiGURES ath The — t 28. Taxon/area cladogram.—29.30. Red Te ona dodoa a ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo.. 70 AS/NA/EU/AS Fig. 37 Fig. 31 Rhyacophila siberica group AS(n) AS (n) /EU/NA(Ce) EU Fig. 38 NA(Ce) Fig. 32 Rhyacophila (branch 6) EU EU/AS/NA/AS Fig. E Fig. 39 No. Linear Pattern AS 37 AS /NA (4/7777 jo AS Fig. 33 Glossosoma 40 AS/ wee — EU/AS 42 MNT TS AS AS/NA/EU/AS 39 EU/AS/NA(w) /-777- "E NA Fig. 40 41 E e EU 8 EU/AS/----* INA AS Fig. 34 Glossosoma Fig. 43 EU EUAS ANA AS Fig. NA Wormaldia NA EU AS Fig. 36 EU/AS /NA (w) /NACe) /EU/AS | NA/EU/AS Fig. 44 Fig. 42 Agapetini FIGURES 31-44, 31-36. Reduced area cladograms of the caddisfly taxa indicated. —31-42. La area cladogram ms.—43. Comparison of the linear iai quential arrangement patterns in SIX monophy taxa.—44. A generalized distribution eme suggested by comparison of the six monophyletic disfly taxa in Figure 43. in the large branch containing the butleri and copula groups." Unis wingly, Ross alluded to the fact that Kybos is a north temperate taxon and that rep- resentatives might well occur in northeast Asia. Ross had certainly seen many eiam this type of Northern Hemisphere distribu wala? GROSATA ated AL EXANDERAE Pr POPUL! ie COPULA ! BUTLER Fa group BUTLER! nate GRIBISA OVALIS is uate? group PETIOLARID nate? TRIFASCIATA ANDRES !^ nate? group HUMILIS (v! NELLA (s) ALBERT tm! Fig. 45 YUKONENS ! $ eer LUCIDAE i CRYSTOL RUFESCENS Swi CLINATA Eu ^ o / Fig. 46 ^ | - T". FIGURES 45-47. 45. A taxon cladogram the h | major groups in the leafhop diet | 47. Taxon/area cladogram butleri (Fig. 46), gribisa, an | in the subgenus Kybos. ra mfascita ata (Fig. 4) BUTLERI EU [Ems NA(w) PETIOLARIDIS NA(e) Fig. 48 GROSATA ALEXANDERAE NA(w) GELBATA NA(e) POPUL I EU z COPULA NACe) NA(Cw) NA(e) NA(w) ALLEN—ARTHROPOD DISTRIBUTION EU/NA(w)/NA(e) Fig. 50 NA(w)/NA(e)/EU NA(w) AMICIS —-NACw) Le group BL. si EU NA (4) /NA Ce) /EU oup i i VIRGATOR EU A wu s oup Fig. 54 NEOMYAS NAle) NAle) MYAS EU(s) EU JE ions AS AS NA(e)/EU/AS ARISTOCHBQA — 45 7) "aded e m Fig, $7 $ AD 48-59. 48,51,54. Taxon/area cladograms orden in the subgenus Kybos. —49, 52, 55. Mia area cladograms of three Kybos lineages il- quere o In. Figures 48, 51, and 54.—50, 53, 56. Se- adi. — 59. Sequential linear arrange- LN the reduced area cladogram of the subtribe from hi 1 a his work with caddisflies. It will be inter- fie Hi 2 see where any north Asian species fit in ture gram should they be discovered in the fien road Robes Major species lineages in Kybos are de- “ined and the distribution ranges are plotted lineages (Figs. 48, 51, 54) that are formative about E xd dud Dships- ; M Kybos lineage 1 containing butleri and tinis. Species (Figs. 48-50); Kybos lineage 2 K ing the copula lineage (Figs. 51-53); and EN RE e an plus i c" 7 g dig Furopean species that occur in the major the i abe (Figs. 54-56). Before considering ibility that these three leafhopper pat- Caddi A in tw “of Pattern (Fig. 44) let us examine patterns GROUND BEETLES SOSUA COLEOPTERA; CARABIDAE) wide Ed beetle family Carabidae is world- distribution with a significant fauna in ON N w NA(sw) AS(EU) GERTSCH EREMELLA SCHUSZTERI OL IGOCENE, . i dnoa6 IMBECILLA NA(n,e) z CHISOS NA(n,e) 22 EARLY LUTO NA(n,e) oa EOCENE ee NOCTURNA EU oD NCOL U = LATE CRETACEOUS A NA(sw) P l NA(sw) g © Uu N 4 m b: EARLY nd EILICA Gondwanaland JURASSIC f Fig. 60 EU NA(e) NA(u)/AS/// SS Fig. 62 Ficures 60-62. 60. Taxon/area cladogram of the spider genus Callilepis.—61. Reduced area cladogram he genus Callilepis.—62. Sequential linear arrange- ment of the reduced area cladogram of the genus Cal- lilepis. North America, Europe, and northeastern Asia. A number of monophyletic taxa in this north temperate fauna have disjunct representatives in NA(e), EU, and AS. To date, only one taxon, the subtribe Myadi, has been objectively analyzed (Allen, 1980). There are six genera in the subtribe Myadi: Neomyas, North America; Myas, Europe; Xe- nion, southern Europe; Aristochroa, Tibet-China; Trigonognatha, China, Japan, Korea, Formosa; and Steropanus, China. Specimens of Steropanus have not been available for study and therefore p esses s 2 E A the relationships of these taxa with their distri- butions has been constructed (Fig. 57). Unfor- tunately, there remains an unresolved trichoto- my at the base of the cladogram, but this does not detract from the information content above this node (Fig. 58). We may see that in the Myadi, North America and Europe appear to have been more recently associated with one another than either area was with Asia (Figs. 57-59). SPIDERS (ARANEAE: GNAPHOSIDAE) Platnick (1975a, 1975b, 1976) has analyzed the cladistic and biogeographic relationships among species in the spider genus Callilepis. Cal- lilepis is presently found in North America (east and west) and in Europe reaching east into north- eastern Asia. This distribution pattern, at first glance, may not appear to be disjunct between Europe and Asia, but when the ranges of the individual species of Ca//ilepis are plotted against the clad m of species relationships and the Callilepis's cladogram is reduced (Figs. 61, 62), a disjunct pattern becomes evident (Fig. 60). It 624 Glossosoma ! EU/AS/NA(w)/----- [--/AS Rhyacophila br. 6 EU/AS/----- /NA(Ce) Myadi NACe) Glossosoma 2 Rhyacophila sib.gr. Wormaldia EU NA(w)/NA(e)/EU FIGURE 63. A comparison of the linear sequential » fe taxa. is true that there are broad areas of sympatry in some taxa, for example C. schuszteri occurs in both Europe and Asia. This apparent sympatry may be due to dispersal at some point after the C. schuszteri lineage had been isolated and had evolved into a distinct species. Now that we have studied a number of north temperate distribution patterns among several arthropod taxa, let us consider all of these pat- terns together. SUMMARY OF DISTRIBUTION PATTERNS We have now examined the distribution pat- terns in seven different arthropod taxa occurring in the temperate Northern Hemisphere. A total of eleven individual monophyletic lineages that have distinct, disjunct representatives in three or more north temperate areas (NA(e), NA(w); EU; AS) have been detected. Using the results of a cladistic analysis of these lineages we may ar- range the individual patterns of these lineages in linear sequences and match the corresponding areas (Fig. 63). The arrangement of the areas in linear se- areas not represented in a particular monophy- letic lineage. But the sequencing of areas must be used in conjunction with the actual clado- grams because phyletic events do not always oc- cur in a precise sequential pattern. Consider the cladograms for the spider genus Callilepis (Fig. 60) and the caddisfly tribe Aga- petini (Fig. 28). The same areas (NA(e), NA(w); EU; AS) are represented in each cladogram but the relationships of the areas are different in each cladogram. In the Agapetini, the area relation- ships occur in a straight linear sequence: Asian forms were isolated first; European forms were isolated second: finally North American forms were isolated into eastern and western segments. In the genus Callilepis, area relationships do not ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 occur in a straight linear sequence: the Callilepis lineage was first isolated into two major seg- ments, Asia plus western North America and Europe plus eastern North America. Species in each of these two major segments were subse- quently isolated into Asian species related to western North American species and European species related to eastern North American species. Thus, the phyletic events depicted in cladograms tell us a great deal about area relationships and must be considered when analyzing area rela- tionships. In all of the taxa we have examined, the phy- letic events depicted in the cladograms imply etc.) and subsequent isolation and (2) the origin of barriers (opening of the North Atlantic) that divided ancient cosmopolitan populations, 16 vicariance. In recent years, there have been 5o geographers that have advocated both dispersal and vicariance explanations. Both of these types of explanations have been proposed in this eer posium to explain the distribution of p rx animals (notably, Iltis as a proponent 0 persalism and McKenna and Allen as phe nents of vicariance). It is not the PE n s paper to give an account of du: 7 w only state that in my estimaul I of dispersal obviously does occur, but the m lution of whether a distribution pattern I$ due dispersal or vicariance can be ascertain after a cladistic analysis of the taxa invo the correlation of this cladistic an Platnick & Nelson, 1978). Hypa va attribute either a dispersal or Maing nation to an individual pattern can e yi is of rated or refuted based on the analysis 9 1983] ulations were divided and isolated during the earth's history, we might ask what the events were that led to division and isolation. In the laxa we have studied we are particularly inter- ested in possible division events affecting the landmasses that are now found in the Northern Hemisphere. The first event relevant to under- standing the history of north temperate taxa was the division of the supercontinent of Pangaea into a northern landmass, Laurasia, and a south- em landmass, Gondwanaland, sometime in the Cretaceous. Raven and Axelrod (1974) suggest that the last possible date for the direct inter- change of taxa between North America, Europe, and Africa was about 180 million years ago (Ma). From the Cretaceous to the present, many pop- ulations have displayed considerable east-west, north-south movements throughout the history of the north temperate areas (the work of Mar- garet B. Davis, this symposium, gives examples of recent movements of some plant species). The older fossil record is indicative of movement and 'orthern Hemisphere. If this is true, then these biotas have been subject to a number of distinct &cological events occurring in the Northern Hemisphere, The fossil and geological record indicates that portions of the Northern Hemisphere have been aci inundated by epicontinental seas. ie €picontinental seas were the Midcon- Seaway that divided North America into i and western sections and the Turgai Straits Were vided Europe and Asia. These seaways Cox a. in the Upper Cretaceous (1 00 Ma). iios ^s suggested that these seas divided an "cci nosaur fauna into distinct groups, an merica Group and an Euramerica Group. (45 aeris area was divided in the Eocene Raven a y the opening of the North Atlantic 1 & Axelrod, 1974; Cox, 1974). The Asia- (40 e area was first severed in the Oligocene a) (Cox, 1974; Colbert, 1973). ipn with the Midcontinental Seaway , Tétaceous) was the beginning of the cOenies that gave rise to parts of the western 1969) E mountain system (Dunbar & Waage, ' + MS western mountain building began in ALLEN —ARTHROPOD DISTRIBUTION 625 the early Cretaceous and continued along the eastern border of the present-day Rockies, ing the intervening basins so that by the Oligo- cene this western area became a flat surface 2,000 to 3,000 feet high. This peneplained, elevated area was not high enough to create an eastern rain-shadow effect. Thus, during the Oligocene the eastern and western biotas of North America y Sub- sided. The European seaway, the Turgai Straits, also subsided and the biotas of Europe and Asia were also reunited. During the Miocene the mountain areas of western North America again became active: Uplifting occurred along the entire Cordilleran range from Alaska into Middle America (Dunbar & Waage, 1969). In addition to this mountain building, climatic changes took place throughout he world. These events eventually led to the separation, once again, of North America into eastern and western biotas and the separation of Asian and European biotas. In the intervening areas separating these continental biotas, exten- sive grassland biomes began to develop. The geological history of the Bering Straits af- ter the Oligocene is one of repeated land con- nections and disjunctions between Siberia and North America (Colbert, 1973). The connections have provided opportunities for plants and an- imals to disperse and establish cosmopolitan populations. The disjunctions would have iso- lated segments of these cosmopolitan popula- tions and provided opportunities for allopatric speciation to occur. The geological events that have just been dis- cussed have been correlated with phyletic events in nine of the area cladograms (Fig. 64) we have studied. (Callilepis and one ofthe K ybos lineages will be discussed separately.) We may see that there is an orderly, sequential occurrence of both phyletic and geological events through time. The connections and disjunctions of continents and continental parts at the different time intervals can be depicted schematically (Fig. 65), illus- trating how different land areas were associated. In the diagram showing the correlation of phy- letic and geological events (Fig. 64), the M iocene mountain orogeny in western North America and the worldwide climatic changes occurring at this time are shown at two points in the cladograms. The occurrence of these Miocene events 1s con- sistent with the cladograms and the geological e 626 z > £ - > m EU e Glossosoma ! Rhyacophila br. 6 Myadi Glossosoma 2 hila NA(w) $SNA(Ce) "n H J| H MEE à NA(w) $ NA(e) EUS AS Wormaldia pene NA(w) BNACe) E EU Agapetini m c > [ —— SKybos 2 z > = m— sKybos 5 ‘JURASSIC 80) a CRETACEOUS (100) EOCENE (45) E 64. The reduced area cladograms of nine e vided and isolated ancient anra iiia lect at the points indicated on the cladogra history. In the two Kybos lineages and in the A sequently divided into eastern and western seg- ments in the Miocene (25 Ma). In these lineages (Kybos 1, 2, Agapetini), additional taxa have not been found in other areas. In the Wormaldia lineage, there was apparently a cosmopolitan population occurring in North America, Asia, and Europe. The North America and Asia-Eu- rope connection was severed in the Oligocene (40 Ma). The now independent North American population was not divided into eastern and western lineages until the Miocene. The Asian and European cosmopolitan population of Wor- maldia was also divided in the Miocene as were populations occurring in the Glossosoma 1 and Rhyacophila branch 6 lineages. The Callilepis area cladogram is a double di- chotomy (Fig. 61) that does not lend itself to inclusion with the linear cladograms of the other taxa. However, the area cladogram of the spider genus Callilepis may be correlated with four of ANNALS OF THE MISSOURI BOTANICAL GARDEN |^ NAw NAe EU AS | EES es Fig. 65 FiGURE 65. A schematic diagram depicting the di- JURASSIC CRETACEOUS EOCENE OLIGOCENE MIOCENE rived from cladograms illustrated in Figure 64. he same geological events we have phe (Fie 60). Callilepis may represent a slightly dif- ferent biogeographic pattern than the pattern ie hibi ted i ane other arthropod taxa we hon "- ied. S among bed in 1 Callilepis: it will be wor " consider’ Platnick’s (1976) discussion? 0 es in greater détail.: jara There are ten species in Callilepis. Their kn distributions. are indicated on the adogram 60). There are two species groups, the e nu group and the schuszteri group. Bot — éd representatives in North America, with the 7 turna group having one species in oa Asia (C. schuszteri). At first, we m t isions (4 that a distinct pattern of population iid is Morus a vicariance pattern mew C . *If we assume that the presen Gom ASB M in Europe is due to dispersa ase e two groups of Callilepis can be ith te sentially vicariant on a global bene North piu group occurring in southwes ca and eastern Asia and the noctu" pur grout occurring in — and northern an North Americ 3 Eur- tas, Platnick keen associated the ciel é america and Asiamerica, with p Š 60). Thes divisions correspond with the Late Ce fi- epicontinental seaways previously iscussed [VoL. 70 ———————————————— ÀÁÉÁM ———— 1983] nally, Platnick correlated the separation of Eu- rope and America (45 Ma) with a separation of the nocturna group into European and eastern North American taxa. The area pattern of relationships in the butleri lineage of the leafhopper subgenus Kybos (Fig. 49) is not congruent with the area patterns we have discussed thus far (Fig. 63). The butleri lin- cage indicates that a cosmopolitan population was first divided into an eastern population and à western North American-European popula- tion. eur the western North American and European population was divided. This se- kv of events in the butleri lineage seems to represent a different biogeographic pattern. It is also possible that the cladistic relationships in the butleri lineage have been incorrectly deter- mined. Whatever the case may be the butleri inconsistency has been noted and future workers my objectively determine the truth of the mat- We may now conclude that of the eleven ar- three biogeographic patterns were detected. The Dhyletic events depicted in the patterns were at- si se the same geological events and there- A patterns resemble each other closely. can say that there is a remarkable degree o vnd among the eleven cladograms. This e mt takes on considerable more sig- si when the number of possible patterns s could occur are considered. E we only considered eleven monophyletic ine $ the summary (Figs. 63—64), the largest patterns we could have detected would are considered, there are two possible cow different cladograms and 12 possible EL ee for one of these cladogram types hun be Nelson, 1978). Thus a total of 13 ifferent patterns exist. The fact that in i s logeographic patterns were detected tages, whi ineages indicates that individual lin- Coin are members of larger biotas, have U this i to historical events in a similar way. E true, then there may well be only a very Put eg of cladistic and biogeographic Stony of th of the world’s biota. A great deal of work farth’s prolutiohacy tintitink: If we choose Me ch for these patterns using an objective, € methodology, then this future work will ALLEN —ARTHROPOD DISTRIBUTION 627 lead us to a more complete and accurate under- standing of the evolutionary process. LITERATURE CITED ALLEN, E a 980. A review of the subtribe Myadi: description of a new genus and species, phyloge- pee MEE and biogeography iru rabidae: Pterostichini). Coleopt. Bull. 34: 1- BALL, 1 1975. Nature and formulation of n durum top Syst. Zool. 24: 407-430. 1947. Tertiary centers and migration 48. 1973. Wandering Lands and Ani- LEP “Dutton & Co., N COX C. B. 1974. Vertebrate Rebum pat- terns and continental drift. J. Bio 5-94. Crisci, J. V. & T. F. STuessy. 1980. ae ing primitive character states for phylogenetic recon- struction. Syst. Bot -135. DUNBAR, C. K. Geology GRAHAM, A. (editor). M. "WAAGE. ES Historical uh ohn Wiley & Sons, N : 1972. Fic DVE and Paleoflo- ristics of Asia and Eastern North rica. Amer- ican Elsevier Publishing Co., ew York. Gray, A. Scientific Intelligence. Analogy be- tween the flora of Japan and that of the United States. Amer. J. Sci. Arts, Ser. 2, 2: 135-1 ; 1866. Generelle Morphologies der Ör- smen. 2 vols. Berlin. PAER to the study of osely related to those of Europe or North cian "Pati J. Fac. Sci. Univ. Tokyo III, 6: 29-96. Pon 2. J. Fac. Sci. niv. Tokyo III, 6: 34 HENNIG, W. 1966. "Phylogenetic Systematics. Univ. Illinois Press, Urbana. 19 KENDEIGH, S. C. 61. Animal Ecology. Prentice- Hall, Englewood Cliffs, New Je Li, H. 1952. Floristic relationships between eastern Asia and eastern North America. Trans. Amer hilos. Soc. 42: 378-4 hospes C.H. The Faunal Connections Be- een Europe and North America. John Wiley & hal New York. Love, A. & D. LÖVE. 1963. North ipea te and Their History. Pergamon Press, NELSON, G. & N. PLATNICK. 1981. Sy ante d Biogeography, Arca od — Vicariance. Colum- bia Univ. dier New PLATNICK, N. 1975a. M poc e of the holarctic spider genus c haa s Shere Gnaphosidae). Amer. Mus -32. ; 1975b. A pee = s spider genus Eilica (Araneae, Gnaphosidae). Amer. Mus. Novit. 2 —19. 1976. Drifting spiders or continents? Vicar- iance biogeography of the spider subfamily Lar- ioniinae (Araneae, Gnaphosidae). Syst. Zool. 25: 101-109. G. NELson. 1978. A method of "€ for historical biogeography. Syst. Zool. 27: POPPER, K. R. 1962. Conjectures and Refutations: ic Books, The Gr Yv New York. 628 RAVEN, P. H. & D. I. AXELROD. 1974. Angiosperm biogeography and past continental movements. Ann. gren Bot. Gard. 61: 539-673. A vicariance modd A Caribbean iie Bia MÀ Sys t. Zool. 24: — ——. 1978. nein i ss Sau ex- planations in biogeography. Syst. Zool. 27: 159— Ross, H. H. 1956. Evolution and Classification of the Mountain Caddisflies. Univ. Illinois Press, Ur- bana. ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo.. 70 1963. An evolutionary outline of the leaf- ley Publ. Co., Reading, Massachusetts. Phylogenetics: The Theory and Practice of Phylogenetic Systematics. John Wiley & Sons, New York. — ————— '"!"'"!"'"—-——-———— OUMMMESAOMLLLDLLAGAGIO" eO 0.0. 0(1A(0101)1 À BIOGEOGRAPHY OF OAKS IN THE ARCTO-TERTIARY PROVINCE! DANIEL I. AXELROD? ABSTRACT duoush d Numerous allied trees, shrubs, and herbs inhabit th conifer-hardwood forests of Horas Their close ancestors occupied middle and high | latitudes into the Middle Miocene (15-16 colder and drier climates NEW Antarctic d (13 m.y.a.), numerous paired Or allied species a n the disjunct segregate forests across Holarctica. duri e Tertiary. The rarity of Tertiary oaks linking Eurasia- North Ameri cor jj day) at the north; fully ries of all oak species occur below latitude 30-35°N and y si seven reach latitude 50°N. P Vo iie into 280 to 300 species was chiefly in response t spreading drier and colder climate iis commenced in e middle Eocene, to increasing topographic diversity in the later Cenozoic, and to fluctua nk Quaternary climate. At times of major climatic shifts or increased relief, oaks responded opportunisticalis, probably by quantum steps (punctuated equilibria) that account for the numerous seri most climates from perhumid montane tropical and temperate regions to the goaded a Hopes and in areas of very d Nos have left a rich fossil record in North erica and Eurasia. This is expectable for the numerous species have hard, durable leaves that vio preservation, each large, spreading tree eds thousands of leaves yearly, of which a few (or many) may contribute to an E record, and they occur in many en aen low to middle-high ae. Some species are present today,’ with their Principal areas of concentration in Mexico-Cen- neris and southeastern Asia, and with Ars me of differentiation in the Mediter- Oak sın and southcentral United States. s range from the montane tropics, where ev- tvi Pay are prominent, to near latitude 50°N, wae s are deciduous; Quercus petraea and hf ave the most northerly distributions, ing latitude 60°N in the Baltic region. The 3i oe, tropics, with a twelve-month grow- n, to near the borders of subalpine and real forests, with a growing season of three n terms of precipitation, they occur in ' This pa "upport is gratefully acknowledged. x s nepartm ent of though some bordi ie state that there are 4 believes that 250 is more nearly an faeere figure. ANN. Missouri Bor. GARD. 70: 629-657. 1983. Paper was written during tenure of a grant from N of precipitation. Some oaks are confined to spe- cific edaphic conditions, as serpentine (Q. dura- ta), sand barrens (Q. incana, laevis), limestone rocks (Q. fusiformis), and swamps (Q. palustris). OLDER RECORDS OF FAGACEAE Megafossils that can reasonably be assigned to Quercoideae were in existence in the Cretaceous is implied also by the occurrence of Trigonobala- nus in the montane rain forest (elev. 1, 500 m of Colombia (A. Gentry, letter of April 1982), disjunct from Yunnan, Thailand, and Borneo; connections probably were via Africa across a narrower Atlantic on islands and old pre-Cam- brian highs that have since been removed by ational Science Foundation (DEB 80-25525); this Botany, University of — Davis, eae 95616. 50 oak s es, C. H. Mueller (pers. comm. March, 1983) 630 erosion (Raven & Axelrod, 1974). The megafos- sil and distributional data agree generally with the pollen record (Muller, 1981). Pollen of the subfamilies Castaneoideae (Castanea, Casta- antonian, a)]. The first reliable pol- len records of Quercoideae are Oligocene, prob- ably because the grains are not especially dis- tinctive; certainly, undoubted oak leaves are in much older rocks. Most Quercus species from older Tertiary rocks can scarcely be referred to modern species, al- though they foreshadow them in foliar type. Members of sects. Leucobalanus and Erythro- balanus were present in the Eocene and Paleo- cene, and members of subsect. Prinoideae were also present. Oaks were already diverse in the late Eocene, as shown by nine species in the Flo- rissant flora (MacGinitie, 1969) that reflect ad- aptation to subhumid climate. Fossil leaves probably assignable to subgenus Protobalanus (cf. chrysolepis) are in the Florissant flora but are not now known from older horizons. Diversity of oaks increased in the Oligocene and by the Mio- cene many essentially modern species were pres- ent. Records of their past distribution provide reliable clues to the Tertiary history of the genus. NEOGENE DISTRIBUTION PATTERNS IN HOLARCTICA To evaluate the historical biogeography of oaks in the temperate forests of the Northern Hemi- sphere, it is appropriate to review the distribu- tion of taxa associated with them and that still live with their descendants. Neogene floras of middle and higher latitud de up of species derived chiefly from mixed deciduous hardwood forests that gave way to conifer-hardwood and then to pure subalpine conifer forests in cooler regions. Many of the fossil floras occur now in areas that support sclerophyll woodland, steppe, grassland, desert, taiga, and tundra vegetation that displaced the rich temperate forests as con- ditions became drier and colder following the middle Miocene. The general composition of the Neogene floras reviewed here is in terms of the geographic occurrences of modern taxa allied to fossil species, with special note taken of the pres- ence of oaks in floras that are scattered across Holarctica. It is these occurrences that clarify their history. Numerous fossil species recorded in Neogene floras of the far West are similar to taxa now in the eastern United States (Fig. 1). These species ANNALS OF THE MISSOURI BOTANICAL GARDEN fras, Taxodium, an [VoL. 70 of Alnus, Acer, Carpinus, Carya, Diospyros, Fa- gus, Hamamelis, Ilex, Juglans, Liriodendron, bricaria, muhlenbergii, prinus, velutina), Sassa- ; d Ulmus occur at many localities. They were associated with forest trees and shrubs similar to those now in the West, distributed in Abies, Alnus, Chamaecyparis, Cornus, Fraxinus, Gaultheria, Picea, Populus, Quercus (chrysolepis, kelloggii, gambelii, lobata), Sequoia, Tsuga, Vaccinium, and others. Associates of the Neogene species in the west- ern United States (Fig. 1) also included taxa whose 4 + By in t 111 V phyllum, Exbucklandia, Ginkgo, Hydrangea, Ilex, Lindera, Paulownia, Phoebe, and Zelkova. Fossil oaks in the Neogene of the West pre viously reported to have Asian affinities (e£. Q. myrsinaefolia, Q. stenophylla) need reexamina- tion; most of the records certainly are not of Asian alliance. Furthermore, the absence of oaks in the Paleogene of Alaska, apart from species of subsects. Prinoideae-Diversipiloseae, raises ? problem as to the identity of Paleogene oaks eh Q. nevadensis Lesq.) previously suggested to ha : southeast Asian affinity. Actually, the do of Q. nevadensis shows it is more nearly al d to Lithocarpus densiflorus than to Q. glauca, v which it has been compared; this implies à DI^ able late Cretaceous connection with east i probably via Europe and across à nar lantic. The Paleogene La Porte, Weaverville, : Goshen floras of western California and Oreg have oaks that have been compared wit! ican species, as Q. durangensis, conzattil, noliaefolia, and xalapensis. In uti Europe, there are fossil pee of species similar to those now in the ag ig. istributed in Acer, : Nf (Fig. 3), distribu Juglans 4 mag- Sabal, Sassafras, Taxodium d 4 now associates included taxa allied to those Ja, Cot- western Europe, distributed in Acer, s on nus, Larix, Pinus, Populus, Quer oa Mediter- pubescens, robur), or in the extended | dont, notably Abies, Celtis, nus, Platanus, Ostrya, Quercus (a cerris, coccifera, ilex. libani, mirbeckii, suber), and Zelkova. ee 1983] AXELROD—OAKS 631 Populus verophylla FIGURE 1. Many taxa now in eastern United States are represented by closely similar species in Neogene floras of the Far West. iss j* Gin Paulownia = V r pictum $ Cep e essai drame A Glyptostrobug Metasequoial_« n e Pseudolarix E burnem lanta (fel Se on E= } Rob a \pse foecacid 7° 9 deum: Ficure 3, Many species now in eastern United S imilar sf 632 Some fossil oaks in western Europe have been compared with species now in the eastern United States (e.g., phellos, virginiana) or Mexico (xala- pensis). These need restudy before they can be accepted as authentic records; judging from the illustrations, they can be matched by Eurasian In western Europe there are also Neogene species allied to taxa now in eastern Asia (Fig. 4), including species of Acer, Betula, Carpinus, Cathaya, Cephalotaxus, Cinnamomum, Eu- commia, Fagus, Ginkgo, Glyptostrobus, Gordo- nia, Juglans, Larix, Liquidambar, Quercus (cf. serrata, variabilis), Sapindus, Tilia, Torreya, : Ikova. They were associated with fossil species related to those now in western Europe and eastern North America (Fig. 3). There also are Neogene fossils in Japan that are allied to taxa now in the eastern United States (Fig. 5). These include species of Acer, Betula, Carya, Carpinus, Cercis, Comptonia, Fagus, Ju- glans, Liriodendron, Nyssa, Platanus, Populus, Robinia, Sassafras, and Taxodium. Some of the disjunct taxa are quite similar, as in Acer sect. Rubra (Fig. 6), having been derived from a com- mon, widespread northern species [A. trilobatum (Sternb.) Al. Braun]. The associates of these species in eastern Asia were primarily taxa re- lated to those now confined to that general area. Among the species of Quercus recorded in China- Korea-Japan are taxa allied to Q. aliena, acutis- sima, dentata, gilva, glauca, mongolica, myrsi- naefolia, phillyraeoides, serrata, and variabilis. hey are especially abundant in the Miocene flo- ras of the region but rare to absent in older floras. Some reports of otherwise wholly temperate American taxa in Japan probably are not valid. Acer sect. Saccharina, recorded as fossil from Washington to Nevada, is reported from samaras in Japan (i.e., A. yoshiokensis, Tanai & Suzuki, 1960), but they are quite similar to those figured as A. pseudogianella; the leaves may represent a species of Acer sect. Campestre. Acer sect. Mac- rophylla, represented only by A. macrophyllum of the Pacific States, has numerous fossil records in Washington, Idaho, Oregon, and Nevada. et al., 1974; Ablaev & Gorovoi, 1974) more probably represent sect. Cerris (cf. Q. cerris L., ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 in Camus, 1936-38, pl. 65—66) as judged from the nature of the leaf lobes. A reported possible red oak (subsect. Rubreae) from Primorye (Al- baev & Gorovoi, 1974) appears to be a terminal leaflet of Acer (cf. negundo). To complete the picture, it is recalled that a number of species in the western Miocene floras from the latitude of central California-Nevada southward have their nearest allies in the sub- humid woodlands that dominated the Madrean province of the southwestern United States and and Sapindus. Many oaks are recorded, includ- ing those allied to Q. arizonica, brandegeei, du- mosa, emoryi, mohriana, and oblongifolia. The Paleogene floras in the interior include taxa sim- ilar to those in warmer, subhumid climates of Mexico and the Southwest. Xeric oaks in the Florissant flora (35 Ma) suggest alliance with Q. brandegeei, chrysolepis, dumosa, incarnata, lobata, and undulata. They formed woodlan bordering temperate mixed deciduous forest that covered moister valleys and slopes that included ies allied to tł in east nd western = i ALU ƏC LLU VV 444 ir E ‘ ie, North America and eastern Asia (MacGinit allied largely to species contributing to vegeta- tion now in west-central Mexico (Jalisco), pt oaks comparable to benthamii and pos and two others of uncertain modern afinity MacGinitie, 1969). As indicated in Figures 1-5, some forest e that occurred widely across Holarctica are å to species now in relict or restricted areas, S e as Ginkgo (Fig. 7), Pseudotsuga (Fig. 8). e (Fig. 9), Liriodendron (Fig. 10). Juglans cine (Fig. 11), idi (Fig 13), Zelkova (Fig. 14), Eucommia (Fig. 15) and Ailanthus (Fig. 16) (for others, see 1963; Walter/Straka, 1970; Meusel et al., 1968). : aes distributions contrast with p currences of widely distributed allied epu i (Fig. 19), Acer (Fig. 20), bag (Fig. 21), | A secti many others. Some modern — 1 peri or subsections) now have more restri the com butions, such as discontinuous between Pop- tinents for Acer sect. Macrantha (Fig pe ulus sect. Aeigeros (Fig. 23), Acer sect. ntrast. (Fig. 24), or Acer sect. Rubra (Fig. 6). bye Acer others are discontinuous on a continent, as pri sect. Saccharina (Fig. 25). Some sections a 1983] AXELROD—OAKS 633 cer \ palmatum Z Actinidia Wisteria FiGURE 4, Many taxa now in the Far East occur in fossil floras of Europe and border areas. tonia Acer saccharihum j Ficure 5, A number of species now in eastern United States have close counterparts in Neogene floras of pycnanthum | à Ficure 6. Disjunct maples of Acer sect. Rubra ser. Rubra imply a wider — for the ancestral taxon, | confirmed by the wide distribution of red maple in Eocene and later floras } 634 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 FIGURE 7. Ginkgo, now confined to eastcentral China (+), had a wide Holarctic distribution from the | Paleocene into the Neogene as indicated by some of its fossil localities (8). | tands (+) in FIGURE 8. Pseudotsuga is widespread today in western North America but has only relictual s eastern Asia. Fossil localities (6) indicate its wider distribution in the past ; . a FIGURE 9. Sequoia, confined now to coastal tad aà Holt Te | distributis as send y by some of i ^ fossil ia Cai SEA AMG, Mdiacent Oregon, 1983] AXELROD—OAKS 635 t GURE 10. Liriodendron is disjunct between the ice United States and China. Some of its fossil localities | Py ok here attest to its former Holarctic occurren ode been recovered at a number of Neogene Sites ed! The distinctive fruits of butternut, Juglans cinerea, ) far removed from its present area in the eastern United Sta | D Cercidiphyllum, confined now to central China (+) and Japan, has been recorded at numerous Holarcti (. ica, some of which are indicated (8). | 636 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vou. 70 Ficure 13. Pterocarya species are confined now to eastern Black Sea region (+), the south shore of the Capa: Sal (+), and to central China and Japan. The fossil species (@) had a wide distribution. w t0 FIGURE 14. Species of Zelkova had a wide distribution into the later Neer (@) but are restricted n0 the eastern Mediterranean-Caspian Sea region (+) and to China, Japan- FiGURE 15. Neogene fossil localities (8) for Eucommia, a relictual genus confined now to central - | 1983] AXELROD—OAKS 637 — f Populus sect. Leuce, most of which are represented by similar FIGURE 17. Distribution of modern species o e flora d America. The fossil record suggests that these diverged from an eM ras of Asia and No | taxon in the Eocene, or possibly earlier. mle C 18. ape josip of io of Populus sect. Tacamahaca. The American species are repre- nee "train fossils o n western North America. Asian poplars of this group are not now | ocu from North jm ` 638 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vot. 70 FiGuRE 19. Modern species of Betula sect. Cailte are disjunct in Holarctica. Asian species are definitely known from western United States. The numerous species in eastern Asia imply evo general area. dere in pe FIGURE 20. Taxa of Acer sect. Negundo are now runs gauger but formerly lived at num the region. Species of the alliance are recorded from me Eocen densiflora KJ thunber giana IO VIS (gore aiwan bis is tropicalis erous sites across et FIGURE 21. The wide distribution of dosi E Pinus subsect. Sylvestres implies an ancient disP* ancestral group, probably in the Late C hh AO: IG OS 1983] AXELROD—OAKS 639 / ^ ^ S máxfmowicz fenisy vanicum. | Gf laxiflorum dii FIGURE 22. Species of Acer sect. Macrantha ser. Macrantha are concentrated in eastern Asia, with a single species in eastern North America. East Asian maples of this alliance are not presently known from western America. North fremontii T - «5. * de tol dés * FIGURE 23. Species of Populus sect. Aigeros are disjunct across the Atlantic. Taxa of both alliances have e recorded as fossil in their respective regions and some have ranged more widely, as fossil species allied to mexicana. Trans-Atlantic connections are implied. circinatum trated in eastern Asia, with one along the Nevada. 4. Species of Acer sect. Palmata ser. Palmata are concen of North America. A species of this series occurs in the Eocene of 640 dtum floridum `| leucoderme ni ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vou. 70 FIGURE 25. Acer sect. Saccharina is represented by disjunct taxa now in the Cordilleran region an aa North America. Ng species allied to saccharum and grandidentatum had a wider distribution in the far into the Miocen sections are localized, as Pinus subsect. Australes (Fig. 26), Acer sect. Campestris (Fig. 27), Pinus subsect. Ponderosae (Fig. 28), as well as others such as Pinus subsect. Cembroides and Acer sect. Glabra of western North America. The fossil rec- ord shows that these (and other) alliances are represented by similar taxa in the same general region as their present areas: they have not ranged widely and seem to be in regions that represent their centers of origin. DISCUSSION The allied taxa in these disjunct areas of Ho- larctica represent the derived, broken links that occidental FIGUR E26. Mem be Is isi Sans fossil pines in western Europe resemble this gro rs of Pinus subsect. Australes are ed to eastern United States and United States, the eastern United States, western Europe, and eastern Asia, and they are not re- corded far from their present areas: this hee pecially true of the oaks (see below). A St statement applies to the Madrean and oe I terfingered with temperate forests t and subtropical ones to the so these regions where ancestral taxa widely, grasslands and deserts have art uth. In each of ranged more | isen more | the Caribbes* 1983] AXELROD—OAKS 641 V S atum Án onos pessu lanum FIGURE 27. Members of Acer sect. Campestris are confined chiefly to the Mediterranean basin apart from A, Fh iie which ranges more widely. Fossil species of this alliance are common in the area of their present Tecentlv + > . * * ently ir former distribut these disjunct areas have allied taxa in numero middle Ma did colder taiga and tundra cli- genera (see Meusel et al., 1965, 1968), acioni mates have disrupted their occurrence at higher sensitive mesic forest herbs that still link the latitudes. Their past continuity is demonstrated widely separated relict forests that were isolated not only by fossil records scattered across Ho- — by spreading drought and cold over middle an larctica, it has been inferred also from the nature higher latitudes following 13 Ma (Axelrod & Ra- of the taxa in the surviving relict forests in the ven, 1972: 227; Axelrod, 1979). oa region (Balkans-eastern Turkey-Cauca- The associated herbaceous understory of the , the Elbruz ruz Mountains, the Himalayan axis, mixed forest is largely unknown in the fossil rec- FS of east-central China, Japan,andthose ord, apart from a few seed floras in western Eu- * eastern and western United States. All rope, Siberia, and Japan. The former do contain 5 Mo ede : ; lawsonii Fund Pca eae a ai qué States by thei “center of origin.” eee pines allied to the group are oder perles in the wes commencing in the Oligocen 642 taxa now in eastern Asia and eastern North America (reviewed in Leopold, 1967, 1969). The eanarata related Y MAR 9vpal atv continents are in Actaea, Anemone, Asarum, ircaea, Dentaria, Erythronium, Gali- Waldsteinia, and many others (see Kornas, 1972; Meusel et al., 1965, 1968). As might be expected, there are more bi-continental species between Europe and Asia than North America (Meusel, 1968, 1969; Meusel & Schubert, 1971). The European species are especially rich in the Colchic forests, Asia that link the European and east Asiatic cen- ters. The present arid central Asian region sup- ported broadl ] mixed decid forests into the middle Miocene (reviewed in Axelrod, 1979). Stebbins noted that there are similar herbaceous groups confined to eastern Asia and eastern America, not only in families regarded as gen- erally “primitive” (Berberidaceae, Liliaceae, Ra- species, as in Antenoron (= Tovara), Brachyely- trum, Cypripedium, and Polygonum, are scarcely different and at most are only varietally or sub- specifically distinct; these herbs are not only an- cient, they have changed but little over the past 15 million years. Clearly, “the evolution of her- baceous elements (in the mixed deciduous for- Sts) has t s slow as that of the woody species” (Stebbins, 1947: 152). Comparison of species in the derived (segre- gated), temperate mixed-deciduous forests of North America, Europe, and Eastern Asia shows that many of the allied taxa (paired species, allied sections, series) have similar ecologic occur- ilar phyto-sociological relationships imply that their ecologic tolerances have not changed great- ly since the middle Miocene (15 Ma), a time when continuous mixed deciduous and conifer- hardwood forests still linked Asia-North Amer- ica (via Beringia) and Europe-North America (via Iceland-Greenland). Clearly, the North Ameri- can and Eurasian paired-taxa are descendants of common ancestral forms that contributed to the Arcto-Tertiary forests (Medwecka-Kornas, 1961; Kornas, 1972). ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 BEARING ON ARCTO-TERTIARY CONCEPT The preceding data, as well as those presented in Figures 1—28, certainly demonstrate the for- mer continuity of a rich temperate forest flora across northern to central Holarctica. It was first termed an “Arcto-Tertiary element" and then formalized into the Arcto-Tertiary Geoflora (re- viewed in Wolfe, 1977). Recorded in the Paleo- cene to Eocene of Spitzbergen, Greenland, Si- beria, Ellsemereland, and northern Canada, by the middle and late Eocene temperate forest woody taxa ancestral to modern species had al- ready occupied mountains in the central Cor- dilleran region (Challi 1Sal floras, Idaho; Yellowstone flora, Montana; Bull Run, Copper asin, Elko floras, Nevada). As climate progres- sively cooled following the late Eocene, evolving taxa in the temperate forest flora shifted into lowland basins, first in the interior and then the coastal strip. By the Miocene many woody taxa were scarcely separable from living species. Reports of the demise of the Arcto- Tonia Geoflora *have been greatly exaggerated. The basis for that belief stems partly from the de- scription of a late Middle and Uppe E tropical flora from coastal Alaska (W olfe, ait 1977), which was then at a latitude near d (Smith et al., 1981). It has been evident j for a decade (unpubl. manuscript) that the fo : which occurs in the marine Kushtaka der is allochthonous. This agrees with the fact Ri later Eocene floras from middle and high !a : um allochthonous (Grantz et al., 982: Hamilton, 1983). An al origin also disposes of the geophysicé wini that an Eocene tropical flora and ¢ maw =. stable, coastal Alaska presents, notably att which “the axis of rotation of the eart rae had a much less— possibly no—inclin (Wolfe, 1977: 37). As for the source of the Arct flora, it is recalled that much of the e woody forest flora has its roots 1n punc subtropical alliances (Axelrod, 1952; the east this symposium). It is probable Maece perit Arcto-Tertiary and late Cretaceous o- Testis M m ‘cal and Wu, 1983] forest flora of conifers and deciduous hardwoods originated over middle (temperate) latitudes, es- pecially in montane areas, and thence sprea northward to high latitudes in response to cooler climate. Documentation of such a history will depend on a closer dating of presently known floras as well as finds of new ones. Certainly the temperate aspect of the early Paleocene Fort Union flora (Brown, 1962) from Montana-Wy- oming-Dakota and northward, with its temper- lerpretation, as is the later occurrence of more modern temperate dicots at higher latitudes. It I5 Supported also by the more archaic character of taxa in the high arctic late Cretaceous-Paleo- cene floras, paralleling the persistence of a flora of"Late Jurassic-Aptian” aspect at high latitudes when early angiosperms were already in the low- lands at lower latitudes and spreading polewards (Axelrod, 1959), The preceding distributional and phyto-socio- logical evidence show that in spite of slight to moderate morphological differences, the ecologic tolerances of scores of temperate taxa have not changed greatly. This statement conflicts with the notion (Mason, 1947; Wolfe, 1969: 197, 1972: 230, 1977: 47) that numerous Tertiary forest taxa have changed their physiology and have shifted ine that they were part of a mixed mesophytic ni d members of the subhumid Madrean Which h yll vegetation of the southern interior to ia € believes they adapted later in the Mio- ©. Actually, nearly all of these examples have the ed in older floras of drier aspect in MR uthern interior. Their occurrence at the ofMi Probably was favored by the construction the onm volcanos along the Cascade axis, with lie »w and breccias providing well-drained, valley f o slopes adjacent to the rich mesic : Orests. Significantly, these taxa are essen- Y absent from large Miocene floras of the Y region (Mascall, Blue Mountains, Spo- AXELROD—OAKS 643 kane, Grand Coulee, 49 Camp, Succor Creek) that bordered local ponds and lakes formed b damming by the Columbia River plateau basalts in a region of generally low reliefand high water- table. The suggestion that the Madrean taxa from the Miocene of northwest Oregon had only a tem- porary, seral relation on drier volcanic slopes in that area is paralleled by distributions that re- sulted from the late Quaternary climatic fluctua- tions. Some Californian taxa that range far north inhabit the warmest, driest available sites that were invaded only since the last glacial. Arbutus menziesii occurs on dark (heat-retaining) meta- volcanic rocks on Vancouver Island, rocks that were covered by ice only 12,000 years ago. Quer- cus garryana is confined to drier slopes in west- central Oregon and southern Washington, an area it invaded in post-glacial time (Hansen, 1942, 1947). The oak groves in the Puget Sound area are being actively invaded by Pseudotsuga forest and have only a seral relation (Franklin ness, 1973: 89). The same relationship is im- plied by Vaccinium ovatum, a common member of the closed-cone pine forest of California that ranges northward into Washington, and was cit- ed by Mason (1947) as an example of floristic instability. Its nearest ally, V. confertum, occurs in the closed-cone pine forests (Pinus sect. Oocarpeae) of Mexico where there are many re- lated species that also make up the Californian coastal pine forests (Axelrod, 1977: 171, 1980). Ecologic studies (Franklin & Dryness, 1973: 85-86) show that V. ovatum borders the coastal forests to the north and has only a seral relation, occupying open, drier sites that presumably are relict from the warmer Xerothermic and more recently from logging and fire. The occurrence of these and many other typically Californian taxa far to the north has been discussed by Detling (1953, 1958, 1961) for Oregon and by Schofield (1969) who reviewed the floristic makeup of the «Mediterranean” climatic portion of southern Vancouver Island and southwestern British Co- lumbia. As they emphasize, the “Californian” species are not an integral part of the climax conifer forests but occur in open, sunnier, drier areas for the most part, and entered this region only since the last glacial, most probably during the Xerothermic (also in Axelrod, 1966: 42- 55). It thus seems unlikely that forest taxa have greatly changed their physiologic requirements, shifting to wholly new associations. Furt er- more, since rapid, severe environmental shifts pt 644 PRINOIDEAE 2h $ád ler idná-é. ontana muh lenber gi ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 E29. Oaks of the allied subsects. Prinoideae and Diversipiloseae are confined now to spen qM den. Fossil species of these alliances were widespread across middle and high latitudes into the Mioce like those of the Quaternary are not documented for the Tertiary, there probably was little inter- change between North America and Eurasia of taxa that represent sections of genera that orig- inated in restricted parts of one continent or the other. This is not to deny that taxa may have shifted to new associations, but these were between the ` same general forest types, as in the shift ofa taxon from the American to the Asian sector of the mixed deciduous or conifer-hardwood forest. Examples of such shifts are not readily demon- strable today because the record is still too in- completely known to determine whether à fossil species in western North America moved to east- ern Asia or. vice versa. Ar ancestral speciei. ed to the impression that the «primitive" nem d grated to give rise to the “derived” taxon. pm ly, the notion that taxa have commonly move into associations widely different from d the ancestral community is not demonstrat of Some taxa have been restricted upward, out FIGURE 30. ier te of species of Erythrobalanus subsect. Rubrae. Taxa similar to coccinea cinea and Pi hav subsects. Laeves. occur in th iocene of Oregon. Oaks of this group Palustres, Pei heeses Pagoda: and Californicae. e also been divided into s 1983] AXELROD—OAKS i FIGURE 31. Lobed white oaks have been divided into subsects. Lobateae, Gambelieae, Albeae. Those of the ormer two occur as fossils in the far West during the Miocene. Members of Albeae have not yet been certainly recorded there. associations now too xeric for them, and others have moved to lower elevations where climates are warmer, but in such cases they are still com- ponents of associations in which they were pres- ent formerly. That is, upland taxa forest earlier reached to lower levels and sclerophyll taxa to higher ones under a climate of ample summer ran and milder winters (Axelrod, 1976, fig. 4). P AST AND PRESENT DISTRIBUTION OF OAKS a preceding data provide a basis for inter- ng the Tertiary history of oaks in the Arcto- F ; IGURE 32. Oaks of sect. Cerris represent two subsections that imply former trans- ive areas during the N Tertiary province. Figures 29 to 37 illustrate the present concentrations of some representative species of several Quercus sections (from Tre- lease, 1924; Camus, 1936-54) and Tables 1 to 3 indicate the present distribution of oaks in the temperate regions chiefly. It is apparent that whereas species of many sections or subsections of dicots have ranged widely across Holarctica (see above, Figs. 1-16), only members of the allied subsects. Prinoideae and Diversipiloseae occurred throughout temperate Holarctica dur- ing the Tertiary and earlier, * in Spitzbergen, Iceland, Alaska, Siberia, and also SS PA N Kl L illiana E o o OF sénescen aquifoloides imotricha Eurasian distribution. F . *ssils of each subsection have been recorded in their respective 646 ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 X laurifolia > * pumila incana FIGURE 33. Taxa of subsect. Phellos, confined now to the southeastern United States, were represented by Bass allied to phellos and imbricaria in the Miocene of Oregon FiGURE 35. Taxa of Quercus ense Ped lat ded as fossil in the Neogene of western Three endemics are in the Caucasus M xd : —— — a Md o — ——À—ÀÁ | 1983] S. tip Tygda Sakhalin Be: Teen, eningrad 60°. robur petraea mongolica - | TE E crispula Papert hartwissia "1 | dentata 1 0°. macratthera | castanae folia bodas FiGURE 36. Northern limit of Eurasian oaks today. at middle latitudes in Eurasia and western North America. The presence of these allied subsec- tions (or series) on separate land masses today and the records of similar fossil species in the far north in the early Tertiary implies a southward shift in response to cooling climate and the re- sultant divergence of two closely allied groups. The alliance was already at middle latitudes by the Paleocene (Brown, 1962, pl. 19) and a fossil species of Prinoideae is in the Bull Run flora (40 m.y.a.) of northern Nevada, where it contributed to conifer-deciduous hardwood forest at an ele- vation close to pure subalpine conifer forest. pecies of most sections and subsections of Quercus are restricted to one continent or to spe- cific parts of it, paralleling those discussed for other dicots (see above). Their close fossil allies ' more widely in the Neogene but most did not range far from their present areas, as Judged from the record in Europe, Asia Minor, m China, and western United States (Figs. Ww. ). Examples of wider distribution in the far iin. are provided by the records of red oaks (aff. um falcata, velutina) and chestnut oaks (aff. or, prinus, muhlenbergii) in Washington, a. and Idaho. Species of sect. Cerris occur Y across southern Eurasia and the species ast and west PP quw" me leer dig, implying an ancestral alliance adapted to Pr nt climates differentiated in them as con- k . changed. Most subsections in the genus -— récognized) display only local occur- nii shown by subsects. Robur (= Pedun- Others qs 35), Phellos (Fig. 33) and many eir taxa have not ranged far from their — areas, which have been restricted by Preading dry climate that compelled forest and Cali fornia Province are distantly related to those AXELROD—OAKS f Atlantic r egion, and the black oaks of the itis B Lake Columbia Winnipeg Gaspe 60? macrocarpa garryana alba 50 rubra ellipsoidalis muhleub. | prinus velut. 40° | | T fer | bicolor , palustris FicunE 37. Northern limit of American oaks to- day. of eastern North America and are remote from those of Mexico-Central America. The relations imply that these oaks originated in relatively lo- cal provinces. Fossil evidence for restriction of species of most subsections (series) to comparatively local re- gions and with little or no continental inter- change during the middle and later Tertiary is consistent with systematic evidence (Tables 1— 3). Furthermore, the restriction of most sections to their present continental areas, or to parts of them, is clearly shown by illustrations of the nu- merous fossil oaks from across the USSR which are of Eurasian affinity (Takhtajan, 1982). As Trelease noted (1924), “the existing oaks of tem- perate regions have developed since the breaking of the circumpolar land connection of Tertiary time: that the genus may have penetrated Europe by way of the Orient in Cerris and Ilex types; and that it may have reached America originally over one or more now vanished land connections Cretaceous links are implied by the occurrence of Lithocarpus and Castanopsis (= Chrysolepis) in California, which are allied to Asian taxa in warm temperate climates. Furthermore, Litho- the late Early Eocene Chalk Bluffs flora. Together with the disjunct occurrence of Trigonobalanus (Y unnan-Thailand-Borneo to Colombia), con- i i tied across the Atlantic (1) when it was narrower than at present and prior to the counter-clockwise rotation of North America, and (2) when more numerous, now-sunken islands linked these regions (Axelrod, 1972, fig. 25, 1979; Raven & Axelrod, 1974). NORTHERN LIMIT OF QUERCUS Available evidence indicates that most oak species have not followed many widespread taxa TABLE 1. Distribution of sections and subsections of Quercus in Holarctica (from Camus, 1936-54). Asia Himalayas Mediter Minor (to Afghan- S. China NE Asia- Europe Basin Persia istan) (Yunnan) C. China Japan Subgenus Cyclobalanopsis x x x x (To Malaysia-Borneo) Subgenus Euquercus Sect. Cerris Subsects. Brachylepides (15 sp.) x x x x (To upper Burma) Echinolepides (3 sp.) X x x Cypriotes (1 sp.) (Cyprus) Cocciferae (4 sp.) x x x Macrolepides (9) x x ee (2) x x X rris (1) x x Sect. Mesos Subse: add (1) x Macrantherae (3) x x x x (excl bog Sect. a Ser. 1. Ger aa Es undae (6) x x x (To Annam) Pachyphyllae (5) x x (To Assam, Siam) Englerianae (1) x Ilex x x x Diversipilosae (9) X x x x Galliferae (5) X x Hartwissianae (1) x Sessiliflorae (6) x x x Lanuginosae (2) x x = Ped latae (7) x x x 8v9 N3G3ÀV5 'TIVOINV.LOS8 PHIOOSSIN JHL JO STVNNV 0L 10A] 1983] AXELROD—OAKS 649 TABLE 2. Distribution of sections and subsections of Quercus in North America (from Camus, 1936-54). Baj Calif. Calif. Ariz.- Texas- Colo. E US- Central Southern US (Canada) Sect. Lepidobalanus Ser. 2. American (relict) (To E. Mexico) ~ e eK x 9» Ke mK OK >< í * Excluding subsects. in Mexico—Central America; also sect. Macrobalanus, first ten subsects. are Madrean in ngin, ^ terms of a broad Holarctic occurrence during the Tertiary; most species and subsections (or Hee iwi had relatively restricted distribu- * Fossil oak remains are not common in ‘tary floras from higher latitudes, as Green- ae Spitzbergen, Ellesmereland, Alaska, or alin. Those that have been illustrated readi- l ^ Y fall into subsects. Prinoideae and Diversipi- aliena, muhlenbergii, mongolica, and their allies. e Vcn the problem as to why other fossil peni T€ rare (or absent) in floras from latitudes Tally above latitude 50°N. Oaks now reach "uy northern limit near latitude 50*N (Figs. 36, 37) and on] ature bes = high latitudes in the Tertiary. Those 16) iui zm rted rich deciduous forests (Figs. 1- Wed. e Middle Neogene (15-16 Ma), yet the N0ideae loc gai de there are those of Pri- latitudes Wversipiloseae. Most oaks occupy Centrati well below 40°N and the highest con- Ons are in latitudes 15-30°N, in Mexico- a pe (with some 150 species), and May be ia China. Judging from Figure 38, light limited e principal factor that limits, and has » the northern distribution of most oaks. In this regard, it seems significant that of the 18 species that now reach latitude 45-50° (Figs. 36, 37), five are members of the Prinoideae-Diver- sipiloseae, which had a wider distribution in the Pale d later hing al latitude 60°N into the middle Miocene. EVOLUTION Oaks comprise one of the more active groups of woody plants in terms of evolution during the Tertiary. There are now about 280 to 300 species and scores of hybrids are recognized. The great plasticity of the group is apparent in the con- vergence in similar leaf form and acorn type among species belonging to different sections of the genus. Oaks proliferated in generally restrict- ed regions, chiefly in response to increasing to- pographic, climatic and ed hic di ity during the middle and later Tertiary. In addition, there were certain times when climatic extremes in- creased sharply, as at 35, 13, and 5 Ma (refs. in Axelrod, 1981; Kennett, 1981), as well as dur- ing Quaternary fluctuations. These major cli- matic shifts may have corresponded with times of especially rapid speciation in oaks (and other taxa) as older alliances disappeared and more open communities became available. An example of such a radiation 1s suggested by the history of Quercus wislizenii and its close uiua k TABLE 3. Distribution of sections and subsections of Quercus in North America (from Camus, 1936-54). 0s9 . Baja Southern Central Eastern NE US California Arizona Calif. N. Mexico US US US & Canada Sect. Protobalanus (6) x x x Sect. Erythrobalanus* Subsects. Durifoliae (1) x x Marylandicae (3) X eK »* * tal * Pad Laeves (1) Pagodaefoliae (2) Ilicifoliae (1) et ia) im = E © ® i -— — » rd x eK x x x x x Californicae (1) x (to S. Oregon & Baja Calif.) NAUVO 'TVOINV.LOS INNOSSIN JHL AO STVNNV * Excluded are all subsects. in subtropical Mexico—Central America. 0L 10A] 1983] AXELROD—OAKS 651 24 > 40° North Latitude Hrs North lat. ) NP 663° ~ North America 1 II 65 x : 52 x / N E g Eureka " E o 3 Ez] u 3 ps Pam 60 g 2 Jw Li E z et Indianapolis z E g E- 7 pgi 2h | S Ns we Philadelphia 8 j 40 Ñ ind 5 pu 30* INN E ^ — 807-20- | Europe S et ges y SEALL Lisbon É qor SEN E N Ir Tip of Italy E 3 S h 3 NN Asia 3 5  a $8 E A tig ai Ankara à N w ~ Baku Peking a b A oM! o s 9g TEES LIAE FIGURE 3R M nct 1 A 1 d rd 1 Pee | g fig ^ a tentation of oaks in the indicated latitudinal belts. (Figure from Strahler & Strahler, 197 oY. Wiley & Sons.) 7, with permission Pie Q. shrevei, parvula, and frutescens (Fig. 18.5 million years. Quercus wislizenoides-wisli- ). In California and Nevada there are large il ified d oal Suites of fossil leaves oftwo members ofthe com- nak zenii have typ grassland wood land vegetation since the later Miocene, shifting from Nevada into California as conditions be- ‘are similar to those produced by the living came drier and cooler to the east as the Sierran | block was elevated. Fossil records of the shrubs are not now known, but their ecologic relations suggest that parvula is older than frutescens. As for their origin, oaks of the wis/izenii alli- ance are related to the series Peninsulares Tre- lease (Q. peninsularis, Q. devia) of Baja Califor- nia. Peninsulares are allied to Durifoliae ( Q. emoryi, durifolia) of the southwestern United fà . | Q Amda (Remington Hill flora) and is now (as l Clip in the outer south Coast Ranges of E^ As climate became drier, wislize- P s replaced shrevoides in western Nevada States and northwestern Mexico, and to the Vi- kee fl Jand then in the lower Sier- mineae (Q. bolanyosensis, duranzillo, balsequil- fà ) i : | a Nevada and the inner Coast Ranges. The shifts Jana, viminea) of the northern Sierra ne daptive cidental. The Peninsulares, Durifoliae, and vaya changes in the upper Miocene (13-12 Vimineae represent taxa adapted to subhumid has and in the latest Miocene (5-6 Ma). There to semiarid areas with summer rain; meae Q. | clearly been a coastward restriction from the — shrevoides and wislizenoides, which lived earlier | Q sh Nevada desert of the presumed ancestral under biseasonal rainfall, have adapted to aa tome oides-shrevei line which has contributed dominant winter-rain climate of Messing = | mesic broadleaved sclerophyll forest for fully ninsulares were restricted to their present area d | | (Verdi, Truc type seem related to major environ- ! 652 ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 COASTAL I-A I-E AK 15E Closed-cone Sclerophyll Oak- pine Chaparral pine forest savanna shrub tree tree shrub parvula shrevei wis lizenii fru tescens 0 f S ge à rap Eu e ur FLORAS F i Napa E ; : Broken Hill, Turlock L. n Pj electo prre Xe j j Mulholland, Oakdale Verdi | of Neogene Cf JS Pi Mt. Eden ba T M/ Piru Gorge, t yess S E Anaverde & Y Y T o3 $y Q > uj s fi Remington Hill, f 10 L p OTY L E Lo qw o p 6 i ie - Forests retreat; E41 Fallon, Stewart Spr. a | exotics reduced f bc elerobdgus Ho b i x - t : 15 te : I E : T Fingerrock k 1 i Middlegate Eastgate E 3 ] OO OOO O l 20 d : FIGURE 39. Inferred relations of oaks of the Quercus wislizenii complex (see text). as the lowlands became drier; they live ina young mountainous region, one elevated only in the last two to three million years (Gastil et al., 1975). The Durifoliae ranged more widely in the past, as shown by records of fossil oaks allied to Q. emoryi in the Tehachapi (17 Ma) and Mint Can-. yon (12 Ma) floras, southern California. The pre- ceding data imply that these series probably are ancient and survived in favorable environments as drier climates spread more widely in the late Cenozoic. To judge from the modern aspect of oaks of subhumid requirements in the Florissant’ flora (34 Ma), the wislizenii complex and its al- lied series may have diverged from an ancestral alliance in Oligocene time. The n "3 HS 1 r j ions of sect. Cerris and Suber across Eurasia and of the A/beae (incl. Lobateae) and Rubreae (incl. Coccineae) across hd Ga S 1 AI = z £0.01 = Iollowing the spread of drier climate and of a shift in sea- far West. Marked sonal rainfall distribution in the a Asi may have been the time of origin of CDM major subsections (or series) of the Mv in the climates commenced to spread rap and differentiated east and west pet sections (series) and many species ompanyiné of new upland terrains and their acc local climatic differences may We tive tad a basis for the origin of numerous as founder populations exploited ™ » vironments. With rapid shifts in € mu have stability in populations, the effect thus si- been the rapid origin of Rer taxa, ©. "T ing the phenomena of punct "n dd phone Cordilleran region of the | —_— — 00 TT 1983] States, sequences of floras from the middle Eocene to the early Miocene (Green River, Florissant, Beaverhead, Ruby, Creede) show that numerous taxa (Le., Cercocarpus, Colubrina, Mahonia Peraphyllum, Philadelphus, Rhus, Sapindus many small legumes) that imply drier climates were spreading over the area. A small-leaved, lobed Eocene white oak (Q. lvratiformis Cock- erell) in the Florissant flora (35 Ma) is the only record of the group in older floras (Fig. 40). It may be suggested that it spread east and west nto moister terrains, developing into larger- leaved species in eastern (alba, macrocarpa, ly- rata) and western (lobata, gambelii, garryana) areas, The problem of the origin and relationships of the lobed red (black) oaks, subsects. Coccineae- Rubreae (incl. Q. kelloggii), is not readily solved y. The oldest black oak is the early Eocene Chalk Bluffs Q. eoxalapensis MacGinitie (1941), but it appears quite unlike an ancestral type for Coccineae-Rubreae. A ~ - died States and to Q. kelloggii of the "Rs. x moderate differences east and west may 9 more than foliar plasticity in the group. Stink: SO evident that in some floras (Mascall, ue ng Water) there appears to have been hy- zation among the lobed oaks. RADIATION Some Natural subsections have radiated into 'verse environments. For instance, species a Chrysolepideae occur in several vegeta- ones. Quercus tomentella is a woodland Pie ene to islands off southern California ne aang Island; it occurs with closed- Islan : Orest on Santa Cruz and Guadalupe ' &uercus cedrosensis is a shrub in coastal n Of sec ev “green forest (sclerophyll forest) of California AXELROD—OAKS 653 and enters the lower part of the adjacent mixed the upper mixed co- nifer forest into fir-subalpine forest and locally reaches close to tree-line. These taxa occur in areas with precipitation as low as 300 mm (12 in.) and as high as 2,500 mm (100 in.) and the montane sites have heavy snow. The growing season may last fully seven months (Ensenada) or be as brief as two to two and a half months (above Twin Lakes, California). In sect. Cerris subsect. Brachylepides, which is largely in the area from central China to Yun- nan, Q. semicarpifolia, aquifoloides, pannosa, and gilliana range up to 3,000 to 4,000 m in Yunnan. Since the age of the mountains is established, the rate of adaptation of these taxa to high montane conditions, and their derivation from ancestral species, can be estimated. Clearly, field studies of radiation of natural cohorts in Quercus can be quite rewarding. DOMINANCE Evidence in northeastern Asia suggests that although Q. mongolica was present there (as Q. protodentata) as a regular member of the Mio- cene rich deciduous forest, pollen studies indi- cate that the present oak forests dominated by . mongolica are late Quaternary (in Ablaev & Gorovoi, 1974). A similar relationship is indi- cated for the dominant lowland oak communi- ties in California and Arizona as judged from late Pleistocene fossil floras and from evidence provided by wood-rat middens. Fossil species allied to Q. agrifolia, douglasii, engelmannii, and wislizenii are recorded in Miocene and Pliocene floras as members of rich sclerophyllous wood- land. The living species that form pure stands at low elevations adjacent to grassland have adapt- ed to these drier areas more recently and rep- dese latest Pleistocene. SUMMARY The preceding review of the present distribu- tion of oaks and areas of their fossil occurrence leads to the following tentative conclusions. 1. The Fagaceae, which probably originated in [Vot. 70 ANNALS OF THE MISSOURI BOTANICAL GARDEN 654 1983] the montane tropics, had migrated across the tropics via the Atlantic and had diverged into the principal living genera by the later Creta- ceous. 2. Following the Eocene, oaks underwent rapid evolution chiefly in regions of mild to warm tem- perate climate as cooler and drier conditions spread and especially in areas where terrain di- i i ing rapidly, as in Mexico-Cen- tral America, the southwestern United States, the Mediterranean basin and southeastern Asia. . Oaks were rare at higher latitudes during the Tertiary even though temperature was mild Vereit YOLLY most oaks evidently was controlled primarily by light (long day) north of latitude 50°. 4. Many sections and subsections (or series) are confined to local sectors of the continents and for the most part their antecedents have inhab- ited the same general areas, although they ranged more widely prior to the spread of grassland and desert climates, This is paralleled by sections of other genera (i.e., Acer, Alnus, Betula, Pinus) that ir locally on the continents, although other sections of these and many other genera have been and are still widespread on the continents. Also linking the present disjunct forests are nu- merous allied forest herbs, many of which have changed Only slightly since the Miocene. To- ther with the trees and shrubs they form the us floristic threads that make up the surviv- "8 Matrix of the temperate Arcto-Tertiary Geo- ora, whose demise has been certainly misrep- Tesented $ - Few oaks have linked North America-Eur- àsta, Consistent with their evolution in generally stricted parts of the continents. Only the allied doi + Diversipiloseae and Prinoideae have oc- ll oc igher latitudes and their derived species Lon] ur on all the northern continents. Taxa nly a few sections have ranged across North d ye or Eurasia, notably A/beae, Coccineae, Cyclo ‘wis In the former, Cerris, Suber, and 6 4nopsis (probably) in Eurasia. siid, rise of oaks to dominate temperate for- woodlands appears to be a phenomenon AXELROD—OAKS 655 of the later Cenozoic, one afforded by selection for greater drought and higher temperature at lower levels (California-Arizona, Spain-Turkey) and by cold in the interior (Manchuria-Primorye, central U.S.). 7. Species of subsections (series) of Quercus radiated from subtropical and warm temperate into subalpine environments as mountains were elevated in the late Tertiary; ecologic studies of these cohorts should prove illuminating. 8. Further progress in understanding oak evo- lution will depend not only on obtaining addi- tional fossil p t d luating tl in old er collections, but also on taxonomists providing a more realistic grouping of taxa into natural alliances. 1 LITERATURE CITED ABLAEV, A. & P. G. GonRovor. 1974. Fossil oaks of Primorye, USSR, and the origin of oak forests. Lethaia 7: 163-169. ABLAEV, A. G., I. V. VASSILIEV & P. G. GoRnovor. 1974 [1976]. Floristic associations between East Asia and North America: a study of the ranges of some recent and fossil plants in Primorye, U.S.S.R. Pa- leobotanist 23: 49-54. AXELROD, D. I. 1952. A theory of angiosperm evo- lution. Evolution 6: 29-60. 8. Evolution of the Madro-Tertiary Geo- ncaster) 24: 433-509. . 1960. The evolution of flowering plants. Pp. 227-305 in S. Tax (editor), Evolution After Dar- win, vol. 1, The Evolution of Life. Univ. Chicago Press, Chicago. . 1966. The Pleistocene Soboba flora of south- ern California. Univ. Calif. Publ. Geol. Sci. 60: 1— . 1972. Ocean-floor spreading in relation to ecosystematic problems. 7n R. T. Allen & F. C T PONE LÓ8S EY A Q M |j + Univ. Arkansas Mus. Occas. Pap. 4: 15-76. . 1976. History of the conifer forests, Califor- nia and Nevada. v. Calif. Publ. Bot. 70: 1-62. . 1977. Outline history of California vegeta- tion. Pp. 139-194 in M. ur & J. Major (ed- itors), Terrestrial Vegetation of California. John Wiley & Sons, New York. — ——, 1979. Desert vegetation, its age and origin. . 1-72 in J. R. Goodin & D. K. Northington (editors), Arid Land Plant Resources. Internat. oe“ ka Maddie] osuere Ficure 40 " : the Eocene. These are from the Fl . : Sac _—1, 4. Q. scudderi Knowlton; —2, 3. Q. rissant flora (MacGinitie, D» ibo di. B IPM RIS Cockerell 656 Center for Arid and Semiarid Land Studies, Texas Tech. Univ., Lubbock, Texas 1980. History of the Maritime Closed-cone Pines, Alta and 2 California. Univ. Calif. Publ. Geol. Sci. 120: 43. . 1981. Role E Volcanism in ios: and Evo- lution. Special Pap. Geo r. 185: 1-59. —— & . RAVEN. 1972. Evolutionary biogeog- raphy vi Pp. 218- 236 in J. A. Behnke (editor), Challenging Biolog- ical Problems, Directions Toward Their Solution. Oxford Univ. Press, New York Brown, R. W 62. Paleocene Flora of the Rocky Mountains and Plains. U.S. Geol. Surv. Prof. Pap. 5: 1-119 CAMUS, A. 1936-54. Les Chenes. Monographie de 39; 4. P. Lechevalier, Pari: COWAN, D. s. e "Geologia evidence for post-40 B. lara. m.y. of part of rie pce Alaska. Geology 10: 309- 313. DETLING, L. E. 1953. pee bec ig xeric flora west of the Cascades in Oregon. Madroño 12: 39-47 : Pednlaviies on the Colas River Gorge flora. Madroño 14: 160-172. 1961. The chaparral formation of south western Oregon, with consideration of its jonga- cial history. Ecology 42: 348-357. FRANKLIN, J. F. & C. T. Dryness. 1973. Natural Vegetation of Oregon and Washington. USDA regen Service pied Technical Rept. PNW 8. Gast, R. G. R. P. PuiLLiPs & E. C. ALLISON. 1975. Reconnaissance Geology of the State of Baja Cal- ifornia. Mem. Geol. Soc. Amer. 140: 1-170. from the Queen Charlotte Islands to the Gulf of Alaska. Geol. Soc. Amer., Abstracts with Programs 14(7): 5 Howtos W. 1983. Cr retaceous and Cenozoic za of the Northern Continents. Ann. Misso Bor Gard. 70: 440-45 HANSEN, H. P. 1942. A pollen study oflake ES Ut in the lower Willamette Moore of western Ore Bull. Torrey Bot. Club 69: 262-280. 1947. Post-glacial pak succession, climate and chronology in the Pacific Northwest. Trans. . Philos. Soc., n.s. 37: 1-130. KENNETT, J. P. : rine tephrochronology. Pp. 1373- 1436 in C. Emiliani (editor), The Sea, vol. The Oceanic Lithosphere Seria J. 1972. Corresponding taxa and their eco- asia and North America. Pp. 37-59 in DH Val- entine — Taxonomy, Phytogeography and Evolution. Academic Press, London and New York. LEOPOLD, E. B. Late Cenozoic yras of T3 extinction. Pp. Hs 246 in P. S. Mart Search for a Ca niv New Haven . 1969. te Cenozoic palynology. Pp. 435 in R. H. Tschudy & R. A. Scott (editors), Aspects of de mes Wiley- Interscience, New York and London ANNALS OF THE MISSOURI BOTANICAL GARDEN 82. The continental margin of [Vor. 70 cames E. ic JR. 1971. Atlas of United States Trees, Vol. . Conifers and Important Hardwoods. U.S. Dept pnis Forest Service, Misc. Publ. 1146. 200 maps. Macon H. D. 1941. A Middle Eocene Flora from the Central "eros Nevada. Publ. Carnegie Inst. Wash. 534: 5 1933. dit pre of the Florissant Beds, NN Publ. Carnegie Inst. Wash. 599: 188 pp. 1947. Evolution of corti floristic associations in western North America. Ecol. Monogr. 17: 201-210. MEDWECKA-KORNAS, A. 1961. Some floristically: » ss Montreal region ‘of Canada and in central Eu pe. Bull. Acad. Polon. Sci., Ser. Sci. Biol. m 255- 260. MeuseEL, H. 1968. — der Holarktis der Pflanzenareale. Feddes Die pflanzengeographische Ozean- E larktis und die Ozeanitäts- Repert. 79: 157-335. i 1969. Beziehungen in der Florendifferenzi rung von Eurasien und Nordamerika. Flora, B. 158: 537-564. . ScHUBERT. 1971 pe zur Pflanzen- eo hie des Westhimalajas. I. Te Areal ; : des eil: Die gesellschaften. III Teil r Gliede- Die pflanzengeographische Stellung und pde estin; siet as. Flora 160: 137-194, 37> 432. ipic d ER E. WEINERT. " chen ba der zeman Fo PA Poe ee T. E WEINERT. hen Vids ee der pie uropáisc Karten, v. 2. J. Fischer, en Flora. son MULLER, J. 1981. Fossil pollen records sof eum giosperms. Bot. Rev. (Lancaster) 47 M. R. Reb & rocks of the Kodiak Islands in latitudinal — à baT Raven, P. H. & D. I. AXEL 197 biogeography an Univ. Press, Lo A SOKOLOV, S. JA., O. A. A UNAM USSR. Areographia Arborum fruticumque * USSR In- Taxaceae-Aristolochiaceae. Acad. Sci. stit. Botanicum Nomine V. L. Kom: 164. Stessins, G. L. JR. 1947. Evidence on beg ad pe lution from the diri a ye STRAHLER, A. St he raphy vie sEnvironmen ent. a New Yor! Fe Plants TAKHTAJAN, jd rae p Fossil Flowering ee ele | 1983] AXELROD—OAKS of the USSR. Vol. 2, Ulmaceae—Betulaceae. Akad Nauk SSSR. Instit. Botanicum Komarovianum. 216 pp. 172 plates. (In Russian.) 60. Miocene maples from —__ YOKOYAMA. 1975. On the lobed oak leaves from the Miocene Kobe Group, Western Honshu, Japan. J. Fac. nen Hokkaido Imp. Univ., Ser. 4, Geol. 17: 129-14 TRALAU, H. 1963. Asiatic dictoledonou affinities in the Cainozoic flora of Europe. Kongl. Svenska Ventenskapsakad. Handl., Fjarde Ser. 9(3): 1-87. ages W. 1924. The American Oaks. Mem. Natl. Acad. Sci. 20: 1-255 657 WALTER, H. 1970. Einfuhrung in die Phytologie, vol. 3, no. 2. He Straka, Arealkunde: oo histo- uttgart. 9. Neogene floristic and vegetational history of the Pacific Northwest. Madroño 20: 83- 110. 1972. An interpretation of Alaskan Tertiary floras. Pp. 201 -233 in A. Graham E Flo- ristics and P. orth America. Elsevier Publ. Co., icis pon don-New York. 1977. Paleogene Floras from the Gulf of Alaska Region. Prof. Pap. U.S. Geol. Surv. 997: -107. GENETIC DIVERGENCE WITHIN THE GENUS LIRIODENDRON (MAGNOLIACEAE)'2 3 CLIFFORD R. PARKS,? NORTON G. MILLER, JONATHAN F. WENDEL AND KAREN M. McDouGAL? ABSTRACT The genus gog E — of a southeast Asian-eastern North American disjunct ate pair, but the genus had a much wider distribution in the Northern Hemisphere during the late Cretaceous bs the Tertiary. kc boue generally similar i in morphology, the two extant species e measurably different. js 1973 and 1977 they as observed in the progeny. After seven years, the interspeci cific hybrids synthesized in 1973 had a mI eres penes than intraspecific d of "a lean The growt wth rate of "interspecific L tulip Acids 4 nannilati as gh dont el of DU geographic or environmental parameters. Whi le ee was s little flavonoid varia atic ion in 1 one > small ve y in China was markedly divergent in its bie e e survey of in secca trees from a few ulations showed L. eh? ense from one locality were identically homozygous, yh but for different alleles at à diverged owever, L. tulipifera is an abundant, jus s of its range, whereas i. chinense has suffered from population poser loss a ‘apa to the degree that it shows inbreeding depression. The genus Liriodendron includes two extant species, L. chinense (Hemsl.) Sarg. and L. tuli- pifera L., and a number of extinct species that are based on fossils of leaves or seeds from var- ious Cretaceous and Tertiary strata. The Chinese tulip tree occurs in widely scattered populations from the northern part of Indochina to the east- ern Chinese provinces of Anhwei and Hupeh (Fig. 1). From conversations with botanists from the People’s Republic of China, it appears that L. chinense exists as a small number of populations each containing only a few individuals. Unlike the Chinese species, L. tulipifera is abundant and often weedy over much of its range, which in- cludes nearly all of the United States east of the Mississippi River (Fig. 1). Fossils in deposits of Tertiary age from Asia, Europe, and North America indicate that Lir- iodendron was once much more widely distrib- uted in the Northern Hemisphere. it survived in Europe at least until the Plioc d early Pleis- tocene (Szafer, 1954), and is represented in west- ern North American Paleogene (Wolfe, wor Neogene sediments (Smiley & Rember, y is Considering this paleobotanical evidence, ran and that its present distribution cu ge restrictions due to local extinction at wench = in the late Cenozoic. There is no evi e ese chinense were ever sympatric, although these we may survive as relicts of antecedents tribution was originally sympat ur, species are highly interfertile (He & erc pers. comm.; Santamour, 1972; js er 1980). In 1973 two of us (Miller & Par lore the a long-range hybridization study to pot anced genetic behavior of the F, hybrid ger is bein generations derived from it. This wo: vraspecilt augmented by analyses of inter- and in logy, bie variability as determined from morphol0&” the Blandy ' The authors would like to thank the staff of the Orland E. White Research Arboretum of of the Uni- Experimental Farm, University of Virginia, Boyce, Virginia, and the Highlands Bici Sta W. Parks, an urnal Series 4 Biological Survey, New ANN. Missouni Bor. GaRp. 70: 658—666. 1983. Carolina, for assistance in obtaining pcm . Hu, and Dr. E. Lee for help in m the Chinese and for information relating to Chinese botany. The technical assistance d y T. Parks is also greatly appreciated. Jo ment of Biology, University of North Carolina, Chapel Hill, North Carolina 2751 i r York State Museum, The State Education Department, Albany, Ne me Liri (€ L. A. Daum, york 1223. [ EU LS nuo i 1983] PARKS ET AL.—LIRIODENDRON 95° 30° 8 80^ 75 7o DISTRIBUTION OF L. TULIPIFER V O 300 400 MILES $ 200 400 600 KILOMETEI | Ficure |. 1971) and L. chinense (Wu & Wang, 1957). Distribution maps of L. tulipifera (Little, 659 660 chemistry, and paleobotany. Our goal is to de- termine the amount of genetic divergence that has taken place in the genus Liriodendron and to gain insight into the evolution of divergence in a pair of long isolated species. In this report we will concern ourselves primarily with the results of the hybridization study. MATERIALS AND METHODS Plant materials. Seven accessions of Lirio- dendron chinense were available for this study. Six of these originated in 1948 from the Lushan Botanic Garden, Kuikiang, Kiangsi Province, China, by way of the Cabot Foundation and the Arnold Arboretum of Harvard University. Of these, five are maintained at the Blandy Exper- imental Farm, University of Virginia, Boyce, Virginia, and one is located in the Coker Arbo- retum on the University of North Carolina cam- pus, Chapel Hill. The seventh L. chinense acces- sion originated from a botanical garden in Hupei Province, China (A. R. Kruckeberg, pers. comm.) and was obtained from the MsK Nursery of Se- attle, Washington. Individuals of L. tulipifera used in the crosses were native trees growing on or near the campus of the University of North Carolina. Crossing experiments and progeny mainte- nance. Interspecific crosses between Lirioden- dron tulipifera and L. chinense were made in 1973 and 1977. The single tree of L. chinense in the Coker Arboretum was used as a male parent because it produced only a few, mostly inacces- sible, flowers. Hand pollinations were made by applying pollen to the stigmas of buds emascu- lated one or two days prior to opening of the perianth. After pollen application, the buds were covered with paper bags for about two weeks to prevent pollen contamination. Self-pollinations were made by opening buds at the same stage, pollinating with pollen from the same tree and covering with bags as with the cross-pollinations. In the autumn fruits were harvested and stratified for at least 10 weeks at 4°C. They were sown in flats in the greenhouse, and seedlings were trans- planted to 4-inch pots when they had reached the appropriate size. The saplings from the 1973 pollinations were planted in an alluvial field in the spring and summer of 1975. Individuals were s ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 container grown under uniform conditions until 1980 when they were planted in the nursery at a site adjacent to the earlier group. iomass measurements. Relative biomass was estimated by parabolic volume of boles (pv = 0.5rr?h, where r = radius and h = height) calculated from t de of the 1973 progeny in July 1982, and measurements made of the 1977 progeny in November 1979. High pressure liquid chromatography (HPLC). Flavonoids were separated using HPLC. Dried leaves collected in late summer and early autumn were extracted in absolute methanol (0.5 g/5 ml), filtered to 0.45 um and used for analysis without further treatment. A Waters Associates HPLC (6000A solvent deliv- ery system, U6K injector, and 4500 variable wavelength injector) was used. An 8-15 ul sam- ple was injected onto a 1.4 mm x 30 cm Ul trasphere-ODS column (5 um particle size, re- verse phase). Separation was obtained using à mobile phase of tetrahydrofuran : 2% acetic acid (22:78) at a flow rate of 1.0 ml/min (2,800 psi) an isocratic modification of gradient techniques used by Asen (1977) and Smith (1980). i was accomplished at 340 nm and 0.1 AUF»: Retention times and compound quantities wer recorded on a Perkin-Elmer Sigma 10. — 4 HCI vUT VA from mature leaves. The methods of ies n traction, gel and electrode buffer composition» and electrophoresis procedures follow 1982). slightly modified from Wendel and Parks ( Subsequent to electrophoresis, gels us 5. and stained for 16 enzyme systems dis and according to methods published in Bau aa Scandelios (1979), Cardy et al. (1980), SIA" Prasad (1970), Siciliano and Shaw (1970), Wendel and Parks (1982). RESULTS fic hybrids Vi early seedling stages. The heterotic resp * in both the rate of growth and in the vegetative parts. The interspecific h the typical bronze coloration O characterizes the paternal parent, way cro but is never seen in L. tulipifera. Two active matographic separations of flavonoid ex EE 1983] PARKS ET AL.—LIRIODENDRON 661 TaBLE l. Seed set and survival from hybridizations between Liriodendron chinense and L. tulipifera. 1973 Pollinations 1977 Pollinations (field grown) (container grown) Self- lf- Open- Polli- L. tulip. L. tulip. olli- Polli- L. tulip. L. tulip. nations x x nations nations x x a’ L. tulip. L. tulip. L. chin. L. tulip. L. tulip. L. tulip. L. chin. Samaraceta harvested 5 7 15 11 38 14 12 Seeds germinated 11 127 273 19 138 231 230 umber germinated samaracetum 2.2 18.1 18.2 1.7 3.6 16.5 19.2 Number Feb. 77 Feb. 77 Feb. 77 Nov.79 Nov.79 Nov.79 | Nov.79 surviving 5 6 172 8 100 132 166 Percent survival 45 50 63 42 72 57 72 developed from a sample of hybrid trees showed essentially a full complementation of flavonoid extractives characteristic of the two parent Species, as has also been reported by Santamour (1972). The heterotic response was so conspic- yo by 1977 that we repeated the crosses with ‘rent parent trees of L. tulipifera to make a that the growth response was not an ar- ; «t of unusually high specific combining ability used in the original crosses. Data on the germination and early survival of rue Progenies from the two hybridization PS: are summarized in Table 1. In Lirioden- disti Samaras with fertile seeds cannot be readily nguished from sterile samaras. Because all Ei of both types were planted, germination i aggregate (samaracetum) rather than ae t germination was recorded. In both years eror combinations resulted in as many 19.2 a seeds per aggregate germinated (18.2 and "^ yersus 18.1 and 16.5 for the two years, re- "Dectively) than intraspecific combinations, and amm survival was greater for the interspe- S. Th S of seedlings (63 and 72 versus 50 and self. S number of viable seeds produced by the mi nations was very small. Most that ger- sin h ied tH Ali g stag ; however, two ix Survive from the first group of crosses. These ave thy trees only a little smaller than the “rage for trees of the same age derived from 75555 between different L. tulipifera parents. 0 To add an additional comparison, open-pol- linated fruits were collected from each of the parent trees used in the 1977 pollinations. Ger- mination was very low (3.6 seedlings per aggre- gate) but the seedlings obtained showed good vigor and survivorship. Since some open-polli- n conte -— vd ou whereas others produced few or none, we suspect that in the spring of 1977 inadequate cross-pol- lination was the rule for the L. tulipifera parents. In 1973 and 1977 open-pollinated fruits were collected from L. chinense, but a total of only ings was obtained. [Open-pollinated seeds of L. chinense from Kuling, China had vi- ability approximately equal to that of open-pol- linated L. tulipifera in America (Johnson, 1948).] Since leaf extracts from the three open-pollinated L. chinense seedlings have flavonoid and iso- zyme profiles typical of the synthetic interspecific hybrids, L. chinense produced no seedlings through self-pollination. s a measure of vegetative heterosis, biomass (bole volume) was calculated for both progenies (Tables 2 and 3). In all cases for the 1973 prog- eny, biomass was significantly greater (P « 0.05) for interspecific trees than for trees resulting from intraspecific crosses. The 1977 crossing series was designed to test whether the heterosis observed in the first series could be repeated. Different L. tulipifera parent trees were used, and an attempt was made to equalize the number of crosses in 662 TABLE 2. 1982 biomass of Liriodendron progenies synthesized in 1973. ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 1979 biomass of Liriodendron progenies synthesized in 1977. Number Number o Biomass? of Indi- Biomass Cross Progeny inm? Progeny viduals in cm’ L. tulipifera (Saunders tree^) Liriodendron tulipifera x L. chinense 46 0.108: chinense (3 different i L. tulipifera (Murphy tree”) L. tulipifera parents) 164 19.23 x L. chinense 93 0.095° L. tulipifera self- He L. tulipifera (Saunders tree) pollinations 10 E x L. tulipifera (Murphy tree) 53 0.074 L. tulipifera open- pollinations 103 13.10 a Biomass: PV = 0.5zr?h. : ^ Large individual parent trees of L. tulipifera on the L. tulipifera University of North Carolina campus. x L. tulipifera * Difference between interspecific and intraspecific (4 trees involved in hybrids significant at the 0.05 level. 5 different combinations) 134 9.25 all combinations. Nearly all pollinated flowers produced fruits, but many were destroyed by squirrels. Table 3 is a summary of measurements made of the products of this second series of crosses. The interspecific hybrids had signifi- cantly greater biomass (P < 0.001) than the in- traspecific hybrids or the open-pollinated group. The progeny of certain L. tulipifera parent trees consistently showed a greater heterotic response than others with respect to both inter- and in- traspecific crosses, but this specific combining ability response did not obscure the rather great vegetative heterosis observed in the interspecific hybrids. In the spring of 1981 several flowers were pro- duced on a few of the interspecific and the in- traspecific hybrid trees generated by the 1973 crosses. The flowers of the Liriodendron tulipi- fera x L. chinense hybrid were, like the leaves, somewhat larger than those of either parent species. Furthermore, orange pigment was pres- ent on most of the outer petal surface in flowers of the interspecific hybrid rather than being re- stricted to a narrow band toward the insertion of the petals as is typical of L. tulipifera. All i en-pollinated fi both the availahlen intra- and interspecific hybrid group (a majority from the intraspecific hybrids) were planted in order to compare growth characteristics of the F; progenies. Four F, seedlings were obtained from the L. tulipifera x L. chinense hybrids, whereas about 40 seedlings were obtained from the hybrids between two L. tulipifera parents. The vigor and variability of both groups are about the same, and the F, seedlings from both groups appear indistinguishable. ——_———_—_—__ a Significantly different from all other progenies at the 0.001 level. Flavonoid and enzyme analysis. Leaf ex tracts from thirty populations (average of five trees each) of Liriodendron tulipifera and from the seven available trees of L. chinense were € amined by high pressure liquid chromatograpiy (HPLC). Fourteen probable, and four known, rat vonoids (absorbent at 340 nm) from the L. lipifera extracts were quantitatively assayed. measurable quantitative differences acteristic of wild trees from the same difer (population), while major quantitative a pu Ah 31 y p ulations. Inspection of the quantitative data (T ble 4) for the 30 populations provided d vid correlation with altitude or geograph! : tion. Furthermore, principal component ps pa of these data failed to reveal any or tering of populations nor were there any tion icant correlations obtained between positions along the firs nents and altitude, latitude or long! five Lit The flavonoid extractives from the dy EX iodendron chinense trees from the pur i perimental Farm were very similar to €ac dif- whereas the tree from the Coker Arbo Taste 4. Distribution of 340 nm absorbent extractives in 30 populations of L. tulipifera. [¢861 dedos Flavonoid and Flavonoid-like Extractives* # Location ! 2 3 4 > 6 7 8 gb 10 11 125 139 2 if 15* 16 17 18 I Macon Co., NC 43 2E 409 20 38 +- O Mà 149: 53 38: 65 20 ^-— 17 A R2" 6 5 Rabun Co., GA = “462 XE -07 59- L0 e DF 100. — ^ Q6. 18. 20 03 03 0a O4. 13 8 Transylvania Co., NC =~: (3030 10. 4S 47 32 29 0 68 03: bS- 44 22 .<— 202 -I3 £0. 0.9 10 Transylvania Co., NC 4.9 6. 4&7 07 04 06 -20 9.7 41:225 65. 16.0 11 = — 18 — 0.3 12 Macon Co., NC Au Oe i9 35 031509. MA 122-20 L6. 37. 218 02 05704: 07 ill 16 Transylvania Co., NC Is 3180 3237 OFr 11.055349 59S 257: 538 25. 90 52 -= L2 2909 — bI 94 17 Henderson Co., NC — AA DS) 14 -45 09 L32 160 $8» a B6. 28 22 :- ^24 —— 29. 1.9 19 McDowell Co., NC 334 MES. 32-0511 ee 150 C3. L7. 37 31 02.11 "R2 = 23° 09 25 Swain Co., NC = 34 OI 393 53 09. 83 Ia 159° 412 3D. 4S 2D “03 209 Yia L5 09 28 raham Co., NC 09 216 21 IB 31.09 "I WO 149: — 2 Ese 32 20 304 :03 %09- L5 08 S 31 Graham Co., NC — UE Ns 24 23 O48 U? DA 122 28 17 35 27 03 08 07:7 £1 0.8 : 32 Swain Co., NC 21 29 17 -—^ L6 V 23 27 AD ee Te 44 QI 0L rO 0307 OT 07 : 33 Macon Co., 58 340 — Di 37 25 28 D3 100 837, 5 2i 19. 5013 708 Q4: L7 D.6 4 41 Transylvania Co., NC 33 398 — 02 28 I9 AS A8 Go 41. T3: 10 26 01702503: BS 0.6 < 44 ; - EI 96 235 38 23 Z4 EE 215 = (IM 34 20 Ci” = à Z4 18 | 46 Buncombe Co., NC 59 D9 B3 16 22 - 35 EM 1069 60 i4 39 39 12 15 I8 39 11t 5 47 Buncombe Co., NC 47 DT MS 22 32 1595545. DO 160 24 08 40 35 01.07 Ss ES DJ S 55 Davidson Co., NC — 49.9 = 34 133 20 F? 0.3 ZI rI = 24 21 = — E — 2.6 X 57 Orange Co., NC I7 235 DD? 19 28 = 29 EB 257 39 J0 Zi I9 = LI 253 AJ 3 S 58 Stokes Co., NC — 3.7 I0 45 49 LT oe m TE 24. 31 56 17 03 08 LI 26 11 E 63 Jackson Co., NC —^ 30 17 05 22 03 IO M7 3u BM L4 28 16 05 05 05 O04 13 S 69 Avery Co., N 43 289 94 40 22 10 R2 40 168 31 30 49 28 O8 1i 15 29 |l 71 Alleghany Co., NC 36.1 Bl 31 23 M XG 100 120 99 31 30 04 04 02 08 07 73 Floyd Co., VA — JA om 9D 30 09 Ge tie 149. — 12 .51 27 — -08 28 38 1.6 79 Greene Co., VA 435 34 719 12 82 >- P s0 1x0 43 O38 46 30 — ZI 13 36 19 83 34 MWI «A L: 44 ^ M DID 150 90 19 87 54 == 2l > 12- 08 84 Snyder Co., P. 4285 93 33 35 $1 20 1.0 60 = O7 13149 39 29- 29 BI 07 89 Lycoming Co., PA 3 WI 43 38 I5 29% RO 10 — 05: t5 29 O01 25 326 44 i7 82A Watauga E = BS 14 UM O! = T93 DJ 2138 64 48 18 97 08 19 09 20 02 82B Montgomery Co., VA 01. D3 27 =æ 1.49 ^ A DM DO 47 21 63 68 03 Ji L4 03 * Values — i - total extract. * Quercetin-3-ru e set 5. S Glano. T ucoside. » Keengéerci- 3- glucoside. nN Os w 664 TABLE 5. Enzymes resolved from Liriodendron leaves. Alcohol dehydrogenase Aspartate aminotransferase talase Diaphorase uorescent e Glutamate dehydrogenase ph nenhate Isocitrate iiia Malate dehydrogenase Peroxidase Phosphoglucose isomerase mutase Shikimic acid dehydrogenase Triose phosphate isomerase ofany single L. tulipifera stand. The seventh tree, from the MsK Nursery, is very different in its flavonoid pattern in comparison to the other six trees of L. chinense. The differences in flavonoid composition between the two accessions of L. chinense from divergent parts of China are at least as great as the maximum difference in fla- vonoids observed between different wild popu- md "- L. tulipifera. vey of Liriodendron leaf extracts (Table 5). While it is uncertain how many genetic loci encode these systems, reasonable estimates can be obtained based upon the patterns of variation observed in the sampled material and upon a knowledge of the quaternary structure of each enzyme. Cross- ing experiments designed to document the ge- netic control of the resolved systems have been initiated, and the results of these experiments will be reported elsewhere (Wendel & Parks, un- pub.). A preliminary assessment of allozyme variability revealed a sharp contrast in the degree of observed heterozygosity for the two species. While approximately 15 percent of the loci were heterozygous in a sample of three populations of L. tulipifera, not a single heterozygote has been observed in any of the seven L. chinense speci- mens examined. Six of these specimens (the Co- ker and the Blandy trees) were allozymically identical, whereas the seventh (the MsK €: was very different. This last tree was apparently equally homozygous, but for alleles esa: than those of the other six trees at a number of loci. ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 DISCUSSION Despite a strong tendency for parthenocarpy (Stairs & Wilcox, 1966; pers. obs.), Liriodendron trees of either ssi set few viable samaras without cross-pollination; however, there is a low, but measurable, sane entne in L. tulipifera (Ta- ble 1; Taft, 1966). It is thought that a large amount of self-pollination under natural conditions is the major cause of poor samara viability in Lirioden- dron(Guard, 1943; Santamour, 1972; Taft, 1966). In a number of studies (Guard, 1943; Santamour, 1972; Stairs & Wilcox, 1966; Taft, 1966) of L. tulipifera, controlled cross-pollination has very substantially increased the number of viable sa- maras per samaracetum. Similar mechanisms may be operating in L. chinense. Santamour (19135 recorded good seed set for L. chinense at the Blandy Experimental Farm, where tulip trees of both species are planted relatively close 10* gether. Our specimen of L. chinense at the Coker Arboretum, however, was isolated from all other individuals of Liriodendron, and we have ob- tained only three viable seeds aie numerous samaraceta collected i in two seas States by Santamour (1972) and i ncn China by He ers. comm.). The rapid growt : uineis ee trees that have reached reproduc etative heterosis has been i i pifera (Stairs, 1968; Stairs & "Wilcox; 19 ae 1966), but the increment of increase is n great as that of the interspecific hybr! id. Fe The production of vigorous F, progenies pe crosses between different species of the ee d nus of woody plants is not in itself un -[ highly significant. Grant (1971) named e of interspecific interfertility found in v ne plant genera the “Ceanothus pattern. € tern is typified by Magnolia, in which F few barriers to crosses between n species same subgenus, despite their great £ page and temporal isolation, but a on longi the successful hybridization of species 197 to the two different subgenera ( Spa an Treseder, 1978). Because these ri brids made and used for horticultural purposes pe by: is little information about the D o brids. beet In woody plants successful crosses p made, or reported, between various 1983] European, or Asian species of Abies (Rohmeder, 1961), Aesculus (A. X carnea Hayne; Upcott, 1936), Betula (Johnson, 1939; Smith & Nichols, 1941), Campsis (C. x tagliabuana (Vis.) Reh- der; Sax, 1933), Castanea (Johnson, 1939), Ca- talpa (C. x hybrida Hort. ex F. L. Späth; Smith, 1941), Juglans (Johnson, 1939), Larix (L. eu- rolepis Henry; Sax, 1932; Smith, 1941), Liquid- ambar (He & Santamour, pers. comm.), Pinus (Mirov, 1967), Platanus (P. x acerifolia (Ait.) Willd.; Sax, 1933), Populus (Johnson, 1939, 1942, 1946), Quercus (Johnson, 1939), and Taxus (Johnson, 1939), Hybrid fertility and F, vigor and variability are better indicators of species relationships in most woody plant genera than the mere fact of s shown by Upcott (1936) to be a Spontaneous allotetraploid with reduced F, fertility, _In the other studies of hybridization of Lir- cad on tulipifera and L. chinense (Santamour, 372; and those made by Prof. Y. Peitzung at ihe Nanjing Botanical Garden in 1963 and 1965, fide Correspondence with Prof. T. T. Yü, Aca- demia Sinica, Beijing, China) F, data have not sg been generated. Our results from the small "inm crop harvested in 1981 lead us to suspect np Such hybrids will show considerable fertility Th Produce a relatively vigorous F, population. “Re high level of heterosis observed in our Lir- pones hybrids coupled with the suggestion ybrid fertility indicates that the Chinese and E species of Liriodendron have not di- lion in d far genetically despite long separa- rently E" and space. This conclusion is cur- 1989 ing further evaluated: In the spring of à much larger number of flowers was pro- ibi Our F, plants, and reciprocal sibling F; Made p backcrosses to L. tulipifera were L rah nfortunately, no reproductive plants of ian "5€ were available at that time.) Fruits report ese crosses have been harvested, and a Dear 9n their germination and growth will ap- at a later date. terms of their geographical ranges, the two Howey of Liriodendron cover similar total areas. ĉr, L. tulipifera is a common and abun- PARKS ET AL.—LIRIODENDRON 665 dant plant throughout much of the southeastern United States, whereas L. chinense is limited to a small number of individuals in several widely separated populations (Shan-an He, Jiangsu In- stitute of Botany, Nanjing, China, pers. comm.; Fig. 1). Individuals of L. chinense have a low degree of vigor (pers. comm.) which approxi- mates that of our inbred L. tulipifera (one gen- eration). Our very limited isozy data on L. chinense indicate that these small pop- ulations have become genetically divergent, most likely as a result of drift. We suggest that the small number of extant trees of L. chinense suffer to some degree from inbreeding depression, a premise supported by the complete absence of isozyme heterozygosity. If this hypothesis is in- deed true, hybrids between individuals from dif- ferent populations should be markedly heterotic. It is our intention to expand the study of vari- ability between and within L. tulipifera and L. chinense in collaboration with Chinese botanists. d flavonoid LITERATURE CITED Asen, S. 1977. Flavonoid chemical markers as an adjunct for cultivar identification. HortScience 12(5): 447-448. Baum, J. A. & J. G. SCANDELIOs. 1979. Develop- tal = > ih "E 4 1 1: H mer superoxide dismutases in maize. Differentiation 13: 133-140. Carpy, B. J., C. W. STUBER & M. M. GOODMAN. 1980. Techniques for starch gel electrophoresis of en- zymes maize (Zea maize L.). Institute of Statistics Mimeographic Series No. 1317, North Carolina State University. CHANEY, R. W. 1959. Miocene floras ofthe Columbia Plateau, Part I: Composition and interpretation. Publ. Carnegie Inst. Wash. 617: 1-134. : GRANT, V. . Plant Speciation. Columbia Univ. Press, New York. . 1943. The development of the seed of 53: 75-77. [Ci eni low-poplar, an annotated bibliography. Entry Dept. Forestry, Clemson University, Clemson, Carolina. ; N .G. 48. Seed quality of the Chinese tulip-tree. J. Forest. (Washington) 46: 459. JoHNSON, L. P. V. 1939. A descriptive list of natural and artificial interspecific hybrids in North Amer- ican forest-tree genera. Canad. J. Res. 17C: 411- 444. . 1942. Studies on the relation of growth rate ood quality in Populus hybrids. Canad. J. Res. i 0 tow T 946. A note on inheritance in F, and F, hybrids of Populus alba L. x P. grandidentata Michx. Canad. J. Res. 24C: 3 13-317. LirrLE, E. L. 1971. Atlas of United States Trees. Vol. 1. Conifers and Important Hardwoods. U.S.D.A. 666 Forest Serv., Misc. Publ. 1146. U.S. Government Printing Office, Washington, D MILLER, N. G. & C. R. PARKS. 1980. A study of the bitypic genus Liriodendron (Magnoliaceae). 2nd of as. Evol. Biol. Vancou- , New ROHMEDER, E. . Praktische Anwendungsmóg- lichkeiten pouce Forschungsergebnisse. Schweiz. Z. Forstwesen 112: 43-71. MEM F. S5 Tr 1972. Interspecific hybrids i in est Sci. 18: 233-236. Sax, H.J. 1932. Chromosome pM in Larix species. J. Arnold Arbor. 13: 368-374 Sax, = 1933. Species hybridsin Platanus and Camp- s. J. Arnold Arbor. 14: 274-278. R. PRASAD. 1970. Starch gel electro- ymes— —a compilation of recipes chemical verification. For- age echo and niques. Vol. 2. Heinemann Medical Books, Lon- d on. SMILEY, C. J. & W. C. REMBER. 1981. Passe Piper d of the Miocene Clarkia Lake (northern Idaho) and its environs. Pp. 551—590 in J. Gray et al. (editors), Communities of the Past. Hutchinson Ross Publ. Co., Stroudsb SmitH, D. 1980. Flavonoid analysis of the Pityro- gramma triangularis complex. Bull. Torrey Bot. Club 107: DUE SMITH, E. C. 1941. Chromosome behavior in Catalpa ybrida abba J. Arnold Arbor. 22: 219-221. & C. NicHOLs, JR. 1941. Species hybrids in trees. J. Arnold Arbor. 22: 443—454. Ameri ld : 312. [Treatment of Liriodendron, pp. 308-312, done in collaboration with N. G. Miller.] ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 Starrs, G. R. 1968. Cold hardiness in VM Ma hybrids of tulip poplar. Proc. Northeastern For. : 15-17. TCited in tated bibliography. Entry 1346 Clemson eH atqiie Clemson, South Carolina.] & J. WiLcox. 1966. Intra-specific hybridiza- tion in DD poplar (Liriodendron tulipifera L.). Proc. Sixth. World For. Cong., drid, Spain, 1446-1448. [Cited in Schoenike, R. E. 1980. Yel- low-poplar, an annotated bibliography. Entry 1347, Dept. Foresty, Clemson University, Clemson, South Carolina.] SzarER, W. 1954. Pliocene flora from the vicinity of Czorsztyn (West Carpathians) and its relationship to the Pleistocene. Prace Inst. Geol. 11. (In Polish, ary.) 66. Cross- and self-incompati- bility and natural selfing in yellow poplar (Lirio- a Spain, 1425-1428. [Cited in Schoenike, R. E. 1980. Yellow-poplar, an annotat bl ography. Entry 1352, Dept sin Clem University, E South Carol p TRESEDER, OF e 1978. Magnolias. Faber and Fa Genetic control of isozyme variation in ronde japonica L. J. 04. Hered. 73: 197- tion of Alaskan Ter- WOLFE, T s 1972. An interpreta jn (edito sor), ew York. ical Plants of the trop! n Vin nan a prelimin inary "and subtropical pen of m ' 183-254. (In Phytotax. Sin. Chinese (ooo —— BIOGEOGRAPHIC, TAXONOMIC, AND CLADISTIC RELATIONSHIPS BETWEEN EAST ASIATIC AND NORTH AMERICAN CRATAEGUS' J. B. PuiPPS? ABSTRACT Thegenus a — some 150 species, more or less, treated. anthae (mainly European, of about 25 species) and of the largest series: Oxyac Azaroli E to Ce dcr Asiatic, of n 12 species) do not at first sight 5 x s C. x of Yunnan be aligned in a North mi ecognized in about 20 series or sections, some of which, whi d Apiifoliae, is Met similar to European Oxyacanthae, aff ana to be not only among the most t nity with Pyracantha and that of other iade with M impinge on the se C. scabrifolia. Most other North r. Thei esp i indicates basal possibilities an -temperate * genus. One is thus led to postulate ap origin that was able to cross Beringia i in the early T Tertiary or i Mi ad Thence d d from a south China relictual base evolution toward th OGI IYU wit spread i into western Eurasia (one major - line); A pes thro ugh Beringia to a North American hA Hm iss are conspecific. Also, except for the postulated ancient relicts, no Chinese stock. closely rela ted. rni Wn deciding what materials on the list ef ascinating and complex genus Cra- rating : t be presented for the St. Louis Sys- ine pasim. in 1982 I had to come to -— Ow to handle incomplete work. My cg sh puaeealy influenced by the view that as received no comprehensive treat- en “a E modern and effective. There- is leue: , ution, which incorporates mod- ow, sai other workers in addition to my tt nts more than a progress re Made in pin to claim that the progress years sigs systematics of Crataegus in recent phic emn one to present a credible biogeo- This ysis of the genus for the first time. Progress has included modern, floristic works Thus, w. parallel e Prion d of. Asian and North American Crataegus, except for relictual taxa, they are not very for almost the whole of the range of the genus, numerous studies on reproductive biology, e.g. introgressive situations (e.g. Byatt, 1975, 1976a; Love & Feigen, 1979), chromosome counts showing three levels of ploidy (2x, 3x, 4x) (Glad- kova, 1968; Muniyamm a & Phipps, 1979a; Dickinson, unpubl.; Smith, unpubl. ) jg ‘au 1979b) eer the extension o i this wuss de a number of taxa a & Phipps, 1984; Dickinson & Phipps. niue, Smith & Phipps, unpubl.), knowledge of the phenetic dissection of apomictic complexes (Sinnott & Phipps, 1983; Dickinson & Phipps, submitted; Smith & Phipps, unpubl.), etc. The specific contributions of this paper will include a numerical taxonomic anal- ysis of the genus that permits of more rational eee for this work was supplied by operating grant A- 1726 from the National Sciences and Engineering f Canada. Council of o The Department of Plant Sciences, The University of Western Ontario, ANN. Missouri Bor. GARD. 70: 667-700. 1983. London, Ontario, Canada N6A 5B7. E a EAE NEE a mos Sei bei ean han regs cie 668 TaBLEl. J.C. Loudon’s Classification of Crataegus in Arboretum et Fruiticetum Brittanicum (1838). L Sect. Coccineae Loud. (p. 816) 2 — ic) II. Sect. eri ha tout (p. 819) 1 species: Cc. erg eios Lodd. IV. Sect. Crus-galli Loud. (p. 820) 3 species, incl. Ü crus-galli L. V t. Nigrae Loud. (p. 822) 2 species: C. nigra Waldst. it. and C. purpurea Bosc. & Ki VI. Sect aee ege Loud. (p. 823) 1 species: C. eciam Sect Flavae Loud. (p. 823) 3 species, incl. ava ES . Sect. Apiifoliae Loud. (p. 824) 1 species: C. apiifolia Michx. (= C. marshallii Egglest.) IX. Sect. Microcarpae Loud. (p. 825) 2 species: C. spathulata Ell. (sic) and C. cordata L.f. X. Sect. Azaroli Loud. (p. 826) 5 species, incl. C. azarolus L. XI. Sect. Heterophylla Loud. (p. 829) 1 species: C. heterophylla Flugge Oxyacantha Loud. (p. 829) 1 species: C. oxyacantha L. [= C. laevigata (Poir.) i Sect. Parvifoliae pd Ae 841) 3 species, incl. C. parvifolia . (= C. uniflora ri nchh.) . Sect. Mexicana Loud. (p. 843) 1 species: C. mexicana M Two Bes sections are not now included in Crataegus species groupings on a world-wide basis than hitherto presented, the biogeographic analysis it- self, and supportive interpretations from cladis- tics, dispersal biology, and to a lesser extent, con- siderations of paleoclimate, paleogeography and some Crataegus fossil evidence. It is my consid- ered view that such a synthesis of mutually sup- portive data from these different viewpoints and workers is vital to disentangling a genus such as Crataegus, which has been called Sonet alarm dign c | "a veritable witches’ —— Th geographical analysis a are: (a) an adequate species study; (d) knowledge of dispersal biology, and (e) knowledge of these topics in times past. I con- sider that (a) and (b) exist to adequate sepsis from literature compilation, a revision of o own (Phipps & Muniyamma, 1980) and etd ium studies on many species. I will present new materials for (c) in this paper (phenetic and cla- ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 distic) and will use available knowledge on (d) and (e). The understanding of the dispersal bi- ology of Crataegus, although sketchy, appears reliable for the arguments to be made, although paleontological data, excepting pertinent paleo- eography, is sparse The genus Crataegus is taxonomically quite well-known at the regional level, although apo- mictic complexes, some active introgressive sit- uations, and some relatively rare segregates from apparently older hybrid situations may compli- cate the picture locally. The genus is most com: mon between 30° and 50° N latitude although some extensions north and south of these limits may be found. In spite of the abundance of species and the ecological significance of Ci buic ac is no rea ment of this vinee that has been presented m erto. Regional floras, e.g. Franco (1968), for Eu- rope, Maire (1980) for North Africa, Browicz (1972) for Turkey, Riedl (1969) for the Aen Iranica region, Pojarkova (1939) for the he USS and Yü and Ku (1974) for China cover Eurasia effectively at the normal level of good floristic in Asia the western half of the United States an me but there is no comprehensive treatmen gir Mexico since Standley (1922). However, in of apparently complete regional covera (uot remains uncertainty about the reliability 0 y ments of certain species groups, especia y of those concentrated in Texas an United States. The situation undoubt real taxonomic problems, as is now ork on cytology ic di i f course, © These taxonomic difficulties can, O nic! ese taxon : to some lack of precision for the 5! surabiliy © of analysis because of the inco omm C taxa of the ‘same’ rank and certain certainties boundaries based on taxonomic un 1983] TABLE 2. C. K. Schneider’s classification of Cra- laegus in Illustriertes Handbuch des Laubholzkunde (1906). Note—authorities as Sinaia d by Schneider are not necessarily correct. Phipps (1983) provides cor- rect citations for serial and sectional names l. Sect. aed Zabel (p. 769) 3 species, incl. C. pinnatifida Bu Sect. end Zabel (p. 771) 6 species, incl. C. sanguinea Pallas . Douglasianae Rehd. (p. 775) 2 species, incl. C. douglasii Lindl. Sect. Tomentosae Sarg. (p. 776) 3 species, incl. C. tomentosa L. 5. Sect. Pentagynae C. K. Schneid. (p. 777) 2 species: " pe Waldst. & Kit. and C. nigra Waldst. nm » >_> 6. Sect. ie Zabel (p. 779) 5 species, incl. C. oxyacantha : . Ori reales Zabel (p. 786) 7 species, incl. C. orientalis Pal 8 Mopac Koch (p. 790) 3 species, incl. C. apiifolia Mic - Sect, on Beadle (p. 791) 2 species, incl. " C. brachyacantha Engelm. & Sa : e Flavae Sarg. (p. 792) 2 species, incl. C. flava npe Set Unifora Beadle (p. 793) 1 species: C. uni- flora Mue 2. Sect. Cineatae Rehd. (p. 793) 1 species: C. cuneata Sieb. & Zuc * Sect. Mexicanae, sect. provis. (p. 794) 1 species: C morna Steud. : ect. Aestivales Sarg. (p. 794) 1 species: C. aesti- : valis Torr, & Gr. + Sect. Aaa Sarg. (p. 794) 2 species, incl. C. a Jacq. wa = Puncta Sect. c alli Sa : tolli 1. galli Sarg. (p. 796) 1 species: C. crus ncn Triflorae Beadle (p. 797) 1 species: C. triflora So. t Molles Sarg. (p. 797) 2 species, incl. C. mollis i st is Sarg. (p. 798) 2 species, incl. C. viridis Sect. Pruinosae Sarg. (p. 798) 1 species: C. prui- nosa K, Koch Sect. Coccineae Sarg. (p. 799) 4 species, incl. C. coccinea L. 2 P. Intricatae Sarg. (p. 801) 3 species, incl. C. Mricata La NO os o os Boe eee ne tate of current taxonomy may be briefly a" here. Loudon (1838) grouped Cratae- In doin 9 14 sections (Table 1) of similar species. lag he presented the standard format rataegus classification that has persisted PHIPPS— CRATAEGUS 669 TABLE 3. Series of the genus Crataegus as proposed by Rusanov (1965) in Dendrologii Uzbekistanii (from Cinovskis, 1971). Ser. 1. Pinnatifidae Ser. 14. Macracanthae Ser. 2. Henrianae (sic) Ser. 15. Punctatae Ser. 3. Microcarpae Ser. 16. a Ser. 4. Oxyacanthae Ser. 17. Intric Ser. 5. Nigrae Ser. 18. E unio nae Ser. 6. Pentagynae Ser. 19. Cordata Ser. 7. Douglasianae Ser. 20. Silvicolae Ser. 8. Sanguineae Ser. 21. Pruinosae Ser. 9. Calpodendra Ser. 22. Molles Ser. 10. Virides Ser. 23. Dilatatae Ser. 11. Cuneatae Ser. 24. Coccineae Ser. 12. Mexicanae Ser. 25. Tenuifoliae Ser. 13. Crus-galli through, for instance, the work of C. K. Schnei- der (1906), whose classification is presented in Table 2, Sargent (1902, 1903), Palmer (1953), Kruschke (1965), Rehder (1940), Rusanov (1965 —see Table 3), Cinovskis (1971), and oth- ers. Thae Hemen vaty in the geographical re d Palmer being North American, Pojarkova Russian, and the others world-wide. They also vary in the predominant rank used (Section, Series, ‘natural group’ or ‘group’ — German ‘Gruppe’) and the number of species known to and accepted by the author. But what these classifications have in common is “flatness.” Neither Loudon's 14 sections nor Rusanov's 25 series claim a chical groupings or interrelationships. Thus, the established view of Crataegus taxonomy, that of many small species groups, represented the most that taxonomists would venture for some 140 ears. However, Cinovskis (1971: 19), in Crataegi Baltici, presented a remarkably complex reticu- late set of interrelationships (Fig. 1) that, al- though possibly close to the botanists' dream of Crataegus, is very confusing and not scientifi- cally very realistic. At the other extreme, El-Ga- zaar (1980) attempted to cut the Gordian Knot by erecting two sub-genera (Fig. 2), on the basis of —: leaf-shape, geographical distribu- tion, and base chromosome number. i cene euius Crataegus represents, more or less, o his criteria, two related sections OL lucas and Azaroli) it is not, as he claims, restricted to Eurasia (note C. marshallii in the United States). His other group, “A mericanae,” purports to rep- resent the remainder of Crataegus. However, the remainder of Crataegus is much more variable 670 CINOVSKIS' (1971) SCHEME OF EVOLUTIONARY RELATIONSHIPS IN CRATAEGUS. CALPODENDRA BREVISPINAE PENTAGUNAE ANNALS OF THE MISSOURI BOTANICAL GARDEN DOUGLASIANAE [Vor. 70 TENUIFOLIAE zu. GRAINERDIANAE ee nio] TTE E 4 ipei ROTUNOIFOLIAE FIGURE 1. in leaf morphology than this simple division im- plies, ** Americanae"-type leaf morphology char- acterising most east Asian species whereas, in quoting Longley's (1924) paper based on hand- sectioned material, El-Gazaar and Badawi (1977) resurrected the totally discredited notion of x = 16 for North American species. El-Gazaar, re- grettably, did little more than reinforce a notion already well understood by Sargent, Schneider, and others that a distinct “Oxyacanthoid” (small leaf, deeply lobed) leaf shape existed, and from there went on to totally misleading suppositions. Thus, prior to the work to be presented here, no credible set of interrelationships within Cra- taegus has been proposed. MATERIALS Some 145 species of Crataegus are held to exist (Appendix A) and it is neither necessary nor con- venient to use all for this study. For the numer- ical taxonomic work all the species groups (e.g. sections, series) were sampled in order to be cer- tain to include all pertinent phenetic variation. This generated a list of 75 species (identified in Appendix A), some quite narrowly defined. This EVOLUTIONARY DIRECTION TO RIGHT Cinovskis’ (1971) scheme of evolutionary relationships in Crataegus. ; distic same set of species is also used for the cla udy. : " CREDE because of the title of the St ae Symposium, the suite of taxa used in w per biogeographic analysis for this paper exc “id 20-25 Oxyacanthae and the ten or so A7 ? Western Asia, Europe, and North Africa an "m few small sections unique to this region. 75 westward demarcation line for the pa longitude. The 62 taxa used in the bio pes study include essentially all eastern Asi mig this paper) and North American gh using a relatively com wha in selecting cladistic set have had to be used a me is operational taxonomic units d se-limis in some cases. Specifically, — vd are still rt g e : es species, an entire series was sore air an OTU. The lack of consistency à x nomic rank of OTUs might pA d nalistically inclined, but in actual fa > : mit for €25" The choice of 75°E long. wonen E this ern Asia is somewhat arbitrary. | | E" } PHIPPS— CRATAEGUS 671 subgenus Crataegus EL-GAZAAR 1977, 1980 subgenus Americanae base chr. no. x=17 x= 16! leaves deeply incised? shallowly lobed? veins to sinuses present absent? distribution Eurasia North America ' quoting Longley, 1924 who used hand-sectioned material—correct is x = 17 (Moffet '31, Gladkova '68, M & Ph ’79). ? Some Eurasian species have subgenus Crataegus foliage type. ? Some N. American species have subgenus Americanae foliage type. 29 FIGURE 2. El-Gazaar’s (1977, 1980) subdivision of Crataegus into two subgenera. nm study limit passing through the currently mpenetrable central Asian massif of Xinjiang- id Proves to coincide with a nat mic and biogeographic partition in Crataegus. DATA , The morphological data used consist of 49 a used for the phenetic studies. For the vits € studies, those 18 characters that could, licis *quate confidence, be polarised, were se- (Table 5). These were appropriately re- rds later discussion). range biogeographical data base consists of 62 ti Range maps made from locality co- by tes (dot maps) of individual records are üt im were substituted, however, because Ot maps are available yet. There is 10 Problem ince vs fi 2 Ta ae floras mentioned, supported by more local » SO long as consistent taxonomic decisions are made in coordinating data from different sources and attention is paid to synonymy when grouping taxa, as discussed earlier. PLAN OF WORK y PCa tudy a pi acters was conducted to test what species groups existed in Crataegus worldwide. For this, 75 species were used, sampling all sections of the genus. The characters used are largely those con- ventional in Crataegus taxonomy, but leaf lobing and associated attributes, as emphasised already by Schneider (1906) and taken up again by El- Gazaar; and thorn type, as emphasised by this author (e.g. Phipps & Muniyamma, 1980) have more prominence than usually accorded by most workers. Due to the exceptionally large number of herbarium specimens available, median or modal values were scored. The OTUs were sub- jected to comparison by euclidean distance and clustered by minimum variance clustering to de- rive a phenogram, and to ordination by the dis- persion coefficient and principal components analysis to derive a scattergram r-type com- ponents analysis indicated character loadings. The biogeographical analysis depended on scoring presence-absence of species on an equi- form grid of 640 km? laid on eastern Asia and North America (Figs. 3, 4). This method is some- 672 ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 TABLE 4. Morphometric data for Crataegus study. TABLE 4. (Continued). abit: 1. tree through small bush wigs: 2. color of l-year old twigs (a) 3. color of 1-year old twigs (b) Thorns: thorns always distinct from leafy short » O0 —-J) ON tA = E o G TH a [72] 2 g & . frequency . Color on 1-year old twigs (a) 10. color on 1-year old twigs (b) Leaves: lla. size: length of blade (in mm) b.*size: breadth at maximum point (in mm) 12a. shape: (char 11a/char 11b as b. shape: distance from base to widest inter- sect on midvein 13. shape: location of widest point: (char 12b/ char 11a as 96) 14. shay i Me POURS 5 À j rO WIUCSL o 15a. oo longest lobe, measured along vein in mm b. bine deepest adjacent sinus (line || base of sinus to tip of lobe, in mm) 16. leafincision index, LII (char 15b/char 15a as 96 17. number of side veins 18. leaf base—angle to widest point of blade 19. extent of lobing: (no. of lobes where LII > 20%) 20. veins to obvi i ( ber of where LII > 20% 21. lobing related to vein no.: (char 20/char 19 as 96) 22. pubescence above 23. glandular teeth on serration tips or peti- oles 24. subevergreen to deciduous po sli si bracteoles at anthesis ie number of flowers 27. pedicel indumentum 28. pedicels punctate or not Flowers: 29. hypanthium pubescence : pe 33. calyx: serration/lo obing 34. calyx: lobes, serrae (f any), gland-tipped 35. flower diameter, cupped (mm) 36. gynoecium: style no. 37. OGI anther no. 38. oecium: anther length 39. ja hE anther color when fresh; (no red, pink to purple) Fruit: a. length, excluding calyx, fresh b. diameter, fresh 4l. UR ARAT (as %) 42. widest part: (in lower '4; middle 1s; upper V. 3) 43. color when fresh and mature (yellow, red, la c 44. undamaged fruit highly pruinose 45. calyx tube (collar) present 46. calyx lobes present 47. nutlet no. 48. nutlet: strongly dorsally grooved 49. nutlet: laterally excavated ipted' characters were re measured, but lat- P he picnic ratios. er bud as redundant because contributing to as is advocated, for instance, in Phipps's “Best- block” (1975b), but the biogeographical data ve (distribution maps with smooth edges) does ei: conveniently permit the use of such sensitiv methods. Nevertheless, if the scale of grid is a ther too large nor too small, pattern is stil] we ta, detected. From the presence-absence em distribution patterns, while additive scori " individual grid squares directly generates CO of areas of high species richness. susti A cautious approach to cladistics W bi id view of the high likelihood of hybrid or ral sf situations (e.g., Phipps, 1984). Phenetic paps of the 25 or so maloidean genera (Table conducted to identify a set of genera most cific transforms may be found in the : mos of the cladistic results. Although availa te on well-known cladistic proarai set o Alldib i vv 1983] PHIPPS— CRATAEGUS 673 Character selection and polarity for cladistic analysis. TABLE 5. Character No. l. OTU identifier Coded descriptors = prim.; +ve and —ve numbers indicate advancement in various directions abit 2: tree (prim.) > small shrub (adv.) Branches 3. thornless with sharp-tipped short shoots (prim.) > short thorns > long thorns (adv.) aves 4 subevergreen (prim.) > early deciduous (adv.) 5 multiveined (prim.) ^ few veined (adv 6 of side-vein with midrib narrow pan ) to wide (adv.) 7 leaves unlobed (prim.) > deeply lobed (a 8 veins to leaf sinuses d (prim.) > Bier (adv.) Inflorescence and flow. 9. pedicel meee aed dam (prim.) > glabrous (adv.) 10. hypanthium indumentum dense (prim.) > psum (adv.) ll. flowers medium in size (prim.) ^ yery npe te p > small (adv.) 12. calyx lobes entire (prim.) — serrate (adv.) : low (adv.) 13. . A carpel number medium (prim.) ^ high (adv,) 14. anthers white or cream (prim.) — purple (adv.) 15. Een number 20 (prim.) ^ 10 > 5 (adv.) 16. iud oblate (prim.) ^ Kcd long (adv.) 17, fruit large (prim.) ^ small (a 18. fruit yellowish (prim.) > reddish > black (adv.) — D9. mutlet dorsally smooth (prim.) > ridged (adv.) they always generate furcations, not necessarily irable in a group such as Crataegus. Instead, Program PRIMO, which successively finds hy- cal t taxonomic units (HTUs) (nodes) unit- ë Most similar pairs of OTUs and HTUs, was = ^um. patristic distances became too long ac- em x e a test criterion and were held likely The pro- Fam als also i incorporates the capability of indicat- al : e two or more lemative fusions show daii patristic dis- likely diagram style, which postulates certain connections and proves to recapitulate the tighter Phenetic groupings. Conformable phenetic, biogeographical, and pro-cladogram patterns are held to be strongly evidentiary of true cladistic relationships. EXTRINSIC DATA ng-distance dispersal potential, and thus Lo mode of propagule dispersal, paleogeography an paleoclimate are critical to a full evolutionary understanding of a genus. T9 ——— are for which they are understood to date. PHENETIC RESULTS Using all 49 morphometric characters or only 14 leaf a ie similar groupings of Cra- taegus emerge —8). Coded on these dia- grams are leaf phe Pets 5—8) and thorn glyphs [Vor. 70 674 ANNALS OF THE MISSOURI BOTANICAL GARDEN ets ses vo mi BASE OM FOR )ASIAN BIOGEC RAPHY. 1982 N | MS - eund pe C | \ N i — \ eo DE euge E. ASIA | FiGure 3. Equiform grid for eastern Asia. (Figs. 5, 6), of which also fit well. The main b) Leaves usually small, broad, dee groupings found are as follows: a) Leaves narrow, multiveined, essentially un- lobed: i) with thorny short shoots: C. mexicana, C. scabrifolia ii) with true thorns: ser. Crus-galli, Puncta- tae veins to leaf sinuses: i) with thorny short shoots: Aza acanthae ii) with true thorns e.g. C C. pinnatifida (China) sh; tro c) Trees of forest shade; fruit ms states | thorns present; distribution SE Uni | roli, OXF marshallii (US) ^--———— GENES cett MEN LLL oet iltl]T A. moO — e ——ÁÀ ——QÀ — 0 o—- — m —— ——— — — 1983] PHIPPS— CRATAEGUS 675 Figu . RE 4. Equiform grid for North America. i ; ) Leaves trilobed: sect. Cordatae (C. phae- ii nopyrum only) LÉ otherwise (ser. Virdes, Parvifo- d) Lea E Brevispinae, etc. €s shallowly lobed, broader, ovate, ob- ovate, broad-elliptic to + deltoid; true thorns present: i) most N. American taxa. This group cor- responds to El-Gazaar’s ‘Americanae’ and it will probably be well justified to erect a 676 OXYACANTHAE (3-25) ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 0 1. C.MEXICANA N A SANGUINEAE ET AL. : S (40-58) : => E an 142 140c FIGURE 5. Plot of q-PCA, first two axes, 75 individuals, leaf data. (OTU numbers cros' Appendix A.) section of this name. However, I feel that this is premature while the boundaries re- main uncertain. ii) Some East Asian taxa (e.g. C. kansuensis, C. maximowiczii). These have shorter thorns than most ‘Americanae.’ e) Some miscellaneous leaf shapes, e.g. C. spa- thulata (US), C. flava (US), C. cuneata (China) In the components analyses (Figs. 5 and 6) zaroli and Oxyacanthae strongly overlap and are opposed on the first two factors by the ‘Ame- ricanae’ and Crus-galli types. Notably, C. mex- icana and C. scabrifolia are closer to Azaroli- Oxyacanthae when all characters are considered than when leaves only are treated, when they are S.E. UNITED STATES DIVERSIFICATION (62-76) © n TA 1982 WORLD CRATAEGUS STUDY. 75 OTUS. 14 CHARS LEAF DA s-referenced to ino- polar to the aforementioned groups. rc Siberian taxa (sect. Sanguineae, et al.) : js the Azaroli-Oxyacanthae using all data hg lap only ‘Americanae’ on leaf data. This " the San- an essentially intermediate position = Siate: the ‘Amer : F and barely overlaps when using all gr merica when leaves only are considered, «t fully Ove nae are not only variable but almo Hn south- | counts: eastern United States groups. Thus, 10 are the the Ameri Oxyacant most contrasted pair, while all the in some ways intermediate between | | SANGUINEAE ET AL. (40-58 ) OXYACANTHAE y 7 g^ Ge . We 1982 WORI LD CRATAEGUS STUDY. 75 OTUS, 49 CHARS ALL DATA A) an sr D P'acanthae-Azaroli. This, of course, has a hg cladistic significance. bw Boios (Figs. 7 and 8) lend detail Notable E oue of the scattergrams. iive isctats itional information involves the rel- Well as E of C. hupehensis and cuneata as gram for wee sect. Flavae. Also, the dendro- bres ta (Fig. 8) shows an interesting Du ation between more glabrous and more The bin. groups, especially in the “A mericanae.” Stood, E significance of this is not under- ysis. en it is it should aid cladistic anal- R- 6) e components analyses indicate (Table ery strong significance of leaf-lobing and PHIPPS— CRATAEGUS E. UNITED STATES E ION RSIFICAT 76 zl (62-76) Y 7n 2 3. A 2. C. SCABRIFOLIA 1. C.MEXICANA FIGURE FIGURE6, Plotofq-PCA, first two axes, 75 individuals, all data. (OTU numbers cross-referenced to Appendix related characters though the loading (72%) of character 10 (pubescence) on root 3 may even- tually appear to be highly significant. When con- sidering all characters (Table 7) the major ones almost equally divide between inflorescence pu- related characters on the one hand leaf morphology and here which accounts for similarities in the and which relate to the observation in the pre- vious ph. Thus, although the presence of some inter- mediate taxa, together with the existence of a broad amplitude for some species groups, do not 101.00 L 1 81.00 VERTICAL SCALE 61-00 41-00 1 21-00 1-00 L OTU NUMBER, left to right 1982 WORLD CRATAEGUS STUDY» 75 OTUS, 14 CHARS LEAF DATA i O y 122 131 140b 129 106 76 n i^ S 3 re iu 48. 45$ 11 al a È ^. 108 61 106 8,9 N3QG3IVO 'TVOINV.LOS DIOOSSIN FHL JO SIVNNV OTU NUMBER, left to right 60 106 65 131 — 130 34 72 142 35 1982 WORLD CRATAEGUS STUDY. 75 OTUS, 49 CHARS,- ALL DATA Ta i 81a 122 71 81b TIG 102 ] 109 126 2 108 59 w 76 8lc 66 z Bo qe 129 B 140a 128 101 z 140b 115 92 2 140c 114 99 E 5 3 95 wu 46 26 88 66 FIGURE 8. Phenogram derived from minimum variance clustering of 75 individuals, all data. (OTU numbers cross-referenced to Appendix A.) [£861 SI523V.LV4O —SddIHd 680 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 TABLE 6. 1982 Crataegus study, r-type components analysis, leaf data. Characters Contributing more than 1096 Each to First Three Roots Percent of Trace a/c by Each Root Root 1 Root 2 Root 3 Root 1 3296 no. 7, 4196 no. 11, 3696 no. 10, 72 Root 2 22% no. of deep lobes per leaf glandular serrations on leaf pubescence above Root 3 13% no. 9, 32% & petiole margin Root 4 10% percent veins ending in no. 8, 31% Roots 5-14 lan lobes veins to sinuses of leaves no. 8, 11% veins to sinuses per leaf no. 5, 11% leaf incision index Character 5, leaf incision index contributes most, 10.3% on all roots combined. generate an utterly sharp taxonomy, there can be no doubt that the taxonomic groups recognized are broadly valid. BIOGEOGRAPHIC RESULTS The phenogram (Fig. 9) derived from the grid- ded distributional data shows that species ranges fall strongly into readily identifiable groups. Bearing in mind that there is a slight overlap between the Sanguineae (e.g. C. sanguinea Pall. ex Bieb.) and some species of the Oxyacanthae and Azaroli in the Soviet Republics of Kazhak, Kirghiz and Tadzhik, the European and west Asian taxa are mainly very strongly cut off from those under discussion here. The first division of our phenogram necessarily differentiates Asiatic from North American taxa because there are nO species common to the two areas. Characteristic Asiatic distributions are repre- sented in Figures 10-12. Basically a northern (N of 40°N lat.) group of species and a southern TABLE 7. 1982 Crataegus study, r-type components analysis, all 49 characters. group may be recognized. In the northern group, the wide-ranging C. sanguinea (Fig. 11) achieves the most northerly distribution of extant Cra- taegus (63°N lat.). Crataegus maximowiczii (Fig. 10) and C. pinnatifida (Fig. 12) of NE Asia and C. chlorosarca (Fig. 10) of the Pacific Rim (N to Kamchatka) are of interest in their approach Beringia. Presumably a quite small climatic shi could permit C. chlorosarca to migrate to the New World. More southerly species are an sented by C. kansuensis (Fig. 11) of north-cen China and the taxonomically isolated C. ae (Fig. 10) widespread in warm temperate P southeast China. Several species of restricted : tribution in western China proper are re which the taxonomically isolated C. scabrifo . ex- (Fig. 12) of Yunnan is a particularly good : taxa ything neoid’ forms and southern and central are largely unrelated to each other or occa j ibuti i ree Roots Percent Trace a/c Major Characters Contributing to First Three Roots by Each Root Root 1 Root 2 Root ? Root 1 18% no. 23, 30% no. 6, 20% no. 15, 17% lobes Root 2 15% hypanthium gland-tipping of calyx lobes number 0 Root 3 10% pubescence no. 17, 14% Roots 4-49 58% peak len lobing/vein number 10096 no. 22, 24% lobing of calyces : dicel no. 16, 1096 no. 4, 17% ear old indumentum veins to obvious sinuses color on n? no. 18, 896 Lio s pubescence upper —— e . M — — ——— — — — 0 OOO | MATCHING COEFFICIENT 1 PE Pah monin. vn FOR CRATAEGUS DISTRIBUTIONS, 1982 ,, "f trise 45 95c 62 48 86 — 143-145 46 94 116 4 49 95d 129 55 112 95 57 113 8l 59 128 3 58 77 106 50 85a 109 | 40 nr.97 114 56 91-93 130-139 51 95a 122 FX 63 124 1 88 n 84 76 100-105 78 72-75 99 l 79 66 85 44 26 140 98 65 120-121 119 n we DJ uH j : NORTHERN SOUTHERN = STW. TEXAS WIDE SOUTH-EAST von NER NORTH-EAST ASIA NORTH AMERICA FIGURE 9. Classification of distributions derived by minimum variance clustering of positive matching coefficient. (OTU numbers cross-referenced to Appendix A.) S123V.LVM2 —SddIHd [£861 189 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 D f INCLUDING C. ne MAXIM. aw d C. CUNEATA sien. & zucc. (O © RIA [n MAXIMICZII C.K, SCHNEID., C. CUNEATA sieB. & zucc P ied > Jp Y Y u C. SANGUINEA PALL. EX BIEB. 4 E C. KANSUENSIS wits.” ‘ Q^ er ul e b pace ER 1€B C. KANSUENSIS wits., C. SANGUINEA PALL. EX SIE ; / pA E. ASIA : C. maxi- FiGure 10. Distribution of Cratae hneid., C. cuneata Sicb. & Zucc. mowiczii C. K. Sc else (e.g. C. hupehensis—not shown, C. scabri- folia and C. cuneata). In North America, western, Mexican, Texan, southeastern, eastern, northeastern and wi ranging types stand out. There are few western species but C. douglasii (Fig. 13), mapped to in- clude the possibly conspecific C. rivularis, is no- table for its extreme northerly range extension in the Anchorage region of Alaska, remarkably near to Beringia. It should be noted that C. doug- lasii is one of the few North American species to share the short stout thorns characteristic of the sanguineoid group from Asia. Crataegus douglasii exhibits the migratory potential of the genus with its disjunct population around the northern Great Lakes. Crataegus columbiana (not mapped) would link C. chrysocarpa (Fig. 17) to the Pacific, if conspecific, which it is close to being. The Mexican group is dominated by the quite wide ranging C. mexicana, which extends through the highlands into Guatemala. It is found under similar latitudinal and climatic conditions to its closest relative, C. scabrifolia, of Yunnan. Texas manifests a rich variety of Crataegus, sev- eral of which, of quite unrelated series (e.g. .C. viburnifolia-Molles, C. tracyi-Crus-galli) are en- oO Ficure 11. Distribution of Crataegus. C. sangul- nea Pall. ex Bieb., C. kansuensis Wils demic there. Another important biogeogra group are the Gulf Coast and pepper is States species. What is notable about ai their lack of relationship to each other or, the notion that they resulted from in these regions. The pertinent United Sas are ser. Aestivales (Fig. 14), Ser. T : mapped), C. marshallii (Fig. He E jac related to European Oxyacanthae, acantha of the Gulf Coast eid ess distinct series Viriaes, Pues and Parvifoliae. Yt is notable that s these are well differentiated from the sn sad ern United States taxa— mainly ‘Amer! i sal and that t somewhat shade tolerant trees. Mid-continent taxa are rep p ee mollis (Fig. 13). The north s i abut inated re ‘Americanae’ speci c, mar . C. macrosperma ipn ne series guretia (Fig. 16), other members of and most of the Silvicolae, Dilatatae, Pr a ei Macracanthae, and Coccineae. It is uth ae to note that taxa with a large e north- a igi uam c C. PINNATIFIDA Bunce i - b SCABRIFOLIA FRANCH. FIGURE folia Fran, 12. Distribution of Crataegus: C. scabri- ch., C. pinnatifida Bunge. in yeti C. crus-galli sens. lat. (Fig. 15) 18) Gib dein isiad (C. punctata + C. collina) (Fig. tn United ^ Americanoid,' nor of southeast- Smaller ser tates relationship. Finally, some E53 te concentrated in the middle lati- | the eastern United States must be re- —— l€— — dl ———— UND —— —— —— ————— — — P n" es Corded. -b noted Noe Triflorae, Intricatae should mape | OPE of these, C. harbisonii (Fig. 19) is i dications that all of these +h, fou ary represent descendants from fairly mericanae’—Flavae introgression. Sev- lomic ria North America pairwise, the taxo- i ationships are as follows: Eur a western Asia/eastern Asia similarity except slight overlap at in- ODE | zm erem Asia/North America imi arit oe 0 arity, except C. marshallii | ‘tern Asia/North America PHIPPS— CRATAEGUS 683 ' RE" us RA 7 ue d EM ^ ! ^ | C. MOLLIS (T. & G.) SCHEELE M NN D riot vt 1 ] — Os, a C. MEXICANA moc. & sESSE ~ ; EA * DUM io . Sie Lu E ee a ee, iH C. MEXICANA moc. & sessé, C. MOLLIS (T. & G.) SCHEELE, C. DOUGLASII LINDL. jme pee sl | Ficure 13. Distribution of Crataegus: C. mexi- cana Moc. & Sessé, C. douglasii Lindl., C. mollis (T. & G.) Scheele. —several similarities: Sanguineae, Ameri- canae, C. scabrifolia, C. mexicana With regard to areas of species' richness, east- ern North America as a w ole stands out as far richer than eastern Asia, western North America or Mexico (Figs. 20, 21). This cannot be ac- counted for merely on the basis of differing taxo- nomic resolution of species’ limits, especially be- | ^w nm UV made for North America by substituting single series for whole sets of species. The extreme rich- ness of Appalachia, however, is partly an artefact of the size of grid square chosen since lowland í ` > 1 la in a cingle aah (sout )st y commingl grid square with predominantly northern (here mid-Appalachian) species. The biogeographic results are rich interpre- mapped). Northward the genus may be largely mean — 10°F isotherm, +10°F in North America (Fig. 22a, b) or, per- haps more adequately, by the southern edge of the coniferous forests. Southeastward, in two con- 684 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 C. AESTIVALIS (waLT.) T. & c. , C. OPACA Hook. FIGURE 14. Distribution of Crataegus: C. o, Hook., C. aestivalis (Walt.) T. & G. SS a as C. sect. Aestivales. tinents, warm climates (? mild winters) and sub- pec : (Fig. 10), and that not in Hainan, while central Florida represents the southeastern range limit for the United States. Although one is tempted to relate the southeastern continental limits of Crataegus to climax habitat type (trend to ev- ergreen forest), in view of the success of other maloids in such areas (e.g. Photinia) this may Obscure a more direct physiological relationship, perhaps that of a lack of adequate winter chilling requirements southward. In the southeast and southern coastal plains of the United States, two ecological-adaptive groups stand out. One is the somewhat shade-tolerant tree-like group of species (e.g. C. phaenopyrum, C. marshallii, C. viridis, C. brachyacantha) of forest, often river and swamp-forest, margins, whereas the other is the more sand-plain habitat characterised by the sect. Flavae. It should be noted that sect. Flavae are trees and shrubs that appear relatively shade-intolerant like their more northern relatives. In cold, semi-arid areas (cen- tral Asia, central United States cordillera —see Figs. 20, 21, 23a, 23b) there is a virtual hiatus A a t j AN Q C. BRAZORIA sARG ~C, GREGGIANA EGGLEST. i IA sar ia ca Bo VE p i , a e vU 4 1 Ge B C. CRUS-GALLI L., C. VIBURNIFOLIA sam. = GR | C. BRAZORIA sarc., C. GREGGIANA EGGLEST - C -gr FiG 15. Distribution of Crataegus: giana Tages, C. viburnifolia Sarg., C. us Sut. C. crus-galli L. : i oit- of Crataegus, indicating little success in expl rainfall maxima, hot sum whereas no such ecological or taxo terpart exists in North America species— C. scabrifolia of Yunnan pet. icana of Mexico-Guatemala—are sm tad with thorns of indefinite growth, narrow, lobed leaves and large pink to yellow d ge occur in similar latitudes (about 20° to pio v broadly similar climatic regions at ur tum e edge of Crataegus inu We wi these apparent vicariants lat ! The cites occupied by AF infrageneri "ed discussed are shown on the map (Fig. ^'^ nomic cou itical CHARACTER STATE POLARITY AND OUTGROUPS ubfamily The maloids (x = 17) are a natural $ which 8 of Rosaceae whose fruit is a prd closes fleshy hypanthial wall overlaps an nmi over the carpels during fruit de velop ling (1964), in discussing rosaceous natural anatomy, agrees that Crataegeae are anml ah n Mespilus were used for PHIPPS— CRATAEGUS l a ape m 7 uet AN KR. [ dl-———— I er : g |j 4 ty n QT og P eg E ( E | : | | EC f i cae NC —i ; e a. a C. CHRYSOCARPA ashe, C. MARSHALLII EGGLEST. ——— a>, bue C. MARGARETTA ashe, C, SALIGNA GREENE | E ER - pida NM | p. o nS sya i —— i ety a - ' i v. (o2, iC. MARSHALLII EGGLEST. Ficure 16. Distribution of Crataegus: C. saligna e, C. margaretta Ashe. FUP. The crataegean genera, Crataegus, Pyra- cantha, Cotoneaster, uM geese Osteomeles, outgroup analyses.’ characteristics (Table 8) are tendency a wall, and bony exocarps generating in es. Mespilus is the most distinct genus and “ts strongly from other maloids in having ich larger leaves and a single-flowered, large- kno og à mi integer hybrids and a graft-chimaera differ in numerous characters of hab- Dhol een-ness, gross leaf-mor- E paring bud, inflorescence (many char- of the fruit, Mere se a to Chacnameles, Cy- has a bag , “hotomanthes may prove to be crataegean but e different fruit. FiGure 17. Distribution of uerus C. chryso- carpa pen C. marshallii Egglest donia, Malus, Pyrus, Sorbus, Photinia, Amelan- chier, Raphiolepis, etc. will underscore this. It is not immediately clear, on a phenetic basis, which of the Crataegeae is the best genus to use as a sister group and, indeed, caution is advised due to the potential for hybridization in the subfamily and the possibility that this existed in creating ancestral Crataegus. The phenetic-bio- geographic patterning within Crataegus as dem- O b y 1Ong 1SO- lation of major sections of the genus and the verified. Indeed this is true for the Maloideae as a whole and very much so for the Crataegeae (Fig. 25). The small Andean genus Hesperomeles, for instance, with one or two species in Costa Rica- Panama, is related to the main core of the Crataegeae. However, it is not very diverse and represents the only native southern hemisphere maloid. It has a specialized distribution (pygmy "ph forest and rocky outcrops at ca. 3,000 m). It is most parsimoniously derived from other Crataegeae as a relictual genus tenaciously hold- gas niche but somewhat anciently crossing Beringia with a rapid move south coin- PAR SN See hh ! C. PUNCTATA Jaco. , C. Ps à JACQ., C ALIM CHAPM ia Caec D el | 3 | FiGure 18. Distribution of Crataegus: C. punctata Jacq., C. collina Chapm. | —.. = rege ager eir \ be T i | | ; | | " | d 1 f, a . + ee $ | s 1 C. HARBISONII sec, C. MCRSPERM asie | MEZ Y | Pp — ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 DENSITY OF CRATAEGUS TAXA IN E. AS IA 2 : NO, OF SPP. OR SPP. GROUPS /640 KM GRID D 4 HEAVY LINES = EDGE OF RANGE k + ur FicurE 20. Map showing richness of species (no. per 640 km?) in eastern Asia. DENSITY OF CRATAEGUS TAXA IN. ie ae l| N. AMERICA f No. OF SPP, OR SPP. GROUPS/40 KM? GRID UT i! | HEAVY LINES * EDGE OF RANGE Wd ——À FIGURE 19. Distribution of Crataegus: C. macro- sperma Ashe, C. harbisonii Beadle : species (09 FicungE 21. Map showing richness of ; per 640 km?) in North America. | PHIPPS— CRATAEGUS 687 i N 2 1| so aa d E. ASI p Y e P d ISOLINES AT 10 F- INTERVALS A A s VY,’ a MEAN JANUARY TEMPERATURE IN | A Uu E. ASIA ANNUAL PRECIPITATION (INCHES) T ASN FIGURE 22a Jan z 4 AA. . uary hr eastern Asia; Note 10°F. FIGURE 23a. Asia. Note arid regions and summer maxima. ean annual precipitation for eastern MC N 7 iy D 5 NK AY a x x mn = tf “ soum or Line 7 E, | vun zone IN SER ft^ . .| NORTH AMERICAN rum : n EE i : ANNUAL PRECIPITA B c X. SOLINES AT 10°F inTeRVaLs EY e (INCHES) m mod = i = EE E - ~an = és el | 3 for Nortt | America. Note 10°F. FIGURE 22b, January isotherms. Mean monthly data Figure 23b. Mean annual precipitation for North America. Note arid regions and winter maxima. TABLE 8. Technical characteristics of the genera of the tribe Crataegeae. Stamen Distribution Habit Thorns Leaves Inflor. Flowers Petals No. Pome Seeds Others No. of Species Mespilus small tree thorny short simple, large single, on large, white round 30-40 medium; 5 pyrenes Europe, W. shoots lea open; Asia 1 shoots brown Crataegus small shrub thorny small, un- corymb small, white + round 5, 10, 15, small to 1-5 pyrenes N. Temp. to small short lobed to 20 medium; sl. ~140 tree shoots deeply open, or short lobed yellow, red shoots or black & thorns Pyracantha shrub thorny small, serrate; corymb small, white + round 20 small; red; 5 pyrenes 2 ovule/ Eurasia 9 short evergreen narrow closed l shoots Hesperomeles shrub or shoots, ` small coria- corymb small, white round 20 small; red or 5 pyrenes 1 ovule/ a. Hic small tree sometimes ceous, or pink dark berry l America 10 thorn- evergreen tipped Cotoneaster small to large thornless small, entire cyme small, white + round 20 small; red 2-5 pyrenes Eurasia 100 shrub or black; closed Osteomeles tree or shrub slight tenden- evergreen, corymb small, white irreg. 20 fleshy 5 pyrenes E. Asia to Po- to short pinnate obovate lynesia ts (Hawaii) 3. 889 N3G3IVO 'IVOINV.LOS TYNOSSIN JHL AO STVNNV 0L 10A] a — ani — 1983] PHIPPS— CRATAEGUS 689 RHAPHIOLEPIS — STRANVAESIA S IN MALOIDEAE UNBROKEN LINES = SEXUAL HYBRIDS BROKEN LINES = GRAFT CHIMAERAS CIRCLES APPROXIMATELY PROPORTIONATE TO GENERIC SIZE AMELANCHIER DICHOTOMANTHES MESPILUS ERIOBOTRYA HETEROMELES tj PHOTINIA CRATAEGUS PYRACANTHA COTONEASTER HESPEROMELES OSTEOMELES e) d CHAMAEMELES / CYDONIA DOCYN Intergeneric crosses recorded in Maloideae (after Robertson, 1974). FIGURE 24, “ident with a climatic deterioration. A variant of this suggested by D. Axelrod (pers. comm.) Will derive Hesperomeles from C. mexicana or in extinct relative, with similar southward spread. northward migration to Panama-Costa Rica “ould be very recent. Pyracantha, the other truly Fei Crataegeae, has about ten species. It is ve Perate tha by “Pilus is Taurian-Anatolian and is biogeo- 4 Phically, as well as morphologically, well re- ved from the core of the tribe. It is a candidate ut Origin genus. The pinnate leaf, s ized t, and remárkable distribution of Osteo- *5 with its Hawaiian disjunction and sea- dore habitat, permits one to discount it as an- p "acantha thus represents the nearest sister "P (see also Fig. 26, Table 8) and the biogeo- SORBUS ARONIA PYRUS NOMELES ae MALUS PSEUDOCYDONIA CHAE n graphic aspect of adaptive radiation in Pyracan- tha-Crataegus-Hesperomeles is most efficiently postulated from a southwest China base with primitive Crataegus emerging in that area, and now being represented only by C. scabrifolia (Fig. 27). Diversification to Western Eurasia has prob- ably only occurred once or twice on the basis of present biogeography and the existence of major barriers in Central Asia, and in this one may consider the early Tertiary Turgai Straits. The Beringian route has been available geographi- cally throughout the Tertiary but available cli- matically only intermittently (Wolfe, 1977), per- haps only in the Eocene, iddle Miocene, and in certain interglacials. The major groups of Cra- taegus are preponderantly different in North America from Eurasia, which together with the diversities in North America itself are clearly indicative of several independent movements Asia to North America via Beringia, followed by southward range collapse in both continental peeo, correlated suites of characters found 690 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vo. 70 m r 20° w [4 Ma 100" 120° no 160° 180° 180° 140° 120° 100° 80* [d mx "OSEE DNE, 7- SIE J| NN : z È RAT : : 3 hr T , K FESS S. ae Á ie pue ry q Rl MN - g " . E D A . — or dw WM NUM 3 p ghi a b tue Ee SR si pe ey A Ee «| i a eas NEN ^ P Á i a, T ^" » i p è VAL | =e A va o A i à on 2 -a ern € umm Ls k m hL m yx ———H \ z en hy 2j um MA ad | 1 ate | DISTRIBUTION OF "dis LEGEND TRIBE CRATAEGEAE ——— COTONEASTER (100) | SEN sasra (60) 2 = — e SALUS soph teats Be AMER , (100) EZZA HESPEROMELES (10) Lid v 20° 40* 60° ar 100° 120° uo Mercator Projection uL EYRACANTUA (8) 180° MESPILUS (1) ——— OSTEOMELES (3) FiGURE 25. Distribution map of crataegean genera. in the majority of crataegean genera are consid- ered ancestral, and, in cladistic terminology, are symplesiomorphic. Apomorphic states are logi- cally derived. Figures 29 to 31 summarize dia- grammatically certain presumptive evolutionary trends in Crataegus, and Table 5 presents the total character set used. CLADISTIC RESULTS Figure 28 is a computed pro-cladogram for the 75 OTUs used in the taxonomic work. The target diagram format is used because the cladistic predicates permit a omputation was stopped, or the results not plotted, when patristic distance exceeded eight units and where several different fusions of sim- ilar parsimony in longer stems (+2 units) were possible. Thus many of the major groups, viz. Chinese and Siberian alliances (including sect. i uth- Sanguineae), and other American and so eastern United States taxa are no central core of the genus. However, the ee centre, including C. scabrifolia and C. -— sit is tied in. Azaroli and Oxyacanthae, ho e seem T-— reliably connected to the co the gen Theiefor, with some exceptions Riot gram proposed unites only the tightly : pue re In doing so, it strongly reflects the phe sults and it also conforms with the biog my the results. — to mindlessly rH ped cladogram through to completion at in my Sides mu be a fundamen iw the relatively few characters I pov the clado- ntal mistake Had cladistic parsimony), in general, o be opment of thorns, of leaf-shape, e t tied into thè eee À— anm | a — 1983] tations, etc., or whether a parallel ‘Americanae’- Sanguineae evolution took place independently on the two continents. In view of the weak re- lationship of * 4mericanae to southeastern United States taxa and to C. mexicana, and in view of the fairly numerous Neogene Crataegus records (of various leaf shape) from the west of the con- tinent, extra Beringian migrations represent an attractive viewpoint. However, the pro-cladogram as currently set up does not make the Sanguineae-‘Americanae’ connection very attractive; but if the ‘America- nae’ were radially raised on the pro-cladogram (=additional phies) a link fi i 7 ) like C. kansuensis, C. maximowiczii or C. au- 'antia becomes extremely attractive. In lieu of attempting the difficult task of polarizing a sig- macant extra number of morphological charac- lers, in the short term it is planned to substitute the use of phenolic assays in the clarification of group structure. These data have been found use- ful in Maloideae by Challice (1981) and Sinnott and Phipps (1983) Note that as fusions are made between cur- id unconnected groups on the cladogram this ae hive? rit bison counted but : are ev- idently highly numerous, but this is conformable With our biological TEENEI of Crataegus reproductive behavi nother reason rez etn in accepting long computed patristic distances lies in the polyploid and hybrid nature of Maloideae. The tribe is of E alloploid origin (maloid x — 17 from 9 ic unions as computed i in standard programs might totally override and obscure actual hybri- dization events. This reinforces the rationale for leaving the published situation at the develop- menta] Stage of the pro-cladogram in 1982 until new data are analyzed. DISPERSAL ABILITY » Pomes of Crataegus, as well as those of other milar berry. like fruit about 1 cm size, are com- monly considered to be bird-dispersed (Phipps Muniyamma, 1980). If this is so, then the min for long-distance dispersal (LDD) im- ely becomes a prominent consideration. PHIPPS— CRATAEGUS FiGure 26. Pyracantha crenulata (lower), Cratae- e. SEA Med left) and C. scabrifolia (top right) e leaf-shape and spiny short-shoots. What is remarkable, however, is the paucity of hard data on Crataegus dispersal (e.g. DeBoer, 1979). Information, such as it is, is related to what eats Crataegus berries, and not how far seeds may travel in the gut of the fructivore. There is no question that several species of pas- serines and game birds eat Crataegus fruit. They may also do so in the gutuma, while on migra- tion B ds g g y lose cour SC, due i bl by being driven. Numerous records attest to such birds ending up on the ‘wrong’ side of the Atlantic or Pacific Oceans. However, were such LDD to be a sig- nificant cause of Crataegus distribution patterns then it is inferred that taxonomically like kinds of Crataegus would be more commonly found than is actually the case on opposing sides of the of Europe), and possessing the smallest size of Crataegus fruit (3 x 6 mm), is aided by LDD concepts for a parsimonious explanation of its ocation. The extreme differences between the North American and European Crataegus floras indicate the effective closure of Atlantic route to — ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 sn eur e 20* 40* 80* 60° 100* 120° uo 160° 180° 160° 40° 120° 100° 80° $i ow i F NUM | | vw. | m i N 4 a H A. B oe P " J at M " N | c V 30 Q E > L S Ju ¥ s 2s, "e Z Eoo b i 9 m. Ho ES | l'OXYACANTHAE pos. p N* we ER i —_ t § (CA. 60)8 b—| ^t AZAROLI (10 se ME C. MEXICANA c. MARSHALLI I | | MAJOR GROUPS OF CRATAEGUS SHOWING POSTULATED MIGRATIONS B DE $ E B: = FK 5o» 33 ee a 100° 120° Mercator Projection uo B w «=O w ur wr M LN wu er trace the FIGURE 27. Major groups of Crataegus showing postulated migrations. Non-bracketed numbers sequence Crataegus during most or all of Crataegus evo- lution. In view of the Atlantic situation, the East ent, rodeni discounted. Both ungulates and ro- ents eat Crataegus fruit and seeds but they can hardly be agents of LDD. Thus, short-distance dispersal (SDD), the ef- fects of which are abundantly observed in nature, for instance in the rapid establishment of new Crataegus populations, is certainly adequate to explain dispersal and radiation between the two northern continents. Short-distance dispersal through Beringia thus remains the only serious candidate migration route and it is proposed as occurring whenever the Beringian climate per- mitted Crataegus to migrate to and thrive in the Chukchi-Alaska region. PALEONTOLOGICAL CONSIDERATIONS The Alaskan-Asian contact in Beringia has re- mained essentially unchanged geographically, from the standpoint of this paper, since wes Tertiary. Rotation of plates on an Alas sies crum, terrane accretion, etc., are of great ips to geologists but of less significance to this The sea gap, however, is of more con and its exact extent in Beringia ui pat factors that affected timing of any Crataegus SDD. Clearly, however, ne ent factor for the Beringian route is paleoch™ di nd It is, therefore, posited that ar er rican or east Asian Crataegus approaching using SDD. The Tertiary and Quaterm climate of Beringia are, therefore, © est. Firstly, Beringia has been at hi throughout the Tertiary. Indeed, d inter- MM ee aa - ed — A a C LTT — — e oo A LEGEND gh ers eu SCALED SEMICIRCLES as DAMURICAM— ÉRA INERDIANAE | Le NO. OF ADVANCED CHARACTER STATES UE * bots = COMPUTED PROCLADOGRAM TU'S & COMPUTED FUSIONS (HTU'S) ? = COMPUTATIONAL RESULTS AMBIGUOU: FOR CRATAEGUS, 1982 = NORTH AMERICAN LINES = EURASIAN LINES 60 N L| 57 Te y AX 9 3 MACRACANTHAEN, A Y * \ "d X — 3 @ 88 TRIFLORA | NS -n s e. e 81 2) 27 PENTAGYNA / n Saga — IN CRUS-GALL I - m. ty x [^ ANCESTRAL : RNS; LEAVES NAR a ain n by te VERY PUBESCENT X FRU > YELLOWISH; WARM desc FIGURE 28. Pro-cladogram of 75 species of Crataegus based on 18 characters. (OTU numbers cross-referenced to Appendix A.) £69 694 POLARITY OF THORN CHARACTERS WITH ENCODING THORNS NOT NUMBER SHORT SHOOTS SHORT-SHOOTS NUMEROUS (1) NOT THORN- TIPPED (1) DEFINITE GROWTH (1) > 6 CM (4) 4 - 6 CM 0 | 2 - 4 CM (2) 1-200) Joo INDEFINITE DO NOT OCCUR (0) DO NOT THO! OCCUR (0) TIPPED (0) GROWTH (0) RE 29. Diagrammatic de ape ag, of polar- ity in Eua characters: (a): t Eocene, the North Pole was in the (present) Chukchi Sea. Data on northern Hemisphere cli- mate changes, however, indicate optima in the Eocene, middle Miocene, and possibly, certain interglacials (e.g. Lamb, 1977). The northern limits of extant Crataegus are, however, well to the south of Beringia, and these seem to be cold- hardy, phyletically advanced species (e.g. C. san- guinea, C. aurantia, C. maximowiczii, C. chlo- rosarca, C. altaica (Asia), and C. douglasii, C. chrysocarpa (N. America)). It is inferred, therefore, that Crataegus crossed (and perhaps re-crossed) Beringia a limited num- ber of times, during climatic optima. The Alas- kan warm-climate interludes are well discussed by Wolfe (1971, 1977). A possible scenario is: a) Eocene event Ancestral Crataegus to Beringia, followed by range collapse southward on both continents with the vicariants C. mexicana becoming an ancient relict now settled in Mexico and C. scabrifolia in Yunnan. b) Miocene event Ancestral sect. Sanguineae types, e.g. C. max- imowiczii types to Beringia, likewise followed by range collapse southwards and extensive radia- tion of ancestral and descendant taxa on both sides of the Pacific Fossil benchmarks would help secure the dates telescoping of the proposed time-scale is possible as long as event (a) is not much later than Mio- cene on general evolutionary considerations. ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 POLARITY OF LEAF CHARACTERS VEINS TO NUMBER OF LOBING BREADTH SINUSES LATERAL VEINS DEEP (3) BROAD (3) PRESENT (2) FEW (2) NONE (0) NARROW (0) NONE (0) NUMEROUS (0) RE 30. Diagrammatic se iret of polar- ity in pan at characters: (b): le FossıL BENCHMARKS Fossil benchmarking appears to require a crit- ical re-evaluation of all proposed Crataegus fossils. It is noteworthy that some for-a-period well-accepted Crataegus have been transferred to genera as taxonomically unrelated as Alnus (e.g. Wolfe, 1971). However the collected pe of reputable palentologists (e.g. Chaney, 192 Lamotte, 1936; Oliver, 1936; Maori 1934; Wolfe, 1977) etc. are bound, in the opin! no this author, to yield a number of genuine ‘Cra taegus records at least as far back as the -— cene. Indeed, the uniq morphology of som Crataegus makes it certain po some existing records are correctly attributed to Crataegus v indeed, they match extant genera. Projection es to the Eocene (e.g. Lamotte, 1952) is far of secure and this period possibly saw the origin 0 Maloideae. DISPERSAL OF OTHER MALOIDS Of the maloidean genera only five penc Sorbus [only subg. Aucuparia], Malus, chier, and Photinia) occur in crocontinents (Table 9). While eis nomeles, Cydonia, and some others same general maloid pomes are small (about u the sane as Crataegus in variability). Their guine ef tribution likewise must surely indica ae er le tribe or poor LDD. Obviously the latter yp. esis conforms best to the Crataegus Indeed for the entire Maloideae, p g à sni tions are cogent for only three species BERMUDES nm 1983] POLARITY OF SOME FRUIT CHARACTERS CHARACTER SHAPE COLOR INDEX CODING 6 O1 NARROW BLACK ADVANCED 2 | RED MIDDLE 1 Y [ BRCAD YELLOW PRIMITIVE 0 9 o» FIGURE 31. Diagrammatic representation of polar- ity in Crataegus characters: (c): fruit sus marshallii, Chamaemeles coriacea (endemic to Madeira), and Osteomeles anthyllidifolia (Ha- Wali) although this latter, disjunct from eastern could have an anthropogenic explanation. TAXONOMY OF UNITED STATES CRATAEGUS Ee I. variety of apparently not closely- ^. orms restricted to the southern United 5 (e.g. Microcarpae, Aestivales, Virides, Fla- "te, Brevispinae) is indicative of a relictual flora t had diversified long in the past. The above LE offer little parallel to Asiatic forms except E Ypothetical symplesiomorphies between "ag and C. hupehensis. The weak relation- open southern United States groups and 5 North American forms, e.g. ‘Americanae,’ oh gang and C. mexicana has already been - It is fascinating to be able to postulate a oup of southeastern United States relicts. TAXONOMY OF CHINESE CRATAEGUS sg Chinese species are fairly closely inter- ; In two subgroups, the northern Sangui- "eae allia n 3 +} LH thern ka 1 1 "E" H- Suensis alliance. It is these two sub-groups that lc viaa closely to North American Cratae- la of c " E oup *Americanae.' Crataegus cunea- faction eastern China, however, is very dis- this ha other Chinese taxa, as is C. hupehensis, having a remarkable leaf-shape and h i. g to be thornless, whereas C. scabrifolia di o. other Chinese species as already EMEN nee a e r PHIPPS— CRATAEGUS 695 TABLE 9. Listing of maloid genera geographically (showing numbers of species per genus). Euras. Am. Tribe Crataegeae Mespilus (1) 1 - Crataegus (150) 55 95 Pyracantha (8) 8 E Hesperomeles (9) — 9 Osteomeles (3) 3 — otoneaster (100) 100 — ?Dichotomanthus 1 _ Tribe Sorbeae Docynia (5) 5 — Chaenom incl. Pseudocydonia 5 = Cydonia (1) 1 Malus (45) 40 5 x Malosorbus (1) 1 E Eriolobus (1) 1 — Pyrus (45) 45 = Eriobotrya (30) 30 — Photinia (50) 45 5 Heteromeles (1) E 1 Stranvaesia (6) 6 = Rhaphiolepis (8) 8 - Sorbus (125) 120 5 Aronia (3) — 3 Chamaemeles (1) 1 - Amelanchier (18) 6 12 Malacomeles (2) — 2 Peraphyllum (1) — 1 east-west Beringian Crataegus migration. Cra- taegus tangchungchangii is inadequately known. SUMMARY Phenetic and biogeographical analysis shows a very strong patterning of Crataegus species- regions. Among these, Azaroli and American ‘Americanae’ in spite of usually sep- arating slightly when subjected to numerical taxonomic analysis. In the United States and Canada a varied group of southern species dis- e M scabrifolia and C. mexicana, without true thorns, 696 ANNALS OF THE MISSOURI BOTANICAL GARDEN BERINGIAN TIME-SCALE AND SUMMARY CLADOGRAM FOR CRATAEGUS, 1 RECENT NORTHERN NORTH AMERICAN DIVERSIFICATION ^ AMERICANAE ' PLEISTOCENE BERINGIAN GAP PLIOCENE = EUROPE MIOCENE SANGUINEAE - E. ASIAN ETC. oe OXYACANTHAE < v < ó o = des a w = ' OLIGOCENE [ (RS ARP 4 1} SOUTHERN U.S, AZAROLI C, MEXICANA T [4 ANCESTRAL CRATAEGUS FIGURE 32. Simplified vicariance Votis ai illus- EOCENE Yp1dlidtiU11J. Cladistic analysis is based on generating ple- siomorphies from the natural tribe Crataegeae and indicates remarkable similarities between racantha, C. scabrifolia, and C. mexicana. The cladistic algorithm was not pursued to comple- tion due to ambiguities about how major groups might be connected. These ambiguities are rein- nee by knowledge of potential for wide hy- tion among Maloideae and also, specifi- es within Crataegus. Biogeographical thinking is thus brought into the cladistic argument, the rea rises resting strongly on the very low likelihood o ae plausible uc events are noted. ias makes [Vor. 70 BERINGIAN TIME-SCALE AND SUMMARY CLADOGRAM FOR CRATAEGUS, 2 RECENT NORTHERN NORTH AMI PLEISTOCENE ERICIN DIVERSIFICATION ' AMERICANAE' Rc UTHERN U.S, DIVERSIFICATION —| (VARIOUS PELICTUAL TAXA) PLIOCENE 7 E. ASIAN ETC. BERINGIAN GAP MIOCENE SANGUINEAE -= MEDIT. TO C. ASIA OLIGOCENE | OXYACANTHAE - EUROPE | AZAROLI LL ———— r lag C. MEXICANA SCABRIFOLI ANCESTRAL CRATAEGUS FIGURE 33. _ Simplified i rego ia EOCENE Jr nation). late Eocene can stand as the midpoint of several possibilities. ith is than left with two broad C , “Americana? being deriv only one successful Beri The procladogram currently favors on gradistic grounds. This may well be ch by further evidence (and is challenged is Alaskan migration potential). Not prn s known about s some v peculine ar specie whic licts be i xa, or re during the Tertiary and oeli climatic optima for SDD to move Crataegus throu Beringia. But oe frequency of this is considered relatively law ber of similar Crataegus phylads on both sides of the Pacific. Fossil evidence, altho ough requiring modern updating, suggests plentiful reir in North America from the Mid-Tert rtiary. The or- igin of the genus is hard to time at present but of older radiation. But not one of these bears 07 the choice between the two models P Land above. A collapsing of the time-scale, whic prove necessary, as evidence accumulates, not of itself affect the form of the dade. Thus, Crataegus has changed, due to ™ um researches, from a genus in which neu PE san to, in some senses, the very opposite —— 1983] digm of an SDD genus with long life history, evolving mainly during the Tertiary, and evi- dencing both evolutionary bursts, probable pe- riods of phenetic gradualism, and, if our feelings about the proposed relicts are correct, almost total stasis in some cases. Overwhelming evi- dence from phenetics, modern biogeography, re- productive biology, dispersal biology, paleo- geography, paleoclimate, fossils, and outgroup analysis, generates a model of Crataegus evo- lution that is the clearest to date, and, in fact, totally novel. LITERATURE CITED BRowicz, K. 1972. In P. H. Davis (editor), Flora of : Turkey 4: 133-147. WATT, I.E 1975. Hybridisation between Crataegus . The struct ture e of some Crat soep pop- astern ASIL ulatianc Belgium. Watsonia 11: 105-115. 7——. 1976b. Pollen morphology of some e European species of Crataegus L. . MunRAY. 1 tergeneric hybrids between Crataegus L. and Mes- L. a ag JE at an old problem. Bot. J. E n. Soc. 74: 329-343. € ICE, J. le duum ed studies in the E mily Rosaceae and the evolutionary origins of : d ami Maloideae. Preslia 53: 289—304. Iv ^ . 1927. Contributions to paleontology. P cology and paleontology of the Crooked Riv- sin with "s Sal reference to the Bridge Creek CHA; : ; 1271. Sidik Pribaltici. Cra- De Bos Baltici. Riga. z Russian. th R J.I. 19 rsal of Crataegus L. (haw- orn) in iwel Ontario. UWO Plant Sci- D *nces Dept. Honours Dt (unpubl.). & J. B. Purp E A. 1984. Taxonomic XIV. € s In Crataegus L. osora Maloideae) bot short leaf heteroblasty in Crataegus ... Tusgalli L. , Sens, lat. Canad. J 4. (in press.) phy — Medi Tax studies i ataegus L. (Rosa e: Maloideae). XIII. Pat- ^. of en variation in Crataegus sec- 9n Crus-galli in On r E — submi Taxonomic studies in aegus A (Rosaceae: Maloideae). XV. The System of Crataegus crus-galli L. sens. A lat. in Ontario, Ame J. Bot. . The taxonomic significance of Jahrb, E be in Crataegus (Rosaceae). Bot. M yst. 101: 457-469. & A. A. Bopawt. “ance of chromosome n mbers we ; Phytol 35: 271-275. (edit 1968. Crataegus. In T. G. Tutin et al. ors), Flora Europaea 2: 73-77. 1977. Taxonomic pe hy PHIPPS— CRATAEGUS 697 GLADKOVA, V. N. 1968. Karyological studies on the edik. (Ma- 1965. Contributions to the tax- onomy of Crataegus. Milwaukee Public Mus. Publ. Bot. 3 0t:3. LAMB, H. H. 1977. Climate, Past, Present and Fail- ure. McEwen, London. Lamotte, R. S. 1936. Contributions to paleobotany. V. The upper Cedarville flora of northwestern Ne- vada and adjacent California. Carnegie Inst. . 1952. Catalogue of the Cenozoic plants of North America through 1950. Geol. Soc. Amer. Mem. 51. LoNGLEY, A. E. a ea studies in the genus Crataegus. int J. Bot. 11: 297-31 LoupoN, J. C. 1838. Arbo did et Fruticetum Brit- tanicum. Lon Love, R. A M. FEIGEN. 1979. Intraspecific hybridisa- tion between native and naturalised Crataegus (Rosaceae) in western Oregon. Madroño 25: 211- MacGinrne, H. D. 1934. Contributions to paleo- botany. II. The Trout Creek Flora of southeastern ee Carnegie Inst. . 1952. The Kilgore Flora. A Late Miocene Flora from Northern Nebraska. U.C.L.A. Public. Geol. Sci. 35: 67-158. McGeary, S. E. & A. BEN-AVRAHAM. 1981. Alloch- thonous terranes in Alaska: implications for the structure and evolution of the Bering Sea shelf. 0. Crataegus. In Flore d'Afrique du Nord 15: 130-142. Publ. Lechevalier, Paris. MUuNIYAMMA, M. & J. B. PuiPPs. 1979a. Pollen mei- osis and polyploidy in Crataegus. Canad. J. Genet. Cytol. 21: 231-241. 1979b. Cytological proof of apo- mixis in Crataegus. Amer. J. Bot. 66: 149-155. & ———. 1984. Studies in Crataegus. pier ——— Amer rica | hawthorns s. Canad. J. Bot. 64. (in Sadi OLIVER, E. Bo Contributions to paleontology. I. A Miocene flora from the Blue Mountains, Oregon. Cara ie Inst. PALMER, E. J. 1953. Crataegus. In H. A. Gleason, New Britton and Brown Illustrated Flora of the Northeastern United States and Adjacent Canada II: 338-375. i : 1975a. Kilometric distance. Canad. J. Bot. 11: 1116-1119. . 1975b. Bestblock: optimizing grid size in bio- geographic studies. Canad. J. Bot. 53: 1447-1452. Sie Cr — d nomenclator for section names. Taxon 32: 598-604. Asie Problems d hybridity and cladistics n F. Gran Douce. LO.P.B. iak cademic in Crataegus. In Conference, Montreal, oe 1980. i of Cra- . Canad. J. Bot. 58: 1621-1699. 1939. Crataegus. In V E Koma- 416-468. Botanical dece Ar AT. MUNIYAMMA. POoJARKOVA, A. I. 2 Flora of the USSR IX: 4 t. USSR, Leningrad. 698 REHDER, E. 1940. eren yt Cultivated Trees and Shrubs merica. 2nd edition. acmillan, New York. . 1969. Siia In K. A. Rechinger (ed- itor), Flora Iranica 66: 49-65. ROBERTSON, K. R. 1974. Genera of the Rosaceae in the southeastern United States. J. Arnold Arbor. —654 . 1965. Introdutsironavye boyarit- schniki botanicheskogo sada an UzSSR. Jn Den- drologie Uzbekistanii. (In Russian.) SARGENT, C. S. 1902. Silva of North America, Suppl. XIII: 31-184. Cambridge ——. 1903. Recently recognized species of Cra- taegus in eastern Canada and New England, I. Rhodora 5: 52-60. SCHNEIDER, C. K. 1906. Illustriertes Handbuch des rr ie ine i bcm. 1001-1008. Fischer, TB OM Jen Sevorr Q. Pe 1983. Variation pat- 1980. Phenology of ring as a mechanism for repro- ductive isclation among related taxa. Botany 80. Vancouver. Abstract. STANDLEY, P. 1922. Crataegus. In Trees and Shrubs of Mexico, pt. 2: 335-336. United Sales National Muse STERLING, ra 1964. Comparative morphology of the rpel in the Rosaceae. III. Pomoideae: Crataegus, repeti Mespilus, Osteomeles. Amer. J. 51: 705-712 TIDESTROM, I. 1933. In H. E. Small, Manual of the Southeastern Flora, ed. 3: 637—644. Wo tre, J. A. 1971. Tertiary climatic fluctuations and methods of analysis of tertiary floras. Paleogeogr. Paleoclimatol. Paleoecol. 9: 27-57. 1977. , Paleogene Floras from the Gulf of à e 1974. Crataegus. In T. Yü, Flora ' of the Peoples' Republic of China 36: 186-206. APPENDIX T RECTE bur OF D CRATA Arranged by sect ion and series. ? — assignment to this section or with same number (e.g. 62, 62b, etc.). Sources: —— floras cited in text except for ooo sd (Harvard University Herbaria), supplemented by a riety of local cae = — used in biogeographic Suavis *- used in taxonomic studie Sect. MN Loud. Ser. Mexicanae LUN ehd. 1*1. C. mexicana hx & Sessé Ser. ‘Henryan T2 C mer ia DRAN ) Rehd. Sect. Oxyacanthae Loud. ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 Ser. Oxyacanthae (Loud.) Rehd. 2: oe a Jacq FA: C. eura Lindm. (7 C. calycina auct. non Pet A D pne arr Hegetschw. 8. C. pseudoheterophvlla Pojark. *9. C. laevigata (Poi 10. C. turcomanica Polar 11. C. ambigua C. A. Mey. ex Beck 12. C. volge Pojark 13. C. transcaspica Pojark 14. C. persica Pojark 15. C.caucasica C. *l6. €: trosanguinea fiia *17. C songarica t 18. €. haemo Pojark. *19; €. miopi C. Koch 20. C. stevenii Pojark. 2Lc 6 n des efus ) Pojark. (= C. pallasii Griseb. in FI. 22.: C. ucrainica Pojark. ; j *23. C. meyeri Pojark. series Erianthae 24. C. eriantha Pojark. Pok C 25. . taurica Pojark. —> Azaro T a e + putative hybrids: C. LI Lindm., ico è . heterodonta Po- ubi rea Re . bile ere e C. insularis Cin.) Ser. Apiifoliae (Loud.) R 1*26. C. marshallii dier Ser. Pentagynae (Schneid.) au *27. C. pentagyna Walds 2 ri ARE A dansi Browicz (close to e pentagyna Fl. Turkey) * 29. C. pscudomelanocarpa M. Pop. giam ounterpart to C. pentagyna—Pojarkova Ser. Azaroli (Loud.) R 30. pycnoloba Boiss & Heldr. $31: c aem 32. *33. E tanaceti folia ( kp ) Pers. oe: L. (incl. C. aronia (L.) Bosc.) *34. C. azarolus *35. C. pontica oc 36. C. szovitsii Pojark 37 sinaica Boiss 38. C. dikmensi. s Po "39: C. heildreichii Boiss. Sect. Sanguineae Zabel ex Schneid. Ser. Pinnatifidae Capea ex Sheet Rehd. t40. C. isis P *43. C. jozana C. K. Schn eid. Ser. Sanguine (Zabel ex ned Rehd. 1*44. anguinea Pall. e ie aa 1*45. C. dahurica Koehne ex C. K- Sch 1*46. C. altaica Lange i korol- 47. C. wattiana Hemsl. ex Lace kowii Henry) | 1983] PHIPPS— CRATAEGUS Ser. ED» 1*48. C. maximowiczii C. K. Schneid. 53. C. remotilobata H. “Raik. ex Pop. Ser. (un-named) *54. C. clarkei Hook. f. P55. C. wilsonii Sarg. 156. C. oresbia W. W. Smith Sect, (un-named) 1*57. C. hupehensis Sarg. 158. C. shensiensis Pojark. Sect. Cuneatae Rehd. ex Schneid. Ps . C. cuneata Sieb. & Zuc 60. C. tangchungchangii Metcalf (?) Sect. Cordatae Beadle ex Eggles . 61. C. phaenopyrum (L. f ) Medic. Sect. Virides Sarg. ex Schneid. Ser. Virides (Sarg. ex Schneid.) Rehd. T*62. C. viridis L. 62a. C. glabriuscula Sarg. 62b. C. sutherlandensis Sarg. 62c. C. anames. g 1*63. C. nitida (Engelm.) Sarg. Ser. Pulcherrimae (Sarg. ex Palmer) Palmer 64. C. pulcherrima Ashe Sect. Microcarpae Loud. EC B ae Michx. EN Pimae Loud. Sect. E V M Cu sto: Muenchh. PRA esti . €X eid. 2. c por CIE Á * & G. A ed opac: B oc “Se ; 75. C. maloides Sarg. (incl. C. fruticosa Sarg., C. luculenta Sarg.) “Brevispinae Beadle ex Schneid. . E cis Engelm. & Sarg. Sect. Douglasii "I e. dou Lindl. 7 C. rivularis Nutt is : C. erythr ipods ho e (incl. C. cerronis Nels.) E — Greene (? correctly placed) Sect, pet lo ea ic m. ) Rehd -galli L., sens. lat. (incl. C. fontane- siana Sach. C. bus C. livoniana ii Sarg., ^ ; C. pia Beadle, etc., etc.) la. C. reverchonii Sarg. 81b. C. cherokeensis Sarg. 81c. C. sublobulata Sarg. she . C. berberifolia T. & G. 183. C. baroussana Egglest. (?) 184. C. parryana Egglest. (? Ser. Punctatae (Loud.) Rehd. . C. punctata Jacq. 185a. C. collina Chapm. 186. C. brazoria Sarg. (?) 87. C. disperma Ashe Sect. Triflorae Beadle ex SUME 1*88. C. triflora Chap 89. austr DENS eet Beadle 90. C. conjungens Sarg. Sect. ‘Americanae (El-Gazaar)’ ies Bracteatae (Palmer) Rehd. C. ashei Beadle e C. harbisonii Beadle 93. C. pearsonii Ashe (?) C xoa UR Sarg. Ser. Molles (Sarg. ex Bap Rehd. T Ser. (un-named 194. C. mollis (T. & G.) Scheele 97. C. pensylvanica Ashe 'egiana Egglest. 1599. C. submollis Sarg. pee Coccineae aon ) Rehd. 3 104. C. fulleriana 2 *105. Ser. Tenuifoliae dera ex i Resear) Rehd. e 1**106. C. macrosperma As 10. C. lucorum Sarg. 1*111. C. flabellata (Bosc.) K. Koch Ser. aE (Egglest. se haste Rehd. 1112. C. columbiana How As . flavida 1116. C. margaretta Ashe 117. C. irrasa 118. C. aboriginum Sarg. 119. C. laurentiana Sarg. +Ser. Intricatae (Sarg. ex Schneid.) Rehd. *120. C. intricata Lange 121. C. foetida Ashe 699 700 Ser. Brainerdianae Palmer succulenta Link TSer. silvicola (Palmer) Hel *130- j dle *132. C. suborbiculata Sarg. 133. C. nitidula Sarg. ANNALS OF THE MISSOURI BOTANICAL GARDEN 134. C. compta Sarg. a 139. C incerta Sa. Q others) Ser. Pruinosae (Sarg. ex Egglest.) R 1***140. C. pruinosa Mac TR: Kod. sens. lat. bs Dilatatae (Sarg. Aa Palmer) R 43. C. dilatata Sarg. (= C. basse Sarg.) "A c: EN Ashe 145. C. glareosa Ashe (?) [VoL. 70 THE DISTRIBUTION OF SCROPHULARIACEAE IN THE HOLARCTIC WITH SPECIAL REFERENCE TO THE FLORISTIC RELATIONSHIPS BETWEEN EASTERN ASIA AND EASTERN NORTH AMERICA! HoNc DE-YUAN? ABSTRACT Based on a general survey of the distribution of Scrophulariaceae in the Holarctic, € Asia is shown to be the richest in members of the family among five regions. Western North America has genera (17/35) common to eastern Asia. genera) shared by North America and Eurasia, four distributional patterns are ' recognized: temperate western North America/Mediterranean disjunction (one genus); temperate Eurasia/North America don (three genera); temperate eastern Asia/eastern North America disjunction Ped genera), and continuous distribution through is Bering Strait or Aleutian Islands (six genera). of plant communicati i i i gration routes of some taxa are traced. Six genera are suggested to have migrated from eastern Asia to North America and four genera are considered to have migrated in the opposite direction, i.e., ee phases o — ti ——— — — NI cá — s from western North America to eastern Asia through Beringia or the Bering Strait. The floristic affinity between eastern Asia and ‘astern North America was first noticed by Lin- naeus (in a dissertation defended by J. Halenius; *e Graham, 1972) as early as 1750 and has been lhe focus of much attention by plant geographers since Gray's (1840, 1846; reprinted in Graham, 1972) brilliant works. The subject has been ex- Sel discussed over the past century (see Li, l raham, 1972, for literature). There have been, however, few, if any, discussions on the Subject Pertaining to individual families, prob- ably due in Part to a lack of floristic work on China, the major part of eastern Asia, in the past. E 1s intended to be an initial effort in i The area covered in this work is the Holarctic | 53% of the genera (16/30) common to eastern Asia, whereas. eastern North America has 48% of the | | | | or al Asia (Takhtajan's region 8 I Region), and Europe and the llérranean together (including the Azores). 4 "ec EN | ——— gi uscript and to Dr. A.-M. Lu for his sugges the ss Sora like to express my ee thanks to nai R.-C. Ching, T. GENERAL ASPECT 1 c large family, containing 18 tribes, about 200 genera, and 3,000 species. It is highly developed in the Holarctic Kingdom with 14 tribes, 110 did and over 2,200 species. The tribes Verbas , Paulow nieae, Hemiphragmeae, Ellisioph licae (if in- cluded in the family), and Collinsieae are endem- ic to the area, and the Antirrhineae, Veroniceae, Rhinantheae, and Digitaleae have most of their members here. Of the 110 genera, 72 are endem- e e, and Aragoeae (Hong & Nilsson, Vis )entirely or mainly confined there. Table 1 shows that of the five regions men- tioned above, eastern Asia has the most tribes, the most genera, the most endemic genera, the most species, and the most endemic species. It is the only subregion with endemic tribes (3). The least number of genera occur in western North America and the fewest species (including en- demic ones) in eastern North America. North America, as a whole, is comparatively poor in scrophulariaceous plants, having only nine tribes (one endemic), 48 genera (16 endemic), and ca. 630 species (ca. 600 endemic). -T. Yü, and Z.-Y. Wu for reading liene of Botany, Academia Sinica, Beijing, People’ s Republic of China. ANN, Missouri Bor. GARD. 70: 701-712. 1983. 702 ANNALS OF THE MISSOURI BOTANICAL GARDEN [VoL. 70 TABLE 1. Taxa in the five regions in North America and Eurasia. Number of Taxa in Eurasia Number of Taxa in North America Europe & E. North W. North Taxon E. Asia W.&C. Asia Medit. Total America America Total Tribe 11 (3) 8 (0) 8 (0) 12 (4) 8 (0) 8 (0) 9 (1) Genus 58 (22) 35 (5) 39 (7) 82 (43) 35 (8) 30 (6) 48 (16) Species 700(550) 600(350) 650(450) 1,750(1,650) 160(100) 500(450) 630 (600) Western North America/eastern Asia have 16 genera in common; and eastern North America/ eastern Asia have 17 genera in common. These paired regions share a greater number of taxa than any ofthe other paired regions. Thus, in the number of genera shared, the two regions in North America are both closer to eastern Asia than to Europe and Mediterranean together and to west- ern and central Asia. All the large genera in North America have the great majority of their mem- bers in the western part of the continent, with few species (Penstemon, Castilleja, and Mimu- lus) or only a single species (Orthocarpus) ex- tending into the eastern part. It is of interest to note that all four of these genera also occur in eastern Asia. Furthermore, 5396 of the genera in western North America are in eastern Asia, whereas 48% of the genera in eastern North America are common to eastern Asia. It is prob- ably reasonable to say that the relationships of the Scrophulariaceae between western North America and eastern Asia are closer than those between eastern North America and eastern Asia. THE DISTRIBUTIONAL PATTERNS OF THE TAXA SHARED BY NORTH AMERICA AND EURASIA North America and Eurasia share six tribes (Anthirrhineae, Gerardieae, Gratioleae, Rhinan- theae, Scrophularieae, and Veroniceae) and 19 genera (Antirrhinum, Bacopa, Buchnera, Castil- leja, Euphrasia, Gratiola, Lagotis, Limosella, Linaria, Linderia, Melampyrum, Mimulus, Or- thocarpus, Pedicularis, Penstemon, Rhinanthus, Scrophularia, Veronica, and Veronicastrum) dis- junctly distributed across the Atlantic and/or the Pacific. Since the genera Bacopa, Gratiola, Li- Gratioleae, is considered because it is best rep- resented in the kingdom, particularly in western North America and shows a special relationship between western North America and eastern Asia. Different i l involved in these disjunct distributions. At the tribal level, a num- ber of paired genera are found disjunctly distrib- uted, one member of each pair in North Amera and the counterpart in Eurasia. In the tribe Ve- roniceae, Veronica on one side and Synthyris and Besseya together on the other are such a case. Veronica is a large and polymorphic genus, with its center from the Pyrenees, along the Alps, nO Anatolia, the Caucasus, the Himalayas, and the Pamirs to southwest China and Tienshan, i all eight sections and the majority of pyar di ist. Only a few species (excluding cosmopo gi and weedy ones) mostly belonging to the secti Veronica, extend into tropical Africa, Australia. and North America. Synthyris and Be gether may be its counterpart in North epi (Fig. 1). The relationships between Veron ae the other two genera are so close that no "i ences, except for habit and chromosome Hor n are useful for distinguishing them. dl gia, Lesquereuxia, and theae provide another example. in eastern Asia, the second in Greece i ABE ern Turkey, and the third in eastern Nos pu ica (Fig. 2). Siphonostegia differs from ssl dit two in having pinnatifid leaves and d si are lobes; Lesquereuxia is different from mainly in its opposite leaves. The three related (Hedge, 1978). e 14 genera comm may be grouped into four categor! to their distributional aw 1) Disjuncts across the Atlantic V of us itae in Europe and the Medie y ont gion and the other in North America). Men genus, Anthirrhinum, falls into this categ ^. consists of 30 species, 17 of them occ rhe oth- the western Mediterranean west of Italy; inents on to the two contine" es acco ean " demens 1983] HONG—SCROPHULARIACEAE 703 almost worldwide, the majority of species and all e cosm | | ee | Distribution i the genera Veronica, — Besseya. Although Veronica is distributed and sections occur in the southern Holarctic Kingdom (the mopolitans are not included in in the map). Syris is teni in western North America and Besseya has only one species in ea a. Synthyri s and Besseya may be regarded as congeneric. Shaded area— S distribution center of Veronica "pack prodiit con dotted area — Besseya. | | êr 13 are restricted to California in the United States (Fig. 3). | ) Gen ba disjunct across both the Pacific and the Atlantic (that is with one part of the area in l Perate Eurasia and the other in temperate | North America). In this category are three gen- _ a, Scrophularia, Linaria, and Melampyrum, "i ‘entered mainly in southern Europe and t €*N. ba Ç large genus with ca. 150 species, of which over 100 grow in the region from the Pyrenees to the Pamirs, and with ca. 20 in a small area in southwest China (Fig. 4). In North America the genus has nine species (or only two species depending on species con- cept). The situation in Linaria appears similar (Fig. 5). It is centered in the Mediterranean re- Ficure 2. Distribution of three very closely related genera, Schwalbea (continuous line), Sip honostegia (broken line), and Lesquereuxia (two black dots). 704 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 FiGURE 3. Disjunct distribution of Antirrhinum. gion (where more than 80 of ca. 120 species oc- cur), particularly in Portugal and Spain. Only one or two species grow in North America. Melam- pyrum is amuch smaller genus with ca. 30 species, two-thirds of which are concentrated in south- east Europe and the Caucasus. Only one poly- morphic species, M. lineare, occurs in temperate North America (Fig. 6). 3) Genera disjunct across the Pacific (with one part in North America and the other in east Asia). Four genera (Penstemon, Mimulus, Veronicas- trum, and Orthocarpus) belong to this category. All but Veronicastrum have the majority of their members in western North America. Penstemon consists of some 220 species, with a great ma- jority concentrated in temperate western North erica, but with one species, P. frutescens, in Kamchatka, the Kuriles, northern Japan, and Sakhalin (Fig. 7). Orthocarpus occurs almost en- tirely in western North America. Only one species extends eastward into the eastern part of the con- tinent; another isolated species grows in Andean merica. A new species, O. chinensis, was T€" cently discovered from central China (Hong, 1979). Thus, Qrtecagust is actually disjunct be- tween western No rica and eastern pe: (Fig. 8). Although the uut Mimulus is widely distributed, abundance and center of divers again in western North America where over TU thirds of the total number of species (ca. lies grow; the Asian part of the range is a strip x the south Kuriles and Sakhalin to the Hina and is distantly disjunct from the Australian ^ African parts of the range of the genus (Fig. 7): FIGURE 4. Disjunct distribution of Scrophularia. The distribution center is shaded. 1983] HONG—SCROPHULARIACEAE 705 PILIS Leh sl E ur i Be Tes (A) TUR 3 | RJ fe £ i genet í 29] A KIA | | i A j * - | é( gus JJ Eo a Yangon: aE r 5 ud Y he - 2 x [^55 "| M [£C S ESL. SS |, | ^ - tad a * D "a i " NY Wes an z c1 $5 | : e ish re > a : a | 3 gu | Pix > Usos o> "à E CN iei Pct ae ws ee s eero a. Meee AA A AEA | ] E i | | 120 150 180 FIGURE 5. Disjunct distribution of Linaria. The shaded area indicates the distribution center. The discontinuous distribution of Veronicas- ‘um was noted by Li (1952) in his discussion of floristic relationships between eastern Asia and ‘astern North America. The genus has 19 species n eastern Asia, from the eastern Himalayas to in, and a single species in eastern North America south of 50°N (Fig. 10). 4) Genera continuously distributed across the Northern Pacific. In this category are six genera, Veronica, Lagotis, Pedicularis, Euphrasia, Rhinanthus, and Castilleja. As mentioned above, Ver onica is a large genus with the majority of its Species in temperate Eurasia, but with 13 species native to North America (excluding the cos- mopolitan species). Seven are endemic, three ( V. alpina, V. fruticans, and V. scutellata) are cir- cumpolar, and V. stelleri occurs from the Chang- Bai Mountain in northeast China to southern Alaska across the Aleutians (Fig. 11). Lagotis is mainly an eastern sian gen ntered in south western China; only two species extend into northeastern Europe and one into the Caucasus. Lagotis glauca, however, reaches Alaska and the Yukon in the northwestern corner of Canada (Fig. 12). Pedicularis, the largest genus in the Scroph- ulariaceae, with ca. 500 species, is typical of the < 5 = ity REGGE: GOERS TX 4 Jae | al E Ll Y a! | xA as yt | poe F3 z Ec f e 4 E: «UE d ee went ab NW — 9; z g E | 9 A 2 aaa sr ee T * 7 E tee | 7 | M BS E | / M. ki bu E | | d 2 Te . EN ~ w d b. d e Zh ^ iin FFI e VD Ur ner. A Í £ P m E pa Al OP ww Ad »---4 E---- Pel lj M E 95 PS EON E STi eu E88 ER. SE | 9 | kern a ^ ! E -T s 4 LM — TH ——Àd-——— TL —T + > a b | o 30 © WM 120 150 180 150 120 Ficure 6. Disjunct distribution of Melampyrum. The distribution center is indicated by the shaded area. 706 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 FIGURE 7. Disjunct distribution of Penstemon. Shaded area indicates distribution center. Only one species, eastern Asia. P. frutescens, occurs in Holarctic Kingdom. Over 300 species are con- centrated in a small area in the eastern Hima- layas, western Sichuan, and northwestern Yun- nan. Among the 32 species native to North America, 12 are common to eastern Asia and eight are common to Europe (Fig. 13). Although Euphrasia is widespread, section Euphrasia, consisting of annuals, is totally within the Hol- arctic Kingdom with the greatest number of species in Europe. Native to North America are seven species, four endemic to the eastern part; two are also in northern Europe and one is also in eastern Asia (Fig. 14). Rhinanthus is to some extent similar to Euphrasia section Euphrasia, but there are two gaps in the continental part v the range in the Far East and Kamchatka an only a single species, R. borealis, occurs in No America (Fig. 15). Castilleja is different from the genera mentioned above in that' it extends to South America and is centered in western No P| | PAREN EN et pee Ha. pompe Ho Veg LE | gta Í i i EO A | eli = Soe Meee HELL E Ty ae ege do (dS | [ EXIT SS FIGURE 8. Disjunct distribution of Orthocarpus. Th Black dot— one single Asian species, O. chinensis Hong. shaded are e distribution center is indicated by the consen — HONG—SCROPHULARIACEAE FIGURE 9, Distribution of renes The distribution center is indicated by the shaded area. All the Asian Species but M. str. rictus, which centered in the Pacific Coast aes northern Eur- America, pith only three sp asia (Fig, 16). and ilt dA typically or pd "veio - The Antirrhineae are mainly in the Med- iterran "sad SART region where there are eight genera Cludino th L4 CHIC 1 icai ? d the Rhinantheae is dest PB. in eastern Asia conspecific with the African plants, belong to the section Paradanthus, also where there are 14 genera (seven endemic); the 13 genera in Eurasia, particularly in the southern part, and four genera in North America. co North America and Eurasia there is also a genus pair distributional pattern, ulin Shono uereuxia. Among 14n ditional pantropical or worldwide genera are out of our consideration) one falls into a Mediter- p 60 90 120 150 180 1 60 'GURE 10. Disjunct distribution of Veronicastrum. Shaded area indicates the distribution center. 708 Zs oct —— : P-— f > LEGS T Ja 41 ; A x«]^ N 20 BE » PPS JU ue 30 ML. P4 5 S d og 2e " d 19 N KEE) 4:5) En UD E S b » D ao T * Pss Xo " e ^R f "SER EMI p N CEST | > j s D ATO : 30 d ) is 150 180 150 120 FicunE 11. Distribution of Veronica stelleri. ranean/western North America pattern; three into a temperate Eurasia/North America pattern; four into a temperate eastern Asia/North America pattern (three are of an eastern Asia/western North America pattern; one is ofan eastern Asia/ eastern North America pattern). The remaining re : a Wu UID : small gaps in the Bering Strait or Aleutian re- gions. One of these is primarily an American genus; five are Eurasian ones (three with eastern Asia as their distributional center). ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 THE PHASES OF PLANT COMMUNICATION BETWEEN NORTH AMERICA AND EURASIA A number of families in Scrophulariales, e.g., Myoporaceae, Pedaliaceae, Selaginaceae, Stil- baceae, and Retziaceae, exhibit a clearly south- ern distribution; in Scrophulariaceae itself there oeae (Hong & Nilsson, un- publ.). Some mainly northern tribes also have well-developed southern elements; for example, the Hebe complex and the genus Detzneria of Veroniceae in Australasia, and Lamourouxia and Euphrasia of Rhinantheae in Australasia and South America. Many tropical and southern gen- era in the family are disjunct, e.g., Bacopa, Lin- deria, Gratiola, Stemodia, and Buchnera in Af- rica, Asia, Australia, and America; Striga and Limnophila in Africa, Asia, and Australasia; and Hebe and Jovellacea in Australasia and South America. On the basis of this distribution, it 1S reasonable to suggest that the Scrophulariaceae originated and differentiated rather early, prob- ably when Pangaea was still extant or at least before Gondwana had broken up. As to North American/eastern Asian distributional pattern, shown by the genera Penstemon, Orthocarpus, Mimulus, and Veronicastrum, it is quite plau- sible that they migrated from North America to Asia or vice versa through Beringia before the PET -SLL zs 2 A g Ie a CN 5 J EE AE , b ut T V d Xon ERI RS * 60 rT E "e asm n d res ' g N |e Was E24 — ET | Ed ue € x 30 EL E usse — | —- c 9 E 5 i in idco H x G We. z ca 5 Qaa : Es = [ x APPI S EER "B [e s OD + - poa s { d = T LIT BENE. Sty umv s; € po pu p 2 | dee » af : T i pas D 22] L M, P, oe —— 30 60 90 120 150 180 150 120 90 FIGURE 12. Distribution of Lagotis. The distribution center is indicated by the shaded area. Only one $f erica. L. glauca, extends into North Am a 1983] HONG—SCROPHULARIACEAE 709 T--4----------4 TS TP. 60 90 120 150 180 150 120 90 60 FiGURE 13. Distribution of Pedicularis. The majority of species (ca. 300 out of 500) grow in the area shaded. end of the Miocene, i.e., some 12 million years — perate (probably before Early Miocene); the time 480, when the climate there was temperate or and route of the second one is rather certain, i.e., cool-temperate (Wolfe, 1972). In pattern 4, rep- through the Beringian region when it was tem- ‘sented by genera with a more or less continuous perate or cool-temperate; the third is even more distribution through the Bering Strait and the certain, through the BeringStrait or the Aleutians *utians, migration via these two routes during in the Quaternary. ite Quaternary can hardly be doubted. E ne. nisu migration pis MIGRATIONAL DIRECTION erica may have tak en place in three phases. The exact time of the Desp yi g the quest ĉarliest one is still vague, either when Laurasia of migrational direction, I would, nevertheless, "as still extant, or later, but when the climate like to speculate about it and suggest hypotheses. ‘round Beringia was warm-temperate or tem- There is not enough information available for +h ELE. 14 4l | B - > | Figure 14, Distribution of Euphrasia section Euphrasia. Shaded area indicates the distribution center. 710 ANNALS OF THE MISSOURI BOTANICAL GARDEN | FIGURE 15. Distribution of Rhinanthus. The distribution center is indicated by the shaded area. determining the center of origin and direction of migration of the tribes Scrophularieae and An- tirrhineae, but I can say something about the Veroniceae and Rhinantheae. Of the 24 genera of Rhinantheae, eastern Asia has 14, of which seven are endemic. There are also 14 genera in the tribe in Europe and the Mediterranean re- gion, but only three are endemic. In western and central Asia occur ten genera of the tribe, only one of which is endemic. In North America as a whole there are only eight genera, two of which are endemic, and the two other are concentrated there. Recently, a new genus of Rhinantheae, Pseu- dobartsia, was described from Yunnan, south- west China (Hong, 1979). A certain number of Si FIGURE 16. Distribution of Castilleja, with the center indicated by the shaded area. Only three species 0€ urope. in northern Asia and northeastern E pe an am 8 nN rt 1983] most genera, the most endemic genera, and also probably the most primitive element of the Rhinantheae. The two endemic genera, however, and the two highly developed genera in North America, Castilleja and Orthocarpus, where the upper lip of the corolla is long and navicular, and the lobe tips of lower lip in Orthocarpus are sac- like, seem to be specialized elements in the tribe; the monotypic genus Schwalbea has its closest relative in eastern Asia. All these facts may in- dicate that the tribe Rhinantheae originated in eastern Asia and its members in North America are derived. Among 15 northern genera of the tribe Veroniceae, 13 occur in temperate Eurasia and nine of them are endemic to the region. All the primitive members of the tribe but Detzne- ria, à New Guinea genus, are here. The two en- demic genera in North America, Synthyris and Besseya are, however, closely allied to Veronica and are apparently derived (Hong, 1984). Figure 17 shows the distribution of the genus Veroni- castrum as an example of migration of scrophu- lariaceous plants from eastern Asia to North America. The migration during glaciation from eastern Asia to North America through the Bering Strait of Pedicularis, Euphrasia, Lagotis, Veronica, and Rhinanthus is almost certain. All eight sections and most of the species of Veronica occur in temperate Eurasia, especially in the southwestern Part of eastern Asia. All 13 native species, apart ftom the cosmopolitan species, of the genus in North America are those adapted to cool or ronica, and Beccabunga. Of these 13 species, four ate common to the extreme northeastern part of “astern Asia. Lagotis has only one species, L. glauca, in Alaska, Yukon, and the adjacent re- Bon of Canada, which also grows in Kamchatka pan. The other genera have similar pat- s. Was plant migration between eastern Asia and North America unidirectional, i.e., only from the former to the latter? Has the opposite migration "Yer taken place and, if so, to what extent? Ac- ing to what has been stated earlier, the fol- Wing four genera may be such cases. Penstemon 5 centered in southwest North America, with only one species extending into central America, ‘nd a single species, P. frutescens, in Kamchatka, the Kuriles, northern Japan, Sakhalin, and the khotsk region (Fig. 7), where it has no close latives. The disjunct occurrence may be the HONG—SCROPHULARIACEAE Yamazaki tion Veronicastrum. result of migration from western North America through Beringia before the Late Miocene, when the Bering Strait had not opened (Durham & MacNeil, 1967) and the climate there was tem- perate or cool-temperate. Orthocarpus, which is centered in the Pacific Coast states, may repre- sent I ple of the migration from west- ern North America to eastern Asia. The only Asian species of the genus, O. chinensis, was de- scribed recently from a single specimen collected half a century ago from central China. No ad- ditional collections have been made. The migra- tion from western North America to eastern Asia of Castilleja, which also has western North America as its distributional center and has only three species in northern Eurasia, is also pre- sumed by Yurtsev (1972). The case of Mimulus is more difficult to explain. In spite of its wide distribution in tropical Africa, Australasia, east- ern Asia, and America, it is best developed in western North America, where nine of the ten sections and over two-thirds of the species occur. All seven species in eastern Asia except M. stric- tus (in northern India and northern Pakistan and conspecific with African plants, and may well indicate another source) belong to the section Paradanthus, which is also highly centered in western North America. The Asian species, ex- cept M. strictus, are very closely related to each other and also to those of the Pacific Coast states; for example, M. sessilifolius in Japan to M. den- tatus on the Pacific Coast, and the Asian M. 712 tenellus to the western North American M. in- conspicuus (Grant, 1924). Another fact connect- e lem in southwestern China and the eastern Hi- malayas, where numerous local races and vari- ants occur. It is usually suggested on the basis of mammals and some other plant groups that migration from Asia to North America was much more intensive than in the opposite direction (Yurtsev, 1972). The conclusion does not seem to hold true as far as the scrophulariaceous plants are concerned. Six genera, Euphrasia, Lagotis, Pedicularis, Rhinanthus, Veronica, and Veronicastrum, are suggested to have migrated from eastern Asia to North America, but four genera, Castilleja, Mi- mulus, Orthocarpus, and Penstemon, may have migrated in the opposite direction. Although Linaria, Melampyrum, and Scrophularia may have migrated from Eurasia to North America, it is doubtful if the event took place from eastern Asia to North America through Beringia. Antir- rhinum presents an even more difficult case in this respect. LITERATURE CITED DURHAM, J. . S. MACNEIL. 1967. Cenozoic migrations of marine invertebrates through the rait region. Pp. 326-349 in D. M. Hop- kins ieai, va Bering Land Bridge. Stanford S ord. e Geography ofthe Flowering Plants. Fo urth mai Lon GRAHAM, A. (editor). 1972. Floristics and Paleofloris- tics of Asia and Eastern North America. Elsevier Publishing Company, Amsterdam ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 GRANT, A. L. 1924 Mur A monograph p the genus Mimulus. Ann. Missouri Bot. Gard. 11: 99-388. Gray, A. 1840. [Review of] Dr. Siebold. Flora Ja- ponica; sectio prima, Plantae ornatui vel usui in- servientes; digessit Dr. J. G. Zuccarini: fasc. 1-10, fol. 100 m: d tab. 1835-1839. Amer. Jour. Sci. Arts 39: 1 dudas between the flora of Japan and that of the United States. Amer. Jour. Sci. Arts, Ser. 2, 2: 135-136. HEDGE, I. C. 1978. 28. Lesquereuxia Boiss. Pp. 781- 782 in P. H. Davis (editor), Flora of Turkey. Vol- ume 6. wae Pi gi Press, Edinb : Hong, D.-Y. . Angiospermae, Dicotyledoneae, d (1). Flora Reipub Mes pepuhri Sinicae 67(2): xvi + 431 pp. 1984. Taxonomy and evo olution of he Ve- roniceae with special reference to palynology. Op- era Bot. —— & S. Nilsson. Aragoeae, tribe state for Aragoa. Unpublished. Li, H.-L. 1952. Floristic relationships between — Asia and eastern North America. Trans. Am Philos. Soc. N.S. 42: 371-429. (Also see Reprint with a foreword, 1971. MELCHIOR, H. 1964. Band II. Angiospermen Übersicht über die Florengebiete der Erde. id A. Engler, Syl- labus der Pflanzenfamilien. 666 TAKHTAJAN, A. ; ages Punt, Origin and Piece Oliver & Boyd, E WOLFE, J. A. 1972. An interpretation of Alaska Ter- tiary sens Pp. 201-233 in A am (editor), Floristics and Paleofloristics of Asia x Eastern North America. Elsevier Publishing Company. msterdam YURTSEV, B. A. 1972. Pieper: of ga. frist Asia and the problem of Transberingia interrelations. Pp. 19-54 in Floristics and Paleofloristics of A North America. Elsevier Publishing Company, Amsterdam ; aj A COMPARISON OF ORCHID FLORAS OF TEMPERATE NORTH AMERICA AND EASTERN ASIA! CHEN SING-CH?2 ABSTRACT A phytogeographic analysis is made of the orchid floras of temperate North America and eastern Asia. The relationships between them are shown characteristically by the disjunctive patterns of distribution in the eight species-pairs of Cypripedium, Listera, Pogonia, Liparis, and Platanthera, as well as in some other genera such as Tipularia, Arethusa (Eleorchis), Aplectrum (partly Cremastra), restricted by ecological conditions. There are a number of closely related orchids extensively or disjunctively distributed between temperate North America and eastern Asia; among them, species pairs are most interesting. These pairs are Similar in morphology to the extent that most Previously have been considered as conspecific. Although such morphological similarities do not necessarily indicate genetic relationships, they often do. No attempt is made here to further *xamine the question of their evolutionary re- lationship, which would require investigations Using other approaches, especially cytogenetic re- "arch. The purpose of this paper is to discuss Now present distribution patterns have been de- veloped. FLORISTIC COMPARISON There is a striking similarity between temper- “e North American and eastern Asian orchid foras, In the latter area, however, there occurs : number of taxa, including many endemics Primitive forms. ) A total of 106 species with many varieties in ^ genera are reported to occur in temperate “orth America (Correll, 1950; Luer, 1975), ereas in eastern Asia there are about 350 ‘Pecies £rouped into 80 genera, of which 62 gen- Neluding 15 species occur in Japan (Ohwi, Arethusa). Consequently, more than two-thirds of the total orchid genera of temperate North America show a phytogeographic link with east- ern Asia. No single genus is found exclusively in North America and Europe, whereas there are genera of epiphytic orchids occur south of the Qin Ling Mountains: Coelogyne, Pholidota, Pleione, Ischno, aris, Dendrobium, Eria, fargesii Finet (33°55'N), Dendrobium hancockii Rolfe (33*55'N), and Cleisostoma scolopendri- folium (Makino) Garay (36°N) (Chen & Tang, 1982). High endemism in eastern Asia is notable. Seven genera, Tangtsinia, Diplandrorchis, Changnienia, Stigmatodactylus, Ephippianthus, Dactylostalix, and Neofinetia (not including Eleorchis and Kitigorchis), are confined to east- ern Asia, whereas three, namely Hexalectris, Calopogon, and Isotria, occur only in temperate - My heartfelt thanks are due to Dr. Tetsuo Koyama, who read the manuscript and offered many valuable ti || | Tügestions o rtl : : : i Institute of Botany, Academia Sinica, Beijing, People’s Republic of China. , ANN, Missouri Bor. GARD. 70: 713-723. 1983. 714 North America. In addition, eastern Asia has more subendemic genera such as Bletilla, Hem- ipilia, Porolabium, Amitostigma, Archineottia, Androcorys, Ischnogyne, and Holcoglossum. The richness of the above-mentioned eastern Asian orchid flora results in part from the wide range of available areas from subtropical to temperate zones, and a number of tropical taxa that have ranges extending northward into this region. However, in North America and northern Africa there is no such range of areas because of the interruption by either ocean or desert, and few tropical orchids there spread northwards as far as in eastern Asia. Similarly, in eastern Asia tem- perate genera such as Cephalanthera, Epipactis, Listera, Pogonia, and Hemipilia extend south- ward to Thailand and Vietnam. In North Amer- ica, there are no examples of such southern mi- gration of temperate orchids in southern Florida and only a few in Mexico. More primitive taxa are seen in eastern Asia (Chen, 1982). For example, a primitive genus Tangtsinia, possessing a regular perianth, a ter- minal stigma, and five staminodes, has been re- in 1958. Another primitive genus, Archineottia, is characterized by an incomplete column, ex- hibiting a terminal stigma and an erect stamen with a free filament attached to the back of the column. The columnar structure found in this genus and the allied Neottia is transitional from one form with a stamen not fully adnated to the style to another in which fusion is complete. The genus consists of four species, two in China and one each in Sikkim and northern India. Its ally, Diplandrorchis, possesses two stamens that are opposite the dorsal sepal and the median petal (lip) respectively. These orchids belong to the subtribes Neottinae and Limodorinae, both hav- ing suspensorless embryo and relatively large chromosomes, characters also found in the Cyp- ripedium group and some other primitive taxa. Additionally, many other primitive genera such as Apostasia, Cephalanthera, Sinorchis, Aphyl- lorchis, and Pogonia, also are found in eastern Asia. DISTRIBUTION PATTERNS The distribution patterns of the orchids indig- enous to North America and eastern Asia can be divided into the following four categories: Continuous distribution in the northern Pacif- ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 ic. Inthisregion, a few orchids are continuously distributed from Japan or northern China via the Aleutian Islands to North America. For example, Orchis aristata Fisch. occurs in Korea, northern apan, Sakhalin, Kamchatka, the Aleutian Is- Cypripedium guttatum Sw. var. yatabeanum (Makino) Pfitz. occurs in northern Japan, Kam- chatka, the Aleutian Islands, and the Alaskan Peninsula (Fig. 2). However, attention is called to Platanthera chorisiana (Cham.) Rchb. f., which is distributed disjunctively in northern Japan, southern Sakhalin, and in the southeastern cor- ner of Alaska and British Columbia, but is absent from the Alaskan Peninsula (Fig. 3). Disjunctive distribution between eastern North America and eastern Asia. Within this category fall species pairs and individual genera. There are at least eight species pairs found in these two regions. They are similar in morphology and, in most cases, were formerly considered to be con- specific. SEM A is a temperate genus of about 30 species. It is found largely in North America and eastern Asia, with two species extending A Eu- rope. Two species pairs are distributed disjunc- tively in eastern North America and eastern om Cypripedium arietinum R. Br. ranges from Illi- nois through New York, and Connecticut to Nova Scotia, southern Quebec, southern Ontario, southern Manitoba; its Asian counterpart plectrochilon Franch., occurs in Sichuan, peo Hubei, and northwestern Yunnan (Fig. 4). ne species often grow in somewhat boggy and pne places. Another pair is of a slightly wider dis 5 bution: the American C. reginae Walt. exten as far south as North Carolina and feum whereas C. flavum Hunt & Summerh. 1$ -— tered from southeastern Tibet through Yunn?^ Sichuan, and Hubei to southern Gansu ( The distribution patterns and environmc! The genus Listera consists of some t for all found in North America and Asia "t ng L. ovata (L.) R. Br. and L. cordata qe cast which also extend their ranges to Europe- s : ]y rec two disjunctive species pairs are generally CHEN—ORCHID FLORAS FiGURE 1. Distribution of Orchis aristata. ognized: L. australis Ldl. and L. japonica Bl. and L smallii Wiegand and L. nipponica Makino. Listera australis ranges from northern Florida and North Carolina northward to the southern | lp of Quebec, whereas its counterpart is confined lo Japan (Fig. 6). The other American species is narrowly distributed from North Carolina to Pennsylvania, with its equivalent L. nipponica occurring in Japan and the far eastern part of the Soviet Union. Listera nanchuanica S. C. Chen, a western Chinese species, closely resembles L. nipponica. In addition, there are also some sim- ilarities between the American L. borealis Mo- rong and the Japanese L. yatabei Makino, as well Q & "oe ^22 Ficure 2. Distribution of Cypripedium guttatum var. yatabeanum. 716 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 FiGURE 3. Distribution of Platanthera chorisiana. Nutt. and the Japanese L. makinoana Ohwi. al, are found in eastern Asia. ong them, E Liparis is a large genus of approximately 250 — /ilifolia (L.) L. C. Richard ex Ldl. and L. mar species. Its epiphyti [ titute the noana Schltr. are similar and constitute a species jority, and center in tropical Asia and Oceania. pair. Liparis lilifolia is found in eastern ar Two terrestrial species occur in temperate North America from Maine and Vermont along the : js ; stri- as between the American L. convallarioides (Sw.) America, and about 25 species, mostly terre h FIGURE 4. Distribution of the species-pair Cypripedium pl, hilon (eastern Asia) and C. arietinum (easte North America). A i 1983] CHEN—ORCHID FLORAS 747 A lantic States southward to South Carolina, Geor- 8a, and Alabama, and westward through the Central and Lake states to Wisconsin, Minne- _ Sota, Iowa, Missouri, and Arkansas, whereas its counterpart is confined to Japan (Fig. 8). Both FiGURE 5. Distribution of the species-pair Cypripedium flavum (eastern Asia) and C. reginae (eastern North merica). species are also similar to L. pauliana Hand.- Mzt., a species widespread to the south of the Yangtze River of China. Another interesting species is L. /oeselii (L.) L. C. Richard, which is the only species in the family confined to North Ficure 6. Distribution of the species-pair Listera japonica (eastern Asia) and L. australis (eastern North America), 718 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 FiGURE 7. Distribution of the species-pair Listera nipponica (eastern Asia) and L. smallii (eastern North ica) America). America and western Europe. However, it shows some similarity to a Japanese species, L. ku- mokiri F. Maekawa. Pogonia has an eastern Asia-eastern North American distribution. Pogonia ophioglossoides (L.) Ker-Gawl. is found in eastern North Amer- ica from Newfoundland to Florida (Fig. 9). The other six species are in Asia, with their southern limits extending to Malaysia. Pogonia ophio- glossoides is most similar to P. japonica Rchb. f., which is widely dispersed from the far eastern part of the Soviet Union and northeastern China A ta: 8. Distribution of the species-pair Liparis makinoana (eastern Asia) and L. lilifolia (castem s merica). 1983] CHEN — ORCHID FLORAS 719 A "2? 743 Ficure 9. Distribution of the species-pair Pogonia japonica (eastern Asia) and P. ophioglossoides (eastern North America). through Japan, southward to Jiangsi, Fujian, and included in Pogonia, from which they were sep- Taiwan, and west to Sichuan and Kweichow in arated mainly on the basis of more advanced China (Fig. 9). Pogonia is a primitive genus pos- compound pollen grains occurring in tetrads. In Sessing single pollen grains. In the New World, Asia, however, no closely allied genus to Pogonia the genus has three allied genera, Cleistes, Iso- is found. : i iria, and Triphora. These genera formerly were — Platanth Or Among s ( à | North FIGURE 10. Distribution of the species-pair Platanthera florentii (eastern Asia) and P. hookeri (eastern No America), 720 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 / P m "et 7% Distribution of the species-pair Platanthera hologlottis (eastern Asia) and P. dilatata (eastern North America). FIGURE 11. its 200 species, about 40 are found in eastern Asia, and 36 in North America. The Japanese P. florentii Franch. & Sav. cl American P. hookeri (Torrey) Ldl., which occurs da and the northeastern United States (Fig. 10). Of special interest is another species pair, P. dilatata (Pursh) Ldl. ex Beck and P. hologlottis Maxim. The former oc- curs in both eastern and western North America, including the Aleutian Islands, whereas the latter extends from the far eastern part of the Soviet Union southwestward to the Yunnan in China (Fig. 11). These species show a transition from the continuous distribution of same species to the disjunction of species pairs between eastern North America and eastern Asia. In addition to Pogonia, there are three more Sichuan and Tibet) (Fig. 12). The map by Li (1952) includes Tipularia camtschatica Spreng., which is not a Tipularia but instead a Platan- thera. The Asian species of this genus are more or less related. However, the American species is rather different from the Asian species with respect to the fugacity of its solitary leaf prior to th fthe infl e. Arethusa and Aplectrum each contain two species that are dis- junctively distributed in eastern North America and Japan. However, among specialists there are different opinions as to whether the Asian and American representatives are congeneric. = (1953) considered Aplectrum unguiculatum (Fi- net) F. Maekawa a member of Cremastra, an F. Maekawa established in 1935 a new genus Eleorchis, based on Arethusa japonica A. Gray. However, the close relationship between the two species of each pair are generally recogni” m Disjunctive distribution between western No ? America and Eurasia. Two closely related oe era, Cephalanthera and Epipactis, rage : this particular pattern of distribution. Con s ing 14 and 20 species respectively, pec are well represented in Eurasia, each with i ; . Gray) species, viz. Cephalanthera a E Hook in western North America tinae is the only saprophytic member of dt nus, which is so unusual in appear ee Gray regarded it as a separate genus, C prit- Epipactis gigantea, scattered from SUM EE a ish Columbia to Mexico is also unusual : pearance. Neither Cephalanthera nor Epi pact | | | CHEN—ORCHID FLORAS 4 ERA FIGURE 12, Distribution of Tipularia. i ; : : s found in eastern North America except Epi- pus followed similar historical successions "B similar ecological conditions. _ Disjunctive distribution in the tropical Pacif- Beo southeastern North America and eastern se ere are two orchid genera, Tropidia and Tro cv. that are disjunctively distributed. : E la consists of about 20 species, including o" gle New World species, 7. polystachya (Sw.) ne from southern Florida, Mexico, E ntral America to the northern end of South eR The remaining members are all trop- bs ELA and Oceanian, with their northern lim- k stribution in Japan (T. nipponica Ma- m.) and Sichuan of western China LT. tmeishanica K. Y. Lang). Tropidia polystachya o EC World genus of some 100 species. Ery- tes querceticola (Ldl) Ames is the only pue that occurs in North America, extending E" Florida to South America. In tropical Asia ad 7" few species of this genus are found, two ^ ch extend northward to China: E. chinen- (Rolfe) Schltr. of Yunnan and Kwangdong and E. /atifolia Bl. of Taiwan and Malaysia. Nei- ther of these genera is found in Africa. DISCUSSION AND INTERPRETATION It is evident from the foregoing data that the distribution of Orchidaceae between the two continents follows the general rule, although it is an advanced family. All taxa mentioned above are terrestrial, be- longing to different groups, most of which are less advanced members of the family. No epi- phytic orchid shows such a pattern of distribu- tion. Although the genus B Ibophyll. I sin gle species occurring in southern Florida (Luer, 1972), and several in eastern Asia, little similar- ity can be found between them. As a pantropical genus, Bulbophyllum consists of more than one thousand species, all of which are epiphytic. The distribution of the species pairs on two continents often show some similarity and cor- relation among them. For example, the Ameri- can species Pogonia ophioglossoides is wide- spread and Listera smallii is of local occurrence. Similarly, their Asian counterparts Pogonia ja- ponica and Listera nipponi idel d nar rowly distributed, respectively. Probably this is due to similar ecological preference and genetic character. However, the American species often show a wider range of distribution and reach 722 higher latitudes, whereas their Asian allies gen- erally grow at higher elevations. Two species pairs of Cypripedium are good examples. The ranges of C. reginae and C. arietinum extend as far north as 50°N in North America, where they are found at an elevation below 1,000 meters; their Asian counterparts are confined to south of the Qin Ling Mountains (33°-34°N) at an altitude be- tween 1,850 and 3,700 meters. Phytogeographically, Japan is significant; the aforementioned taxa are found except for Cypri- pedium plectrochilon, C. flavum, and E throdes. In some cases, species have continuous ranges that extend from Japan or China through coastal areas and continental islands to North America. Examples include Orchis aristata, Cyp- ripedium guttatum var. yatabeanum, and Pla- tanthera chorisiana, although the last species is absent in the Alaskan Peninsula, showing a de- gree of disjunction in its range. A highly inter- esting pattern of distribution is exhibited by the species pair Platanthera dilatata and P. hologlot- tis, which seem to have been differentiated more tribution would approach that of Cypripedium reginae and C. flavum. Thus there is a series of transitions in distri- butional types; Orchis aristata type (continuous distribution in the northern part of the Pacific at the species level); Platanthera chorisiana type (distribution somewhat disjunctive in the north- ern part of the Pacific at the species level); Pla- tanthera dilatata—P. hologlottis type (species pairs between North America and eastern Asia); Lis- tera australis—L. japonica type (species pairs be- tween eastern North America and western China); Tipularia type (disjunctive distribution between eastern North America and eastern Asia at the generic level); and Tropidia type (disjunctive dis- tribution in the tropical Pacific at the generic level). It might be assumed from these types that the migration of orchids between these two con- The Orchis aristata type and the Platanthera chorisiana type must represent more recent mi- gration. Slightly earlier would be those of species pairs, such as the Platanthera dilatata—P. ho- loglottis, the Listera australis—L. Japonica, and the Cypripedium arietinum-C. plectrochilon ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 types, but migration does not seem to be as early as the Paleo ity from its Asian allies. ferentiation, however, does not necessarily show a direct correlation to the time because the rate of evolution is different in different groups under different environmental conditions. More data and further investigations are necessary. Furthermore, distribution of some very prim- itive orchids indicate that there was an ancient connection between these two continents. For instance, Apostasia and Selenipedium are both relicts and are related to each other in possession ofa three-celled ovary, seeds with highly sclerotic and opaque testa, and habits. Apostasia is found in tropical Asia, with one species A. nipponica Masam., extending to southern Japan, whereas Selenipedium is confined to Central America and northern South America. Both the mode of their ringia. During the E T A. Wolfe (1972), tropical forests did exist 1n Alaska. At the same time, the subtropical zone had reached as far north as 42°N in China (Hsü & Li, 1980), and subtropical forests seem to have covered nearly the whole of Japan (Tanai, 1972). Such a climate lasted for a long time until the second half of the Pliocene. f The dissimilarity between orchid floras © western and eastern North America is notable. In western North America there are only ten gen- era, of which only Cephalanthera and Epipactis are absent from the East Coast. The flora of Cal- ifornia, as pointed out by Raven and ee : (1978), is characterized by a high degree wired mism. No exception is found in Orchidaceae.’ a example, Cephalanthera austinae, CM californicum A. Gray, and Platanthera geen peris all are very distinct taxa, to which no c 4 allies are found in eastern North America an Eurasia. eee There is little indication of more recent id change of orchids between western Europe North America, although it is gene nized that, during the Late Cretaceous Tertiary, European floras were similar eastern North Amrica, but not to (Hallam, 1981). Only one species, selii, is exclusively indigenous to easte — = e SONO" — — 1983] America and western Europe, whereas many Eurasian species, such as Epipactis helleborine, Cephalanthera longifolia (Huds.) Fritsch, Epi- pogium aphyllum (Schmidt) Sw., Gymnadenia conopsea (L.) R. Br., Cypripedium macranthum Sw., Listera ovalis (L.) R. Br., and Neottianthe cucullata (L.) Schltr., are not found in North America. It deserves mentioning that Epipactis helleborine has grown well and spread rapidly since it was introduced into eastern North Am ica some 100 years ago, indicating that its ab- sence from eastern North America is not due to ecological but to phytogeographical causes. n summary, the distribution patterns of Or- chidaceae, as in most other angiosperms, show EE ue null America and eastern Asia. Between these two regions there is a series of transitional patterns af distribution from continuous ranges of the species vt and species. It is rational to dies, Pret that the disjunctive distribution between fastern North America and eastern Asia was re- alized as the result of the migration via Beringian and the coastal areas as well as through the continental islands, although these were not nec- ‘sarily the only migration routes. From these data, of course, it seems to be of little significance to assume the time of their interchange, the or- igin, and the direction of migration, because there are only few fossil records available for the Or- Chidaceae (only Protorchis from the Eocene in urope is considered to be credible, and the oth- *ts such as Paleorchis and Orchidacites, are more or less doubtful), and their genetic relationships àre not clear. As an advanced family, however, Such data would be of s some help in elucidating Ttain hn temperate North America and eastern sia, CHEN—ORCHID FLORAS LITERATURE CITED CHEN SING-CHI. 1982. origin and early differ- entiation of the Diae, Acta Phytotax. Sin 20: e e NG TsıN. 1982. A general review of or- chid gs of China. Pp. 39-81 in J. Arditti (anion. Orchid Biology, Review and Perspectives, II. Cor: nell Univ. Press, Ithaca and London. ConRELL, D. 1950. Native Orchids of North merica. Chronica Botanica Co., Waltham, Mass. liso A. 1. Relative importance of plate vements, eustasy, and climate in controlling jor ł SOZOIC. Pp. 304-334 inG. Nelson & D. E "Rose (editors), Medias giro dE A Critique. Columbia Univ. Press ork. Hst JEN & Li SHI-YING. von Development = evo- lution of Cenozoic vegetation in Chin . 61- 68 in Wu Cheng-i aiian, The E > China. The Science Press, Peki Li Hur-LiN.. 1952. Piorstic- relationships betw merica. Temis: Amer. Philos. Soc. 42: 371-429. 1972. The Native Orchids of Florida. W. owell, Ipswich. 1975. The Native iea of ahe United States owell, I I. The Wild Orchids of Japan in Colour. Shibundo, Tokyo. Onwi, J. 1953. Flora of Japan. Shibundo, Tokyo. RAVEN, P. H. . I. AXELROD. 19 Origin and Relationships of the California Flora. Univ. Calif Press, Berkeley, Los Angeles, London 1972. evened history of vegetation i in Ja- r2 n G. Graham (editor), Floristics "e Palcolrisi of Asia and Eastern North Am Pss r Publishing Co., Amsterdam, fonder M ak Wow J. A. idi] da peg gps of Alaska Ter- ary floras. Pp. 201-203 in G. Graham (editor), Plocistics and Paleoficristice of Asia and Eastern North America. Elsevier Publishing Co., Amster- dam, London, New Yor Wu pares ap od Ho-SHEN, acteristics Chin 82-114 in Wu Cheng-i -i (editor, e" Spits du 4 China. The Science Press, Pek 1 980. Floristic aged PHY TOGEOGRAPHIC RELATIONSHIPS OF PTERIDOPHYTES BETWEEN TEMPERATE NORTH AMERICA AND JAPAN! MASAHIRO KATO? AND KUNIO IWATSUKI? The pteridophyte flora of North America com- prises 325 species (Tryon, 1969) and the non- endemic species among them are referred to four geographic groups: circumpolar, amphioceanic, tropical, and Mexican. Japanese pteridophytes American, although the Japanese Archipelago is much smaller. The Japanese species are also more compl ically and phytogeographically (Tagawa, 1959; Ohwi, 1965). The distribution pattern of the Japanese ferns is considered to be related to climatic zones by some authors in- cluding Nishida (Graham, 1972). They can be arranged roughly in the following phytogeo- graphic groups: ci polar, Eurasian, trans-Pa- cific (related to North America), Sino-Himala- yan, and tropical Southeast Asian. The phytogeographic relationships between North America and Japan are found primarily in the temperate to boreal species. Since Asa Gray’s time, phytogeographic rela- tionships between eastern North America and eastern Asia have been discussed by many au- thors for pteridophytes as well as for flowering and non-vascular plants (Hultén, 1958, 1962; Li, 1952; Graham, 1972; Tryon, 1969; Tryon & Tryon, 1973). In spite of the accumulation of data, there still is no complete phytogeographic analysis of the biological relationship between eastern North America and the Far East. The complex geohistory often is taken into account and the floristic comparison often is made rather theoretically. Changes in land and climate in the Northern Hemisphere that have occurred since the Tertiary have given rise to replacement of boreal, temperate, or warm floras in the areas concerned, and have enabled plants to migrate between the North American and Asian conti- nents via the Bering Strait (chiefly dry land in the Tertiary) and Aleutian chain. This migration probably occurred several times and in either eastward or westward directions. Similar events ana won lA ha between temperate and tropical regions. Specia- tion and gene-flow caused by the geographic iso- lation and reunion might have taken place in both those areas. The fern floras of North Amer- ica and eastern Asia include identical or related species. Some of the identical species might be younger and widespread in the Northern Hemi- sphere, whereas some related species might be older and become restricted to either North America or eastern Asia. Phytogeographic relationship f fern floras re- sult from species migration that is initiated pri- marily by spore dispersal. Spore viability varies from about a week to more than several years; green spores are especially short-lived compared with non-green spores. It is reported that fern spores resist, to a considerable extent, physical and chemical environmental stress from low temperature, desiccation, UV- and X-rays, and other factors that operate during the transport (Page, 1979). Tryon (1970) argued the effective- ness of long dispersal of fern spores in island fern floras and estimated that “distances up to 300 miles are only a slight barrier to the migration of a fern flora and that a distance of 500 miles is not a significant barrier.” In spite of this, con P" : a E nct ot y a limited number have wide ranges. This means that the dispersal or migration of most ferns 1S regulated by some unknown factors. Page (1 pointed out that many biotic and non-biotic eco- logical relations are involved in migration. In his study on genetic features of disjunct fern popu- lations, Klekowski (1972) considered that - long-distance establishment and future inar tionary processes of fern populations are depen vs. intergametophytic fertilization) meson "pne ic Pene variability as well as geographic T dpi d spore dispersal and species establish new locality is still insufficient to fully 9 ei stand the phytogeographic relationships "a fern flora in question, for which further exper mental and field studies are desired. : ents. : We thank M. G. Price for reading the manuscript and D. I. Axelrod for his review with valuable peer , Botanical Gardens, Faculty of Science, University of Tokyo, 1842 Hanaishicho, Nikko 321-14. Ja Botanical Gardens, Faculty of Science, University of Tokyo, 3-7-1 Hakusan, Tokyo 112, Japan. ANN. Missouri Bor. GARD. 70: 724—733. 1983. — = | 1983] In this paper the phytogeographic relation- ships between temperate North America and tween the two areas, it is intended to elucidate l) the similarity or difference between the two areas, especially in the composition of their flo- ras, and 2) the taxonomic relationship between the apparently identical or vicarious species known in the two areas. FLORISTIC SIMILARITY Of 116 species occurring in eastern North America (Morton, 1968) 52 species (45 percent) are also native to Japan, and the figure ap- proaches that for the bryophytes that have the similar dispersal means of airborne spores. Of the 52 common species, 45 species (39 percent) are either widespread in the temperate to boreal Northern Hemisphere or are cosmopolitan, and seven species (cited below) are restricted to east- erm Asia, including Japan. The species common lo these two areas have their center of distribu- lion in Japan in central and northern Honshu Tican species are distributed also in Europe. Northwestern North America has a pterido- Phyte flora consisting of 81 species (Hitchcock & Cronquist, 1955), of which 37 (46 percent) also in Japan. Of these, 34 species are circum- borea] and three are among the seven species of the eastern North America-eastern Asian distri- bution pattern. Li (1952) stated that the distribution patterns 9f the pteridophytes are different from those of KATO & IWATSUKI—PTERIDOPHYTES 725 NE. N. AMERICA 116 The number of common species among FIGURE 1. Hokkaido, northeastern North America, and Europe. For explanation see the text. the seed plants that have different means of re- production and dispersal, and that whereas the floristic relationship of the seed plants between temperate eastern North America and eastern Asia is primarily generic, few fern genera, if any, indicate such a relationship. By contrast, Tryon and Tryon (1973) compared the fern species in both areas and pointed out that the phytogeo- graphic relationship of pteridophytes i th areas is strong and that about 20 percent of the species in eastern North A ica show disj ffiniti with identical or related species in eastern Asia. Phin ue n : x 4 D t most of the apparently identical species (some of which (ron in every area of eastern North America, Hok- kaido, and Europe. The other several species are restricted to both areas in question. The rela- tionships clearly imply a former, more wide- spread, similar temperate fern flora that was broken by climatic change in the late Tertiary (post-Miocene). IDENTICAL AND VICARIOUS SPECIES The phytogeographic relationship between temperate North America and eastern Asia is strongly suggested by the apparently identical or vicarious species that are confined to the two areas. The seven apparently identical species are: Lycopodium obscurum, L. sitchense, Osmunda cinnamomea (eastern; extending into tropical America as var. imbricata), O. claytoniana (east- ern), Adiantum pedatum, Onoclea sensibilis (eastern), and Polypodium virginianum — A few eastern p su YY ve northwestern part of North America, Alaska, and Pacific Canada: Lycopodium chinense, Selagi- nella sibirica, and Mecodium wrightii. Tryon and Tryon (1 973) cited examples of sev- eral vicarious species between North America ANNALS OF THE MISSOURI BOTANICAL GARDEN 7 30 Cae FIGURE 2. Distribution of Equisetum hyemale.—A. var. hyemale;—B. var. affine. (After Hauke, 1963.) and eastern Asia: Dennstaedtia punctilobula of eastern North America vs. D. appendiculata and D. scabra of eastern Asia; Thelypteris novebora- censis of eastern North America and T. neva- densis of western America vs. T. beddomei and T. nipponica of eastern Asia; and T. simulata of eastern North America vs. 7. glanduligera and T. japonica of eastern Asia. Asplenium rhizo- phyllum of eastern North America and A. ru prechtii of eastern Asia are a well-known example ofa vicarious species pair. They are os eee as comprising a separate genus Camptosorus be cause they are typologically different, aoe in their reticulate venation, from the other species of Asplenium. The following is a discussion of some identical or vicarious species of phytogeo- graphic interest. Equisetum hyemale. Of the ten species and several hybrids of Equisetum that occur in tem- perate North America, eight are also native to Japan (Hauke, 1963, 1978): E. sylvaticum, E. pratense, E. arvense, E. fluviatile, and E. palustre of the subgenus Equisetum and E. hyemale, E. variegatum, and E. scirpoides of the nce Hippochaete. Equisetum hyemale has two eties, vars. hyemale and affine, of which the vues shows the North America-eastern Asian re : tionship. i hyemale var. affine is dis tributed i whereas var. hyemale occurs in Eurasia, a ing eastward to Asiatic Russia and westernm' China (Fig. 2). The former variety is japos ward direction (Hauke, 1963). Hauke (1963) speculated on the phytogeo ibgenu Hippoc pochaete, It on a phylogeny inferred from _morphology. pni originated in South America and radiated Y : erly; several species were differentiated 1 process. The most primitive member 0! i dd genus is E. giganteum, distribut ted in sell America. An E. giganteum-like ancestor di gave rise to E. ramosissimum with two SUP cies, the Eurasian ssp. ramosissimum, Miro zi from Europe, Mediterranean Africa and East Africa to Japan, and ssp. debile ranging (rte i KATO & IWATSUKI—PTERIDOPHYTES FIGURE 3, line, var. Squamigera; dash and dot, var. palustris. (After Tryon, 1971 from India to Fiji through Southeast Asia. From the E. ramosissimum complex originated two American species, E. myriochaetum of Central and South America, and E. /aevigatum of Western North America. “Equisetum hyemale male" (Hauke, 1863) whereas var. affine crossed the Aleutian chain to eastern Asia. Thelypteris palustris. This species is world- wide in distribution except in South America E 3) and is diversified in several characters. Consists of the following four varieties: var. Scens of northeastern North America and U “astern Asia: var. haleana of the southeastern Ed States and Bermuda; var. palustris of Urasia: and var. squamigera of Africa, India, Burma, Thailand, Sumatra, Borneo, New Guinea, Distribution of Thelypteris palustris. Solid line, var. apie dashed line, var. haleana; dotted ) € New Zealand. The classification of this di- species varies according to the author. Beca (1971) followed Fernald (1929) in com- paring the spore characters of four varieties of T. palustris. She suggested that continental southeastern Asia is the center of radiation of this species. It might in turn be suggested that the distribution range of var. pubescens has been established by eastward migration from Asia to North America. Thelypteris noveboracensis and T. nippo- nica. These two species are often referred to as representing a nina example of vicariants with disjunct distribution (Tryon & Tryon, 1973). void noveboracensis occurs in eastern America and 7. nipponica is distributed in prionai the Kuriles, Hokkaido, northern and central Honshu, Korea, and central and western China. They belong to Parathelypteris sensu Holttum and are similar to each other in phenetic features with the exception of the scale features and soral position. The chromosome ANNALS OF THE MISSOURI BOTANICAL GARDEN L L RENAE QNNM UGERANT IN o FiGURE 4. Distribution of Asplenium rhizophyllum (A) and A. ruprechtii (B). number of T. noveboracensis is reported as n — 27 (Tryon & Tryon, 1973) and that of T. nip- ponicum as n — 62 (Hirabayashi, 1970). The latter is often considered to be related to the Asian T. beddomei, the chromosome number of which is n = 31, and T. beddomei is a suggested vicariant of 7. nevadensis of western North America. In spite of their phenetic resemblance, it is necessary to revise them, and so we refrain here from discussing the geographic relationship between T. noveboracensis and T. nipponica. The eastern North American T. simulatais also com- parable with the Asiatic 7. glanduligera, al- though we have no conclusive remarks on the phytogeographic relationship between the two species Asplenium rhizophyllum and A. ruprechtii. As- BG . L: + ae Ri S TUE Pee. | of eastern North America from Quebec to Min- nesota, Georgia, Alabama, Kansas, Oklahoma, and Michigan, and A. ruprechtii is found in east- ern Asia in eastern Siberia, northern and north- eastern China, Korea, and Japan (Fig. 4). These species are isolated from the other species of idi plenium, especially in their special habit of sine miferous tips of the simple fronds (often call walking) and anastomosing sagenioid venation, the two species are discriminated by the form > the leaf base. Both often grow on limestone. though they can grow also on non i rocks. It is certain that these two are Pm " lated, although we have no fossil Mig c^ umenting the migration and speciation of a. Deparia acrostichoides. f . icar- i to eastern Asiatic VIC 5). It is closely related ac of section widely spaced leaves. Section Las gy its center of distribution in eastern many diverse species and therefore it | | 1983] KATO & IWATSUKI—PTERIDOPHYTES 729 bo pue pem E ep TR T FIGURE 5. Distribution of Deparia section Lunathyrium. See text for discussion. ected there. It seems likely that a part of the : Pycnosora group migrated in an eastward di- on and gave rise to D. acrostichoides. Diplazium pycnocarpon. This species is the y member of Diplazium from temperate east- . 8m North America (Fig. 6), although it is still commonly placed in Athyrium. It possesses sev- E alized aspects, including a temperate E. ution (in contrast to a tropical and sub- » Pical distribution of the majority of Dipla- um), and frond morphology (particularly in the i Vei : : | E which reach the concave pinna margin). . . Dlazium pycnocarpon seems to be, as stated &ro : open : l^ Which has similar anastomosing venation te f 4 "x POMPA ENPT: y: . 3 Dinn I v I1 'UIICGI5 $ nae (Fig. 7), and thereby has been separated as Dictyodroma by some taxonomists. Kato (1977) supposed that the D. javanicum group and the D. heterophlebium group are related to tropical American D. marginatum (most often known as Hemidictyum marginatum) with a more elabo- rate reticulate venation (Fig. 7). These species are distinct not only morphologically but also cytologically. It is reported that the basic chro- mosome number of the D. javanicum group and the D. heterophlebium group is x = 41, a com- mon number of the genus Diplazium, whereas D. pycnocarpon has x = 40 and D. marginatum x= 31 (Love et al., 1977). It is possible that D. pycnocarpon is derived from a D. marginatum-like ancestor, but more likely it was established as a specialized temper- ate species with a closer affinity to the Asiatic species. Gymnocarpium dryopteris. The genus Gym- nocarpium comprises six species that are clas- sified into two subgenera, Gymnocarpium of five 730 ANNALS OF THE MISSOURI BOTANICAL GARDEN mou ea TTT MEM Al! )) : — Ficure6. Distributi £ Tal "WP 1 P Gald ine „D. pycnocarpor d p i is E T 3*1 MA dashed line, D. j iopsis); vertical lines, D. heterophlebium group (Dict) dot, D. marginatum rini see 4 s Saghalien, closely related species, and Currania with one junctum hasa rn North species, G. oyamense. Three of the species are restricted to Asia: G. remote-pinnatum in Tai- wan, G. fedtschenkoanum in Tadzhikistan, and G. oyamense in Nepal, continental China, Tai- wan, Japan, Luzon, and New Guinea, whereas the other three are widely distributed in Asia, America, and Europe (Sarvela, 1978). From its nated and diversified into two subgenera in Asia and spread to Europe and North America. Cy- tological studies show that G. dryopteris mp. dryopteris has 2n = 160 and ssp. disjunctum 2n 80, G. jessoense ssp. jessoense 2n = 80, G. nd ertianum 2n — 160, and G. oyamense 2n = 80 and 160, suggesti ofthe genus arose ; mostly at the diploid level. The tet- raploid G. dryopteris ssp. dryopteris has a cir- cumboreal distribution and the diploid ssp. dis- hatka, and northweste southern Kamtc seed ide : suggested th America (Fig. 8). It is sugg t of a recent distribution of the species is sa resul cia, and five species that include a adire putative hybrid origin (Kato & Sahashi, Onoclea consists of two species, O. —— O. orientalis. The former is separa y varieties, var. sensibilis of eastern North ecd ica and England (probably introduced), pang interrupta of eastern Asia, which differs — 2m sensibilis in possessing more wide gem tile pinnules (Fig. 9). Paleobotanical eV! vary th shows that in the Cretaceous and Tert ddr Onoclea sensibilis group was diversified i tributed more widely than the present, an tates, known from Alaska, Canada, the United S to —M 3 1983] KATO & IWATSUKI—PTERIDOPHYTES 731 FIGURE 7. Venation of Diplazium pycnocarpon and presumably related species. —A-F. D. (Dictyodroma) . heterophlebium:—A-E. Juvenile plant.—G-J. D. (Diplaziopsis) cavalerianum:—H-J. Juvenile plants.—K. D. (Hemidictyum) marginatum.—L. D. pycnocarpon. A-J, L x1, ] | | - as js EJ oo Ficure 8. Distribution of Gymnocarpium dryopteris. Solid line, ssp. dryopteris, dashed line, ssp. disjunctum. 732 ANNALS OF THE MISSOURI BOTANICAL GARDEN EC en abe cn m A = Distribution of onocleoid ferns. Solid line, Onoclea subsibilis; vertical lines, O. orientalis; dotted FIGURE 9. line, Matteuccia hintonii; dashed line, M. struthiopteris. central and southern Europe, and Japan. Most of the fossils are fragments of leaves with retic- ulate venation, however. Regarding the taxonomic relationships, Kato and Sahashi (1980) considered that each species of the onocleoid ferns is distinct, O. sensibilis and Matteuccia hintonii are relatively primitive, and O. orientalis and M. struthiopteris are de- rived. Further, that Onoclea sensibilis and Mat- teuccia hintonii originated from a common ancestor is suggested by the resemblance in the sagenioid venation of sterile leaves, the com- poundly pinnate fertile leaves, and spore orna- mentation. From the distribution pattern and the onocleoid ferns originated somewhere in Central or North America. Onoclea orientalis might have developed in Asia from an O. sen- sibilis-like ancestor whereas M. struthiopteris might have originated in North America from a M. hintonii-like ancestor and later spread widely in the Northern Hemisphere. DISCUSSION between these two areas, as is the case of ies seed plants. It is, however, rather difficult to o phytogeographic relationships between any ues fion k J MENR i dam the terns of the extant species. In mos P ini distribution of the plants is in good win und with the geohistory of the areas concerne» 3 8 P ry ¥ ae henner this respect, re. The distri" despite their easy dispersal by spo: vironmental factors, especially by such gun conditions as temperature and moisture. is sis in turn influence pteridophytes during the € as life cycle from spore germination throu io rophyte maturation. Moreover, the ee ij of the pteridophytes is strongly dependen n w. the vegetation of the areas in which they gro —" coil í ——" 1983] The climatic factors influence the vegetation and thus indirectly delimit the distribution. Because the climate in the Far East is more similar to that in eastern North America than to that in western and central North America, at least a part of the geographic similarity of Far East- eastern North American pteridophytes is regu- lated by such climatic and ecological conditions. We examined the taxonomic relationships of some apparently identical or vicarious species occurring in temperate North America and Ja- pan, and from these deduced the phytogeograph- ic relationships of each species, pru the as- sumptions are derived from i ence. The indication is that the apparently iant and vicarious species under consideration have their own phylogenetic background, that the sim- ilarity of their present distribution patterns is due to dispersal with allied forests, and that during that time there has been some evolutionary ge. Further elucidation of the taxonomic re- lationships of those species with reference to the morphology, cytology, genetics and ecology sure- ly will contribute to the understanding of the dynamic aspects of the geographic relationships of the fern floras of temperate North America and eastern Asia. LITERATURE CITED FERNALD, M L 1929. A study of Thelypteris palus- 7-36. 1972. Floristics and Paleoflo- ristics of Asia and eastern North America. Elsev- ier, Amsterdam HAUKE, R. 1963. onomic monograph of the genus Equisetum subgenus Hippochaete. Beih. Nova H wigia 8. E 1978. A taxonomic monograph of Equisetum Subgenus Equisetum. Nova Hedwigia 30: 385-455. HiRABAYASH Hi H. I romosome numbers in Several species of the Aspidiaceae (2). J. Jap. Bot. 45: 45-52. Hitcucock, C. L. & A. CRoNQUIsT. 1955. Flora of E, Pacific Northwest. Univ. Washington Press, tle. KATO & IWATSUKI—PTERIDOPHYTES 733 HULTÉN, E. Sve 1958. The amphiatlantic n Kongl. nska Vetenskapsakad. Handl. Ser. 7(1). Kato, M. 1977. Classification of Athyrium and allied genera of Japan. Bot. Mag. (Tokyo) 90: 23-40. & N. SAHASHI. 1980. Affinities in x ono- cleoid ferns. Acta Phytotax. Geobot. 31: 127-138. KLeEKOwsKI, E. J. 1972. Genetical features of ferns as sted to seed plants. Ann. Missouri Bot. Gard. 59: 138-151. Li, H.-L. 1952. Floristic relationships between east- ern Asia and eastern North America. Trans. Amer. je rus Line 42: 371-429. [Reprinted as: Morris um Monograph Edition. 1971. With a Med d and additional literature citation i-ii. = . Löv ERMOLLI. 1977. Cyto- mical Atlas of = Pteridophyta. Cramer, MORTON, C. V. 1968. Pteridophyta. Jn H. A. Glea- son, The New Britton and Brown Illustrated Flora of is Northeastern United States and Adjacent iW er, New York. Manion. y": An outline of the pet i of J apanese fins Pp. 101- -105 in y, am td or), Fl ia and e: rn North America. Elsevier, Amste crit iit: J. 1965. Flora of Japan. English edition. Smithsonian Institution. Washington, D.C. Pace, C. N. Ex xperimental aspects of fern ecol- nocarpiu TABANA M. 1959. "Coloured stration of the Jap- se Pteridophyta. H a, Osaka. RYON, A F. 1971. bel ison pea variation in spores of Thelypteris palustris. Rhodora 73: 444-4 ON. 1973. Thelypteris in northeast- ern North America. Amer. Fern J. 63: 65-76 TRvoN, R. 1969. Taxonomic problems in the geog- raphy of North American ferns. BioScience 19: 790-795 1970. Development and evolution of fern flor. as of oceanic islands. Biotropica 2: 76—84. TUTIN, T G. et al. (editors). 1964. Flora Europaea. Vol. 1. Lycopodiaceae to Platanaceae. Cambridge Univ. Press, Cambridge. EASTERN ASIAN-EASTERN NORTH AMERICAN FLORISTIC RELATIONS: THE PLANT COMMUNITY LEVEL PETER S. WHITE! ABSTRACT z 1 x a M, £41 g £ orth A a the systematic or floristic level. Data on vegetation pattern and structure, when pecie t, have been equally interesting non- random "pin aa m mere fos ra: sid were more likely to be herbaceous of the flora as a whole in terms of degr — tab t open ness and substrate type. When gradient distributions were corrected for aids sad ca distributions, absent taxa showed their nus importance on xeric sites. Taxa in wide-ranging north temperate genera (817 species or 67 pne o the flora) generally constituted a random sample of the flora as a whole; they showed peak im at higher elevations. These nce data are used to make inferences about the distribution of species hien in the Smokies landscape. Evidence that the eastern Asian tem rate deciduous forests were the mos diverse temperate forests on earth is also reviewed. Alternative hypotheses concerning the nature of this heightened diversity are discussed. Investigations of the floristic relationship be- tween temperate eastern Asia and eastern North America generally have treated such relation- logical correlates to this relationship that have gone, for the most part, untested. This investi- gation uses a well-known eastern North Ameri- can landscape, the Great Smoky Mountains, to examine ecological correlates of the floristic re- lationship. Although others have qualitatively described many of these patterns (cf. Hu, 1936; Cain, 1943; Li, 1952; Wang, 1961), quantitative studies have been lacking. An objective of this research, then, was to develop explicit tests for the distribution of ecological properties in the flora as a function of phytogeography. The land- scape distribution of the Great Smoky Moun- tains flora was used to develop and discuss com- munity level hypotheses on the eastern Asian- North American floristic relationship. NORTH TEMPERATE DISJUNCTION AND DIVERSITY The disjunctions between eastern Asia and eastern North American are generally at the ge- neric level, although a few plants are conspecific and others are disjunct only at subfamily oF subgeneric levels (Good, 1927; Li, 1952, 1972; Sharp, 1953; Graham, 1972; Wood, 1970, 1972; see also other papers, this volume). About 1 À percent of the genera in the southeastern Unite States is involved (Wood, 1972). Some 62 — are strictly disjunct; another 58 genera are nearly so, being strongly discontinuous in their so (Woo d, 1972; Li, 1952). There are an additions i: and climatic cooling during the last — mi lion years. These events caused fragmenta rse the Arcto-Tertiary geoflora, the most stem eastern Asia tandi in eastern North an less important in the latter area than papa? Arcto-Tertiary centers (e.g., Europe an western North America). Asia The ecological parallels between eastern REA win Field Research Lab, Great Smoky Mountains National Park, Twin Creeks, Gatlinburg, Tam ANN. Missouri Bor. GARD. 70: 734—747 1921 ee — il dka RI iat, — D a 1983] TABLE 1. Diversity ratios for the genera and fam- ilies of eastern Asian—astern North American disjunct vascular plants (data calculated from Li, 1952). WHITE—PLANT COMMUNITY LEVEL stern Eastern North Diversity Ratio n Asia America Genera/family 23 3.6 2.4 Species/genus: Temperate genera 37 27 1.9 Temperate-tropical genera 19 13.8 3.2 All genera 56 6.5 2.3 Species/family 23 23.4 55 and eastern North America have been studied toa lesser extent, but are also intriguing: 1) woody, broad-leaved, deciduous plants are prominent in the disjunct groups and dominate the forested landscapes of both areas (Li, 1952; Wang, 1961; Wood, 1972); 2) disjunct herbaceous taxa tend to be rhizomatous or tuberous, spring-blooming, early-leafing ephemerals or shade-adapted geo- . Phytes (Li, 1952); 3) broad-scale generic turnover ong environmental gradients is similar in the two areas (for example, both areas have Quercus- Acer-Betula-Picea south-to-north latitudinal gradients (Wang, 1961; Braun, 1950); and 4) Overall climatic regimes also are similar (e.g., precipitation is available year round). The sim- llarities in life history and physiognomy are pre- Sumably the result of descent from common ancestors that has taken place in similar envi- Tonments (Li, 1972) rather than the result of con- vergence because of similar environments. Nonetheless, these functional parallels have ap- rently survived some 50 million years of iso- tion, taxon AB E g t the specific] L and episodes of worldwide climatic change. Li (1952) also noted phylogenetic correlates to the disjunctions: Ranalian families are prominent in the list of disjunct groups. By far the most important observation at the . Community level is the comparative species rich- . Ress of the various remnants of the Arcto-Ter- tiary flora. Most of the disjunct genera have more Species in eastern Asia than in eastern North America (Li, 1952; Table 1). In addition, the families with disjunct genera in both areas often ve more total genera in eastern Asia than in ‘astern North America. The result is that there âre over four times as many species in the dis- Junct groups in eastern Asia (Table 1). In addi- tion, temperate Asian families and genera that ccm |f A MN. eren ttt a me Aa ll, arnt 735 presently are not native in eastern North Amer- ica are known from the fossil record there (no- tably Ginkgo, Metasequoia, Cercidiphyllum, Dipteronia, Pterocarya, and Zelkova; Axelrod, 1984). These same statements apply to the con- trast between eastern North America and Eu- rope: North America is richer in deciduous forest taxa than is Europe and some of the presently disjunct American-Asian taxa are known as fi sults from differing Pleistocene extinction rates in the three areas. Prior to the advent of man, eastern Asia harbored the world’s most diverse temperate forests. We should be cautious in accepting this ap- parently higher species richness in eastern Asia, however. Although disjunct groups are more di- verse in eastern Asia, other non-disjunct groups can be richer in North America (for example, this seems to be true in Asteraceae). Commu- nities from the two continents must be carefully compared. For example, Wang (1961) used the term “mixed mesophytic" forest in China for the Mf | raan Ai id rw f T ri gr u erly latitudes. The “mixed-mesophytic” forest of eastern North America is almost entirely decid- uous (Braun, 1950). Further, eastern Asia has a much more complete and complex temperate to tropical gradient on mesic sites (this is a gradient of increasing species richness); the gradient in m might also be differences in taxonomic treatment in the two areas. Despite these caveats, some of the increased richness in eastern Asia is hardly debatable (e.g., the striped-bark maples, Acer sect. th the distincti demic Asian gen- era already noted). If the heightened species richness is real, its resolution is fundamental to community level comparisons between eastern Asia and eastern North America. How is the added species rich- ness partitioned in the landscape? There are three possibilities (Whittaker, 1972): 1) higher alpha diversity (within-community species packing or point diversity); 2) higher beta diversity (closer species packing along gradients); and 3) higher gamma diversity (closer species packing across whole landscapes). Three kinds of gamma di- versity are: a) that caused by different habitat diversities within one landscape; b) that caused by the interplay of natural disturbance and 736 succession — I ana c) that €— by Y the "development of sibling species across geo- graphic barriers). At present, a complete resolution of these questions is impossible: quantitative data on Chinese forests are scarce. There is some ques- tion whether a resolution will ever be possible: that country has been under intense pressures of population growth for centuries (Smil, 1981) and has only recently formulated conservation poli- cies (Wang, 0). his paper, then, addresses these questions through an analysis of one relatively well-known eastern North American landscape, the Great Smoky Mountains, in North Carolina and Ten- nessee, which lies near the center of diversity of the eastern deciduous forest (Braun, 1950; Whit- taker, 1956; Monk, 1967). I examined the com- munity level using two aspects of the flora of this region: autecology (how the plants behave —their life history, growth form, gradient distribution), c as a function of phyto- geographical category. Primary questions were: Are the disjunct taxa a random assortment of the life histories and growth forms of the flora as a whole? Are the disjunct taxa randomly distrib- uted in the Smokies landscape? How is species richness distributed in the landscape? Although many of the questi ad been q ly ad- dressed previously (see Li, 1952; Cain, 1943), they previously had not been tested explicitly against appropriate random models THE STUDY AREA: THE GREAT SMOKY MOUNTAINS Great Smoky Mountains National Park (here- after, GRSM: 208,000 ha) is the largest United States National Park in a deciduous forest land- scape. It is located at 35?30'N latitude in the southern Appalachian Mountains (Fig. 1), near the richest part of the deciduous forest (Braun, 1950; Monk, 1967). Elevation ranges from 260 to 2,021 m. The climate is continental: July tem- perature averages 22°C at 445 m and 13.5°C at 1,920 m; January temperature averages 4.5°C at 445 m and —0.8°C at 1,920 m (Shanks, 1954). Precipitation averages 147 cm at 445 m and 231 cm at 1,920 m. Deciduous forest is dominant on mesic to subxeric sites below 1,900 m. Spruce- fir evergreen needle-leaved forest is dominant ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 above this elevation. Pine forest dominates the driest sites at low and mid elevations and hem- lock dominates cool stream valleys (Whittaker, 1956). The mountains are not high enough for a climatic tree line, but two kinds of tree-less com- munities (grassy balds and heath balds) occur; these are probably a result of disturbance and/ or exposure. METHODS The 1,211 native vascular plants of Great Smoky Mountains National Park (White, 1982) are divided into three phytogeographical cate- gories (this is usually done at the generic level— see Wood, 1972): 1) taxa strictly disjunct (¢.g., Caulophyllum) or strongly discontinuous (e.g. Aesculus), found in eastern Asia, eastern North America, and occasionally in one or two other Arcto-Tertiary refuges (e.g., Pacific northwestern North America and mountainous southeastern Europe); 2) genera widely distributed in the No Temperate zone (e.g., Quercus); and 3) North merican genera absent from eastern Asia (e.g.. Phacelia). There is continuous gradation from disjunct to discontinuous to wide-ranging gen- era; for disjunct taxa, the floristic judgments of Wood (1970, 1972) and Li (1952, 1972) are fol- owed. plants were coded for a series of auteco- logical vigi e E history, growth form, deciduousness, hei mature, life form (Raunkiaer, 1934), ae aie Community data developed through a yegetinon monitoring computerization of herbarium labels bet pur to code presence of the plant : species alon dominant s: 1) AID. 2) site moisture class (hydric, mesic, submesic- o distur- dients are strongly correlated with plant distri- bution in the southern Appalachians (Whittaker, 1956; Golden, 1974). _ 2 sapie versity) and along gradients (beta diversity) ‘vid derived from the distributional data descr above. These data were also used for gauging th effect of disturbance patches on species Ames Landscape (gamma) diversity was approach "t using an analysis of plant abundance and d bution; all species were coded for abundance ver evenness of distribution throughout the Fá KENTUCKY | 7 VIRGINIA ¢ 7 f TENNESSEE Æ GREAT SMOKY MTS. nyA NAT'L. PARK g NORTH CAROLINA ————————- L ES od ens di g 4" + GEORGIA {SOUTH CAROLINA n D\Elkmont cr n LeConte I Pal Fa NEWFOUND CAP 8!s aW Ü uu Cove i : j Spence Field — Trjj, TE ae. CLINGMAN Russell Field 9380 ee “Euan Cy Nun DOMES) gr ` D, Twontymi le FIGURE |. Location of Great Smoky Mountains National Park, North Carolina and Tennessee. [£861 THAT ALINAWWOO .LNV'Id —4.LIHAA LtL 738 (both taken from White, 1982). Only species richness was analyzed; dominance relations of species within communities were not addressed. Finally, a phylogenetic code was developed to test the distribution of taxa by presumed distance from the Ranales. Higher groups were er to three categories (Stebbins, 1974): Primi- tive—Superorders. Magnoliidae and eel dae; 2) I osidae, Ha- mamelidae, Dilleniidae, and Caryophyllidae, and m Bod is Commelinales, and Juncales; and ced — Superorder Asteridae, and Orders Orchidales Cyperales, and Poales. Plants in the three phytogeographical cate- gories were contrasted with the total native flora for each of the phylogenetic, natural history, and habitat variables. The null hypothesis in each case was that the plants in the ens geograph- ical categories represented a random sample of the flora as a Nue the probability level of 0.05 was used as the nificance. The data were analyzed in two ways: "m percent distribu- tion of ecological attributes within phytogeo- graphical categories (i.e., normalizing to a com- mon species pool, 100 taxa, within each category); and 2) percent distribution of phytogeographical gory to a particular ecological category. In the former case, the spectrum of ecological charac- teristics within a phytogeographical category was compared to that spectrum in the flora as a whole; in the latter case, the percent contribution of the floristic category to a particular ecological cate- gory was compared to its percent contribution to the flora as a whole. Statistical analysis and graphic presentation were done at the University of Tennessee Computing Center. RESULTS Thirteen percent of GRSM’s native vascular plants belong to genera that are disjunct or dis- continuous (hereafter called “disjunct” plants); 19 percent belong to genera that are absent from eastern Asia (hereafter “absent” plants); and 68 percent belong to genera with wide north tem- perate distributions (hereafter “no. temperate" plants) (Table 2). Because the latter group of plants also occurs in eastern Asia, 81 percent of the GRSM flora belong to genera that occur in east- ern Asia. Twenty-five percent of GRSM genera are in the disjunct category; these genera account for about 60 percent of the disjunct genera of Wood (1972). The disjunct genera are poorer ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 (average — 1.5) in species than the total native genera (average — 2.8) (Table 2). The absent gen- era (average — 1.9) also are relatively species- poor, whereas the widely distributed north tem- perate genera are relatively species-rich (aver- age — 3.5). The disjunct taxa have a lower species/ genus ratio than that extracted from Li (1952) for all of eastern North America (average = 1.9 species/genus, Table 1). The distribution of taxa in the likely to be **Ranalian" than the total native flora; disjunct taxa contribute 47 percent of all native taxa in this category, although they account for only 13 percent of the flora as a whole. By con- trast, only one percent of absent taxa are “Ran- alian," whereas 65 percent (the highest of any of the groups) are in the third, most advanced, phy- logenetic category. Disjunct taxa are strongly un- d in the latter aeo (14 percent e category 3; four a le of all phylogenetic uod plants are in As might be expected, then, Poaceae (Koyama & Kawano, 1964) and Asteraceae are relatively poor in disjunct taxa, but relatively rich in taxa in the absent category (Table 3). Taxa in widely distributed north temperate genera are generally a random selection of the flora as 4 whole; the only significant departure from a ran- dom model is an under-representation in the *Ranalian" phylogenetic category. AUTECOLOGY Life history. The disjunct taxa are more strongly perennial (98 percent of all disjuncts) and woody (51 percent of all disjuncts), than the flora as a whole (Table 4). Ninety-nine percent of the annual and 91 percent of the biennial plants are contributed by the non-disjunct phytogeo" graphic categories. The absent taxa are strongly herbaceous (92 percent); except for woodiness, however, these taxa are essentially random sam- but otherwise also are randomly compared to the flora as a whole. Basics the disjunct taxa have a significantly higher s centage of evergreens (eight percent), despite common observation that tiae. are associ- ated with deciduous forest landsca 42 ct Growth form and mature MM Disjun 1983] WHITE—PLANT COMMUNITY LEVEL 739 TABLE 2. Phytogeographical categories in the native flora of Great Smoky Mountains National Park (the distribution of the genera does not sum to 100 percent because of taxa that are disjunct at other than the generic level). Taxa Genera Species/ (n) % (n) % Genus Absent 232 19 125 29 1.9 Disjunct 162 13 108 25 I5 North temperate 817 68 231 53 3.5 Native flora 1,211 100 437 100 2.8 taxa are dominantly phanerophytes (42 percent) and geophytes (33 percent); these percentages are higher in the disjunct category than in any other phytogeographic category (Table 5). By contrast, taxa belonging to genera absent from eastern Asia are dominantly hemicryptophytes (54 percent) and have a relatively low percentage of phaner- ophytes (seven percent). The absent taxa are over- represented in therophytes (25 percent) and geo- Phytes (25 percent), whereas the disjunct taxa are over-represented in the phanerophyte (33 Percent), chamaephyte (40 percent) and geo- Phyte categories (23 percent). Disjunct taxa have a higher percentage distri- bution in the taller height classes (34 percent of all disjunct taxa reach at least 4 m in height; 55 Percent of all plants surpassing 4 m in height are disjunct taxa) than any other floristic category | (Table 5). Most taxa absent from eastern Asia _ 4re in the smaller height classes (95 percent reach | less than 2 m tall), and this phytogeographical , Category makes its strongest contribution to the flora in the smaller sizes (Table 5). Phenology and flower color. Seventy-two percent of the disjunct taxa bloom in spring and less than one percent in autumn (Table 6). The ak flowering time for absent taxa is summer (58 percent). Percent contribution to the phen- logi iods al flects this: disj ttaxa make their strongest contribution to the spring flora (22 percent), whereas absent taxa make their pe contribution to the aestival flora (29 percent). Flower color is also skewed: the most important color for disjunct taxa is white (44 percent of all disjunct taxa). White is also the most important flower color for absent taxa (28 percent), but that group has the highest percentage yellow (23 per- cent) and blue-purple (19 percent) flowers of any floristic category and makes a relatively strong total contribution to those categories (30 percent of all yellow and 34 percent of all blue-purple flowered taxa are absent from eastern Asia—the yellow-flowered taxa occur mostly in the Aster- aceae, a family poor in disjunct taxa, as noted above). Disjunct taxa also have a significantly (P < TABLE 3. Distribution of taxa in phylogenetic and phytogeographical categories (see text for an explanation pue gr Oups). The random model in Part A below is perc art B it is the percentage of the phytogeographic category in entage of native taxa in the phylogenetic category; in the flora as a whole. ** = P < 0.01; * « 0001; ++ = P < 0.0001; ns = not significant at the 0.05 level. Ec Phylogenetic Categories Phytogeo i i 2 3 Poaceae Asteraceae graphic Total Native 1 e1 Category Taxa (n Flora n = 62 n = 534 n = 545 n = 92 n = 138 A. Distribution within phytogeographic categories: pt 232 1+ 34** 65** 9 ns 20** Disjunct 162 19+* 67 14^ m 2+ North temperate 817 5 ns 47 ns 48 ns 8 ns 11 ns Native flora L211 6 46 48 8 11 B. Distribution within phylogenetic categories: 2 Absent 19 3+ 15** 28** 22 ns 34 Disjunct n o ud 20° 4 3" 2 68 50* 65 ns 68 ns 75 ns 64 ns . North temperate EL mu TABLE 4. Distribution of taxa in life history and phytogeographic categories (see text for an explanation of the latter). The random model in Part A below is the percentage of native taxa in the life history categories; in Part B it is the percentage of the phytogeographic category in the flora as a whole. * = P < 0.05; ** = P < 0.01; * = P < 0.001; ** = P < 0.0001; ns = not significant at the 0.05 level. Total 06 Life History Woodiness Deciduousness xa Native Perenn. Annual Bienn. Woody Herbac. Decid. Evergreen Phytogeographic Category (n) Flora n= 1,045 n — 131 n= 35 n = 246 n = 965 n= 1,137 n= 74 A. Distribution within phytogeographic categories: Absent 232 83 ns 14 ns 3 ns pr 92** 96 ns 4 ns Disjunct 162 9p" ist Sae 49++ 92 ns 8* North temperate 817 85 ns 12 ns 3 ns 28 ns 82 ns 94 ns 6 ns Native flora LZH 86 11 3 20 80 94 6 B. Distribution by life history: Absent 19 18 ns 25" 20 ns qe Ee 20 ns 12 ns Disjunct 13 15* le 9 ns 39** a 13 ns 20 ns North temperate 68 67 ns 74 ns 71 ns 60* 70" 67 ns 68 ns TABLE 5. Distribution by growth form and phytogeographic ene (see text for an explanation of the latter). The random model in Part A below is the percentage of native taxa in the life history categories; in Part B it is the percentage of the phytogeographic category in the flora as a whole. * = P < 0.05; ** = P < 0.01; + = P «0.001; ** = P < 0.0001; ns = not significant s. the 0.05 level. Raunkiaer Class Height when Mature (m) Total 96 emi- Phytogeographic Taxa Native Phaner. Chamae. crypt. Geo. Thero. L} 1-2 2-4 4-8 28 Category (n Flora n=235 n=10 n=605 n=229 n=132 n=799 n-212 n=76 n= 21 n= 103 A. Distribution within phytogeographic categories: Absent 232 T 0.4ns 54ns 24* 14 ns 72 25° 3" 0.4** ptt Disjunct 162 apt 3 ns 16t* 33 Vr 44** or Iz Uer Y. de North temperate 817 17 ns OGns .50** i ES ei 12 ns 69 ns 18 ns 6 ns 0.6** T Native flora 1,211 19 0.8 50 19 11 66 17 6 2 9 B. Distribution by growth form: bsent 19 Tati 10 ns 21 ns Fiabe 2° 20 ns 25" or o" Et unct 13 3317 Do" 4t 230 E 9+ yt 26 21v durs North temperate 68 60* 50 ns IFF -PAg 73 ns 71% 68 ns 65 ns 27 58* OtL N3GIV5 'IVOINV.LO8 PEIOOSSIN JHL JO STVNNV 0L 10A] - din — - — tet te oe it LE 6. Distribution ^ Phenology, flower color, and phytogeographic category — TY — MÀ asi i — to 100 Peranuse. E plants bloom - decade the year). The random model in Part A below is the percentage of native taxa by fl in P. g g category in the flora as a whole. * = < 0.05; ** = P < 0.01; + = P < 0.001; ++ = P < 0.0001; ns = not i a at the 0. 05 level. % Phenology Flower Color Total Na- Late Blue- Pink- Phytogeographic Taxa tive Vernal Vernal Aest. Autumn. Brown Green purple Red Orange Yellow White Category (n) Flora n=79 n=562 n=467 n=77 n=21 n=343 n=130 n=91 n=13 n=178 n=365 A. Distribution within phytogeographic categories: Absent 232 qme 334 58** 4 ns 2 ns I5 19** 11 ns 2 ns 23°" 28 ns Disjunct 162 14++ fees 25 Dg p 27 ns t fs 9 ns 0.7 ns or 44+ North temperate 817 6 ns 48 ns e e tag o7 36t* 11 ns 7 ns | ns 15 ns 31* Native flora 1271 T 46 39 6 2 30 11 8 l 16 32 B. Distribution by floral characteristics: Absent 19 13 ns 3 29r 13 ns 24 ns 10** 34** 28 31 ns 30** 18 ns Disjunct 13 Zo 16* per iis 526 13 ns 3» 14 ns 8 ns gu Oo North temperate 68 59* 71 ns P ana 86+ 24** TI GL 58* 61 ns 62 ns 63* TABLE 7. Distribution by habitat and PoE de p (see text for an explanation of the latter). The random model in Part A below is the percentage of native taxa in the life history categories; in Part B it is the percentage of the phytogeographic category in the flora as a whole. * = P < 0.05; ** = P < 0.01; ** = P < 0.0001; ns = not significant at k 0.05 deis Habitat (Substrate) Categories Total % Rack Habitat Openness Phytogeographic Taxa Native Terrestrial Wetlands Outcrops Epiphytic Forest Intermed. Open Categories (n) Flora n= 932 n = 230 n= 36 n= 13 n — 488 n= 396 n 327 A. Distribution by phytogeography: Absent 232 82* 14” 3 ns 0.4 ns 41 ns 34 ns 25 ns Disjunct 162 91+ ot 1” 0 ns 66** 28 ns 6** North temperate 817 io 22* 4 ns pt 35 ns 33 ns 32* Native flora E21 77 19 3 0.5 40 33 27 B. Distribution by habitat: Absent 19 20 ns 14* 19 ns 7 ns 19 ns 20 ns 17 ns Disjunct 13 15* ya 4* 0* Zo 12 ns 3** North temperate 68 55* 8I: TH 93 ns 3921 68 ns 80** T3A31 ALINQWWOO LNV'Id ^ 3.LIHAA [£861 Ivl 742 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 A. % TOTAL ABSENT TAXA B. % TOTAL DISJUNCTS 2000 2000+ 1 6000 6000 2 5000 5000 g| 1500 = z| 1500F z 2 = z z $ 5 9 5 E 4000 |& E 4000 | q S = S S WW ul Lu ZI NE i Em c ü Lu uJ W 1000 3000 1000F 3000 2000 6 62 48 12 42000 Hydric | Mesic Submesic— subxeric Xeric Hydric Mesic Submesic— subxeric Xeric MOISTURE GRADIENT MOISTURE GRADIENT C. % TOTAL NATIVE FLORA 2000 6000 3 5000 g| 1500 È z z o o = 4000 |5 E E: Lu Lu PS m Lu Lu 1000 3000 2000 Hydric | Mesic Submesic — subxeric Xeric MOISTURE GRADIENT urs dt 2. Percent distribution of GRSM — along a bona anon bcp" field. ec ^ are y ten percent classes; represen ta are cited for low ations an generalized hydric Dm tland), m C. Total native flora (1,211 tax 0.001) earlier peak leaf period (13 percent of all disjuncts and 31 percent of all disjunct herbs have peak leaf periods in spring) than either the absent (eight percent of all taxa, 25 percent of herbaceous taxa have peak leaf periods in spring) or north temperate groups (seven percent of all tativ , and high elev fe ). mesic, submesic-subxeric, and xeric sites. A. Absent taxa (232 taxa). B. Disjuncts (162 taxa a). taxa, 15 percent of herbaceous taxa have peak leaf periods in spring). However, m p er t the peak geographical categories pe Rest dominantly summer (no significant : from the total native flora value of 85 perc? are found). 1983] WHITE—PLANT COMMUNITY LEVEL 743 ABSENT 3 B DISJUNCTS % Flora/site f %_Flora/site : 2000 46000 13 13 6 io os um 45000 E 1500F 10% = E i500} 10% pru 5 - = 20% = $ 20% i m z z= È o = o 1 4000 s g E W 22 H 2096 [m ul i d Í le 8 d 30% 24 17 n 1000+ 43000 1000+ 22 32 10% 14 8 2000 Hydric | Mesic Submesic— subxeric Xeric | Hydric | Mesic Submesic—subxeric Xeric | MOISTURE GRADIENT MOISTURE GRADIENT EAST ASI NORTH TEMPERATE C. % FAM. js D. % Flora/site 2000 65 77 81 E = g| '5oor = e = z z o 5 2 $ c as = m eed s 13 3 E a uo — 4 ù d a ` 64 62 d 1000F 70% 7i 62 63 LN Woran rer Hydric | Mesic Submesic- subxeric Xeric | mesic — su MOISTURE GRADIENT oU GRADIENT FIGURE 3. ge cane of GRSM E id L4 total native flora is similar to that of the absent E S Rol -— d taxa, in that there is a submesic-subxeric peak Pi S at low elevations and a gradual decline toward er 2 : is moisture extremes. = Bp g GL o a Two simplifications are possible to better con- gs3 S8/9225 trast the three floristic groups in terms of land- > sS |. 40 scape distribution. First, because disjunct taxa E E > E are unimportant on open sites (Liriodendron is E = be m So E an interesting exception), the data base was re- PE be E 8 g mes E duced to plants of forested sites only. Second, E i S e b FEES because I overs? sperma richness gradient par- > 2 E z allels that 253 (Fig. hen the data were v expresiédia asa percent Oss a n ontribution to the total flora expected on a given $ 9 ^ E E he = E e site (Fig = n this manner, most of the S E ^ olas Rede a information on disjuncts available in the 3-di- B Wd mensional array (elevation, site moisture, habi- a 3 S tat openness) is portrayed in two dimensions. cee $342|22-zz en the data are expressed in this way, the Se M R E Sison absent taxa show a peak importance in xeric sites, E E il a secondary peak in hydric sites, and a trough in 2 s t mesic sites (Fig. 3A). Disjuncts show a very dif- 22 = ferent distribution-their peak importance is on Ioe mesic sites (Fig. 3B). Taxa of wide-ranging north È ra is z : temperate genera show yet a third pattern, with ^ $a 2 5 peak importance at high elevations (the only "3 E 2 g group to show such an increase with elevation; € 3 = E] £ 5 E Fig. 3C). The north temperate taxa are least im- 2% xs S E G g z > portant on xeric low- to mid-elevation sites; these É ^s V 2s È = 53 taxa have a secondary peak on hydric sites *& ROLLARI cm — 1983] WHITE—PLANT COMMUNITY LEVEL 745 TABLE 9. Narrowly distributed and widely distributed species and an index of beta diversity for two broad- scale gradients. Narrowly Restricted Widely Distributed Species Turnover Phytogeographical Category e Species Ratio A. Elevation gradient: Absent 115 127 0.9 Disjunct 60 102 0.6 North temperate 429 388 Li Native flora 604 607 1.0 B. Site moisture classes: Absent 133 95 1.4 Disjunc 86 76 1.1 North temperate 505 312 1.6 Native flora 724 487 1.5 Disjunct taxa and taxa of wide-ranging genera both occur in eastern Asia. The importance of these two groups is summed in Figure 3D, which Shows the percent of forest flora belonging to genera common to eastern Asia as a function of elevation and site moisture class. Eastern Asian site moisture classes. Xeric sites have the lowest Importance in distribution of eastern Asian gen- era. Overall, 85 to 95 percent ofthe taxa on mesic E belong to genera that also occur in eastern sia If life history and environmental distribution are combined, these patterns are reinforced. For- ty-two percent of all woody taxa on mesic low- to mid-elevation forested sites belong to disjunct Benera; 95 percent of all woody plants on mesic low- to mid-elevation forested sites belong to &enera that also occur in eastern Asia. RARITY AND PARK DISTRIBUTION Tiie ntract The ed in rarity and park distributibar- in order to gauge the effect of sporadic distributions on the richness gradients as a whole (Table 8). Disjunct laxa have a lower percent rarity, both in terms of the rarest elements of the pax fiors and the sum of very rare, rare, Disjunct E are over-represented among frequent and mon plants. Disjunct plants are also well- Bud throughout the Smokies (75 percent of all disjuncts are found throughout the park, the highest percent of any of the phytogeographic groups). The absent taxa are generally a random sample of the flora as a whole, but intermediately common taxa are dis ctii and common taxa are under-represe Another aspect of epa breadth was also analyzed: the phytogeographical categories were contrasted in the percentage of taxa restricted to single habitat classes along the three dominant environmental gradients. Two of these gradients were already coded as broad habitat classes: site moisture class (hydric, mesic, submesic-subxe- ric, xeric) and habitat openness (forested, inter- mediate, open). Elevation data were summarized in three belts: low (260-760 m; 850-2,500 ft), mid (760-1,370 m; 2,500-4,500 ft), and high (1,370-2,021 m; 4,500—6,621 ft). The results show that disjunct taxa are narrowly distributed along only one of the gradients: habitat openness (Ta- ble 8), as noted earlier. In other ways (elevation, site moisture class), the disjunct taxa are less restricted to habitat classes than would be ex- pected based upon chance alone. The other phy- togeographical categories are essentially random samples of the flora as a whole in terms of habitat restrictions (Table 8). The data on gradient restrictions were also used for an index of gradient (beta) diversity. The number of species replacements between the habitat classes (the number of new restricted species encountered in each habitat class) was compared to the number of widely distributed taxa (Table 9). The north temperate taxa have the highest species turnover ratio, whereas the disjunct taxa have the lowest. All phytogeo- turnover ratios for (TR ntl 746 the site moisture class gradient compared to the elevation gradient. DISCUSSION Taxa belonging to disjunct genera in GRSM are clearly a non-random sample of the native flora: they are much more likely to be perennial, woody, primitive or intermediate in phyloge- netic terms, and earlier in blooming and leafing. Disjunct herbaceous taxa are more likely to be geophytes than are non-disjunct taxa. The taxa belonging: to disjunct genera are also non-ran- likely to be found on terrestrial, remis and mesic sites than, are non-disjunct tana. Disjunct d more widely distributed in the park than their non-disjunct counterparts. Unexpectedly, taxa belonging to genera absent from eastern Asia proved just as interesting as the disjunct taxa. For many attributes they also are a non-random assortment of the flora as a whole: absent taxa are more advanced in phy- logenetic standing, more herbaceous, less apt to be phanerophytes, shorter in mature height, later blooming (with a stronger representation of yel- low flowers), and later leafing. Absent taxa are, however, essentially a random sample of the flora as a whole with regard to substrate and habitat openness. Absent taxa also parallel the whole native flora in gradient distribution: they show a peak occurrence in low elevation submesic- subxeric sites. When the gradient patterns are corrected for the overall distribution of species richness, absent taxa have their peak importance on xeric sites where they make up 20-35 percent of the flora. McVaugh (1943) found that xeric site species (on southeastern United States gran- ite outcrops) had strong — ties to western North America. Alt the xeric sites analyzed in GRSM are don RM by pine forests, rather than open outcrop communities, the pattern is similar. The taxa of genera absent from eastern Asia are often ones with south- western United States relatives. These facts support earlier observations on prominent features of the disjunct taxa: their woodiness (Li, 1952; Wood, 1972), primitive- ness (Li, 1952; Sharp, 1953), and habitat distri- bution (Cain, 1943). Further, there is a clear re- lationship between life history traits and habitat. Early leafing, geophytic, herbaceous plants are adapted to mesic deciduous forests; in this way ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vor. 70 the disjunct taxa parallel the gradient distribu- tion of life history traits in the Southern Appa- lachians as a whole (Thomas Givnish, pers. comm.). The importance of large reves is also related to habitat: mesic, low- to m elevation forests have the largest biomass es any GRSM communities. The importance of pha- nerophytes in favorable environments is one of the broad-scale generalizations of the geography of life forms. Several inferences concerning the contribution of disjunct taxa to the pattern of species richness in the GRSM landscape can be made from these data. Disjunct taxa do not contribute to the sev- eral kinds of gamma (whole landscape) diversity. Disjunct taxa are unimportant in open, distur- bance-produced habitats. They are relatively fre- quent plants with uniform distributions through- out the park. They do not increase GRSM richness through occurrence in rare habitats; neither are they rare in common habitats. As a component of the flora, they make no disproportionate con- tribution to landscape richness per se: they are not patchily distributed. Disjunct taxa are also relatively unimportant in their contribution to species turnover along gradients (beta diversity). Disjunct taxa decrease more sharply in richness along gradients than the flora as a whole. They do not disproportionately increase richness on extreme moisture class sites or at high elevations (where total diversity 1s rel- atively low). Individual taxa are not more nar- rowly distributed on gradients than non-disjunct taxa, except for habitat openness. The strong in- ference is that disjunct taxa do not have a higher turnover of taxa along gradients than the flora as a whole, and do not contribute dispropor- tionately to beta diversity. The disjunct plants are, however, concentrated in mesic low- to mid-elevation forests. These are relatively species-rich habitats for the GRS i indscape as a whole. The contribution 0 — '5 was, in fact, the concentration of Arcto : plants in mesic low- to mid-elevation forests tha led Cain (1943) to infer 60 million years of stasis in Southern Appalachian vegetation (since sd futed, Delcourt & Delcourt, 1979, 2: che can summarize t increase the richness of peti that - relatively rich in non-disjunct taxa. This € contri 1983] bution is habitat concentrated, rather than dif- fused across the landscape or along gradients. These conclusions only can be extended be- yond the study area with caution, however. Only 60 percent of Wood’s (1972) disjunct or discon- tinuous genera occur in GRSM. For all disjunct genera, Li (1952) reported a higher species/genus ratio than that reported here. This is true for North American species of these genera (1.9 species/genus vs. 1.5 for the Smokies) as well as for eastern Asian species, as discussed earlier. Thus, there are, on the continental scale, addi- species richness. Finally, if the eastern Asian flora was less affected by Pleistocene changes than that of eastern North America (Delcourt & Delcourt, 1979, 1981), there are implications for distri- bution of species richness. Less environmental change might leave ecological barriers to migra- tion intact and might pose less of a threat to rare, patchily distributed species. These questions will be resolved only as quantitative data become available from the temperate deciduous forests of eastern China, the richest such forests on earth. LITERATURE CITED AXELROD, D. I. ree fie Sie. of oaks in the o-Terti ce. Ann. Missouri Bot. Gard. 70: 629-657. BRAUN, E. L. Deciduous Forests of Eastern ele ertiary Arbo of the cove h ardwood forests of Great Smoky Mountains Na- tional Park. Bull. Torrey Bot. Club 70: 213-236. DELCOURT, P. A. & H . R. DELCOURT. Pleist the deciduous forest in the southeastern United States. Veróff. Geobot. Inst. ETH Stiftung Rübel ES ia 68: 79-207. . 1981. Vegetation maps for eastern E North America: 40,000 yr b.p. to the present. Geo- Forest vegetation and site re- 1974. lationships in the central portion of the Great k tai . Disser- e Goop, R. 1927. A summary of discontinuous dis- tributions in the Angiosperms. New Phytol. 26: 249-259, GRAHAM, Á. (editor) 1972. Floristics and Paleo floristics of Asia and Eastern North America. she tni sevier Publ. Co., New York. Hu, H. H. 1936. A comparison of the ligneous flora of China and — North America. Bull. Chin Bot. Soc. 2: 67- WHITE—PLANT COMMUNITY LEVEL 1979. Late of 747 Koyama, T. & S. Kawano. 1964. Critical taxa of America. Trans. Amer. 9 astern Pisalo North America species pair in wider ranging genera. Pp. 65-78 in A. Graham (editor), Floristics and Paleofloristics of Asia and Eastern North America. Elsevier, New Yor MCVAUGH, R. 1943. The vegetation of the granitic flat-rocks of the southeastern United States. Ecol. —166. 67. Tree species diversity in the east- ern deciduous forest with particular reference to north central Florida. Amer. Naturalist 101: 173— 187 RAUNKIAER, C. 1934. The Life Forms of Plants and Statistical Plant Geography. Clarendon Press, Ox- 4. Climates of the Great Smoky 361. a of Mexico world distribution v the woody dicotyledonous families n of modern vegetatio Ecol. 41: 3 Smit, V. 1981. Land use and management in the People’s Republic of China. Environ. Managem 5: 301-311. STEBBINS, G. L. 1974. Edere Pies. Evolution fis, ridge and gen WaNG, CHi-Wu. 1961. The Forests of China. Maria 0:5. I atu ende Vation in China: the present situation Parks 5: 5-10. WhiTE, P. S. 1979. Pattern, process, and natural dis- turbance in fé crisis Bot. Rev. (Lancaster) 49: 29-299. 1982. The flora of Great Smoky Mountains National Park: an annotated checklist of the vas- cular p USDI, National Park Service, Reseaich/Re- sources Management Rept. SER-55. WHITTAKER, R. 1956. Vegetation of hurt Great Smoky Mountains. Ecol. Monogr. 26: 1-80. 1972. Evolution and measurement of species diversity. Taxon 21: 213-251. p, C. E, Jn. 1970. Some credant Panonio and w > between the Southern aegis —404 stitute and State University. Res. Div. Monogr. 2. irginia. . 1972. Morphology and phytogeography: t Peces approach to the study of d. n. Missouri Bot. Gard. 59: 107-124. ISOENZYME VERIFICATION OF AMERICAN-CHINESE HYBRIDS OF LIQUIDAMBAR AND LIRIODENDRON' SHAN-AN HE? AND FRANK S. SANTAMOUR, JR.? ABSTRACT Control-pollinated interspecific hybrids between Liguidambar Jorat ke and L. formosana and between Liriodendron tulipifera and L. chinese were verified by co rison of isoperoxidase banding patterns developed by starch gel electrophoresis of cambial iinet The relationships between some of the trees of eastern Asia and eastern United States (Li, 1952) are especially evident in Liriodendron, in which the only two species, the American Lir- iodendron tulipifera L. and Liriodendron chinese Helmsley, occur in these regions. There are only iquidambar acalycina and Liq- uidambar formosana Hance are native to China. A fourth species is found in Turkey. se two genera were among the first to be investigated after the initiation of a project on cytogenetics, breeding, and improvement of landscape trees at the U.S. National Arboretum in 1967. Santamour (1972a) reported successful crossing among three Liguid. pecies, using leaf morphology to verify interspecific hybridity. Because verification of young hybrids would be more difficult in Liriodendron, in which leaf shape of the two species is very similar, Santamour used the biochemical evaluation of leaf flavo- noids (Santamour, 1972b) to establish the true hybridity of seedlings from crosses of L. chinese made in 1970. This was thought to be the first controlled hybridization between these species. However, similar research in China was un- known in the West. According to the Nanjing Technological College of Forest Products (1980), they and the Jiangsu Institute of Botany had suc- cessfully crossed the two tuliptrees first in 1963, and Un several times between 1963 and 1980. hey rybrids in growth to L. ae at 12 years of db" In addition, Huang and Chen (1979) verified both natural and artificial interspecific hybrids ! The authors gratefully "crines id bee technical assistance of Alice Jacot McArdle rden M Yat-Sen, Nanjing, People’s Republic of China. Visiting Scholar * Nanjing Botanical bise rris tional Arboretum, 1981-1982. by analyses of isoenzyme banding patterns from dormant bud extracts. They found that by ac- rylamide gel electrophoresis of peroxidase iso- enzymes they could distinguish the hybrids that contained the isoenzyme bands of both parents as well as two new bands. Inasmuch as our joint research efforts at the U.S. National Arboretum involved considerable work in isoenzyme elec- trophoresis, we decided to investigate the poten- tial of using this technique beg Men ss hier ambar an U.S. National Arboretum and the eee hybrids produced in China. MATERIALS AND METHODS All of the parent trees and the progenies de- rived from controlled crosses at the National Ar- boretum were available for study. The first ma- terial received from China included dormant twigs from one tree each of the parent species and one putative hybrid. A second shipment con- pei dormant twigs from three putative hy- ahead most isoenzyme studies in plants have utilized leaf tissue, Santamour and Demut (1980) found that cambial tissue was equally as effective for isoperoxidases. Furthermore, stud- ies of cambial isoenzymes can be made at any time of the year and dormant twigs can be shipped SPRE méthddy VS Ur faik "- electropho- resis and staining of cambial peroxidase isoen- zymes were the same as those reported by San- tamour (1982) and have been successful in a wide range of woody genera (Santamour & Dem uth, 1980, 1981). at the S. National psa USDA Agricultural Research Service, Washington, D.C. ANN. Missouri Bor. GARD. 70: 748-749. 1983. 1983] ui o F xu: Ts mat. ,, Cambial Mr Ves banding aperiat in rin. Ao Liriodendron chinese, ee Liriodendron tulipifera, (D) Liquidambar formosana, mbar styraci- flua x * Liquidambar Jormosana, and (DL iquidambar styra RESULTS AND DISCUSSION Isoperoxidase banding patterns of cambial tis- interspecific hybrids showed many, but not all, of the bands found in the parent trees, but the hybrids were easily distinguished. The lack of Some parental bands in the hybrids Lush be expected in view of our lack of knowledge con- cerning the nature and dominance of the genes Involved. Houston and Hood (1982) were able to determine the inheritance patterns « of only six of the 13r y found in leaf tissue of tuliptrees, and those were gov- €rned by a single pair of alleles. With cambial isoperoxidases, it is important that the original parents be used in comparison to the hybrids, and not just an individual of the parent species. We found that it was impossible to verify the tuliptree hybrids created in the Peo- ple's Republic of China by cambial isoenzymes because the parent trees were not available for study. However, as Huang and Chen (1979) not- ed, the hybrids could be verified by dormant bud HE & SANTAMOUR —ISOENZYME 749 analysis. We also analyzed dormant buds of the Chinese material, and although we could distin- guish hybrids, the isoenzyme bands were not very clear. The use of cambial peroxidase isoenzyme di g patt i pt bl thod for ver- ifying interspecific hybrids in the cases studied here, especially when material of both parents is also available for study. Our work in Acer (San- ta have similar patterns of isoenzyme variation. Thus, an individual of one species may or may not possess a different banding pattern than an individual of another species. Two individuals of the same species, however, may have suffi- ciently different banding patterns that would al- low the detection of intraspecific hybrids be- tween them. LITERATURE CITED Houston, D. B. & S. K. Hoop. 1982. Genetic vari- ation in peroxidase isozymes of Liriodendron tu- lipifera L. J. Hered. 73: 183-186. HuANG, MINREN & DAOMING CHEN. 1979. An iso- zyme analysis of tulip-tree hybrids (Liriodendron chinese x L. tulipifera). J. Nanjing Technol. Coll. Forest Products 1-2: 156-158 (In Chinese, with Trans. NANJING TECHNOLOGICAL COLLEGE OF FOREST s. 1980. Genetics and breeding of forest trees. Science Press (In rere SANTAMOUR, FRANK S., JR. 1972a. ‘Interspecific hy- bridization in ose a Forest Sci. 18: 23~ ——. ` 1972b. Interspecific hybrids in Liriodendron and their chemical verification. Forest Sci. 18: 233~ 236. . 1982. Cambial peroxidase isoenzymes in re- lation to systematics of Acer. Bull. Torrey Bot. 109: 1 1980. Identification of idase isozym lery banding a ; Hered. 7 —— & 1981. Variation in cambial per- oxidase isozymes in Quercus species, prove- nances es. Northeast. Forest Tree Impr. Conf. nd ^T 55-65 (1980). Systematics Symposia Published In 1953 the Missouri Botanical Garden began holding a series of annual systematics symposia each fall. Since 1969, with the sixteenth symposium, the Garden has published the proceedings of most in its quarterly series, ANNALS OF THE MISSOURI BOTAN- ICAL GARDEN. Those published symposia which are available are listed below, and the price of all except the twenty-fourth is $10.00. The twenty-fourth, which was reprinted separately from the ANNALS, is available for $6.00. Many of these symposium issues of the ANNALS contain other papers, but each issue listed below contains the complete proceedings of the particular symposium. The proceedings of the 1981 symposium will be published during the second half of 1982, and if you wish to purchase a copy of that symposium, you may subscribe at this time. Upon publication, your copy will be sent. Sixteenth Annual Symposium, 1969—TROPICAL ISLAND BIOGEOGRAPHY. 56(3). 127 pp. $10.00. Seventeenth Annual Symposium, 1970—HYBRIDIZATION, EVOLUTION, AND SYSTEMATICS. 59(3). 142 pp. $10.00. Twentieth Annual Symposium, 1973—PLANT-ANIMAL COEVOLUTION. 61(3). 132 p. $10.00. Twenty-first Annual Symposium, 1974—BIOGEOGRAPHY. 62(2). 143 pp. $10.00. Twenty-second Annual Symposium, 1975—EVOLUTION AT THE POPULATION LEVEL. 63(2). 100 pp. $10.00. Twenty-third Annual Symposium, 1976—CHEMOSYSTEMATICS. 64(2). 98 pp. $10.00. Twenty-fourth Annual Symposium, 1977 —SYSTEMATICS STUDIES IN AFRICA. 223 pp. $6.00. Twenty-fifth Annual Symposium, 1978 —ÉPALYNOLOGY AND SYSTEMATICS. 66(4). 248 pp. $10.00. Twenty-sixth Annual Symposium, 1979 — EVOLUTION AND SYSTEMATICS OF THE GRAMINEAE. 68(1). 104 pp. $10.00. Twenty-seventh Annual Symposium, 1980— m STRATEGIES IN PLANTS AND ANIMALS. 68(2). 146 pp. $10.0 Twenty-eighth Annual Symposium, 1981 DOE, STUDIES IN CENTRAL AMERICA. 69(3). 304 pp. $15.00. Twenty-ninth Annual Symposium, 1982—BIOGEOGRAPHICAL RELATIONSHIPS BETWEEN TEMPERATE EASTERN ASIA AND TEMPERATE EASTERN NORTH AMERICA. 70(3 & 4). 327 pp. and map. $15.00. To place an order use this form or a photocopy of it. Orders should be prepaid; a $1.00 fee will be added to orders requiring invoices. No shipments are made until payment is received. Mail form with your check or money order, payable to Missouri Botanical Garden s Department Eleven Missouri Botanical Garden P.O. Box 299 St. Louis, MO 63166-0299 U.S.A. Please send the SYMPOSIA checked above to: Name O Payment enclosed. O Send invoice ($1.00 fee will be added to total) Address Postal Code Country Contents continued from front cover The Distribution of Scrophulariaceae in the Holarctic with Special Ref- erence to the Floristic Relationships between Eastern Asia and East- ern North America Hong De- Yuan A Comparison of Orchid Floras of Temperate North America and East- ern Asia Chen Sing-chi Phytogeographic Relationships of Pteridophytes between Temperate North America and Japan Masahiro Kato & Kunio Iwatsuki .. Eastern Asian—Eastern North American Floristic Relations: The Plant Community Level Peter S. White Isoenzyme Verification of American-Chinese Hybrids of Liquidambar 734 and Liriodendron Shan-An He & Frank S. Santamour, Jr. ...... 748