JOURNAL OF THE ARNOLD ARBORETUM HARVARD UNIVERSITY VOLUME 70 NUMBER 1 ISSN 0004-2625 Journal of the Arnold Arboretum The Journal of the Arnold Arboretum (ISSN 0004-2625) is published quarterly in January, April, July, and October for $70.00 per year, plus $10.00 or $15.00 postage for addresses outside of the United States, by the Arnold Arboretum of Harvard University. It is printed and distributed by Allen Press, Inc., 1041 New Hampshire Street, Lawrence, Kansas 66044. Second-class postage paid at Lawrence, Kansas. POSTMASTER: send address changes to Journal of the Arnold Arboretum, % Allen Press, Inc., P.O. Box 368, Lawrence, Kansas 660 Subscriptions and remittances should be sent to Journal of the Arnold Arboretum, 1041 New Hampshire Street, Lawrence, Kansas 66044, U.S.A. Claims will not be accepted after six months from the date of issue. Volumes I-51, reprinted, and some back numbers of volumes 52-56 are available from the Kraus Reprint Corporation, Route 100, Millwood, New York 10546, = A EDITORIAL COMMITTEE S. A. Spongberg, Editor E. B. Schmidt, Managing Editor P. S. Ashton K. S. Bawa P. F. Stevens C. E. Wood, Jr. Printed at Allen Press, Inc., Lawrence, Kansas COVER: The stylized design appearing on the Journal and the offprints was drawn by Karen Stoutsenberger. THIS PUBLICATION IS PRINTED ON ACID-FREE PAPER. JOURNAL OF THE ARNOLD ARBORETUM VOLUME 70 JANUARY 1989 NUMBER | TAXONOMIC STUDIES IN THE PIPERACEAE~—I. THE PEDICELLATE PIPERS OF MEXICO AND CENTRAL AMERICA (PIPER SUBG. ARCTOTTONIA) ALLAN J. BORNSTEIN! The taxon 4rcfottonia, originally proposed by Trelease as a new genus to accommodate the northern element of Neotropical pipers with pedicellate Ottonia were previously allied is considered to be a homoplasious character (Callejas, 1986) that occurs in three distinct evolutionary lineages. As currently circumscribed, Piper subg. Arctottonia is distinguished by its pedicellate flowers arranged in a loose raceme, its tendency toward adnation of the stamens to the ovary, its palmate-veined leaves, and its distribution from central Mexico to northern Central America. Fourteen species are recogn nized, oF which five are described as new. A key is given for the subgenus; ustrations and specimen citations are provided for the species; and an eee is pre- sented for recognizing Arctottonia and all other segregates from Piper sensu lato at the rank of subgenus. The Piperaceae Agardh are a large pantropic family of well over 3000 de- scribed species, with major centers of diversity in northern South America and Central America in the New World, and Malaysia in the Old (Yuncker, 1958). The family can be easily recognized on a gross morphological basis. The leaves are always simple and, with rare exceptions, entire; they are usually alternate, although the arrangement is somewhat variable. The flowers are minute, are perfect or unisexual (the plants then typically dioecious), and lack a perianth. The inflorescence (or infructescence) is a compact (or sometimes loose) spike or rarely a raceme. Each flower is subtended by a small, variable bract, and the ovary is unilocular with a single basal ovule. The fruit is a small drupe 'The Arnold Arboretum of Harvard University, 22 Divinity Avenue, Cambridge, Massachusetts 02138. Present address: eae of Biology, Southeast Missouri State University, Cape Gi- rardeau, Missouri 63701-479 © President and Fellows of Harvard College, 1989. Journal of the Arnold Arboretum 70: 1-55. a 1989. 2 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Despite a considerable amount of attention during the past century and a half, the Piperaceae continue to be taxonomically difficult and enigmatic. The present chaotic taxonomic state of the family is largely a result ofits unfortunate past treatment. Nowhere is this more true than for Piper L. sensu lato, which is perhaps the largest genus of flowering plants devoid ofa natural classification. The primary reason for this situation is the large number of described species, conservatively estimated at over 1500 for Piper alone. Many of these species, especially those established since the turn of the century, have been described without any reference to their putative position within the genus. Several factors have contributed to this artificially inflated species number. First, the flowers are small and the vegetative and floral morphology is uniform throughout the genus (Burger, 1972). Second, species descriptions have often been based on insignificant characters of little taxonomic value, or on fragmentary material often lacking flowers or fruits essential for identification. Third, new species are established based on their occurrence in a distinct political distributional unit, despite their similarity with previously described species. Both C. de Candolle and Trelease (two individuals who had a significant impact on the taxonomy of the family) were especially guilty in this respect, a situation that Standley and Steyermark (1952, p. 229) discussed as follows: In later years, however, [C. de Candolle], with more abundant material for study, became too enthusiastic in proposing new continental species, and named a very that the species of Piperaceae were extremely local, as many of them are. However, he went still further, and assumed that those of one country were distinct from all those (except a very few widespread and more or less “weedy” species) of the adjoining countries. Such narrow limits of distribution are not found in any other group of Central American plants, nor, in the opinion of the writers, does it exist in the Piperaceae. In his recognition of limiting geographic units, Trelease was scarcely consistent. He treated the whole of Mexico as one Piperaceous unit, and each of the Central American countries as a unit of equal standing. Now each o these Central American countries is about equal in area to one of the states of Mexico, obviously making such an fre improbable Within the last 25 years there have been significant advances toward com- bating the nomenclatural problems created by this overabundance of names. Much of this progress has been achieved through floristic treatments of specific geographic and/or political units (e.g., in the Neotropics, Burger, 1971; Yunck- er, 1972, 1973, 1974; Howard, 1973; Steyermark, 1984). Less frequent have been revisions of small segregates from Piper sensu lato (Steyermark, 1971; Smith, 1975). While the former method is perhaps more significant from a nomenclatural standpoint in that it tends to include far more taxa (and therefore more names), the latter can accomplish similar nomenclatural results while also addressing questions concerning evolutionary relationships within Piper. This paper, which treats the pedicellate pipers of Mexico and Central America, uses the latter approach. The goal is to provide a much-needed update of the taxonomy of this natural group of closely related species, with the ultimate intention of making progress toward a phylogenetic classification of Piper sensu lato 1989] BORNSTEIN, PIPERACEAE 3 TAXONOMIC HISTORY Trelease (1921) considered the pedicellate pipers of Mexico and Central America as the North American counterpart to a group of primarily South American species known as Ottonia Sprengel or Piper subg. Ottonia (Sprengel) C. DC. In 1930 he segregated these same northern constituents as the new genus Arctottonia on the bases of their distinctive palmate leaf venation, their tendency toward adnation of the stamens to the ovary, and their disjunct distribution. The first arctottonioid piper was described by Opiz (1828) as Piper stipu- laceum. Kunth (1839), segregating genera from Piper sensu lato, transferred this species to his new genus Enckea, even though he characterized Enckea, in part, by sessile flowers and fruits. Miquel (1843-1844), in his monographic treatment of the Piperaceae, shared Kunth’s opinion that Piper should be divided into several small genera and consequently maintained the placement of Enckea stipulacea (Opiz) Kunth. Either Kunth and Miquel did not believe Opiz’s original description to be correct with respect to the pedicellate fruits, or they simply chose to ignore this fact. The only other significant nomenclatural change before C. de Candolle assumed the mantle of authoritative monographer of the Piperaceae was Presl’s (1851) first combination ofan arctottonioid piper in the genus Oftonia, creating Ottonia stipulacea (Opiz) Presl. In his monograph for the Prodromus (1869), De Candolle relegated all of Kunth’s and Miquel’s segregate genera to sectional status and/or submerged them in synonymy. His sect. Enckea, for example, comprised Kunth’s genus Enckea and Miquel’s genus Callianira. His circumscriptions allowed for the inclusion of both pedicellate- and sessile-flowered species in the same section, because he placed more emphasis on the number of stamens and the position of the inflorescence (axillary vs. terminal and leaf opposed). This resulted in the assignment of newly described arctottonioid pipers to his sects. Enckea (Kunth) C. DC. (Piper muelleri C. DC.) and Steffensia (Kunth) C. DC. (P. neesianum C. DC.). He also (p. 367) placed P. stipulaceum in his category of “species non satis notae, amenta oppositifolia, flores hermaphroditi, ameri- canae,” even though Presl’s 1851 description of Ottonia stipulacea is rather extensive. For the next 50 years little change in the taxonomy of Piper occurred with respect to taxon Arctottonia. During this period only one arctottonioid piper was described; P. yucatanense was established by De Candolle and assigned to his sect. Enckea. Indeed, since the taxonomy of Piper remained largely in the hands of De Candolle, his conservative viewpoint prevailed. Even in his later years, De Candolle’s static views changed significantly only to the extent that he recognized Ottonia as a distinct section that combined the pedicellate members of his sects. Enckea and Steffensia. With the influx of material collected in Mexico and Central America during the beginning of the twentieth century, new arctottonioid species were discov- ered. De Candolle described four such Mexican species, all of which were placed in section Ottonia, in 1920 (published posthumously). In his final treatise (De 4 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Candolle, 1923), the key to all known piperaceous taxa, eight species referable to Arctottonia were included in Ottonia, which he now treated at the level of subgenus. Trelease (1934, 1935), subsequent to his creation of Arctottonia, described three additional species of the genus, all supposedly new to science. Aside from A. stipulacea (Opiz) Trel., no other valid combinations have ever been pub- lished in Arctottonia. In fact, Lundell (1937) soon transferred two of the “new” species to Piper because he did not consider the characters used to distinguish Arctottonia significant enough to warrant its continuation as a distinct genus. This sentiment regarding the status of Arctottonia has since prevailed. Yun- cker (1950) eventually submerged Arctottonia under Ottonia, stating (p. 71) “‘it is not believed that the differences upon which Trelease based the genus 4rc- fottonia are of generic status.” Trelease and Standley (in Standley & Steyermark, 1952) did describe another arctottonioid piper as Piper guazacapanense but indicated in the key that it belonged, along with two other taxa, to subg. Arctottonia. CLASSIFICATION OF PIPER SENSU LATO A recent cladistic analysis of the tribe Pipereae (Callejas, 1986 and pers. comm.) confirmed the notions of Kunth and Miquel, among others, that Piper sensu lato 1s a highly heterogeneous assemblage from which segregate taxa should be recognized. One such group is the arctottonioid pipers of Mexico and Central America. Although they share a relatively unusual feature (pedi- cellate flowers) with Ottonia, which undoubtedly led De Candolle, Yuncker, and others to consider them as representatives of the same taxonomic unit, it is now apparent that the arctottonioid pipers are more closely related to another group of Neotropical pipers (i.e., sect. Enckea), with which they share the campylodromous or acrodromous leaf venation and a tendency toward ad- nation of the stamens to the ovary. The racemose inflorescence that had so influenced the thinking of previous monographers is apparently a parallel de- velopment that has arisen in three different lineages within Piper sensu lato (4rctottonia, Ottonia, and Zippelia Blume, a monotypic Old World taxon). From the brief history presented in the previous section, it should be ap- parent that the hierarchical status of segregates from Piper sensu lato remains unresolved. The following comments constitute an argument for recognizing these groups at the level of subgenus. On the one hand, all currently accepted segregates (i.e., Lepianthes Raf. (= Pothomorphe Miq.), Macropiper Miq., Sarcorhachis Trel., and Trianaeo- piper Trel.; excluding, for the moment, Ottonia, which has recently been revised by Callejas (1986)) have been recognized at the generic level. This has served to emphasize the distinctiveness of their axillary (or seemingly axillary) inflo- rescences in contrast to the terminal, leaf-opposed inflorescences for the re- maining taxa within Piper. However, if the development of axillary inflores- cences is such an important evolutionary step, then one would expect these four taxa to be fairly closely related to one another, perhaps defining a distinct lineage within Piper sensu lato. To the contrary, current evidence (see Callejas, 1986) indicates that these groups are quite distantly related and that the de- 1989] BORNSTEIN, PIPERACEAE 5 velopment of an axillary inflorescence is another example of evolutionary parallelism, similar to the situation for pedicellate flowers. Therefore, if one recognizes these four groups as genera, then all other segregates from Piper sensu lato must be similarly treated. This, of course, is a valid option, but one that carries with it enormous nomenclatural consequences (hundreds of new combinations would be required) in a family already burdened by such prob- lems A more practical option is to consider all segregates as subgenera within Piper. This system will not only be the most convenient but will also reflect the evolutionary relationships within Piper sensu lato as accurately as a clas- sification at the generic level. The level of recognition here is not nearly as important as recognition itself. As long as the classification system 1s a relatively stable one with inherent meaning, the hierarchy within that system becomes merely a matter of choice. Therefore, I have chosen to recognize the pedicellate pipers of Mexico and Central America as Piper subg. Arctottonia, which agrees with the choice of Callejas (1986 and pers. comm.) for the ottonioid pipers (Piper subg. Ottonia) and thereby initiates what we hope will be the beginning ofa more stable and meaningful infrageneric classification for the genus Piper. HABITAT AND GEOGRAPHIC DISTRIBUTION Members of the genus Piper are common understory shrubs or small trees of various habitats but are best represented in lowland tropical rainforests. As such, most species prefer moist forest formations below 1000 m alt. Never- theless, a considerable number of them, including the majority of species in Piper subg. Arctottonia, can tolerate the seasonally dry subdeciduous to decid- uous tropical forests. With a few notable exceptions the species in subg. Arctottonia occupy rela- tively open, often disturbed sites at the edge of primary forests, being frequent inhabitants of steep banks and stream sides. Because they are able to withstand these disturbed conditions, and with the increased destruction of the tropical forests by man, many of these species are able colonists of what would be considered second-growth forests. The arctottonioid pipers can be subdivided geographically into two major groups. The vast majority of species occur in a zone between the coast and the foothills of the western Sierra Madre, from Sinaloa in the north to Chiapas and Guatemala in the south. The remaining three species occupy the eastern portions of Mexico, within or east of the Sierra Madre Oriental from central Veracruz to the Yucatan Peninsula, continuing into neighboring Honduras, with isolated collections from the lowlands of southeastern Nicaragua and central Panama. This geographic distinction is highly correlated with leaf shape: broadly ovate to suborbiculate and commonly cordate at the base in species from the ‘“‘western assemblage,” and narrowly ovate-lanceolate or narrowly elliptic with a long-acuminate apex in those of the “‘eastern assemblage”’ (see FIGURE l|) For the four generic segregates mentioned, most of the ca. 40 species within these groups have been treated at one time or another as members of the genus Piper. Therefore, relatively few new combinations would be required to treat these groups appropriately at the rank of subgenus. 6 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 At low elevations near the coast in Nueva Galicia (see Rzedowski & McVaugh, 1966, fora definition of this area), several members of the ‘western assemblage” (Piper brachypus Trel., P. brevipedicellatum Bornstein, P. rosei C. DC., and P. stipulaceum) are common inhabitants of the palm forests dominated by Or- bignya cohune (C. Martius) Dahlgren ex Standley. At slightly higher elevations and often farther inland, these same species are joined or replaced by other pedicellate pipers (P. abalienatum Trel. and P. novogalicianum Bornstein) and occupy sites in the disturbed semideciduous forests dominated by Brosimum alicastrum Sw., Hura polyandra Baillon, and species of Ficus L. and Bursera Jacq. ex L. Sull other species in this group, including P. co/otlipanense Bornstein from the state of Guerrero and P. guazacapanense from Chiapas and neigh- boring Guatemala, occur 1n seemingly stable and mesic environments in more remote localities. Within the “eastern assemblage,” the dominant and widely distributed Piper neesianum C. DC. and P. yucatanense are native to the “‘ramonales’’ forests calcareous soils and dominated by the breadnut tree, Brosimmum alicastrum. However, the influence of man in these once relatively unoccupied regions has all but destroyed these primary forests, at least in Mexico. Fortunately, these two species (especially P. neesianum) have adapted as able colonizers of dis- turbed sites to remain fairly conspicuous elements of the second-growth forests that now dominate this area. HABIT With the single known exception of Piper guazacapanense, all members of Piper subg. Arctottonia are erect plants. They are 0.5-6.5(-9) m tall and 0.5- 7.5 cm in diameter at the plant base; the most common forms are shrubs and slender treelike specimens. All of the species are woody with conspicuously nodose stems (a common feature for the genus), although the amount of wood developed in some species, especially the prostrate P. guazacapanense, is min- imal. Since the inflorescences in subg. 4rctottonia are terminal and leaf opposed, continued apical growth is assumed by the apical bud, resulting in a sympodial growth pattern. MEASUREMENTS AND TERMINOLOGY Information provided in the key and descriptions was obtained from the study of herbarium specimens, fluid-preserved materials (especially flowers and fruits), and notes made in the field. Voucher specimens of my own collections have been deposited at MICH, with complete or partial sets distributed to the following herbaria (cited according to Holmgren et a/., 1981, and in order of distribution): MEXU, CHAP, INIF, GH, NY, CAS, DUKE, F, and Mo. The terminology used to describe the various plant features generally follows that of Radford et al. (1974). With the exception of a few pickled fruits, all measurements have been taken directly from herbarium specimens. In some BORNSTEIN, PIPERACEAE 1989] } j B / f Major leaf shapes in Piper subg. Arctottonia. A, B, broadly ovate to sub- ee commonly cordate at base: A, P. guazacapanense; B, P. abalienatum. C, D, El. narrowly ovate- conn or narrowly elliptic with long-acuminate apex: C, P. yucata- nense, D, P. nees 8 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 instances flowers were placed in a weak detergent mixture to soften them in order to count the number of stamens and stigma lobes more accurately. Whenever possible, ten measurements were made from each collection num- ber in an attempt to obtain a reasonable sample size and to document any variation that may exist within a single collection. Obviously immature or incomplete structures were avoided. Measurements of inflorescence length in- clude the peduncle, since it was not always possible to separate the peduncle from the flowering rachis accurately. A scanning electron microscope was used to observe details of the leaf surface (e.g., hair type) and various floral characteristics. This was facilitated by critical- point-drying the ais pickled in the field in order to retain their shape during the drying proc Certain terms ae to the hairs and their overall appearance are used here in a relative manner: puberulent—having very short (generally less than 0.1 mm), glandular or eglandular hairs barely visible at 10x magnification. pubescent—similar to puberulent, but with the eglandular hairs noticeably longer (ca. 0.1-0.3 mm long) and generally visible to the unaided eye. long-pubescent—similar to pubescent, but with the hairs 0.3-0.5 mm long. densely pubescent —similar to pubescent, but with the hairs more numerous and therefore easily visible to the unaided eye. (This is obviously a sub- jective distinction. Quite often, densely pubescent and long-pubescent are correlated features.) MORPHOLOGY LEAVES The leaves of Piper subg. Arctottonia, like those of the vast majority of pipers, are alternate, simple, and entire margined. As briefly discussed above, two distinct groups occur with respect to shape: those with leaves conspicuously longer than broad and lance-elliptic to elliptic-ovate, and those with leaves more nearly equal in length and width and broadly to very broadly ovate or suborbiculate (see FiGuRE 1). The apex is predominantly acuminate, either long or short, while the base varies from cuneate to rounded or cordate. In some specimens of P. yucatanense and P. neesianum, the base is slightly oblique. Leaves are membranous in all species of subg. Arctottonia except P. neesianum, in which they are coriaceous, especially upon drying. According to the terminology associated with Hickey’s (1979) classification of the architecture of dicotyledonous leaves, the most prevalent venation pat- tern in subg. Arctottonia is acrodromous, with basal position and perfect de- velopment. In some species with cordate leaf bases, especially Piper guaza- capanense, campylodromous venation is commonly found. The number of major veins varies from 3 to 9 (to 11), depending to a large degree on the width of the leaves, but only the three most central ones project into the acumen. The secondary veins are irregularly spaced, although in some species they form a series of interconnecting arches near the margin. Tertiary and higher-order veins form a random reticulation. 1989] BORNSTEIN, PIPERACEAE ) Ficure 2. A, pollen grain of Piper abalienatum, representative of subg. Arctottonia, x 3850. B, C, hair types in Piper subg. Arctottonia: B, bicellular, glandular hair with short stalk cell and ellipsoid gland cell (P. novogalicianum), x 550; C, multicellular, uniseriate, eglandular hair (P. stipulaceum), x 275 INDUMENTUM Two hair types occur in subg. 4rctottonia that I term glandular and eglan- dular. The glandular hairs are bicellular, with a small stalk cell subtending a larger, ellipsoid gland cell (see FiGURE 2B). These hairs are quite small (up to 5 wm) and are therefore easily overlooked. On mature leaves they are often absent or very sparse. They usually occur on both surfaces of the leaves, and 10 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 70 Ficure 3. Distribution of hair types in Piper subg. Arctottonia, as seen on leaves of P. space A, upper surface, immature specimen, x 60; B, lower surface, immature specin x 60; C, upper surface, mature specimen, numerous eglandular hairs and sparse eadabe ones (compare with A), 5 cm long, many-flowered. 9. Prophylls, petioles, and leaf dees Eee (to puberulent or very rarely sparsely pubescent). 10. Rachis and floral bracts glabrous to sparsely puberulent, rarely with few scattered eglandular hairs; pedicels usually 1.5-2.5 mm long. . Bie uae igh ind eaites, evleea at etuam yeaa aees ee eee eae ap michelianum. 10. Rachis and floral bracts puberulent to more commonly pubescent or densely pubescent with eglandular hairs; pedicels 0.5-1.5 mm long, rarely longer, or flowers subsessile. 11. Flowers subsessile, the pedicels to 0.5 mm long, ey longer. .. Rings eienpactaueea 2 ootaa tao eaiel coon es le ee Ds P. brachypus. 11. Flowers with pedicels 0.5-1.5 mm long. ......... 12. P. rosei. Prophylls, petioles, and leaf blades pubescent to densely pubescent or to- mentose (rarely only puberulent). 12. Pedicels generally < 0.5 mm long, stout, ma : long as broad. Getineer cite teeth ts Mls uss ous a tendeks Mee eles 2chiethen es Be aioe P. brevipedicellatum. 12. Pedicels generally > | mm long, as long as—or, more commonly, distinctly longer than—broad. . Filaments 0.5-1 mm long; young twigs, petioles, leaf blades, and prophyll densely long-pubescent to tomentose. ............... br i ndeenct nea abate toc aah aegaraee atatetece ae sea eee P. colotlipanense. Filaments to 0.5 mm long; young twigs, petioles, leaf blades, and prophylls puberulent to densely pubescent. ................... ~~ \O — ies) — ies) 1. Piper abalienatum Trel. Amer. J. Bot. 8: 216. pi. 8, fig. J. 1921. Type: exico, Colima, Colima, July and Aug. 1897, Palmer 100 (holotype, us!; isotypes, A!, MICH!, UC!). FIGURE 5. Piper el-muletonum C. DC. Candollea 1: 72, 246. 1923, nomen nudum. Piper el- muletonum C. DC. ex L. J. Schroeder, Candollea 3: 137. 1926. Type: Mexico, Michoacan, El Muleto, 850 m, 30 June 1898, Langlassé 215bis (holotype, c!). Medium-sized shrub, sparsely branched from base, appearing treelike, 1.5— 4 m tall; youngest twigs lime green, densely pubescent, becoming glabrous, brown, and eventually light gray with warty lenticels; older stems with con- spicuous swollen nodes, maximum diameter to 6 cm. Prophylls to 1.5 cm long 18 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 at maturity, base truncate, apex acute, dorsally densely pubescent to tomentose. Leaves with petiole 1—4.5 cm long, stout, densely soft-pubescent with eglandular hairs; blade broadly ovate to orbiculate to occasionally depressed-ovate, 7.1- 20.4 x (6.5-)7.6-19.8(-21) cm, rounded to distinctly cordate at base, gradually to abruptly acuminate at apex, (5- to) 7- to 9- (to 11-)nerved, puberulent to pubescent (rarely almost glabrous) above, densely soft-pubescent to tomentose (rarely glabrous to puberulent) beneath with eglandular hairs more prominent on slightly elevated major veins. Inflorescences (4.8-)5.7-15.2 cm long, green when young, turning yellow-white at maturity; peduncle stout, pubescent at base, remaining rachis glabrous or minutely puberulent with glandular hairs; bracts to 1.25 mm long (usually ca. 0.75 mm), narrow petiolate basal portion broadening gradually to abruptly into spatulate to spatulate-cupulate apical region, abaxially sparsely pubescent to pubescent. Flowers with pedicel (0.5-) 1-2(-3) mm long, to | mm in diameter, glabrous to sparsely puberulent with glandular hairs; stamens (4 or) 5 to 7 (or 8), adnate to near middle of ovary, occasionally appearing as if in 2 alternating whorls, the filaments elongate (to 1 mm), the anthers equal to or slightly shorter than filaments, dehiscing some- what to conspicuously divergent upward from lateral; ovary ovoid-ellipsoid, 1- 2 tumes as long as style, puberulent with glandular hairs, the style columnar to slightly conical, (0.5-)1-1.5(-2) mm long, persistent in fruit, the stigma 3- to 5-lobed. Fruits globose, to 2 mm in diameter, drying brown to black, with glandular hairs still present. DISTRIBUTION. Confined to small area in west-central Mexico from south- western Jalisco, along western side of Sierra Madre, through state of Michoacan (see Map 1). Although not known from Guerrero, to be expected there, espe- cially at mid-elevations in Sierra Madre del Sur. HABITAT. Dense shade of mixed deciduous forests, 100-1000 m alt.; often associated with various Leguminosae, Moraceae (especially species of Brosi- mum Sw. and Ficus), Euphorbiaceae (Hura L.), Bursera, and Malpighia L. Always grows where water plentiful at some time of year; most abundant at edge of ravines above intermittent mountain streams. PHENOLOGY. Flowering late May through July, most reliably in June. Young fruits beginning in late June; specimens with mature fruits collected as late as mid-October EPRESENTATIVE SPECIMENS.” Mexico. JALIsco: | km E of El Divisadero (1 km W of Villa ae 18 km ESE of Tomatlan, //tis & Nee 1650 (Ny), Estacion de Biologia de Chamela, Pérez & Sarukhdén 1354 (mexu, MIcH); 15 mi SSE of Autlan on trail between Chante and Manantlan, Wi/bur 2105 (micu). CoLima: 2.3 km Ixtlahuacan, Bornstein 90 (micH); mountain trail S of Los Tepames, Bor ees & Hensold 18 (micu); 5.7 km S of Los Tepames on gravel road to Las Tinajas, Bornstein & Hensold 2] (MICH); just beyond km 190 and before Rio Tuxpan bridge on old Hwy. 110 between Colima and Pihuamo, pals 91 (MICH); ca. 18 road mi E of Colima, near bridge of Rio Tuxpan, km 189, McVaugh 15498 (micu); ca. 11 mi SSW of Colima on Manzanillo road, McVaugh 15564 ree gorge of Rio Cihuatlan, near bridge 13 mi N of Santiago, McVaugh 15809 (micH). MIcHOACAN: 3.5 mi N of San Juan de Lima on Hwy. 200 ‘A complete list of collectors for the entire study is given in the appendix. 1989] BORNSTEIN, PIPERACEAE 19 Figure 5. Piper abalienatum: A, flowering branch, x 0.2; B, inflorescence, x 0.8; C, flower with bract, x 12; D, fruits with bract, x 5; E, prophyll with emergent young leat -3:9% Bornstein & Chase 94 (micH); Coalcoman, Hinton 13855 (ENCB, ILL, NA, NY, PH, us); 32 km N of Playa Azul (region of Los Encinos), King & Soderstrom 4929 ee TEX, s). Piper abalienatum is most often confused with P. stipulaceum, because in both species the stamens are adnate to near the middle of the ovary. The differences between these two closely related taxa are discussed following the description of P. stipulaceum (and see TABLE). iper abalienatum also resembles P. colotlipanense, especially with respect to stylar features and filament length. It is distinguished from that species in having the stamens adnate to the middle of the ovary (vs. hypogynous or irregularly adnate to the lower portion of the ovary). 2. Piper brachypus Trel. Amer. J. Bot. 8: 215. p/. 6. 1921. Type: Mexico, Colima, Manzanillo, 2-18 March 1891, Palmer 1332 (holotype, us!; isotype, GH!). Shrub to 4 m tall; twigs and older branches glabrous to very sparsely pu- berulent with glandular hairs. Prophylls to 0.7 cm long, base truncate, apex 20 JOURNAL OF THE ARNOLD ARBORETUM i) [vOL. 7. Distribution of Piper abalienatum (solid squares), P. brevipedicellatum (cir- r). Map I. cles), P. colotlipanense (open squares), and P. mevaughil (sta Characters used to distinguish between Piper abalienatum and Piper stipulaceum. CHARACTER SPECIES Piper abalienatum Piper stipulaceum ves Average size (cm) Surface Stamens Arrangement Filament length (mm) Ovary shape (in flower) Style Shape Length (mm) To ovary ratio In flower In fruit ca. | 1:2-1:3 12.6 x 11.8 Pubescent to tomentose especially beneath Irregular; occasionally in 2 0.5—1; equal to or exceeding anthers Ovoid-ellipsoid Columnar to slightly coni- ca (0.5-)1-1.5(-2) 10 x 8.7 Puberulent to pubescent, especially beneath In | whorl 0.5; shorter than or nearly equal to anthers Ovoid to globose Depressed-ovoid to conical To 0.5, rarely slightly longer 1: 1-1 1:3-1: 1989] BORNSTEIN, PIPERACEAE 21 Ficure 6. Piper brachypus: A, fruiting branch, x 0.4; B, portion of infructescence, x 6. acute, glabrous or puberulent with short eglandular hairs. Leaves with petiole 0.7-1.6 cm long, narrow, glabrous or sparsely puberulent with glandular or short eglandular hairs; blade ovate to broadly ovate to occasionally suborbic- ulate, 4.4-8.6(-9.6) x 3.3-6.6(-9.1) cm, rounded to truncate to shallowly cor- date at base, acuminate at apex, 5- to 7- (to 9-)nerved, glabrous to puberulent (especially at blade-petiole junction) with predominantly short eglandular hairs. Inflorescences (5.9-)6.8-11.6 cm long; peduncle glabrous to commonly pu- berulent to sparsely pubescent, remaining t to densely pubescent with long eglandular hairs; bracts sessile to subsessile, broadly spatulate to 22 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Map 2. Distribution of Piper brachypus (open squares), P. cihuatlanense (star), and P. novogalicianum (circles). cupulate, to 0.5 mm long (rarely longer), pubescent to densely pubescent (es- pecially abaxially) with eglandular hairs. Flowers with pedicel to 0.5(-1) mm long and in diameter, glabrous to puberulent with eglandular hairs; stamens 4 or 5 (or 6), hypogynous, to | mm long, the filaments to 0.5 mm long, the anthers small, dehiscing laterally; ovary ellipsoid to ovoid, puberulent with glandular hairs, the style depressed-ovoid to very short-conical, to 0.5 mm long, usually quite conspicuous in both flower and fruit, the stigma 3- to 5-lobed. Fruits ellipsoid to more commonly ovoid to globose, 1-1.75 x 1-1.75 mm, puberulent, style persistent but sometimes obscured by expanding ovary. DISTRIBUTION. Growing in narrow band from Manzanillo, Colima, north to Just south of Puerto Vallarta, Jalisco (see Map 2). Habitat. Essentially unknown (Mexia and Koelz collections from wooded slopes, De/gado et al. from an oak forest). Near sea level to 450 m alt. PHENOLOGY. Presumably follows phenological pattern of piper relatives in Nueva Galicia, therefore flowering May to July: specimens with fruits collected in March, mid to late July, November, and December SPECIMENS EXAMINED. Mexico. JALISCO: crossing puente of Rio Las Juntas, on road to Puerto Vallarta, near El oe Delgado, Hernandez, & Trejo 204 (micn); Quimixto, Mexia 1177 (a, GH, Mo, Jc). COLIMA: vic. of Manzanillo, SW edge of Cuyutlan lagoon, Ferris 6186 (A); rn cee near Rio Cihuatlan, N of Playa de Santiago, Koe/z 34228 (ENCB, MICH). 1989] BORNSTEIN, PIPERACEAE 23 Piper brachypus is distinguished by the combination of broadly ovate, gla- brous leaves; pubescent rachises; densely pubescent bracts; subsessile flowers; hypogynous stamens; and small, usually ovoid to globose, conspicuously short- styled (= stylopodium, as described by Trelease) fruits. 3. Piper brevipedicellatum Bornstein, sp. nov. FIGURE 7. Frutex 2.5—5(—9) m altus, ramunculis pubescentibus; prophyllum 5—10(-13) mm longum, dorsaliter pubescens; folia laminis late ovatis vei orbiculatis, (6—) 7.3-12.7 x (4.7-)5.6-12.8 cm, breviacuminatis vel acuminatis, basi cordatis, utrinque pubescentibus; inflorescentia 6—-11.5(—13.4) cm longa, rhachidi pu- berula vel pubescenti pilis elongatis; bracteae usque ad | mm longae, spathu- latae vel spathulicupulatae, abaxialiter pubescentes; pedicelli raro usque ad | — stylus depresso-ovoideus vel breviconicus, usque ad 0.5 mm longus; fructus globosus, 2 x 2 mm, puberulus pilis glandulosis, stylo disciformi instructus. Type. Mexico, Jalisco, 7.9 km WNW of Hwy. 80 along Hwy. 200 toward Puerto Vallarta, 250 m alt., 27 June 1983, Bornstein & Weber-Bornstein 80 (holotype, MICH; isotypes, CHAP, GH, INIF, MEXU, NY). Medium-sized treelike shrub, moderately branched from base, 2.5—5(-9) m tall; youngest twigs pubescent with eglandular hairs; older stems gray-brown with swollen nodes and warty lenticels, glabrous, maximum diameter to 5 cm. Prophylls 5—10(-13) mm long, base truncate, apex acute, abaxially puberulent to densely pubescent with eglandular hairs. Leaves with petiole 0.7-2.5 cm long, stout, puberulent to pubescent with eglandular hairs; blade broadly ovate to orbiculate, (6-)7.3-12.7 x (4.7-)5.6-12.8 cm, cordate at base, short-acu- minate to acuminate at apex, 5- to 7-nerved, puberulent to more commonly pubescent above with predominantly eglandular hairs, pubescent beneath with both hair types (eglandular more abundant and conspicuous on major veins, glandular more predominant in areolar regions). Inflorescences 6-1 1.5(—13.4) cm long, green becoming yellow-white to white; peduncle stout, puberulent to pubescent with eglandular hairs, remaining rachis sparsely puberulent or sub- glabrous to pubescent with eglandular hairs; bracts to 1 mm long, sessile or subsessile, broadly spatulate to broadly spatulate-cupulate, 0.5-1 mm long, sparsely pubescent to pubescent with eglandular hairs abaxially, occasionally with some pubescence adaxially. Flowers with pedicel rarely to | mm long and wide, glabrous to puberulent with glandular and eglandular hairs; stamens 4 to 6 (or 7), hypogynous although often (especially in fruit) appearing as if adnate to base of ovary, filaments and anthers + equal in length, anthers occasionally dehiscing slightly upward from lateral; ovary ovoid to globose, puberulent with glandular hairs, the style depressed-ovoid to short-conical, to 0.5 mm long, persistent in fruit, the stigma 3- to 5-lobed. Fruits globose at maturity, to 2 mm in diameter, puberulent. DISTRIBUTION. Central Colima north to southwestern Jalisco from near sea level to ca. 1200 m alt. (see MAP 1). Also known from a single disjunct collection 24 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Figure 7. Piper brev seis ga A, fruiting pear a x 0.4, with detailed inset of petiole-blade junction showing indument; B, portion of young inflorescence, x 5.6; C, ei of infructescence, x 5.6; D, prophyll with a young leaf and inflorescence, 1.9, ca. 500 km north in Sinaloa, which represents the northernmost extent of subg. Arctottonia. Hasirat. Tolerant of a variety of conditions, including shaded, low-elevation palm forests dominated by Orbignya cohune, deciduous to subdeciduous, low- 1989] BORNSTEIN, PIPERACEAE 25 elevation forests near intermittent mountain streams, and deciduous, mid- elevation forests probably dominated by Quercus spp. The only potential com- mon denominator of these sites is disturbance, which indicates that this species colonizes second-growth forests. PHENOLOGY. Flowering late May through June and possibly early July. Fruits appear in late June; collected as late as December SPECIMENS EXAMINED. Mexico. SINALOA: Sierra de Pinal, just E of El Espinal, Kimnach & Mejorada 2266 (MEXU, US). JALISCO: along Hwy. 80 between Autlan and Melaque, 4 km N of side road to Casimir Castillo, Bornstein 75 (micH); 7.9 km WNW of Hwy. 80 along Hwy. 200 toward Puerto Vallarta, Bornstein & Weber-Bornstein 81 (micw); 0.9 km de la carretera Puerto Vallarta—Barra de Navidad, ca. 7 km al SE de la Estacion de Biologia Chamela, sobre el camino a “El 25” (IMSS communidad), Lott & Magallanes 611 (micw); Estacién de Biologia, Los Pozos, Arroyo Colorado, Lott 719 (mic); ca. 8 mi S of Autlan toward La Resolana, Wi/bur 1625 (MICH). CoLima: 0.5 km S of Comatlan on Hwy. 98 to Minatitlan, Bornstein 87 (micH); ca. 16 mi WNW of Santiago, road to Cihuatlan, McVaugh 14983 (micH); km 293, ca. 15 mi SE of Manzanillo, McVaugh & Koelz 1623 (MICH). Piper brevipedicellatum and P. brachypus both have flowers with hypogynous stamens and fruits crowned by a “stylopodial disk.”” However, in the former species there is a strong tendency toward adnation of the stamens to the ovary and the rachis appears less pubescent than in P. brachypus. Unfortunately, neither of these features is consistently different. On the other hand, the two species are quite distinct vegetatively since P. brachypus 1s essentially glabrous, whereas P. brevipedicellatum has pubescent twigs, prophylls, and leaves. 4. Piper cihuatlanense Bornstein, sp. nov. FiGuRE 8. Frutex usque ad 6 m altus, ramunculis et ramis glabris; prophyllum usque ad 6 mm longum, glabrum vel puberulum; folia laminis ellipticis vel anguste ovatis, 4.9-7.3 x 2.7-3.9 cm, acuminatis, basi rotundatis, glabris vel sparsim puberulis; inflorescentia 6.5—9.1 cm longa, rhachidi dense pubescenti pilis eglandulatis longis; bracteae 0.25-0.5 mm longae, spathulatae vel cupulatae, dense pubescentes pilis eglandulatis longissimis; pedicelli (0.5—)1(-1.5) mm longi, pubescentes; stamina hypogyna; stylus usque ad 0.25 mm longus, de- presso-ovoideus, inconspicuus; ovarium et fructus ellipsoideus vel obovoideus vel globosus, 1-1.5 =x 0.75-1.25 mm, puberulus. Type. Mexico, Colima, steep, heavily wooded, moist ravines in gorge of Rio Cihuatlan, near bridge 13 mi N of Santiago, 200-300 m alt., 27 July 1957, McVaugh 15838 (holotype, MICH). Treelike shrub to 6 m tall; twigs and older branches brown (occasionally silvery due to detaching epidermis), glabrous or with few locally concentrate eglandular hairs, especially at branch points, with inconspicuous round, pale brown lenticels. Prophylls to 0.6 cm long, base unknown, apex acute, glabrous to puberulent with very short eglandular hairs. Leaves with petiole 0.4-1.1 cm long, appearing glabrous but usually sparsely puberulent with glandular or very short eglandular hairs; blade elliptic to narrowly ovate, 4.9-7.3 x 2.7-3.9 cm, round at base, acuminate at apex, 5-nerved, glabrous to sparsely puberulent 26 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Se ARK, SR ype a ee we hat FiGure 8. Piper cihuatlanense: A, fruiting branch, x 0.45; B, portion of infructes- ence 6.8. (especially at blade-petiole junction) with short eglandular hairs. Inflorescences drooping, 6.5—9.1 cm long, color unknown; peduncle puberulent, passing abruptly into rachis, this densely pubescent with elongate eglandular hairs; bracts broadly spatulate to cupulate, < 0.5 mm long, usually ca. 0.25 mm, densely pubescent with very long eglandular hairs on both surfaces. Flowers with pedicel (0.5-) 1.5) mm long, narrow, usually 1.5-2 times as long as broad, puberulent to distinctly pubescent with eglandular hairs; stamens 4 to 6, hypogynous, the filaments < 0.25 mm long, the anthers subglobose, to 0.5 mm long, dehiscing laterally; ovary ellipsoid to obovoid to subglobose, puberulent with glandular hairs, the style depressed-ovoid, < 0.25 mm long, inconspicuous in flower 1989] BORNSTEIN, PIPERACEAE 21 totally obscured in fruit, the stigma 3- or 4- (or 5-)lobed. Fruits similar to ovary in shape, |-1.5 x 0.75-1.25 mm, puberulent. DisTRIBUTION. Known from two localities in northwestern Colima (see Map 2). Habitat. Apparently restricted to barrancas of Rio Cihuatlan at 200-300 m alt. (This area is quite moist during the rainy season and supports a dense tropical subdeciduous vegetation.) PHENOLOGY. The McVaugh and Koelz collections suggest a flowering period that lasts through the rainy season (mid-May through October). More direct evidence is needed. SPECIMENS EXAMINED. Mexico. Cotima: barranca of Rio Cihuatlan, N of Playa Santiago, Koelz 34223 (ENCB, MICH). Piper cihuatlanense is easily distinguished by its small, elliptic to narrowly ovate, subglabrous leaves, its very densely pubescent rachis and bracts with relatively long eglandular hairs, its often-pubescent pedicels, and its virtually nonexistent style (especially in fruit). If it is true that the racemes droop, as is indicated on the label of the type specimen, then this would also be a unique feature (for the entire subgenus). 5. Piper colotlipanense Bornstein, sp. nov. FIGURE 9. Frutex usque ad 2.1 m altus, ramunculis fulvis dense tomentosis; prophyllum usque ad 0.8 cm longum, dorsaliter dense pubescens vel tomentosum; folia laminis late ovatis vel fere orbiculatis, (5.3-)6.6-13.1 x 4.6-11.6 cm, brevi- acuminatis vel acuminatis, basi rotundatis vel leviter cordatis, supra pubes- centibus, subter pubescentibus vel tomentosis; inflorescentia 8.1—-10.1 cm lon- ga, rhachidi pubescenti vel sparsim pubescenti; bracteae 0.5-1 mm longae, spathulatae vel spathulicupulatae et breviter vel longe petiolatae, pubescentes vel dense pubescentes; pedicelli (0.5—)1.5(-2.5) mm longi; stamina hypogyna filamentis elongatis, 0.5-1 mm longis; stylus depresso-ovoideus vel conicus; ovarium et fructus ovoideus vel globosus, puberulus. Tyre. Mexico, Guerrero, 2 mi W of Colotlipa, 2700 ft alt., 22 June 1953, > Rowell & Kubicek 3666 (holotype, MICH). Medium-sized treelike shrub to 2.1 m tall; youngest twigs densely tawny- tomentose with eglandular hairs, becoming silvery and glabrous; older stems drying brown to gray, with swollen nodes and warty lenticels. Prophylls to 0.8 cm long, base truncate, apex acute, abaxially densely pubescent to tomentose with eglandular hairs. Leaves with petiole 1.1-2.3 cm long, narrow, vaginate at base or occasionally for nearly '2 length of petiole, densely soft-pubescent to tomentose with eglandular hairs; blade broadly ovate to nearly orbiculate, (5.3-)6.6-13.1 x 4.6-11.6 cm, rounded to shallowly cordate at base, short- acuminate to acuminate at apex, 5- to 7- (to 9-)nerved, pubescent above (es- pecially along major veins), densely pubescent to tomentose beneath (especially along major veins), the hairs long, eglandular. Inflorescences 8.1—10.1 cm long, color unknown; peduncle pubescent to tomentose at base with eglandular hairs, 28 JOURNAL OF THE ARNOLD ARBORETUM [vVoL. 70 Ficure 9. Piper colotlipanense: A, flowering branch, x 0.42, with detailed inset of lower leaf surface showing indument; B, portion of inflorescence, x 6.3; C, portion of semimature infructescence, x 6.3 gradually transitional to sparsely pubescent for remaining rachis; bracts sessile or subsessile to distinctly long-petiolate, spatulate to spatulate-cupulate at apex, 0.5-1 mm long, pubescent to densely pubescent with eglandular hairs, most often on abaxial surface. Flowers with pedicel (0.5-)1.5(—2.5) mm long, up to 0.5 mm in diameter, glabrous to puberulent with glandular hairs, occasionally sparsely pubescent with eglandular hairs; stamens (4 or) 5 or 6, hypogynous, the filaments 0.5-1 mm long, the anthers to 0.5 mm long, rarely slightly longer, dehiscing laterally to slightly divergent upward from lateral; ovary ovoid to globose, puberulent with glandular hairs, the style depressed-ovoid to conical, 1989] BORNSTEIN, PIPERACEAE 29 in young fruit 4-4 length of ovary, the stigma 3- or 4-lobed. Fruits (when submature) ovoid to globose, style very evident DisTRIBUTION. Guerrero, in low mountains of Sierra Madre del Sur, south of Chilpancingo de los Bravos (see Map 1). Hasitat. Herbarium label data indicate that this species occurs near edges of streams or rivers, or on hillsides on rocky soil at 150-850 m alt. PHENOLOGY. Flowering May through June, possibly into early July. Young fruits in early June, expected into July and August. REPRESENTATIVE SPECIMENS. Mexico. GUERRERO: S of Chilpancingo, Clark 7141 (mo, Ny); Wallecitos, Montes de Oca, Hinton 10261 (ILL, K, NA, NY, PH, TEX, US); Pie de la Cuesta, 10 mi N of Acapulco, MacDaniels 201 (Ff); 1 mi W of Achuitzotla, Rowell & E. H. Cooper 2507 (micu); 1 mi W of Colotlipa, Rowell & Irby 3523 (micH). LOCALITY UNKNOWN: Sessé & Mocifio 318 (F), 328 (F), 329 (F), 339 (ILL), 348 (F, ILL). Piper colotlipanense is distinguished by its densely long-pubescent prophylls and leaves and its hypogynous stamens with long filaments 6. Piper guazacapanense Trel. & Standley, Fieldiana, Bot. 24: 301. 1952. Type: Guatemala, Santa Rosa, along Avellana road, S of Guazacapan, 150 m alt., 12 June 1940, Standley 79461 (holotype, F!; isotype, NY!). Ficure 10. Procumbent perennial, occasionally rooting at nodes; stems largely herba- ceous, becoming woody with age, glabrous. Prophylls to 8 mm long, base truncate, apex acute, abaxially puberulent to hirtellous. Leaves with petiole 4— 17 mm long, slender, hirtellous with both glandular and eglandular hairs; blade broadly ovate to orbiculate-ovate, (3.5—)5-9 x 4-8 cm, cordate with rounded lobes at base, gradually to abruptly acuminate or long-acuminate at apex, 5-nerved (3 principal veins persistent almost to apex), dark green above and sparsely puberulent, paler beneath and puberulent to hirtellous with eglandular hairs + restricted to veins and glandular ones to areoles. Inflorescences 1-3 (—4.2) cm long, yellow-green when young, turning white at maturity; peduncle and rachis glabrous to more commonly puberulent; bracts spatulate-cupulate to broadly cupulate, rarely to 1 mm long, abaxially puberulent with eglandular hairs (rarely glabrous). Flowers with pedicels 1-2.5 mm long, glabrous or oc- casionally minutely puberulent; stamens 4 to 7, hypogynous, filaments 0.25 mm long, anthers dehiscing laterally; ovary ovoid to globose, minutely pu- berulent with glandular hairs, the style 2 to equal the ovary in length, becoming obliterated in fruit, the stigma 3- or 4-lobed. Fruits globose, 2-2.5 mm in diameter, green when young, translucent pearl white at maturity, puberulent. DIsTRIBUTION. Southwestern Chiapas on northwestern edge of Sierra Madre de Chiapas, the north-central plateau in Chiapas, and southeastern Guatemala on southern edge of Sierra Madre (see MAP 3 HasiratT. Moist, shaded ravines. Commonly found as dense groundcover along banks above small streams in tropical subdeciduous forests at 150-1000 m alt. [VoL. 70 JOURNAL OF THE ARNOLD ARBORETUM eit oe: mia ay xt rey Me g: eR v PL abs? pee (SS 2S / net ay LESS ——+ ++ i NON sy i Hi Th ype tee ry, Se FiGure 10. as aoe apanense: A, flowering branch, x 0.4; B, inflorescence, x 1.6; flower with brac ; D, immature and mature fruits, < 8.6; E, prophyll with x 13, C, young leaf, young res and succeeding prophyll, x PHENOLOGY. Flowering mid-February to late May or early June, predominantly in April and May. Fruits collected in October, December, and February. EPRESENTATIVE SPECIMENS. Mexico. CHIAPAS: 12 km S of Hwy. 190 near Rizo de Oro behind microwave tower of La Mina in moist shaded ee acl tatl Chase, & Born- stein 59 (micu); 13 mi N of Arriaga along Hwy. 195 0 eep slopes along ravines, Breedlove 25520 (EN F, MEX ICH, MO, NY, TEX); ee river below Huitiupan, Breedlove & A. C. aaee 32576 ( Piper guazacapanense 1s easily recognized by its prostrate habit, its small, heart-shaped leaves, its short, few-flowered inflorescences, and its globose, >M MEXU). translucent fruits. FIGuRE I 1. 7. Piper mcevaughii Bornstein, sp. nov. Frutex ramunculis fere glabris; eae usque ad 0.3 cm longum, aa 5.5-7.6 c berulum vel brevipubescens; folia laminis late ovatis, 5.1-8 x breviacuminatis, basi cordatis, supra bai subter sparsim puberulis 1989] BORNSTEIN, PIPERACEAE 31 ie) 100 200 300 400 500 KM - 1 it 1 1 J T T T fo) 100 200 300 Mi 105° 100° 1 1 Map 3. Distribution of Piper guazacapanense (stars), P. muelleri (squares), and P. neesianum (circles). florescentia 6.7-8.8 cm longa, rhachidi sparsim puberula pilis glandulosis; brac- teae usque ad 0.75 mm longae, late spathulatae vel valde cupulatae, sparsim puberulae vel puberulae; pedicelli 0.5(—0.75) mm longi, sparsim pubuleri, pilis glandulosis; stamina ovario adnata; stylus depresso-ovoideus vel brevicol- umnaris, usque ad 0.75 mm longus; ovarium ovoideum-ellipsoideum; fructus globosus, usque ad 2 mm diametro, puberulus. Type. Mexico, Jalisco, coastal plain near hwy. to Autlan, 4 mi N of Bahia Navidad, 40 m alt., 8 Nov. 1960, McVaugh 20817 (holotype, MIcu). Shrub of unknown dimensions; youngest twigs drying brown, with sparse glandular hairs; older branches silvery, with brown, warty lenticels, glabrous. Prophylls to 3 mm long, base truncate, apex acute, abaxially puberulent to short-pubescent with glandular or short eglandular hairs. Leaves with petiole 1.1-1.8 cm long, narrow, appearing glabrous but sparsely puberulent with glandular hairs; blade broadly ovate, 5.1-8 x 5.5-7.6 cm, cordate at base, short-acuminate at apex, 5- (to 7-)nerved, puberulent above with predomi- nantly glandular hairs (especially at petiole-blade junction and along blade margin), sparsely puberulent beneath on major veins, with mostly short eglan- dular hairs. Inflorescences 6.7—8.8 cm long, color unknown, with flowers widely spaced at maturity; peduncle and rachis sparsely puberulent with glandular hairs; bracts sessile or subsessile, broadly spatulate to occasionally strongly cupulate, to 0.75 mm long, sparsely puberulent to puberulent with glandular or short eglandular hairs. Flowers with pedicel ca. 0.5(-0.75) mm long and broad, sparsely puberulent with glandular hairs; stamens 4 or 5, adnate from 32 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Figure 11. Piper mcvaughit: A, fruiting branch, x 0.4; B, portion of infructescence, x 6. near base to middle of ovary, filaments ca. 0.25-0.5 mm long, anthers unknown; ovary ovoid-ellipsoid in late flower-early fruit, puberulent with glandular hairs, the style depressed-ovoid to short-columnar, to 0.75 mm long, the stigma (3- or) 4- or 5-lobed. Fruits globose at maturity, 2 mm in diameter, puberulent, style persistent and conspicuous. DISTRIBUTION. Known only from type locality, southwestern Jalisco near Co- lima border (see Map 1) Hasirat. Type collection from second-growth deciduous forest with associated species of Cordia L., Trichilia P. Browne, and Sideroxylon L. PHENOLOGY. Relatively unknown. Holotype, collected in November, has pre- dominantly mature fruits. Based on flowering times of neighboring pipers, this species probably flowers from late May through June, at the beginning of the rainy season. Eponymy. Named in honor of Rogers McVaugh, Professor Emeritus at the University of Michigan, for his contributions that have increased our knowl- edge of botany in Mexico, and especially for his work on the flora of Nueva Galicia. Superficially similar to Piper brachypus, P. mcvaughii differs in having sta- mens adnate to the ovary (vs. hypogynous) and the rachis and bracts puberulent BORNSTEIN, PIPERACEAE 1989] 12. Piper michelianum: A, sae branch, x 0.42, with detailed inset of petiole-blade junction to uberulence; B, portion of inflorescence in late fower— early fruit, x 6.3; C, portion of young ee x 6.3; D, prophyll with emergent young leaf and inflorescence, x FIGURE with predominantly glandular hairs (vs. pubescent to densely pubescent with eglandular ones). 8. Piper michelianum C. DC. Annuaire Conserv. Jard. Bot. Genéve 21: 317. 1920. Type: Mexico, Michoacan, El Muleto, 850 m alt., 30 June 1898, Langlassé 215 (holotype, G!; isotypes, GH!, K!, us!). FIGURE 12. Piper mas Trel. Amer. J. Bot. 8: 215. pl. 7, fig. 2. 1921. Type: Mexico, Michoacan, El uleto, 850 m alt., 30 June 1898, Langlassé 215 (holotype, us!; isotypes, G!, Gu!, 34 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Medium-sized shrub, sparsely branched from base, appearing treelike, 2-5 m tall; youngest twigs very minutely and sparsely puberulent, predominantly with glandular hairs, soon glabrescent; older branches brownish, with warty lenticels; stems ash-gray, with swollen nodes, maximum diameter 6 cm. Pro- phylls to 1.2 cm long, base truncate, apex narrow and acuminate, both surfaces glabrous. Leaves with petiole 0.9-2 cm long, narrow, glabrous to puberulent (especially near blade junction) with eglandular hairs; blade round-ovate, (6-) 7,.2-10.9(-13.4) x 6.1-8.6(-12.3) cm, shallow to moderately cordate at base, acuminate at apex, (5- to) 7-nerved, puberulent above with glandular and eglandular hairs (the latter generally restricted to veins), puberulent to sparsely and transiently pubescent beneath (especially on proximal '2) with glandular and eglandular hairs. Inflorescences 8.5-15.7 cm long, greenish when young, cream colored and fragrant at maturity; peduncle stout, glabrous to very mi- nutely puberulent with mostly glandular hairs, remaining rachis glabrous to puberulent with glandular or very short eglandular hairs; bracts variable, rang- ing from sessile and strongly cupulate to distinctly petiolate with petiole grad- ually to abruptly expanding into spatulate to spatulate-cupulate apex, 0.251 mm long, abaxially glabrous to puberulent to very sparsely pubescent with eglandular hairs. Flowers with pedicel 0.5-2.5 mm long (depending on ma- turity), rarely to 0.5 mm wide, glabrous to puberulent with glandular hairs; stamens 4 (or 5), hypogynous, the filaments to 0.5 mm long, the anthers broadly obovoid, 1-3 times length of filaments, dehiscing slightly to conspicuously divergent upward from lateral; ovary ovoid to subglobose, puberulent with glandular hairs, the style conical to short-columnar, to 0.75 mm long, equaling to slightly exceeding ovary in length when in flower, the stigma 3- or 4-lobed. Fruits ovoid to subglobose at submaturity, 1.5-2 x 1-1.5 mm, style persistent but somewhat obscured by expanding ovary. DIsTRIBUTION. Foothills of Sierra Madre, from southeastern Jalisco through northern Colima to southwestern Michoacan, 800-1800 m alt. (see MAP 4). Habitat. Collected on steep hillsides above coffee plantations in association with species of Juglans L. and Ficus; in “precipitous wooded ferny barrancas”’; in densely wooded arroyos with Ficus and various lianas; and on more open rocky slopes and ledges with Bursera, Acacia Miller, Agave L., woody Um- belliferae, and various orchids and bromeliads. All of these habitats feature steep terrain in an area where water is generally abundant during the rainy season but limited or absent during the dry season (mid-elevation subdeciduous forest). PHENOLOGY. Flowering late May through June. Semimature fruits collected in mid-August. REPRESENTATIVE SPECIMENS. Mexico. JALisco: | km ESE of Ayotitlan (27 km por camino al E de Cuautitlan), Cochrane et al. 10877 (Ny); 3.3 km SSE of El Chante, halfway to Tecopatlan Valley, //tis & Nee 1416 (Ny); 0.5-1 km NW of Rincon de Manantlan (13 km S$ of El Chante), //tis et al. 2559 (Ny). CoLIMa: just S of San Antonio on road to Comala, Bornstein & Chase 95 (micu); dirt road 3.5 mi N of San Antonio, Chase 83112 (mIcH); 1-1.5 mi S of Hacienda San Antonio, SW foothills of Nevado de Colima, McVaugh 1989] BORNSTEIN, PIPERACEAE 35 AP 4. Distribution of Piper michelianum (stars), P. rosei (circles), and P. stipula- ceum (squares). 16107 (micH). MICHOACAN: San José, in distr. of Coalcoman, Hinton 13836 (ENCB, ILL, ). TAXONOMIC NOTE. Piper michelianum and Piper mas were described on the basis of the same type material and are therefore nomenclatural synonyms. Piper michelianum is distinguished by the combination of shallowly to mod- erately cordate leaf blades puberulent above and puberulent to transiently pubescent beneath, glabrous to puberulent rachis and floral bracts, narrow and elongate (usually 1.5—2.5 mm long) mature pedicels, and hypogynous stamens (usually Potentially confused with Piper rosei where their ranges overlap, P. miche- lianum lacks the conspicuous pubescence of the rachis and bracts and has narrower and slightly longer pedicels, as well as acuminate (vs. long-acuminate), basally cordate (vs. rounded) leaves that are more conspicuously pubescent The IItis collections from Jalisco are sterile and are therefore included here ona tentative basis. In sterile condition Piper michelianum resembles P. brachy- pus and P. mcvaughii, both of which are also glabrous to puberulent vegeta- tively. The only means for separating the three species morphologically is by examining the inflorescences. I have placed these two collections under P. michelianum on the basis of their habitat and distribution. 9. Piper muelleri C. DC. in A. DC. Prodr. 16(1): 243. 1869. Type: Mexico, Veracruz, Orizaba, 1853, Mueller 180 (holotype, BR!; isotypes, . GHI, ILL!, LeE!, Ny!). FIGURE 13. 36 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 70 FiGure 13. Piper muelleri: A, fruiting branch, x 0.4, with detailed insets of lower leaf surface showing indument, and prophyll and young inflorescence emergent from next- oldest prophyll (absent); B, portion of semimature infructescence, x 6; C, portion of mature infructescence (note recurved stigmas), x 6. Habit unknown, probably small to medium-sized shrub; young branches puberulent with eglandular hairs; older branches glabrous. Prophylls 0.5—1.2 mm long, slender, base truncate, apex acute, abaxially puberulent to pubescent with eglandular hairs. Leaves with petiole 0.6-1.8 cm long, stout, puberulent to densely pubescent with eglandular hairs; blade lance-elliptic to lance-ovate 1989] BORNSTEIN, PIPERACEAE 37 to rarely ovate, (7.3-)8.8-15.3 x 3.4-6 cm, rounded and symmetrical at base, acuminate to long-acuminate at apex, occasionally drying rugulose to bullate, 5- to 7-nerved (3 most central protracted to apex), glabrous above, densely pubescent beneath (especially along major veins) with eglandular hairs. Inflo- rescences 5—8.3 cm long; peduncle narrow, puberulent to pubescent, remaining rachis puberulent with eglandular hairs; bracts subsessile to minutely petiolate, 0.5-1.25 mm long, the apical region broadly spatulate-cupulate to cupulate, abaxially pubescent with eglandular hairs. Flowers with pedicel 0.5-2. long, sparsely puberulent to puberulent with eglandular hairs; stamens 5 to 7, hypogynous, the filaments to | mm long, equaling or slightly surpassing ovary length in flower or young fruit, the anthers subglobose, dehiscing laterally; ovary depressed-ovoid, glabrous, the style elongate, 1-1.5 times ovary length in flower, becoming largely obliterated during fruit maturation, the stigma (2- or) 3- or 4- (or 5-)lobed, with lobes recurved, persistent in fruit. Fruits de- pressed-ovoid when young, becoming globose at maturity, glabrous. DISTRIBUTION. Eastern slopes of Sierra Madre Oriental, from Coscomatepec south to Orizaba in Veracruz (see Map 3 Hasitat. Essentially unknown, although two collections (Matuda 1515 and Seler & Seler 5172) are from high elevations on mountains. PHENOLOGY. Flowering in May, perhaps also in late April and early June. Fruits occurring from May until August or later. SPECIMENS EXAMINED. Mexico. VERACRUZ: Orizaba, Botteri 192 (= 1156) (GH, K, US (2 sheets)); Cerro Punta Coscomat, Matuda 1515 (A, MICH, MO, Ny); Cérdoba, Cerro de Chocaman, Seler & Seler 5172 (Gu, us). Piper muelleri is distinguished from its closest relatives, P. neesianum and P. yucatanense, by its pubescent (vs. glabrous) se its comparatively long staminal filaments, and its uniquely shaped gynoecium Piper muelleri is perhaps the rarest species of mn subgenus. It has not been collected since 1937, despite my efforts in recent years. With the continued destruction of tropical forests in many areas of Mexico, including the area surrounding Orizaba, it would not be surprising if this species were extinct. However, because I had very little information to guide me in my quest, it may simply be a matter of looking in the wrong place. I suspect that an ex- haustive search of the region between Coscomatepec and Orizaba during the month of May could be rewarding and could thereby provide more detailed information concerning this species. 10. Piper neesianum C. DC. in A. DC. Prodr. 16(1): 256. 1869. Type: Mexico, locality unknown, 1841-42, Karwinski 823 (holotype, LE; isotypes, G!, H!). FiGureE 14. Arctottonia sempervirens Trel. ex Lundell, Publ. Carnegie Inst. Wash. 436: 302. 1934, nomen nudum. Arctottonia sempervirens Trel. ex Standley, Field Mus. Nat. Hist., Bot. Ser. 12: 405. 1936. Piper sempervirens (Trel. ex Standley) Lundell, Publ. Car- negie Inst. Wash. 478: 208. 1937. Type: Guatemala, Petén, Uaxacttin, 16 April 1931, Bartlett 12563 (holotype, MicH!; isotypes, ILL (2 sheets)!). 38 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 APM EOS i KAMAL J dl ux 4 SEES ~ a tT ma Figure 14. Piper neesianum: A, flowering branch, x 0.4; B, inflorescence, x 1.7; C, flower with bracts, x 10.5; D, immature and mature fruits with bracts, x 8.4; E, prophyll with emergent young leaf and inflorescence, x Slender shrub, multibranched from base, 1-6.5 m tall; bark green on young branches, becoming brown and warty with age. Prophylls to 0.5 mm long, base truncate, apex acute, appearing glabrous but with glandular hairs. Leaves with petiole 2-12 mm long, stout, with glabrescent glandular hairs; blade lance- elliptic to elliptic to ovate-elliptic, 5.2-12.6(-15.2) x 1.9-6.1 cm, rounded to more commonly acute, slightly unequal and occasionally auriculate at base, abruptly acuminate to long-acuminate at apex, stiff and coriaceous, 3-nerved (central nerve protracted to apex), dark shiny green above, gray-green to silvery beneath, both surfaces appearing glabrous but with glandular hairs (at least when young). Inflorescences (2.4-)3-6.9 cm long, white at maturity; peduncle slender, rachis appearing glabrous but with glandular hairs; bracts occasionally petiolate, narrowly spatulate to spatulate to rarely spatulate-cupulate, 0.25-0.5 mm long, glabrous. Flowers with pedicel 0.5—2.5(—3.5) mm long, lax, divaricate, appearing glabrous but with very sparse glandular hairs; stamens (3 or) 4 or 5 (or 6), the filaments ca. 0.25 mm long, the anthers subglobose, dehiscing lat- erally or occasionally slightly upward from lateral; ovary ovoid to rarely ellip- soid, puberulent with glandular hairs, the style conical, 4 to equal ovary in length, indistinct when young, becoming more distinct with age, the stigma (2- or) 3- to 5- (or 6-)lobed, persistent in fruit. Fruits ovoid-ellipsoid to globose, 1989] BORNSTEIN, PIPERACEAE 39 2-2.5 x 1.5-2 mm, conspicuously ribbed on drying, with sparse glandular hairs. DisTRIBUTION. Largely restricted to Yucatan Peninsula, especially central and southern portions; also extending into Chiapas and Veracruz (see Map 3). HasitatT. Originally occurring in “‘ramonales,” upland climax subdeciduous forests developing on calcareous soils and dominated by the breadnut tree, Brosimum alicastrum (Moraceae). This vegetation type was typical of the Yu- catan Peninsula until 20-25 years ago, when agricultural demands resulted in destruction of large areas of natural vegetation. Currently, therefore, Piper neesianum is usually found at the edge of disturbed secondary forests, at 20- 400(—700) m alt., where it is one of the more common understory shrubs. In more remote areas of the Yucatan Peninsula, such as portions of the Petén region of Guatemala, one should still be able to find this species in its undis- turbed native habitat. PHENOLOGY. Flowering February to June, but most prominently in “ and May; specimens with mature fruits collected from June to November REPRESENTATIVE SPECIMENS. Mexico. VERACRUZ: Papantla, Liebmann 18 (G, kK, us); Lo- mas al sur del Poblado 11, ca. 30 km E del campamento La Laguna, seain rene Wendt et al. 2633 (ENCB, MEXU, NY). mes 20 mi N of Ocozocoautla, near km 31 marker, toward Apitpac, Croat 4064 amino a Jato del Tigre, cerca Fe Galon, Miranda 6245 (MEXxU); near El Ocote, 30 km NW of Ocozocoautla, Miranda 6252 (MEXU, US). Tapasco: 100 m del limite con México sobre carretera 15-E, Novelo et al. 178 (MEXxv). CAMPECHE: along Hwy. 186 between Escarcega and Xpujil at km 87 marker, Bornstein & Chase 69 (micH); Campo Experimental Forestal Tropical “El Tormento,” km 5 on 1553 (MEXU), E\S- 1631 (MEXU); zona arqueologica de Becan, Loma central, Cowan 2997 (ENCB, MEXU); Tecolutla, Liebmann 17 (G, GH); Reforma, Lundell 834 (F, ILL, MICH, MO, us); Pixoyal, Ross 128 (GH, MEXU, US); 14 km S of Xpujil, Shepherd 68 (ENCB, MICH). YucaTAn: Chichén Itza, Lundell 7406 (F, GH, MICH (2 sheets), NY, TEX, US). QUINTANA Roo: ee Bacalar shoreline pkwy., 1.1 km S of school, Bornstein & Rutledge 27 (micu); 1.9 km S of Bacalar-Cenote Azul oe along Hwy. 307, Bornstein & Rutledge 30 (mIcH); 0.3 km SE of km 20 marker along Hwy. 196, Bornstein & Rutledge 33 (MICH); along Hwy. 186 between Escarcega and Chetumal, |.9 km E of km 225 marker, Bornstein & Rutledge 38 (micH); Lake Chichankanab, Gaumer 23678 (F, GH, MO, Ny, US); 26 km S of Tulum, Téllez & Cabrera 2606 (MEXxU); 8 km NE of Felipe Carillo Puerto, oe ; ; ‘ R Georgeville on road to Augustine, Croat 23440 (ENCB, F, GH, MEXU, MO, TEX); Chalillo Crossing, Lundell 6510 (Ff, GH, MICH, NY, TEX, US). Guatemala. nN Uaxactin, Bartlett 12326 (ILL, MICH, NY, UC, US); 10 km E of Remate: Contreras 916 (F, G, MEXU, TEX, US); W of Remate, Contreras 1033 (MICH, TEX (2 sheets), Us); Tikal Natl. Park, Contreras 1246 (G, MICH, TEX, US); Lake Macanche, Contreras 5518 (F, MO, NY, US); Lake Petén Itza, Santa Elena, Contreras 5763 (F, GH, TEX); Macanche, Contreras 5840 (F, GH, (ILL, MICH, TEX); 35 km E of Santa Elena, Molina 15482 (F, Ny); Santa Elena, km 6 on road to San Andres, Ortiz 1978 (F, MO, NY, US); 2 km NE of Uaxacttin, Ortiz 2609 (F, MICH, US); bordering Lake Macanche, 35 km from Flores, Ortiz 2696 (F, MICH, US); Cerro 40 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Ceibal, between mouth of Rio Santa Monica & Rio San Martin, Satis 46061 (F, MICH AS sheets), NY, US). ALTA VERAPAZ: Cerro Chinaja between Finca Yalpemech and Chinaja, Stevermark 4558] (F, Micu). IZABAL: Cienaga, on pe Gn tea road, Con- treras 10851 (MEXU, MO, TEX, US). Piper neesianum 1s commonly confused with its sympatric congener, P. yu- catanense. Both species have leaves distinctly longer than broad, but P. nee- sianum has a more dense, shrubby habit; firm, shiny, nonrugulose (vs. mem- branous, dull, and rugulose or bullate) leaves; smaller, narrower floral bracts; and conspicuously ribbed (vs. smooth) fruits (especially when dry). A third species with lance-ovate leaves, P. muelleri, has rarely also been mistaken for P. neesianum. However, it can be easily distinguished by its rather heavy indument of eglandular hairs along the veins on the leaf undersurface, its pubescent floral bracts and rachis, its long staminal iene and its uniquely shaped gynoecium. 11. Piper novogalicianum Bornstein, sp. nov. FIGURE 15. Frutex 1.5-5 m altus, ramunculis puberulis ad dense pubescentibus; pro- phyllum usque ad 1 cm longum, puberulum vel dense pubescens; folia laminis ovatis vel late ovatis vel suborbiculatis, 4.9-13.2 x 4.1-11.5(-—13.5) cm, bre- viacuminatis vel acuminatis, basi rotundatis vel cordatis, supra puberulis vel pubescentibus, subter puberulis vel pubescentibus vel tomentosis; inflorescen- tia 6.2-14.3(-16) cm longa, rhachidi sparsim ad dense pubescenti; bracteae usque ad 0.75 mm longae, spathulatae vel spathulicupulatae, sparsim puberulae vel pubescentes vel dense pubescentes; pedicelli 0.5—1.5(—2) mm longi; stamina hypogyna vel interdum ut videtur basi ovaril adnata; ovarium ellipsoideum vel subglobosum; stylus depresso-ovoideus vel breviconicus vel brevicolum- naris, 0.5-0.75 mm longus; fructus immaturus transverse-ellipsoideus, pu- berulus. Type. Mexico, Jalisco, 5.1 km W of Hwy. 80 along Hwy. 200 toward Puerto Vallarta, 300 m alt., 20 June 1983, Bornstein 99 (holotype, MICH; isotypes, CHAP, GH, INIF, MEXU, NY). Small to medium-sized shrub, often treelike, 1.5—5 m tall; youngest twigs puberulent to densely pubescent, glabrescent, the hairs predominantly eglan- dular; older stems gray-brown to ash-gray, with swollen nodes and round, warty lenticels, maximum diameter to 6 cm. Prophylls to 1 cm long, base truncate, apex acute, abaxially puberulent to densely pubescent with eglandular hairs. Leaves with petiole 0.4—2.7 cm long, narrow to moderately broad, pubescent to densely long-pubescent with eglandular hairs; blade ovate to broadly ovate to suborbiculate, 4.9-13.2 x 4.1-11.5(-13.5) cm, truncate to rounded to cor- date at base, gradually short-acuminate to acuminate at apex, 5- to 7- (to 9-)nerved, puberulent to pubescent above with both hair types in + equal abundance, puberulent to pubescent to tomentose beneath with eglandular hairs predominant along and + restricted to major veins. Inflorescences 6.2—14.3(-16) cm long, yellowish white to white; peduncle puberulent to pubescent with eglandular hairs, remaining rachis sparsely pubescent to pubescent with eglan- 1989] BORNSTEIN, PIPERACEAE 41 Ficure 15. Piper novogalicianum: A, flowering branch, x 0.42, with detailed inset of lower leaf surface to show indument; B, portion of inflorescence, x 6.3; C, portion of infructescence, x 6.3 dular hairs; bracts sessile to subsessile, spatulate to spatulate-cupulate, to 0.75 mm long, abaxially sparsely puberulent to densely pubescent with eglandular hairs. Flowers with pedicel (0.5—)1—1.5(—2) mm long, usually puberulent with glandular and/or eglandular hairs; stamens 4 to 6 (or 7), hypogynous or oc- casionally appearing as if adnate to base of ovary, filaments and anthers + equal in length (to 0.5 mm), anthers dehiscing slightly to conspicuously upward from lateral; ovary ellipsoid to subglobose, puberulent with glandular hairs, the style depressed-ovoid to short-conical or short-columnar, 0.5-0.75 mm 42 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 long, ca. 2 to equal length of ovary when in flower, the stigma 3- to 5-lobed. Fruits (when partially mature) transversely ellipsoid, ca. 1 x 1.5 mm, puber- ulent, style persistent. DISTRIBUTION. Southwestern Jalisco and neighboring Colima; disjunct in state of Mexico (see Map 2). HAsirTAat. Steep hillsides in subdeciduous forests, from near sea level to 1350 m alt.; apparently also in second-growth woodlots, especially inland and at higher elevations. PHENOLOGY. Flowering May to mid-July or into August. Fruits appearing as early as mid-June, but most commonly July to September SPECIMENS EXAMINED. Mexico. JALisco: | km NE of Cihuatlan along small stream parallel to Rio Cihuatlan, Bornstein & Hensold 5 (micu), 2 km § = a Manzanilla on road to Tamarindo beach, Bornstein & ye & (MICH); along . 80 between Autlan and Melaque, 4 km N of side road to Casimir Castillo, Eon "16 (MICH); ca. 11 km S of Autlan on Hwy. 80 to Barra de Navidad Bornstein 96 (mIcH); Santa Cruz, Delgado 348 & Herndndez 2599 (ENCB, MEXU); ca. 22 mi SSW of Autlan, Wilbur 1589 (micu); 0.5 mi N of La Resolana, Wilbur 1597 Hoar 2 mi W of Autlan, Wilbur 1675 (micu). COLIMA: 2.5 km S of coastal Hwy. 200, W of Santiago along gravelly dirt road to Vida del Mar, Bornstein & Hensold 2 (micw); 5 mi W of Santiago, then 2.5 mi on road to Pefia Blanca, McVaugh oo (MICH). Mexico: Tejupilco in district of Temascaltepec, Hinton 4287 (ILL, K, MICH, S). In my dissertation (Bornstein, 1985) I recognized the high-elevation collec- tions (Bornstein 96, Hinton 4287, Wilbur 1675) as a distinct species, Piper tejupicoanum Trel. ex Bornstein, with the Hinton collection at Micu designated as the holotype. At that time I also noted my concern about the recognition of this taxon, since the characters that serve as the basis for the distinction (1.e., leaf size and pubescence, rachis and bract pubescence, and pedicel length) are usually rather variable. After reconsidering this situation, I have decided to recognize a single variable species, P. novogalicianum, while noting that the populations at high elevations apparently represent one extreme in which the leaves are smaller and the plant parts more pubescent. This matter will only be settled with additional collecting (especially of flowers and fruits from the same plant), as well as field studies of populations at various elevations. This species, as presently circumscribed, is distinguished from Piper brevi- pedicellatum on the basis of having longer and narrower pedicels. 12. Piper rosei C. DC. Annuaire Conserv. Jard. Bot. Genéve 21: 316. 1920. Type: Mexico, Sinaloa, Colomas, 13-20 July 1897, Rose 3234 eels here chosen, us 302199!). E 16. Ge ac DC. ex Trel. Amer. J. Bot. 8: 215. 1921. Type: Mexico, Sinaloa, Colomas, 3-20 July 1897, Rose 3234 (holotype, us!). Aas aerated Trel. Amer. J. Bot. 8: 215. 1921. Type: Mexico, Jalisco, Diguet s.n. (holotype, Medium-sized treelike shrub, 2-5 m tall; youngest twigs glabrous to puber- ulent with short eglandular hairs; older branches soon glabrous, with brown, 1989] BORNSTEIN, PIPERACEAE 43 FiGureE 16. Piper rosei: A, flowering and fruiting branch, x 0.42; B, portion of inflorescence, x 6.3; C, portion of infructescence, x 6.3. elliptic, warty lenticels, maximum diameter to 7.5 cm. Prophylls to 10 mm long, base truncate, apex acute, glabrous to puberulent with short eglandular hairs. Leaves with petiole 0.7—1.9(—3.1) cm long, narrow, glabrous to puberulent with glandular and eglandular hairs; blade ovate to most commonly broadly ovate to suborbiculate, (5.3—)6.7-13(-15.6) x 4.2-12.1 cm, rounded to shal- lowly cordate at base, short- to long-acuminate at apex, 5- (to 7-)nerved, gla- brous to puberulent, the hairs largely restricted to petiole-blade junction, glan- dular and short eglandular. Inflorescences (5—)6.3—15.5 cm long, greenish yellow when young, white to cream colored in full flower; peduncle glabrous to sparsely pubescent, remaining rachis sparsely to densely pubescent with eglandular hairs; 44 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 bracts sessile to subsessile, spatulate to cupulate, 0.25-0.75 mm long, sparsely pubescent to pubescent (often on both surfaces, but especially abaxially and marginally) with eglandular hairs. Flowers with pedicel 0.5-1.5(-1.75) mm long, usually glabrous to puberulent, but occasionally sparsely pubescent near base with both glandular and eglandular hairs; stamens (3 or) 4 or 5 (or 6), hypogynous but often appearing adnate to base of ovary, filaments and anthers to 0.5 mm long, anthers dehiscing laterally or conspicuously upward from lateral; ovary ovoid-ellipsoid to widely depressed-ovoid to subglobose, puber- ulent with glandular hairs, the style depressed-ovoid to short-conical to short- columnar, from 4 to equal length of ovary when in flower, the stigma 3- to 5- lobed. Fruits subglobose to globose, 1-2 x 1-2 mm, puberulent, style becoming obscure with fruit maturation. DisTRIBUTION. Southern Sinaloa south through southwestern Jalisco and north- ern Colima (see MAp 4). Apparently absent from Nayarit (probably a reflection of limited collecting throughout Nueva Galicia). Hasirtat. Prefers steep hillsides or slopes in relatively open, disturbed, semi- deciduous or deciduous forests. At low elevations found with other pipers in the coastal palm forests dominated by Orbignya cohune. At high elevations occurring in oak forests, above coffee plantations, and in forests dominated by Bursera spp. From near sea level to 1400 m, but mostly below 1000 m alt. PHENOLOGY. Flowering late May to early July. Young fruits appear by mid- June; mature fruits collected in September and October. A second period of flowering may also occur at this time (end of rainy season). REPRESENTATIVE SPECIMENS. Mexico. SINALOA: Colomas, Rose 1657 (GH, K, Us); Copalita, Moran 13310 (uc), JALIsco: ca. 2 km S of La Manzanilla on road to Tamarindo beach, Bornstein & Hensold 6 (micu); ca. 15 km § of Autlan on Hwy. 80 toward Barra de Navidad, Bornstein 97 (micu); 5.1 km W of Hwy. 80 along Hwy. 200 toward Puerto Vallarta, ae 100 (micH); 4 mi NE of Talpa, Chase 83127 (mMicu); 5 mi S of Talpa on road to Tomatlan, Chase 83129 (micH); 16 km N of Mascota on road between Mascota and San Sebastién, Lott et al. 1349 (micu); ca. 5.5 mi N of Tecalitlan, McVaugh 15436 (MICH); ca. 16.9 mi NW of Hwy. 80-200 junction (near Melaque) along Hwy. 200, Stevens & Fairhurst nae soe TEX), 8 mi S of Autlan toward La Resolana, Wilbur 2190 (micu). ‘ota: 1.3 mi W of San Antonio, Chase 83113 (mic). — TAXONOMIC NOTES. Piper rosei C. DC. and P. rosei C. DC. ex Trel. are inde- pendent descriptions based on the same type material; i.e., they are nomen- clatural synonyms. De Candolle examined the Rose collections at us in the late 1890's, apparently supplying a name for the collections but not publishing it until over 20 years later. Trelease saw these same Rose collections while preparing his first publication on the Piperaceae and adopted the herbarium epithet suggested by De Candolle. De Candolle’s treatment of the Mexican Piperaceae was published several months prior to Trelease’s treatment of the North American pipers of sect. Ottonia, and therefore Trelease’s name is an illegitimate later homonym. The characters cited by Trelease to distinguish Piper diguetianum and P. rosel are not reliable. The racemes in the former are shorter because they are 1989] BORNSTEIN, PIPERACEAE 45 obviously broken. The pedicel length difference (1.5 vs. 1 mm, respectively) is a result of the level of floral maturation: the flowers of the Rose collections are considerably younger, and therefore the pedicel has not fully elongated. The third distinction, stamen number (5 vs. 3 or 4), has been shown to be a poor character on which to base species differences. Again examining the Rose material, I have seen numerous flowers with five stamens in addition to those with three or four. In my original investigation of this species (Bornstein, 1985), I separated the low-elevation collections (Stevens & Fairhurst 1868; Bornstein & Hensold 6, , & 9) as a distinct taxon, Piper palmicola Bornstein, with the Stevens & Fairhurst collection at TEx designated as the holotype. This species was rec- ognized as distinct on the basis of its less-pubescent floral rachis and bracts, its short-acuminate to acuminate leaf apex, and its shorter (0.5—1 vs. 1-1.5 mm) pedicels. After further consideration, I have decided that these features either are essentially continuous (pedicel length) or are too variable (pubescence and leaf shape) to warrant the recognition of P. palmicola, which remains an unpublished name, and I have consequently interpreted P. rosei in a rather broad sense. The low-elevation collections form a group that tends to obscure the boundary between P. brachypus and P. rosei, strongly suggesting that these two species are closely related. 13. Piper stipulaceum Opiz in Presl, Relig. Haenk. 1(3): 160. 1828. Type: Mexico, locality unknown, Haenke s.n. (lectotype, here oad PR!; isolectotypes, B (destroyed?), PR! FIGURE 17 Enckea stipulacea (Opiz) Kunth, Linnaea 13: 600. 1838. Ottonia stipulacea (Opiz) Presl, Epimel. Bot. 231. 1851. Arctottonia stipulacea (Opiz) Trel. Proc. Amer. Philos. Soc. 75: 716. 1935. Piper aguilanum C. DC. Annuaire Conserv. Jard. Bot. Genéve 21: 318. 1920. Type: Mexico, Guerrero, Santa Rosa prés d’Aguila, 650 m alt., 16 July 1898, Langlassé & (holotype, G!; isotypes, GH! k!, us! Piper albicaule Trel. Amer. J. Bot. 8: 216. ry 8, fig. 2. 1921. Type: Mexico, Guerrero, Santa Rosa prés d’Aguila, 650 m alt., 16 July 1898, Langlassé 248 (holotype, us!; isotypes, G!, GH!, K!). Medium-sized treelike shrub, sparsely branched from base, 2—4 m tall; youn- gest twigs pubescent to densely pubescent, glabrescent, ash-gray and with con- spicuous warty brown lenticels, to ca. 5 cm in diameter. Prophylls 6-—10(-15) mm long, base truncate, apex acute, abaxially pubescent to densely pubescent to occasionally tomentose with eglandular hairs. Leaves with petiole 0.6-3.5 cm long, stout, pubescent to densely soft-pubescent with eglandular hairs; blade ovate to broadly ovate-suborbiculate, (4.5—)6.3-13.7(-14.6) x (2.4-)4.8-12.6 (-14) cm, rounded to truncate to mostly shallowly cordate at base, short- acuminate to acuminate at apex, (5- to) 7- (to 9-)nerved, puberulent to pu- bescent above, puberulent to densely pubescent beneath especially along major veins (both hair types present, with glandular ones more prevalent in areolar regions and eglandular ones + restricted to veins). Inflorescences (5.8—-)7-12.9 cm long, yellow-green when young, whitish in full flower; peduncle stout, 46 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Ficure 17. Piper stipulaceum: A, fruiting branch, x 0.43; B, inflorescence, x 8.6: C, flowers and associated bracts, x 8.6; D, portion of young infructescence, x 8.6; E, mature fruit, x 12.9; F, prophyll with emergent young leaf and inflorescence, x 8.6. puberulent to pubescent at base with glandular and eglandular hairs, gradually transitional to glabrous to puberulent to sparsely pubescent rachis; bracts oc- casionally sessile, but usually with narrow stalklike basal portion that broadens gradually to abruptly into narrow to broadly spatulate to spatulate-cupulate apical region, 0.5-0.75(-1.25) mm long, abaxially puberulent to pubescent. Flowers with pedicel 0.5—2 mm long, ca. 0.5 mm wide, glabrous to puberulent with glandular hairs; stamens 4 or 5 (or 6), adnate to ovary from near base to 1989] BORNSTEIN, PIPERACEAE 47 middle, the filaments to ca. 0.5 mm long, the anthers depressed-ovoid, usually longer than filaments, dehiscing laterally or occasionally slightly divergent up- ward; ovary ovoid to subglobose, puberulent with glandular hairs, the style depressed-ovoid to conical, to 0.5 mm long, from 4 to nearly equal length of ovary when in flower, the stigma 3- or 4- (or 5-)lobed. Fruits ovoid to globose to slightly obloid, usually ca. 2 x 2 mm, drying brown to black, with glandular hairs, the style usually persistent to some degree, but not more than 4 length of ovary. DISTRIBUTION. Southwestern Jalisco south to northwestern Guerrero, near Zi- huatanejo, in narrow band from Pacific Ocean to foothills of Sierra Madre (see AP 4) Hasirtat. Variable: at low elevations near coast, a relatively frequent understory shrub in palm forests dominated by Orbignya cohune; at higher elevations farther: inland, found along oe of ravines above intermittent mountain streams in ests usually dominated by Brosimum alicastrum, Ficus spp., or Bursera spp. Fionn near sea level to 650 m alt. PHENOLOGY. Flowering May to July, rarely later. Young fruits appearing late May or early June; mature fruits collected from August to September, but probably persisting on plants into dry season SPECIMENS EXAMINED. Mexico. JALISco: 5.1 km W of Hwy. 80 along Hwy. 200 toward Puerto Vallarta, N side of road, Bornstein 79 (mic); 7.9 km WNW of Hwy. 80 along Hwy. 200 toward Puerto Vallarta, S side of road, Bornstein & Weber-Bornstein 83 (MICH); ca. 12 km NE of Nacastillo on road to San Miguel Purificacién and Purificacion, oe & Nee 1531a (Ny); 1-4 km E of Casimir Castillo, ca. 20 km SSW of Autlan, “A od Tacubaya,” at base of “Cerro la Petaca,” //tis et al. 296 (NY); ca. 2 km N of division line of UNAM, road to Nacastillo, Magallanes 754 (TEX). COLIMA: 7.8 km N on Hwy. 98 toward Minatitlan, E side of road, Bornstein 86 (micH); 0.5 km S of Comatlan on Hwy. 98 toward Minatitlan, E side of road, Bornstein 98 (MICH); Playa. de Santiago, on road to Manzanillo airport, McVaugh 15594 (MICH). MICHOACAN: 2 mi N of San Juan de Lima (on road to Coahuayana), Hansen, Hansen, & Nee 1442 (MEXU, MICH, US, WIS); La Placita (Rio de Aquila), Turner 2035 (micu); 1 mi E of La Placita, Turner 2071 (MICH). GUERRERO: Vallecitos, Montes de Oca, Hinton 11372 (F, K, MEXU, MO, NY, US); 2 mi E of Vallecitos de Zaragoza (between kms 52 & 53) on Hwy. 200 toward Ciudad Alta- mirano, Murray & Johnson 1433 (MICH). TAXONOMIC NOTE. As in the Piper michelianum—Piper mas and Piper rosei situations, both Casimir de Candolle and William Trelease published new species, Piper aguilanum and Piper albicaule, respectively, on the basis of the same type material. These taxa are thus nomenclatural synonyms. Piper stipulaceum is most often confused with Piper abalienatum because in both the stamens are adnate to the ovary (only P. mcvaughii also shares this character, and it is easily distinguished by its puberulent vegetative parts, very- short-pedicellate flowers, and smaller leaves). The characters and character states that I use to distinguish these two species are summarized in the TABLE. When the opposite extremes of these characters are expressed, which is fre- quently the case, these species are quite different and readily separable. There are, however, a few specimens that approach the established limits of the range 48 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 of the character states, often appearing to bridge the gap that separates the species. Usually all or most of these characters are affected simultaneously in a given specimen, possibly suggesting that introgressive hybridization might be the cause rather than abnormal extremes of variation. Although I have not observed these species growing together, their geographic ranges do overlap in portions of Jalisco, Colima, and Michoacan. In addition, habitat preferences of the species are similar, although P. stipulaceum does seem to be more restricted to cool, moist sites. Therefore, the potential for hybridization exists, as do the newly disturbed, unstable habitats that hybrids may need to become established. However, detecting and documenting introgressive hybridization would require more extensive collection of specimens in the areas of geographic overlap. In the meantime, I have placed those “intermediate” specimens in one or the other species on the basis of overall similarity using the criteria available. A somewhat different problem occurs in trying to distinguish Piper stipu- laceum and P. novogalicianum. As presently circumscribed, the taxa differ significantly only in the position of the stamens relative to the ovary (adnate and hypogynous, respectively). As I have mentioned previously for several taxa, P. novogalicianum among them, there is a strong tendency for the stamens to appear to be adnate to the base of the ovary. This feature is most apparent in the fruiting condition, where the ovary has expanded, thereby partially incorporating the upper portion of the pedicel (the receptacle). This, in turn, causes the stamens seemingly to ‘“‘shift’’ to the adnate position. Since the re- maining features of these two species are quite similar, this apparent change can lead to identification problems, especially with pressed specimens. To enable distinction between the taxa, flowering material should be obtained, preferably in pickled form. Fortunately, most of the specimens that I have examined have had some flowers, so that when they are restored to their original state with a detergent solution, it is possible to assign them to the appropriate species. 14. Piper yucatanense C. DC. Linnaea 37: 334. 1872. Type: Mexico, Yucatan, July 1840, Linden s.n. (holotype, G!). FiGur_E 18. Piper thiemeanum Trel. Amer. J. Bot. 8: 214. pl. 5, fig. 2. 1921. Ottonia thiemeanum uncker, Ann. Missouri Bot. Gard. 37: 7. 1950. Type: Honduras, Cortés, San Pedro Sula, June 1887, Thieme 5455 (holotype, us!; isotypes, GH!, K!, UC!, US (3 sheets)!). Piper tatet Trel. Amer. J. Bot. 8: 215. 1921. Type: Nicaragua, 1867-8, Tate 367 (holotype, k!). Piper chichankanabanum Trel. J. Wash. Acad. Sci. 19: 330. 1929. Type: Mexico, Quintana Roo, Lake Chichankanab, April 1917, Gaumer 23699 (holotype, F!). Arctottonia pittiert Trel. Proc. Amer. Philos. Soc. 73: 328. 1934. Type: Panama, Canal Z Alhajuela, 11-14 May 1911, Pittier 3458 (lectotype, here chosen, F!; isolec- totypes, F!, us!). Arctottonia tuxpenyana Trel. ex Lundell, Publ. Carnegie Inst. Wash. 436: 311. 1934, nomen nudum. Arctottonia tuxpenyana Trel. ex Standley, Field Mus. Nat. Hist., Bot. Ser. 12: 406. 1936. Piper tuxpenyanum (Trel. ex Standley) Lundell, Publ. Car- 1989] BORNSTEIN, PIPERACEAE 49 Ficure 18. Piper yucatanense: A, flowering branch, x 0.4; B, ee x 8; “ flower with bracts, x 10; D, immature (unfertilized?) and mature fruits with bract, 10; E, prophyll with emergent young leaf and inflorescence (detached os displaced ee normal position), x gie Inst. Wash. 478: 208. 1937. Type: Mexico, Campeche, Tuxpena, 7 Feb. 1932, Lundell 1300 (holotype, micu!; isotypes, F!, ILL!, NY!). Slender, loosely branched shrub 0.5-3 m tall, appearing glabrous throughout. Prophylls to 10 mm long, base truncate, apex narrowly acute, with very sparse glandular hairs. Leaves with petiole 1-10 mm long, slender, with sparse glan- dular hairs at least when young; blade lanceolate-ovate to oblong-ovate to ovate-elliptic, 6.5-15 x 2.2-5.5 cm, acute to more commonly rounded (rarely shallowly cordate), slightly unequal and occasionally minutely auriculate at base, gradually acuminate to very long-acuminate, occasionally cuspidate at apex, 5- to 7-nerved (3 most central protracted to apex), membranous, drying thin and generally rugulose (rarely bullate), appearing glabrous but with glan- dular hairs when young. Inflorescences 2—5.5(-6.3) cm long, white at maturity, fragrant; peduncle slender, rachis appearing glabrous but with glandular hairs; bracts broadly spatulate-cupulate to strongly cupulate (occasionally appearing 50 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 ae i a Et } 100 200 300 400 500 KM ° 100 200 300 MI Map 5. Distribution of Piper yucatanense; inset of Central America shows isolated collecting localities in Nicaragua and Panama. petiolate), 0.25-1.5(-2) mm long, nearly glabrous. Flowers with pedicels di- varicate, |-4 mm long, very sparsely puberulent with glandular hairs: stamens 4 to 8, hypogynous, the filaments less than 0.25 mm long, the anthers subglo- bose, dehiscing laterally; ovary obloid to ovoid-ellipsoid, puberulent with glan- dular hairs, the style from 3 to nearly equaling length of ovary, becoming largely obliterated in fruit, the stigma lobes (2 or) 3 or 4 (or 5), persistent in fruit. Fruits ovoid to ovoid-ellipsoid, to 3 x 2 mm at maturity, with numerous glandular hairs. DisTRIBUTION. Most prevalent in Yucatan Peninsula, where sympatric with Piper neesianum, ranging naturally into western Honduras. One collection known from Nicaragua, one from Panama (see Map 5). Hasitat. Originally most often in “ramonales”; currently in second-growth forests with Piper neesianum. From near sea level to 1000 m alt. PHENOLOGY. Flowering February to late June or early July, most commonly late April to early June. Mature fruits from August until early in next flowering period. REPRESENTATIVE SPECIMENS. Mexico. CHIAPAS: Yaxchilan, Breedlove 42858 (MEXU, MO). Tasasco: 2 km [?] de Teapa, Cowan 316] (ENCB, MEXU); San Isidro near Balancan, Matuda 3382 (F, GH, MEXU, MICH, NA, NY). CAMPECHE: 3.5 km § of Xpujil, Bornstein & Chase 58 (MICH); ruinas de Chicana, Hopelchén, Chan & Ucdn 994 aside 2km S of Xpujil, Shepherd 25 (F). YUCATAN: Xcan, Miranda 8248 (MEXU, US). QUINTANA Roo: 1.6 km S$ of Santa Cruz Chico along Hwy. 307, Bornstein & Rutledge 23 es 0.3 km SE of km 20 marker along Hwy. 196, Bornstein & Rutledge 34 (micu); along Hwy. 186 toward 1989] BORNSTEIN, PIPERACEAE Dl Chetumal, 1 km E of intersection with Hwy. 307, Bornstein & Rutledge 37 (micu); 1.9 a S of Bacalar-Cenote Azul turnoff along Hwy. 307, Bornstein & Chase 71 (micn), 8 m S of Xactun, Felipe Carillo Puerto, Pérez 418 (mic); 25 km N of Coba, Téllez & Cabrera 2534 (MEXU). Belize. ORANGE WALK: Honey Camp, Lundell 312 (F, GH, ILL, Ny). BELIZE: mi 24-25 on Northern Hwy., Dwyer & Liesner 12037 (GH, MO); Maskall, ans 1246 (GH, ILL, MICH, NY). Cayo: 9.5 mi S of Georgeville on road to Augustine, Croat 23337 (F, MO); vic. of Cuevas, Sopa Croat 23629 (Mo); 4 mi S of Grano de Oro on road to La Flor, Dwyer 10906A (F, Ny); Cuevas, Dwyer 11572 (MEXU, Mo); Vaca, Gentle 2511 (GH, MICH, TEX); 6 mi S orc Grano de Oro, La Flor camp, Lundell 6289 (F, GH, MEXU (3 sheets), MICH, NY, TEX, US); Cohune Ridge, Lundell 6499 (Gu, Silk Grass Creek Reserve, Gentle 3102 (micu); between Silk Grass and Railway, Sampson 12 (K (2 sheets)). TOLEDO: Columbia Forest Reserve (1-2 mi N of entrance), Croat 24174 (mo); Punta Gorda-Seven Hills road, Gentle 5040 (F, TEX); near San Antonio, Gentle S)5 trail from Columbia Forest =e to Esperanza, Gentry 8171 (F, MO); Forest Home, Schipp 506 (F). District UNKNOWN: Prospecto—Maskall road, Gentle 895 (GH, ILL, MICH, MO, NY). Guatemala. PETEN: La. tiberad Aguilar 251 (ILL, MICH, MO, NY); Dolores, Contreras 2557 (Ny, TEX); Lacandén, Contreras 3412 (TEx); 1 km S of Lacandon, Con- treras 3519 (TEX); 6 km SSW of Lacand6én, Contreras 3538 (F, G (2 sheets), MICH, TEX); Macanche, Contreras 5807 (GH, TEX); Tikal Natl. Park, Gentry 8327 (F, GH, MO); El Paso, Lundell 1487 (F, GH, MICH, MO, NY, US); Santa Elena, at km 38 near road to Remate, Ortiz 2048 (F, MO, NY, US). ALTA VERAPAZ: finca Argentina above Papalja, 15 mi W of Teleman on road from Tucuru to El Estor, Croat 41539 (mo); Cacao, Trece Aguas, Lewton 333 (F). Izapat: 6 km from La Ruidoza, Contreras 10796 (MEXU, MO, TEX); Cienaga, on Petén—Guatemala road, Contreras 10853 (MEXU, MO, TEX), ENE of finca El Rita and Copan ruins, Molina 30605 (F (2 sheets), MO). SANTA BARBARA: W side of Lake Yojoa, Williams & Molina 11423 (F, GH, MO, Us); near village of Las Vegas, Yong et al. 85 (Mo). Cortés: valley W of San Pedro Sula, Bangham 345 (F, GH); canyon of Rio Piedras, 3 km from San Pedro Sula, Molina 3432 (F, Us). COMAYAGUA: near La Libertad, Molina 7009 (F (2 sheets), TEX); valley of Jamalteca, Molina 7185 (F (2 sheets), TEX), Jardines, Barkley & Herndndez 40847 (Gu). Nicaragua. ZELAYA: mountains and rain forests between Toro Bayo and Esquipulas, Shank & Molina 4637 (F). Piper yucatanense is, at first glance, vegetatively similar to P. muelleri, but it is easily distinguished by its glabrous appearance. However, several collec- tions (Burch 6094, Contreras 10796, Gentry 7830, Molina 30605, Standley 24675, Williams & Molina 11423, and Yong et al. 85), primarily from the Guatemala-Honduras border, do have considerable pubescence on the pro- phylls, petioles, leaf blades, rachises, and bracts, reminiscent of P. muelleri. Fortunately, these two species can be distinguished on the basis of filament length and fruit shape, and they are geographically separated as well. Piper yucatanense has also often been confused with its sympatric neighbor P. neesianum. It is distinguished from that species by its smaller, less dense habit, its membranous, rugulose (vs. coriaceous, nonrugulose) leaves, and its longer, broadly spatulate bracts. In addition, it occurs in more open areas of the forest and is not a dominant member of the understory. Yuncker (1950), when making the new combination Ottonia thiemeana (Trel.) Yuncker, cited Pittier 3458 as the type collection. This is clearly incorrect because the type must be that of the basionym, Piper thiemeanum (Thieme 5455). 52 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 ACKNOWLEDGMENTS Iam most grateful to W. R. Anderson, who unwittingly first introduced me to the family Piperaceae, for his guidance during the course of this study, for his valuable editorial assistance, merciless and otherwise, and for correcting the Latin descriptions. To W. H. Wagner, Jr., I offer special thanks for his assistance On numerous technical matters and his constant encouragement. Helpful oe of the manuscript was provided by W. C. Burger, C. J. Hum- phries, and M. Tebbs. I am particularly indebted to L. M. Villarreal de Puga, and her staff and students at the Herbario, Instituto de Botanica, Universidad de Guadalajara, for their gracious hospitality and assistance during my field studies. Thanks are also due to N. Hensold, M. Chase, and especially S. DeVivo, who accom- panied me on several field excursions. I thank the curators of A, BR, C, ENCB, F, G, GH, H, ILL, K, LE, MEXU, MICH, MO, NY, PR, TEX, UC, and us for the loan of herbarium specimens used in this study. For their artistic skill in rendering excellent illustrations from material ad- mittedly less than ideal, J. Charles, A. Eisenberg, and especially P. Kernan are to be commended. I also gratefully acknowledge the expert editorial assistance of E. B. Schmidt in the preparation of this manuscript. This paper represents a portion of a dissertation submitted in partial fulfill- ment of the requirements for the Ph.D. degree (Botany) at the University of Michigan. Financial support for this project was provided, in large part, by grant DEB-8116762 from the National Science Foundation to the University of Michigan. I also received support in the form of grants and fellowships from the Department of Botany, the Division of Biological Sciences, and the Horace H. Rackham School of Graduate Studies (all at the University of Michigan), as well as from the Michigan Chapter of Sigma Xi. Their help is much appre- ciated. Last, | want to thank my wife, Nancy, for her continuous support and patience during this study. LITERATURE CITED BAILLON, H. 1872. Histoire des plantes. Vol. 3. L. Hachette, Paris. BornsTEIN, A. J. 1985. A revision and systematic evaluation of Piper subgenus Arc- tottonia (Piperaceae). Unpubl. Ph.D. dissertation, University of Michigan, Ann Ar- bor, Michigan. BurGeR, W. C. 1971. Piperaceae. /n: Flora Costaricensis. Fieldiana, Bot. 35: 5-218. . 1972. FA a Central American species of Piper (Piperaceae). Brittonia 24: 356-362 CALLEJAS, R. 1986. oak revision of Piper supa Ottonia (Piperaceae). Un- publ. Ph.D. dissertation, City University of New York. CANDOLLE, C. DE. 1869. Piperaceae. Pp. 235-471 in A. L. P. P. pE CANDOLLE, ed., Prodromus systematis naturalis regni vegetabilis. Vol. 16, pt. 1. V. Masson, Paris. 20. Piperaceae Mexicanae. Annuaire Conserv. Jard. Bot. Genéve 21: 305— 322 1923. Piperaceum clavis analytica. Candollea 1: 66-415. 1989] BORNSTEIN, PIPERACEAE oe) Eames, A. J. 1961. Morphology of the angiosperms. McGraw-Hill, New York. ae G. 1966. Pollen morphology and plant taxonomy. Angiosperms. Hafner 1. Co., New Yor eae e 1977. Anatomy os seed plants. ed. 2. John Wiley & Sons, Inc., New York. Hickey, L.J. 1979. Are classification of the architecture of dicotyledonous aes Pp. 25-39 inc. R. SS & L. CHALK, Anatomy of the dicotyledons. ed. 2. Vol. 1. Clarendon Press, Oxford. HotmareEN, P. K., W. Keuken, & E. K. SCHOFIELD. 1981. Index herbariorum. ed. 7. ) Howarp, R. A. 1973. Notes on the Piperaceae of the Lesser Antilles. J. Arnold Arbor. 54: 377 Jounson, D. S. 1902. On the development of certain Piperaceae. Bot. Gaz. (Craw- fordsville) 34: 321-340. Kuntu, K. S. 1839. Bemerkungen iiber die Familie der Piperaceen. Linnaea 13: 561- Lawrence, G. H. M. 1951. Taxonomy of vascular plants. Macmillan, New York. LUNDELL, C. L. 1937. The vegetation of Petén. Publ. Carnegie Inst. Wash. 478: 1-244. MiqueEL, F. A. G. 1843-1844. Systema Piperacearum. Kramer, Rotterdam Murty, Y. S. 1959. Studies in the order Piperales V. Proc. Indian Acad. Sci. 49: ee . 1828. Piperaceae. Pp. 150-164 in C. B. Presi, ed., Reliquiae Haenkeanae. 1. J. G. Calve, Prague. ee ee B. ae a botanicae. A. Haase, Prague. RapFoRD, A. E., W. C. Dickison, J. R. Massey, & C. R. Bett. 1974. Vascular plant systematics. Harper and ee New York. RENDLE, A. B. 1925. The classification of flowering plants. ed. 1. Vol. 2. Cambridge University ee Cambridge, England. RzepowskI, J., & R. McVAuGH. 1966. La vegetacién de Nueva Galicia. Contr. Univ. Michigan Herb. 9: 1-123. SmituH, A. C. 1975. The genus Macropiper (Piperaceae). J. Linn. Soc. Bot. 71: 1-38. ls. STANDLEY, P. C., & J. A. STEYERMARK. 1952. Piperaceae. /n: Flora of Guatemala. Fieldiana, Bot. 24(3): 228-337. STEYERMARK, J. A. 1971. Notes on the genus Sarcorhachis Trel. (Piperaceae). Pittieria 3: 29-38. 1984. Piperaceae. Jn: Flora of as Vol. 2, pt. 2. Instituto Nacional de Parques, Ediciones Fundacion, Cara TRELEASE, W. 1921. North American ee of the section Ottonia. Amer. J. Bot. 8: —. 1930. The geography of American peppers. Proc. Amer. Philos. Soc. 69: 309- 327. —. 1934. New amplifications of the North American Piperaceae. [bid. 73: 327- 330. 1935. The pedicellate peppers of South America. /bid. 75: 691-716. Tucker, S.C. 1982. d flower development in the Piperaceae III. Floral ontogeny of Piper. Amer. J. Bot. 69: 1389-1401. WALKER, J. W. 1976. Comparative pollen morphology and phylogeny of the ranalian complex. Pp. 241-299 in C. B. Beck, ed., Origin and early evolution of angiosperms. Columbia Univ. Press, New York. & J. A. DoyvLe. 1975. The bases of angiosperm phylogeny: palynology. Ann. Missouri Bot. Gard. 62: 664-723. YUNCKER, T. G. 1950. Piperaceae. Jn: R. A. ae & R. W. Scuery, eds., Flora of Panama. Ann. Missouri Bot. Gard. 37: 20. 1958. The Piperaceae—a family is net 10: 1-7. 54 JOURNAL OF THE ARNOLD ARBORETUM 1972. 366. ee 1-167. [VOL. 70 The Piperaceae of Brazil—I: Piper, groups I, I, II], 1V. Hoehnea 2: 19- The Piperaceae of Brazil—II: Piper, group V; Ottonia; Pothomorphe; Sinorhachis Ibid. 3: 29-284. figs. 168-292. Ae ae 293-459 The oo of Brazil—III: Peperomia; taxa of uncertain status. /bid. APPENDIX. Index to collectors.* Aguilar 25/ (14) m . Hernandez 40847 (14) Bartlett 12326, 12563, 12683 (10) Boege 3249 (10) Bornstein 75 (3); 76 (11); 79, 86 (13); 8 (3); 90, 97, 93 (1); 96 (11); 97 (12); sf (13); 99 (11); 700 (12) Bornstein & Chase 68 (14); 69 (10); 70, 71 (14); 94 (1); 95 (8) Bornstein, Chase, & C. Bornstein 59 (6) Bornstein & Hensold 2 2,5 (11); 6, 7 (12); 8 (11); 9 (12); 15, 16, 17, 18, 19, 20, 2 1) Bornstein & Rutledge 23 (14); 27, 28, 30, 31, 32, 33 (10); 34, 37 (14); 38 (10); 45 46, 47, 74 (6) Bornstein & Weber-Bornstein 80, 8/ (3); 83 (13 Bottert 192 (= 1156) (9) Breedlove 25520 (6), 33953, 42858 (14) Breedlove & A. C. Smith 32576 (6) Breedlove & Thorne 20619 (6) Chan & Ucan 994 (14) Chase 83112 (8); 83113, 83127, 83129, 83130 (12) Chavelas ES-932, ES-1041, ES-1553, ES- 0) 7141 (5) Cochrane et ‘al. 10877 Contreras 405 (14); 9/76, 1033, 1048, 1050, 1126 (10); 1200, 1216, 1244 (14); 1246, 1354 (10); 2557, 3412, 3519, 3538 (14), 5518 (10); 5562, 5566 (14); 5763, 580] : 23440 (10); 23629, 23793, aa 24680, 40640 (10): 41539 (14 Delgado 348 & R. Hernandez 2599 (11) Delgado, R. Hernandez, & Trejo 204 (2) Diguet s.n. (12) Dwyer 10766 (10); 10802, 10906A (14); 10940, 11240 (10); 11572 (14); 12480, 12482 Dwyer & Liesner 12027 (14) Dwyer & Pippin /02// (10) Egler 42-234 (10) Ferris 6/86 (2) Gaumer 23678 (10), 23699 (14) Gentle 895, 1246, 2511, 3102, 5040, 5435, 5455, 6181, 7530, 7679 (14) — 7830, 8171 (14); 8297 (10); 8327 ae sn. (13) Hansen, Hansen, & Nee 1442 (13) Harmon & Dwyer 2764 (10) Hinton 4287 (11); 10261, 10266 (5); 11372 (13); 13836, 13839 (8); 13855 (1) Hubbell s.n., 1956 (10) Iltis & Nee 1416 (8), oe (13), 7650 (1) Ilts et al. 296 (13), 2559 (8) Jones, G. C., & i any (14) Salles 823 (10 ach & Mejorada 2266 (3) , & Soderstrom 4929 (1) Koelz Ey. 223 (4), 34228 (2) Langlassé ar Ke 215bis (1), Lewton 333 Liebmann : a (10) Liesner & Dwyer 15/4 (10) Linden s.n. (14) Lott 719 (3) Lott & Magallanes 6// (3) Lott et al. 1349 (12) Lundell 3/2 (14); 834 (10); 7300 (14); 1457 (10); 1487 (14); 3500 (10); 3657 (14); 3828 ee 6289, 6499 (14): 6510, 7406, 7467, | , 15925, a eo en H., 207 (5) Magallanes 754 (13 Matuda /5/5 (9), 3382 (14) McVaugh 14983 (3); 15436 (12); 15498, 15564 (1); 15594 (13); 156917 (11); 15809 (1); 75838 (4); 16107 (8); 20817 (7) 248 (13) — — *The designations in parentheses refer to the corresponding taxa in the text. 1989] McVaugh & Koelz /623 (3) Mexia 1177 (2) Miranda 6245, 6252 (10); 8248 (14) Molina 3432, 7009, 7185 (14); 15482 (10); 15760, 15777, 30605 (14 Moran /3310 (12) Mueller, F., /80 (9) Murray & D. Johnson /433 (13) Novelo et al. 778 (10) Ortiz 169 (10); 337, 403 (14); 427 (10); 469 (14); 1715, 1978 (10); 2048 (14); 2109, 2609, 2696 (10 Palmer, E., /00 (1), 1332 (2) see 418 (14 érez & Sarukhan /354 (1) ee 3458 (14) Proctor 30064 (14) ose, J. N., 1657, 3234 (12) Ross, G. B., /28 (10) Rowell & E. H. Cooper 2507 (5) Rowell & Irby 3523 (5) Rowell & Kubicek 3666 (5) BORNSTEIN, PIPERACEAE Sampson /2 (14) Schipp 506 (14) Seler & Seler 5172 (9) Sessé & Mocifio 3/8, 328, 329, 339, 348 (5 ) Shank & Molina 4637 (14) Shepherd 25 (14); 68, 68a (10) Standley 23952, 24675 (14); 79461 (6) Steere 1308, 2359 (10) Stevens, W. D., & Fairhurst 1868 (12) Stevermarke 45581, 4606] (10 Tate, R., 367 (14) Téllez & Cabrera 2534 (14); 2606, 2838 (10) Thieme 5455 (14) Turner, B. L., ee ae (13) Wendt et al. 2633 Wilbur /589, pas ne 1625 (3); 1675 CD) 2705-Ch)p 2220 (02) Williams, R. O., & Molina 1/423 (14) Wunderlin et al. 382 (10 Yong et al. 85 (14) AL-SHEHBAZ & SCHUBERT, DIOSCOREACEAE a7 THE DIOSCOREACEAE IN THE SOUTHEASTERN UNITED STATES! IHSAN A. AL-SHEHBAZ AND BERNICE G. SCHUBERT? DioscorEAcEAE R. Brown, Prodr. 1: 294. 1810, ‘“‘Dioscoreae,”” nom. cons. (YAM FamILy) Twining [rarely erect] herbs [lianas or subshrubs] with rhizomes or fleshy [rarely woody or corky] tubers, the tubers derived from the hypocotyl, the internode above it, or both; plants with raphides in mucilaginous idioblasts, frequently rich in steroidal sapogenins, and usually accumulating chelidonic acid and lactone alkaloids. Stems smooth, winged [or spiny]; vascular bundles closed, arranged in 2 [or 1] ring(s), the ee restricted to the roots, stems, and petioles, with scalariform perforation plates; sieve-tube plastids with cu- neate, proteinaceous inclusions. Leaves alternate, rarely opposite or whorled, long petiolate, simple [rarely palmately compound, with 3-7 leaflets], usually cordate, entire, undivided [or palmately lobed], often with embedded muci- laginous pits or nectaries, the tips usually with a distinct pore; venation palmate, with 3-13 converging main veins and anastomosing lateral veinlets; stomata anomocytic, rarely different; trichomes unicellular, eglandular, simple [some- times peglike, furcate (T-shaped), or stellate], often confined to the abaxial surface and along the veins, rarely occurring on the petiole or stem. Inflores- 'Prepared for the Generic Fl fthe Soutt United States, a long-term project made possible by grants from the National Science Foundation and currently supported by BSR-8415769 (C. E. Wood, Jr., principal investigator), under which this research was done, and nae aa sear (N. G. covered by the Generic Flora includes North and South Carolina, Georgia, Florida, Tennessee, Alabama, Mississippi, Arkansas, and Louisiana. The descriptions are based primarily on the plants of this area, with information about extraregional members of a family or genus in brackets. The references that we have not verified are marked with asterisks. We are most grateful to Carroll Wood for his support and continuous help during the preparation of this paper, and particularly for his critical review of the manuscript. We should also like to thank Norton Miller for reviewing the paper, Rogers McVaugh for help in lectotypifying Dioscorea, Charles Jarvis for notes and photographs of the type of D. villosa, and Barbara Nimblett for typing the manuscript. We are also grateful to Elizabeth B. Schmidt and Stephen A. Spongberg for their editorial advice. Ficure 1, band k, were drawn by the late Dorothy H. Marsh (DHM), e and m-o by Ihsan Al-Shehbaz (IAS), and the remainder by Karen Stoutsenberger (KS) under earlier grants. Carroll Wood and Kenneth R. Robertson prepared the material and supervised the illustrations. Preserved material, as well as herbarium specimens in the Arnold Arboretum and the Gray Herbarium, was used as the basis for the drawings. ?Arnold Arboretum, Harvard University, 22 Divinity Avenue, Cambridge, Massachusetts 02138. © President and Fellows of Harvard College, 1989. Journal of the Arnold Arboretum 70: 57-95. January, 1989. 58 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 cences axillary racemes, spikes, panicles, or cymes, many flowered, usually bracteate. Flowers trimerous, actinomorphic [very rarely zygomorphic], 1m- perfect (the plant dioecious [rarely monoecious]) [or perfect], subtended by | small [to large] bract and | [rarely 2] bracteoles. Tepals 6, in 2 whorls, distinct for connate], the perianth rotate [tubular to campanulate]. Stamens (absent in the carpellate flowers but often represented by staminodia) 6, in 2 whorls [the inner 3 sometimes staminodial or obsolete]; filaments distinct [connate at base, or fused into a tube]; anthers 2-locular and the lobes contiguous or separated, tetrasporangiate, dehiscing by longitudinal slits, introrse [or extrorse]; micro- sporogenesis simultaneous; pollen grains binucleate, 2- to 4- [rarely 1-]sulcate for 4- or 5-foraminate]. Gynoecium (rudimentary in the staminate flowers) 3-carpellate; ovary inferior, 3-locular, the placentation axile, septal nectaries usually present; styles 3, connate at base [or distinct]; ovules 2 [to many] in each locule, anatropous, bitegmic, crassinucellate, endosperm development nuclear, the megagametophyte (embryo sac) of the Polygonum type. Fruits triangular, 3-winged capsules [l-seeded samaras, or berries]. Seeds flattened [or globose], winged [or wingless]; embryos small, well differentiated, with a subterminal plumule and a broad, flat, nearly lateral cotyledon embedded in the hard, copious endosperm that contains lipids and aleurone, as well as hemicellulose deposited in its thick cell walls. Base chromosome numbers 9, 10, 12. (Including Androsynaceae Salisb., Cladophyllaceae Dulac, Stenomer- idaceae J. G. Agardh, Tamaceae Gray.) Tyre GENus: Dioscorea L. A family of seven genera and approximately 900 species of tropical, sub- tropical, or rarely warm-temperate plants. Averra Perr. (monotypic) is endemic to Madagascar, while Borderea Miégeville (two species; Heywood) is restricted to the Pyrenees. See, however, Miége (1986; reference under Dioscorea) for assignment of D. Gillettii Milne-Redhead (Kenya, Ethiopia) to Dioscorea sect. BorDEREA (Miégeville) Bentham. Epipetrum Phil. (three species; Reiche) is confined to Chile, Rajania L. (25 species) is native to the West Indies, Steno- meris Planchon (two species) is indigenous to Malaysia (Borneo, Malay Pen- insula, Philippines, Sumatra), and Tamus L. (five species) 1s widespread in the Mediterranean area (southwestern Asia; southern, central, and western Europe; northwestern Africa) and Macaronesia. Dioscorea, the largest genus of the family, has approximately 850 species distributed on all continents except Antarctica. The limits of the Dioscoreaceae are controversial, and Knuth (1930) and Burkill (1960) recognized ten and six genera, respectively. Both authors and Cronquist have retained 7richopus Gaertner (monotypic; Sri Lanka, southern India, Malaysia) in the family, while Ayensu (1966, 1972), Hutchinson (1973), Dahlgren (1980, 1983), and Dahlgren, Clifford, & Yeo placed it in a monotypic family, the Trichopodaceae Hutchinson, assigned to the Dioscoreales, a view with which we agree. 7richopus differs from genera of the Dioscoreaceae in having an erect (nontwining) herbaceous stem with one or a few flowers borne opposite the leaf, glandular hairs, successive microsporogenesis, and long-ap- pendiculate anthers. It also differs in nodal anatomy and chromosome number (see below). The monotypic Peftermannia F. Mueller (New South Wales and Queensland, Australia), which was retained in the Dioscoreaceae by Knuth 1989] AL-SHEHBAZ & SCHUBERT, DIOSCOREACEAE 29 (1924, 1930), is now believed to represent a unigeneric family, the Peterman- niaceae Hutchinson, somewhat related to the Dioscoreaceae (Conran; Dahl- gren, Clifford, & Yeo; Tomlinson & Ayensu). However, Hutchinson (1973) assigned this family to the Alstroemeriales, while Cronquist placed Peterman- nia in the Smilacaceae Vent. We prefer to associate the Petermanniaceae with the Dioscoreales for reasons discussed by Conran. Burkill (1960) reduced both Borderea and Epipetrum to sections of Dioscorea. The presence in Borderea of wingless seeds, nontwining stems, unbranched leaf veins, and a base chromosome number of 12 supports its recognition as a distinct genus. Furthermore, in Epipetrum the lack of wings around the seeds, the presence of well-developed stylar rudiments in the staminate flowers, and the spiral twisting of pedicels of the carpellate flowers are sufficient grounds for its maintenance at the generic rank. On the basis of their perfect flowers, prominent staminal appendages, and thin, nonfleshy rhizomes, both Avetra and Stenomeris were considered to be anomalous in the Dioscoreaceae. Avetra was placed in the Trichopodaceae by Hutchinson (1973), but the climbing habit, three-winged fruits, and nodal anatomy (Ayensu, 1972) strongly support its retention in the Dioscoreaceae, as was initially recognized by Perrier de la Bathie. Stenomeris also deviates from the rest of the Dioscoreaceae in having linear fruits with numerous seeds. It was placed in the Stenomeridaceae by many authors, but the overwhelming anatomical evidence favors its placement in the Dioscoreaceae. Dahlgren, Clif- ford, & Yeo assigned Avetra and Stenomeris to subfam. Stenomeridoideae and suggested that they should perhaps be treated in two unigeneric subfamilies or even in an independent family. Knuth’s (1924, 1930) division of the Diosco- reaceae into the tribes Dioscoreae and Stenomerideae Planchon may be useful only after the exclusion of Petermannia and Trichopus from the latter tribe. Avetra may represent a monotypic tribe. However, reorganizing the subfamilial classification of the Dioscoreaceae is beyond the scope of this flora. The remaining genera of the Dioscoreaceae are easily distinguished by their fruits. Rajania has one-seeded samaras, Tamus has berries, and Dioscorea has three-winged, few-seeded capsules like those of Borderea and Epipetrum. Hi- ginbothamia Uline, a monotypic genus described from the Yucatan, was said to differ from Dioscorea in having four instead of two seeds per locule. It is now recognized as a section of Dioscorea with two species, one of which is endemic to Belize (see Schubert (1966) under references for Dioscorea). The evolutionary relationships among the genera of the Dioscoreaceae have not been fully resolved. Because of their perfect flowers and long-appendiculate stamens, Avetra and Stenomeris are probably basal (Burkill, 1960). The re- maining genera of the family have imperfect flowers (the plants dioecious), and with the exception of Dioscorea they all have a restricted distribution. Ac- cording to B. W. Smith (see under Dioscorea), these small genera have arisen from Dioscorea. However, this view implies that Dioscorea is paraphyletic. The development of one-winged fruits in D. cyphocarpa Robinson ex Knuth and D. tacanensis Lundell may suggest a closer relationship between Rajania and Dioscorea. It is possible, however, that samaroid fruits evolved independently in the two genera. 60 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 The ordinal disposition and familial relationships of the Dioscoreaceae are controversial. A few authors (e.g., El-Gazzar & Hamza) advocated that the family be placed in the dicotyledons near the Aristolochiaceae Juss. because of its twining stems, petiolate, exstipulate leaves, trimerous, epigynous flowers, anatropous ovules, loculicidal capsules, endospermous seeds, and binucleate pollen. Furthermore, the presence in both families of PIIC-type sieve-tube plastids, which is universal in the monocotyledons, may be considered as evidence supporting this view. However, because of the presence of other plastid types in the family, Dahlgren & Rasmussen believe that the occurrence of these plastids in only two genera of the Aristolochiaceae is by convergence. Similarities between the Dioscoreales and the Magnoliales were summarized by Dahlgren & Clifford, who suggested (p. 342) that the extant forms of these orders do not represent “missing links or relicts of the common ancestors of the monocotyledons.” awton & Lawton, who studied the seedlings of five species of Dioscorea, suggested that the first seedling leaf is homologous to a modified second cot- yledon, whereas the first cotyledon remains in the seed as an absorbing struc- ture. However, P. M. Smith (reference under Dioscorea) did not find any structure comparable to a second cotyledon in the Dioscoreaceae. The slightly lateral position of the cotyledon in the family is considered to be ancestral among the monocotyledons, which otherwise have terminal cotyledons (Dahl- gren, Clifford, & Yeo) The close relationship between the Dioscoreaceae and the Smilacaceae was pointed out as early as 1810 by Brown. Most recent authors (e.g., Cronquist; Dahlgren, Clifford, & Yeo; Takhtajan) accept such a relationship. They and Thorne also associate the Dioscoreaceae closely with the Stemonaceae Engler (see Rogers), the Taccaceae Dumort., and the Trichopodaceae. However, these authors have variously placed the Dioscoreaceae under the Dioscoreales, the Liliales, or the Smilacales. The Dioscoreales (sensu Dahlgren, Clifford, & Yeo) are somewhat homogeneous and are apparently monophyletic. On the other hand, raising the Trilliaceae Lindley, the Taccaceae, and the Smilacaceae- Petermanniaceae to three orders, as suggested by J. W. Walker (pers. comm.), may be more appropriate. Evidently, the placement of the Dioscoreaceae with 14 other families in the order Liliales (Cronquist) needs further stu monocotyledons (Dahlgren, Clifford, & Yeo). However, since various families of the Dioscoreales have several specialized features, particularly tetramery, epigyny, complex nodal anatomy, and fleshy fruits, a critical evaluation of the order is needed. According to Coursey (1967; see under Dioscorea), the Dioscoreaceae likely evolved before the end of the Cretaceous, and Dioscorea was probably wide- spread before the separation of America from Africa. Pax listed Majanthe- mophyllum petiolatum O. Weber from the Miocene of Germany and Dioscorites resurgens Saporta from the Tertiary of southern France. Daghlian suggested that fossils described as Dioscorites Saporta from the Upper Cretaceous (Maas- trichtian), as well as from many other Tertiary sediments, are disputable as to 1989] AL-SHEHBAZ & SCHUBERT, DIOSCOREACEAE 61 familial and generic identity. Pollen of Rajania described from the Miocene of Veracruz (Graham), if indeed dioscoreaceous, may well belong to Dioscorea. The identity of the North American fossil D. cretacea Lesq. is questionable. Pollen of the Dioscoreaceae is heterogeneous. 7amus has reticulate, 2-colpate pollen (Clarke & Jones). In Dioscorea pollen is monocolpate in the presumably primitive sect. STENOPHORA Uline and 2-colpate in four other sections (Su; see under Dioscorea). Monosulcate, multiaperturate pollen has been found in D. polygonoides Humb. & Bonpl. (Zavada). Avetra deviates strongly from the rest of the Dioscoreaceae in having spinulose, 4- or 5-foraminate pollen. Cytology of the Dioscoreaceae has not been surveyed adequately, and chro- mosome numbers are known for only about nine percent of the entire family. Petermannia (x = 5) and Trichopus (x = 7), which are now placed in monotypic families, are evidently different not only in base number but also in chromo- some size (Ramachandran, 1962, 1968). Diploid counts of 2” = 24 in Borderea support its separation from Dioscorea, a genus uniformly based on either nine or ten. The closely related Rajania is apparently a polyploid based on nine, but only R. cordata L. (2n = 36) has been studied cytologically. Polyploidy based on x = 12 probably played an important role in the evolution of Tamus, as 1s evidenced by its occurrence in the tetraploid 7. communis L. (2n = 48) and the octoploid 7. edulis Lowe (2n = 96). Goldblatt (1980) suggested that the base chromosome number for the Dios- coreaceae may be seven and that such a number is found in the relict Trichopus, as well as in the families Stemonaceae and Taccaceae. As indicated above, however, Trichopus does not belong to the Dioscoreaceae, and it is more appropriate to assume that the base number for the family is ten. No chro- mosome counts are known for Avetra or Stenomeris. The Dioscoreaceae are characterized by three anatomical peculiarities that are unusual among the monocotyledons. First, xylem and phloem glomeruli are present at the nodes. These represent an interlacing mass of prosenchy- matous xylem that encloses one to many phloem glomeruli, each with inter- lacing sieve tubes and sieve plates of various sizes (see Behnke (1965) for further details). Karnick (1970; see under Dioscorea) believed that this ‘nodal plexus” is the site of various chemical activities, including the biosynthesis of diverse constituents. Second, vascular bundles of the stem are often arranged in two circles. Those of the outer circle have a V-shaped group of metaxylem vessels and tracheids, with two phloem units terminating the flanges and a third one at the converging ends of the V. Bundles of the inner circle have an elliptic arrangement of metaxylem vessels and tracheids, with one or two large phloem units on the inner side of the innermost pair of the larger metaxylem vessels and at least one at the outer end (Ayensu, 1972). Third, the sieve tubes are rather large and have highly oblique, compound sieve plates and numerous sieve areas. Because of their size, sieve elements could easily be mistaken for vessels. The presence of these three anatomical features in Avetra and Stenomeris supports their placement with Dioscorea, Rajania, and Tamus in the Dios- coreaceae, instead of in two other families (Ayensu, 1972). Furthermore, 77i- 62 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 chopus and Petermannia have nodal and vascular-bundle anatomy quite dif- ferent from that of the Dioscoreaceae and are therefore appropriately assigned to monotypic families (Ayensu, 1966, 1972; Tomlinson & Ayensu). Vessels of the Dioscoreaceae have scalariform perforations; these are present in the roots, aerial stems, and petioles but absent in leaf blades, bulbils, rhi- zomes, and tubers. Although most fibers of the Dioscoreaceae have tapered ends, a large proportion have square ones (Ayensu, 1972). Stomata of the Dioscoreaceae are anomocytic and are not associated with morphologically differentiated subsidiary cells. Furthermore, the contiguous epidermal cells have no distinctive arrangement (Ayensu, 1972; Stebbins & Khush). However, Patel and others (e.g., Ekundayo; Ling et a/.; Purnima & Srivastava; G. L. Shah & Gopal; references under Dioscorea) have indicated that other types of stomata may also be found. The stomata are distributed primarily on the abaxial leaf surface, but in a few species (e.g., D. bulbifera L.) they are present on both surfaces. Dispersal in the Dioscoreaceae is primarily by wind. In Rajania and Dios- corea, which produce samaras and variously winged seeds, respectively, short- distance dispersal is usually accomplished by seeds or fruits gliding or whirling from carpellate plants that can climb to heights of more than 30 feet. The red berries of Tamus are dispersed in the Mediterranean region by birds (Burkill, 1937), whereas in species of Epipetrum the capsules become buried in the vicinity of the mother plant Numerous species of Dioscorea are important medicinal or food plants. They are discussed in some detail under the generic account. The remaining genera of the Dioscoreaceae have little or no economic value. REFERENCES: ALLARD, H. A. The direction of twist of the corolla in the bud, and twining of the stems in Conv ees: and ae Castanea 12: 88-94. 1947. ARBER, A. Monocotyledons. A morphological study. Frontisp. + xiv + 258 pp. Cam- bridge, Eneland. 1925. ee Tamus. | Ayensu, E. 8. Taxonomic status of 7richopus: anatomical evidence. Jour. Lin Bot. 59: 425-430. p/. 1966. [Anatomy supports the exclusion of Trichopus a the Dioscoreaceae and its placement in a monotypic family. Comparative vegetative anatomy of the Stemonaceae (Roxburghiaceae). Bot. Gaz. 129: 160-165. 1968. [Similarity of Sfemona to Dioscoreaceae. . Aspects of the complex nodal anatomy of the Dioscoreaceae. Jour. Arnold Arb. 50: 124-137. 1969. [Dioscorea, Tamus; xylem and phloem glomerules unique to the family.] 2 Dioscoreales. In: C. R. MeETCALFE, ed., Anatomy of the monocotyledons. Vol. xii + 182 pp. + /6 pls. London. 1972 BATE- nee E.C. The phenolic constituents of plants and their taxonomic significance. Il. Monocotyledons. Jour. Linn. Soc. Bot. 60: 325-356. 1968. [Dioscorea, ten species; Tamus communis. | BEHNKE, H.-D. Uber das Phloem der Dioscoreaceen unter besonderer aie eae ihrer Phloembecken. I. Lichtoptische Untersuchungen zur Struktur der Phloe becken und ihrer Einordnung in das Sprossleitsystem. (English summary.) Zeitschr. Pflanzenphysiol. 53: 97-125. 1965. [For part II, Electronenoptische Untersuchungen zur Feinstruktur des Phioembeckens, see Ibid. 214-244] 1989] AL-SHEHBAZ & SCHUBERT, DIOSCOREACEAE 63 Uber den Aufbau der Siebelement-Plastiden einiger Dioscoreaceen. (English summary.) Ibid. a 243-254. 1967. [Sieve-tube plastids; six species of Dioscorea and one of Tamu. . Siebelement- piden Phloem-Protein und Evolution der Bliitenpflanzen: II. BENTHAM, G. Dioscoreaceas, In: G. BENTHAM & J. D. Hooker, Gen. Pl. 3: 741-746. 1883. [Eight genera recognized. ] Berry, E. W. Miocene plants from Colombia, South America. Bull. Torrey Bot. Club 63: 53-66. 1936. [Dioscorea, 55.] BOLKHOVSKIKH, Z., V. GriF, T. MATVEJEVA, & O. ZAKHARYEVA. Chromosome numbers of flowering plants: A. A. Feporov, ed. (Russian and English prefaces.) 926 pp. Leningrad. 1969. [Dioscorea, Tamus, Trichopus, 258, 259; chromosome counts through 1964. Boyp, L. Monocotylous seedlings. Trans. Proc. Bot. Soc. Edinb. 31: 1-224. 1932. [Dios- coreaceae, 81-83. Braun, H. J. Die Leitbiindelbecken in den Nodien der Dioscoreaceae, mit besonderer Berticksichtigung eines neuartigen Typs assimilateleitender Zellen. Ber. Deutsch. Bot. Ges. 70: 305-322. 1957. [Dioscorea, Tamus.] BREWBAKER, J. L. The distribution and phyl penta ees of binucleate and trinucleate pollen grains in the angiosperms. Am. Jour. Bot. 54: 1069-1083. 1967. pean eee 1078; Dioscorea, Rajania, and Tric neon have binucleate pollen.] BRouWER, R. The arrangement of the vascular bundles in the nodes of Dioscoreaceae. Acta Bot. Neerl. o 66-73. 1953. [Vascular elements of adjacent internodes connected by nodal glomeruli. ] Brown, R. Prodromus florae Novae Hollandiae et Insulae Van-Diemen. Vol. 1. viii + 145-590 + 2 unnumbered pp. London. 1810. [Dioscoreaceae (as Dioscoreae), 294, 295.] BUCHERER, E. Beitrége zur Morphologie und Anatomie der Dioscoreaceen. Bibliot. Bot. 3(Heft 16): 1-34. pls. 1-5. 1889. [Dioscorea, Tamus.] BurKILL, I. H. The life-cycle of Tamus communis L. Jour. Bot. London 75: 1-12, 33- 43, 65-74. 1937. [Fruit maturity and dispersal, poisons, seed germination, growth, ee pollination; relationship of Tamus to Dioscorea.] wth and tensions between the nerves in the leaf-blade of Tamus communis en bid. 77: 325-333. 1939. . The distribution of raphides in the leaves of Tamus communis Linn. Ibid. 78: 17-19. 1940. . Biological flora of the British Isles. Tamus communis. Ecology 32: 121-129. 9 Plie s of the family Empididae and other insect-visitors of the flowers of Tamus communis. Proc. Linn. Soc. London 157: 99-102. 1946. Dioscoreaceae. in i Os G. G. J. van Steenis, ed., Fl. Malesiana 4: 293-335. 1951. [Detailed account; Dioscorea (59 species in seven Sections): Stenomeris, Trichopus.] ———. The organography and the evolution of Dioscoreaceae, the family of the yams. Jour. Linn. Soc. Bot. 56: 319-412. 1960. [Detailed theoretical study of the family, with emphasis on the Old World taxa; comparative morphology and anatomy of selected groups; distributions of certain sections; keys to the genera and to the Old World sections of ee eecuegaues among those sections; poor coverage of the New World taxa: 59 es.] CHEADLE, V. I. The eee a types of vessels in the various organs of the plant in the Monocotyledoneae. Am. Jour. Bot. 29: 441-450. 1942. [Dioscoreaceae, 448.] CLarRKE, G. C. S., & M. R. Jones. The Northwest European pollen flora, 23. Diosco- reaceae. Rev. Palaeobot. Palynol. 33: 45-50. 1981. [Tamus communis: pollen 2-colpate, reticulate. ConrAN, J. G. Embryology and possible relationships of Petermannia cirrosa (Peter- 64 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 manniaceae). Nordic Jour. Bot. 8: 13-17. 1988. [Family isolated in the Dioscoreales but probably near the border of the Asparagales Cronauist, A. An integrated system of classification of flowering plants. Frontisp. + xviii + 1262 pp. New York. 1981. [Dioscoreaceae (including Stenomeridaceae and Trichopodaceae) after Smilacaceae in the order Liliales, 1228-1231.] pases C. P. A review of the fossil record of monocotyledons. Bot. Rev. 47: 517- 1981. [Dioscoreaceae, 522; Dioscorea, 541, 546.] Sea R. M. A revised system of classification of the angiosperms. Bot. Jour. Linn. Soc. 80: 91-124. 1980. [Dioscoreales, L8: Dioscoreaceae aie Steno- meridaceae es Trichopodaceae), Taccaceae oomiaceae) and Trilliacea . General a of angiosperm evolution and macrosystematics. Nordic Jour. Bot. > 119-149. 1983. [Dioscoreales, 145.] _T. CurrorD. The monocotyledons: a comparative study. xiv + 378 pp. Bees and other cities. 1982. [Dioscoreales: limits, 27; delimitation, 303; com- parison with ha 340—342.] ; J&P. EO. The families of the monocotyledons. Structure, evolution, and taxonomy. xi + 520 pp. Berlin, Heidelberg, New York, and Tokyo. 1985. [Dioscoreales includes Stemonaceae, inca Taccaceae, Dioscoreaceae, Petermanniaceae, Trilliaceae, and Smilacaceae, | 10-128.] & F. N. Rasmussen. Monocotyledon evolution: character and aan estimation. Pp. 225-395 in M. K. Hecut, B. WALLACE, & G. T. PRANCE Evolutionary biology. Vol. 16. New York and London. 1983. [Cladistic vee Dioscoreales, 363, 364. ——_S. RosENDAL-JENSEN, & B. J. NIELSEN. A revised classification of the ae te with comments on correlation between chemical and other characters. Pp. 149-2 in D. A. Younc & D. S. SEIGLER, eds., Phytochemistry and angiosperm phylogeny. New York. 1981. [Dioscoreaceae: tropane eas 180; oxalate raphides, 182; chelidonic acid and steroidal saponins, 183, Davis, G. L. Systematic embryology of the ee viii + 528 pp. New York, London, and Sydney. 1966. [Dioscoreaceae, 107. Eames, A. J. Morphology of the angiosperms. Frontisp. + xiii + 518 pp. New York, Toronto, and London. 1961. [Dioscoreaceae with a vestigial second cotyledon, 343, 360.] EL-Gazzar, A., & M. K. HAMza. On the monocots-dicots distinction. Publ. Cairo Univ. Herb. 6: 15-28. 1975. [Dioscoreaceae associated with Aristolochiaceae, 24, 2 ERDTMAN, G. Pollen morphology and plant Mavroe es Angiosperms. Frontisp. + xu1 + 553 pp. New York. 1971. Wena oneetes 149, 150. Gaussen, H. Révision des Dioscorea (Borde ee Bull. Soc. Hist. Nat. Tou louse 100: 383-398. pl. 1965S. eee. Chouardii, B. pyrenaea, generic lias: morphology, cytology.] Gipss, R. D. Chemotaxonomy of flowering plants. 4 vols. 2372 pp. Montreal and London. 1974. [Dioscoreaceae, 3: 1927, 1928; tests for various compounds, number of species tested in each genus; numerous references on the family. Vol. 4 includes bibliography, index, and addendum. ] Go.pBLaTT, P. Polyploidy in angiosperms: monocotyledons. Pp. 219-239 in W. H. Lewis, ed., Polyploidy, biological relevance. New York and London. 1980. [Dios- coreaceae, 224, 234] _ed. Index to plant chromosome numbers 1975-1978. Monogr. Syst. Bot. 5. vil + 553 pp. nr [Dioscorea, 204.] _ed. Index to plant chromosome numbers 1979-1981. /bid. 8. vii + 427 pp. 1984. ee 160, 161. _ed. Index to plant chromosome numbers 1984-1985. /bid. 23: [ix] + 264 pp. 96.] 1988. [Dioscoreaceae (including Petermannia), Dioscorea, Tamus, 1989] AL-SHEHBAZ & SCHUBERT, DIOSCOREACEAE 65 GoncuHarov, N. F. Dioscoreaceae. Jn: V. L. KoMAROV, ed., Fl. URSS 4: 494-498. 1935 (in Russian); Fl. USSR 4: 377-380. 1968 (English translation by N. LANDAU). [Dios- corea, three species; Tamus GraHaM, A. Studies in Neotropical paleobotany. II. The Miocene communities of Venera Mexico. Ann. Missouri Bot. Gard. 63: 787-842. 1976. [Rajania, 801, 812, 814, 826.] GuepeEs, M. Sur la morphologie de la feuille de deux a Phyton Austria 12: 216-227. 1967. [Dioscorea bulbifera, Tamus commun HARTWELL, J. L. Plants used against cancer. A survey. Llovdia 32: 79-107. 1969. [Dios- coreaceae, 102, 103.] HeGnauer, R. Chemotaxonomie der Pflanzen. Band 2, Monocotyledoneae. 540 pp. Basel and Stuttgart. 1963. [Dioscoreaceae, 133-152, 476, 477; supplement, 7: 609- 618. 1986.] Hestot, H. Le nombre chromosomique des Dioscoréacées pyrénéennes et leur rattache- ment au genre Borderea Miégeville. Compt. Rend. Acad. Sci. Paris 237: 433, 434. 1953. [Chromosome numbers support the maintenance of Borderea as distinct from Dioscorea.] Heywoop, V. H. Dioscoreaceae. /n: G. TuTIN et al., eds., Fl. Europaea 5: 84, 85. 1980. ‘LBorderea, Dioscorea, Tam Hooker, J.D. Dioscoreaceae. FL. British India 6: 288-297. 1892. [Dioscorea, 25 species; Trichopus zeylanicus. | Howarb, R. A. Dioscoreaceae. Fl. Lesser Antilles 3: 504-512. 1979. [Dioscorea, Ra- Jania. HUuBER, H Die Samenmerkmale und Verwandtschaftsverhaltnisse der Liliifloren. Mitt. Bot. Staatssam. Miinchen 8: 219-538. 1969. [Dioscoreales includes Dioscoreaceae, Stenomeridaceae, and Trichopodaceae Hutcuinson, J. The families of flowering plants. II. Monocotyledons. xii + 243 pp. London. 1934. [Dioscoreales, 141-149; ed. 2, 652-660. 1959.] —. The families of flowering plants. ed. 3. xviii + 968. Oxford. 1973. [Dioscoreales includes Stenomeridaceae, Trichopodaceae, Roxburghiaceae, and Dioscoreaceae, 806-8 14. IRELAND, C. R., W. W. Scuwase, & D. G. Coursey. The occurrence of batatasins in the Dioscoreaceae. Phytochemistry 20: 1569-1571. 1981. [Distribution of four phe- nolic compounds in 13 species of Dioscorea and in one each of Rajania and Tamus. | Kate, N. N., & R. M. Par. The floral anatomy of Trichopus zeylanicus Gaertner. Proc. Indian Acad. Sci. B. 88(part 2): 63- we 1979. [Floral anatomy supports the exclusion of the genus ee the Dioscoreacea Kimura, Y. Systéme et phylogénie des s aeeemiedones Not. Syst. Paris 15: 137-159. 1956. Tate ore Tes, 139: includes Dioscoreaceae, Petermanniaceae, Stenomerida- ceae, REST aas order originated from Liliales.] ay Knutu, R. Dioscoreaceae americanae novae. Notizbl. Bot. Gart. Berlin 7: 185-222. 1917, rank. of numerous new taxa; Dioscorea, 123 species; Rajania, 17 species. ] . Dioscoreaceae. /n: A. ENGLER, ed., Pflanzenr. IV. 43(Heft 87). 387 pp. 1924. [Comprehensive treatment of all known species of nine ee oer two tribes; Dios- corea, 614 species in 60 sections and four subgenera; 69 figur ioscoreaceae. Jn: A. ENGLER & K. PRANTL, Nat. Pflanzenfam. ed. 2. 15a: 438-- 462. 1930. [Family divided into the tribes Dioscoreae (six genera: Borderea, Dios- corea, Epipetrum, Higinbothamia, Rajania, Tamus) and Stenomerideae (Avetra, Peterma nnia, Stenomeris, Trichopus), Dioscorea includes four subgenera and 60 1924 Kuntu, C. S. Dioscorineae. Enumeratio plantarum 5: 323-456. 1850. [Dioscorea, Hel- mia, Oncus, Rajania, Tamus, Testudinaria. KuprIANOVA, L. A. Pollen morphology and phylogeny of the monocotyledons. (In 66 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 on ) Acta Inst. Bot. Acad. Sci. URSS 1. Syst. 7: 163-262. 1948. [Dioscoreaceae, 234. Lionor. J.R.S. A note on callose distribution in the phloem of Dioscoreaceae. Zeitschr. Pflanzenphysiol. 55: 287-291. 1966. [Dioscorea alata, D. bulbifera.] J wTon. The morphology of the dormant embryo and young seedling of five species of Dioscorea from Nigeria. Proc. Linn. Soc. London 178: 153-159. 1967. [D. bulbifera, D. hirtiflora, D. odoratissima, D. praehensilis, and D. Preussti believed to have two cotyledons.] Metcuior, H. Dioscoreaceae. a 533, 534 in H. Meccuior, A. Engler’s Syllabus der Pflanzenfamilien. ed. 12. Vol. 2 Berlin. 1964. [Tribes piOseorede Gnchides Borderea, Dioscorea, E pipetrum, ele and 77 Stenomeris and Trichopus) recogniz Moore, R. J., ed. Index to aa chromosome numbers 1967-1971. Regnum Veg. 90: 1-539. 1973. [Dioscorea, 132.] ,ed. Index to ee chromosome numbers for 1973/1974. Ibid. 96: 1-257. 1977. ees 93,9 Mutter, J. Fossil a records of extant angiosperms. Bot. Rev. 47: 1-142. 1981. NETOLITZKY, F. Anatomie der Angiospermen-Samen. Handb. Pflanzenanat. 10. vi + 364 pp. 1926. [Dioscoreaceae, - 84. Orr, M. Y. On the secretory organs of Dioscoreaceae. Notes Bot. Gard. Edinb. 15: 133-146, p/. 275. 1926. [Extrafloral nectaries, types, distribution among species and sections: Borderea, Dioscorea, Stenomeris; D. — lacks these nectaries. PATEL, R. J. Epidermal structure and ontogeny of stomata in some Dioscoreaceae and Taccaceae. Flora 160: 562-572. p/. 10. 1971. (Nine types of stomata in seven species of Dioscorea.] Pax, F. Dioscoreaceae. Jn: A. ENGLER & K. PRANTL, Nat. Pflanzenfam. II. 5: 130- 137. 1888. [Tribe Dioscoreae includes Dee. es Rajania, Tamus, and Testudinaria; tribe Stenomerideae includes Oncus, Petermannia, Stenomeris, and Trichopus; fossil genera Dioscorites and Majanthemophyllum.] PERRIER DE LA BATHIE, H. Un nouveau genre de Dioscoréacées. Bull. Soc. Bot. France 71: 25-27. pl. 71. 1924. [Avetra.] RAMACHANDRAN, K, Studies on the cytology and sex determination of the Dioscoreaceae. Jour. Indian Bot. Soc. 41: 93-98. p/. 3. 1962. [Dioscorea, Trichopus, chromosome numbers, sex chromosomes in Dioscorea. . Cytological studies in Dioscoreaceae. Cytologia 33: 401-410. 1968. [Chromo- some numbers for 7richopus and ten species of Dioscorea.] Rao, A. N. Embryology of Trichopus zevlanicus Gaertn. Jour. Indian Bot. Soc. 34: 213- 1955. ReicHe, K. Zur Kenntnis der Dioscoreaceen-Gattung Epipetrum Phil. Bot. Jahrb. 42: 178-190. 1908. [Three species; morphology, ecology, eae , key.] ReNDLE, A. B. The classification of flowering plants. ed. 2. . Gymnosperms and monocotyledons. xvi + 412 pp. Cambridge, England. ice face 309- 313. Rocers, G. K. The Stemonaceae in the southeastern United States. Jour. Arnold Arb. 63: 327-336. 1982. [Relationship to the Dioscoreaceae, 329.] SCHLITTLER, J. Die systematische Stellung der Gattung Petermannia F. v. Muell. und ihre phylogenetischen Beziehungen zu den Luzuriagoideae Engl. und den Diosco- reaceae Lindl. Vierteljahrsschr. Naturf. Ges. Ziirich 94: 1-28. 1959. [Petermannia considered a member of the Luzuriagoideae that approaches the Dioscoreaceae in w characters. ] STAUDERMANN, W. Die Haare der Monokotylen. Bot. Arch. 8: 105-184. 1924. [Dios- coreaceae, 142, 143, 182.] 1989] AL-SHEHBAZ & SCHUBERT, DIOSCOREACEAE 67 StrepBins, G. L. Flowering plants. Evolution above the species level. xviit + 399 pp. Cambridge, Massachusetts. 1974. [Dioscoreaceae in the Liliales, 355.] & G. S. Kuusu. Variation in the organization of the stomatal complex in the leaf epidermis of monocotyledons and its bearing on their phylogeny. Am. Jour. Bot. 48: 51-59. 1961. [Stomata of Dioscoteaceae lack subsidiary cells.] TAKHTAJAN, A. L. Outline of the classifi of flowering plants (Magnoliophyta). Bot. Rev. 46: 225-359. 1980. [Dioscoreaceae (including Stenomeridaceae and Tricho- podaceae), 312, 313. Tan, K. Dioscoreaceae. Jn: P. H. Davis, ed., Fl. Turkey 8: 552-554. 1984. [Tamus.] THORNE, i = cage classification of the Angiospermae. Pp. 35-106 in M. K. Hecut, W. C. STEERE, & B. WALLACE, eds., Evolutionary biology. Vol. 9. New York and ee 1976. Rpereacae in the order Liliales, oe Tomutnson, P. B. Development of the stomatal complex as a taxonomic character in the monocotyledons. Taxon oe 109-128. 1974. ioeeene 126.] & E. Notes on the vegetative peas ogy and anatomy of the Petermanniaceae cine iedones) Bot. Jour. Linn. Soc. 62: 17-26. pl. 1969. [Petermannia; anatomy supports its removal from a Dioscoreaceae and its closer association with the Smilacaceae. ] Uuing, E. B. Eine Monographie der Dioscoreaceen. Bot. ae a 126-165. 1898. [Two tribes (Dioscoreae includes Borderea, Dioscorea, Epi. mn, Rajania, and Tamus, while Stenomerideae comprises Oncus, Petermannia, eee ae Trichopus), Dioscorea divided into 51 sections; keys to genera and sections. WAGNER, P. Vessel types of the monocotyledons: a survey. Bot. Not. 130: 383-402. 1977. [Dioscoreaceae with scalariform vessels, 385, 397, 402.] WENDELBO, P. Dioscoreaceae. /7: K. H. RecHINGER, ed., Fl. Iranica 104: 1, 2. 1973. [Dioscorea, Tamus. ZAVADA, M. 8. Comparative morphology of monocot pollen and evolutionary trends of apertures and wall structures. Bot. Rev. 49: 331-379. 1983. [Dioscoreaceae, 363.] Dioscorea Linnaeus, Sp. Pl. 2: 1032. 1753; Gen. Pl. ed. 5. 456. 1754. Twining [rarely erect] herbs [or lianas]; glabrous or with unicellular, eglan- dular, simple [or variously branched] trichomes. Underground organs rhi- zomes, woody corms, or usually fleshy, single or clustered, edible [or poison- ous], smooth [or spiny] tubers with thin [or thick to corky] skin. Stems almost always produced annually from underground structures, smooth, winged [or armed], terete or angled, dextrorse or sinistrorse. Leaves alternate, opposite, or whorled, petiolate, simple [or palmately compound, with 3-7 leaflets], cor- date to nearly deltoid, undivided [or palmately lobed]; main veins 3-13, con- verging, the lateral ones reticulate, usually horizontal; petioles usually with a pulvinus at both ends, sometimes with a bulbil at the axil. Inflorescences axillary, many-flowered, bracteate racemes, panicles, spikes, or cymes. Flowers imperfect (plants dioecious or—abnormally— monoecious), actinomorphic, with rotate [funnelform or campanulate] perianth; tepals 6, in 2 similar [or sharply differentiated] whorls, distinct or variously connate; staminate flowers usually grouped in cymes borne on racemes, often with stylar rudiments, the stamens usually 6, in 2 whorls, all fertile [or the inner 3 either staminodial or absent], the filaments free [or connate], very short [or as long as the tepals]; carpellate flowers usually in spikes or spikelike racemes, epigynous, with 6 [3 or 0] stam- inodia; styles 3, branched. Fruits dehiscent, 3-winged, leathery or membran- 68 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 70 aceous capsules. Seeds flattened [or not], reticulate or smooth, broadly [to narrowly] winged all around [or only at | end, rarely wingless], 2 [or 4] per locule. Base chromosome numbers 9, 10. (Including Androsyne Salisb.; Bo- tryosicyos Hochst.; Elephantodon Salisb.; Hamatris Salisb.; Helmia Knuth; Higinbothamia Uline; Hyperocarpa (Uline) Barroso, Guimaraes, & Sucre: Mer- ione Salisb.; Nanarepenta Matuda; Polynome Salisb.; Rhizemys Raf.; Sismon- daea Delponte; Strophis Salisb.; Testudinaria Salisb.. Ubium casa Lectotype: D. bulbifera L.; see Green in Hitchcock & Green. (Name com- memorating the Greek Pedanios Dioscorides (fl. ca. A.p. 40-80), sete physician, and officer in the army of Nero. His De Materia Medica, which was published in five volumes during the period A.p. 50-70 (Riddle), was the leading text on pharmacology for 16 centuries and contained medicinal properties of more than 1000 drugs derived primarily from plants.)— YAM A genus of approximately 850 species, of which the majority grow in the humid tropical and subtropical areas of the world, and only several occur in the warmer parts of the temperate zones. Small’s estimation of 160 species for Dioscorea is well below the mark. As was suggested by B. W. Smith, nearly 50 percent of the species of Dioscorea are found in South America. There are approximately 130 species in Brazil, 120 in Central America and Mexico, 160 in Africa, and 250 in Asia (authors’ compilation). The genus is poorly repre- sented in Europe and adjacent Asia (three species), Australia (four), and the United States and Canada (three). Three of the six species growing in the southeastern United States are naturalized weeds of Asiatic origin. The sectional classification of Dioscorea is controversial. About 75 sections, including 60 in Knuth (1924), have been recognized by various authors. Knuth ‘The lectotypification of Dioscorea has not been fully resolved. Britton & Brown, who were first to eee pify the genus, chose D. sativa from the eight species described by Linnaeus (Sp. Pl. 2: 1032, 1033. 1753). Prain & Burkill (1919) clearly demonstrated that Linnaeus, in his citations ne D. sativa, ee elements now believed to belong to five species of Dioscorea, as well as one each of Cardiopteris Wall. ex Blume (Cardiopteridaceae Blume) and Tinospora Miers (Menispermaceae Juss.). Hooker's suggestion (p. 291) that “the plant figured in Hortus Cliffortianus must be accepte as [D.] sativa Linn.” was followed by Prain & Burkill (1919), who also sv that the name D. sativa must be sandone because the figure in Hortus Cliffortianus repres a plant with leaves from D. villosa L. and fruits from another species. The single specimen ae me in the Linnaean herbariun which was ae by Linnaeus as D. sai was collected by Kalm from eastern North peers and is generally agreed (Prain & Burkill, | to belong to D. villosa. As indicated by Jackson’s letter to Bartlett (see Bartlett, p. 8), “I t find any specimen aac by 1 herbarium, but as sativa is an | East inion: species, and the specimen is of Kalm’s collection, it is patent that there is a blunder.” If D. sativa is lectotypified on Kalm’s specimen above, we would be forced to replace the well-known North American D. villosa by D. sativa. This, however, would be a serious mistake because Linnaeus had neieates in ue Oneal apscripHon of - Sativa that the plant grows in the “Indiis” (East Indies; see | Dioscorea villosa is neither cultivated nor edible. Evidently D. sativa was based on materials inadequate and highly controversial with respect to citations, diagnosis, illustrations, and specimens. No matter how the species is typified, there would still be several unresolved problems. On the basis of i facts, D. sativa could be maintained as the lectotype of the genus only if it is satisfactorily typified. It is doubtful, however, that this will ever be seh nares In fact, the report of the Standing Committee on Stabilization of Species Names (Taxon 24: 171- 1975) listed D. ae as a nomen peaebieutn to be rejected under Article 69 of the ICBN. eee g D. bulbiferal yI of the genus, as was proposed by Green (in Hitchcock & Green), 1989] AL-SHEHBAZ & SCHUBERT, DIOSCOREACEAE 69 has often been criticized for his narrow sectional concept, particularly with respect to the New World taxa. Burkill (1960) assigned the Old World species to 23 sections (including Borderea) and placed particular emphasis in his keys on aspects of the underground parts. On the other hand, Knuth (1924, 1930, family references) divided Dioscorea into four subgenera based on the position of the seed wing, relying heavily on characters of the staminate flowers and fruits in his sectional divisions of subgenera. Some of the sectional realignment proposed by Ayensu (1972) on the basis of anatomy should be taken into consideration. In our opinion, the number of New World sections accepted by Knuth (1924) should be reduced. However, it is beyond the scope of this flora to do that. Burkill (1960) and Prain & Burkill (1939) suggested that Dioscorea evolved during the Cretaceous in what are now the temperate areas of the Far East, where the presumably primitive sect. STENOPHORA Uline is primarily distrib- uted, and that the genus later migrated to the tropics. Burkill (1960) believed that the formation of the Atlantic rift was responsible for the separation of the range of Dioscorea into two portions, each of which evolved independently because no sections or species are now common to the New and Old worlds. Section STENOPHORA is considered primitive on the basis of having primarily diploid taxa, monocolpate pollen, and rhizomatous underground organs (Bur- kill, 1960; Pei et a/.). The assignment of the eastern North American species of Dioscorea to sect. STENOPHORA by Burkill (1960), Coursey (1967), and Prain & Burkill (1936) is unacceptable because seeds of these species are broadly winged all around, while those of sect. STENOPHORA are winged above (distally). We follow Uline (1897) and Knuth (1924) in placing these species in sect. MAcRopopDA Uline but disagree with the latter author on the sectional limits. Uline admitted only three species in this section: D. villosa and two unnamed ones that were said to grow in the Caucasus and the Himalayas. Therefore, D. villosa has to be considered as the sectional type. Coursey (1967) has suggested that the eastern North American taxa evolved from ancestors that migrated from eastern Asia across the Bering Strait land bridge. However, there is no evidence at present that supports this speculation. Section MAcROPODA Uline (stems sinistrorse; staminate flowers grouped in verticillate to subcapitate cymes borne on racemes; fertile stamens 6, inserted at base of perianth, the filaments short; capsules large, obovate) is represented in the United States and Canada by three species that also grow in the Southeast. Dioscorea quaternata J. F. Gmelin (D. villosa L. subsp. quarternata (J. F. Gmelin) Knuth, D. paniculata Michx., D. villosa subsp. paniculata (Michx.) Knuth, D. paniculata var. glabrifolia Bartlett, D. villosa subsp. paniculata var. glabrifolia (Bartlett) Knuth, D. villosa var. glabrifolia (Bartlett) Blake, D. villosa f. glabrifolia (Bartlett) Fern., D. glauca Muhl., D. villosa subsp. glauca (Muhl.) Knuth, D. quaternata var. glauca (Muhl.) Fern., D. villosa var. glabra C. G. Lloyd, D. Lloydiana Krause), 2” = 36, 54, 60, 1s the most variable and widely distributed of the North American species. It is distributed primarily east of the 95th meridian and between 27 and 45° north latitude. It does not grow in New England and is very rare in New Jersey, New York, Wisconsin, Minnesota, 70 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 Iowa, Kansas, Oklahoma, and Texas, but 1s quite common in the remaining states that fall within its geographic range. It is sporadic in southern Ontario, Canada. Dioscorea quaternata grows in moist, rich or rocky woods and thickets, on limestone or talus slopes, in creek bottoms, and along roadsides, railroad tracks, and borders of swamps, ponds, and marshes. Dioscorea villosa L. (D. hirticaulis Bartlett, D. villosa subsp. hirticaulis (Bart- lett) Knuth, D. villosa var. hirticaulis (Bartlett) Ahles) is distributed from south- western Massachusetts (Bristol County), Rhode Island (Washington County), and western Connecticut (Fairfield, Hartford, Middlesex, New Haven, and New London counties) southward along the Coastal Plain of all states to northern Florida. We have not seen any material from Delaware, and the records from New York and New Jersey are based on old collections. Furthermore, D. villosa is disjunct and probably introduced to Kentucky; it is first recorded here (Beck- ett 651, GH; Clark County). The species is common in Virginia and the Car- olinas, where it grows in bogs, peaty depressions, and swamps. Linnaeus (Sp. Pl. 2: 1033. 1753) stated that it occurs in Florida, but this has not previously been confirmed. We are recording D. villosa from Alachua (Arnold s.n., 1931; FLAS), Jackson (Godfrey 76493; Fsu), and Leon (Lazor 922; Fsu) counties, Florida. Dioscorea villosa is easily recognized by its narrowly winged, polygonal (8- to 14-angled) stems (FIGURE 1m), alternate lower leaves, and densely grouped glomerules on short (usually 2-4 cm) staminate inflorescences. The closely related D. quaternata has terete, wingless stems, whorled lower leaves (three to nine per node), and lax, longer inflorescences. Field notes on the phyllotaxy of lower nodes are useful for separating the two species. Dioscorea quaternata is the earliest published binomial for the North American dioscoreas with whorled lower leaves. The identities of Dioscorea hirticaulis, D. quaternata, and D. villosa have been confused. A few authors (e.g., Ahles, 1968; Gleason; MacRoberts; E. B. Smith) suggested that the first and/or second be reduced to synonymy of the last. As shown above, however, D. quaternata and D. villosa are very distinct and should therefore be maintained at the specific rank. Dioscorea villosa was typified by Blake on Clayton 94 (Bm) from Virginia. Charles E. Jarvis (pers. comm.) has indicated that there are two Clayton sheets in the British Museum (photos, A!), one bearing staminate material and the other a carpellate plant. The staminate material, annotated as Clayton 94 and with a handwritten note by Gronovius that male and female flowers are borne on separate plants, is the type. It has narrowly winged, 8- to 14-angled, very sparsely pubescent stems and alternate lower leaves. Previously these features have been attributed to D. hirticaulis. Therefore, what has been called D. hirticaulis by all North American authors is in fact D. villosa. Accordingly, the range of D. villosa becomes restricted to only the coastal states above, and all records from the other states should belong to D. quaternata. Evidently, Deam’s record of D. villosa (as D. hirticaulis) from six counties in Indiana belongs to D. quaternata. Both Dioscorea villosa and D. quaternata are highly variable in stem and leaf pubescence. Forms with densely pubescent to completely glabrous leaves that are either glaucous or nonglaucous occur in populations of both species. 1989] AL-SHEHBAZ & SCHUBERT, DIOSCOREACEAE 71 Bartlett’s recognition of four species in this complex was based primarily on differences in the branching of rhizomes and in pubescence, color (green vs. glaucous), and number of leaves per node. These alleged distinctions are far from being realistic. In fact, characters of rhizome branching, thickness, and surface configuration can be modified by soil texture, habitat, and environment. Furthermore, petiole length, leaf pubescence and glaucescence, and fruit length show continuous variations of no taxonomic value. Dioscorea floridana Bartlett (D. villosa subsp. floridana (Bartlett) Knuth, D. villosa var. floridana (Bartlett) Ahles) is the most distinctive of the North American species. It is readily distinguished by its two or three staminate inflorescences in the axils of upper leaves, longer (ca. 0.4 mm), inwardly curved filaments, connate anther lobes (see FiGuRE In, 0), and oblong tepals. Both D. quaternata and D. villosa have single staminate inflorescences in the axils of upper leaves, shorter (ca. 0.2 mm), straight filaments, separated, didymous anther lobes (see FiGure le), and ovate tepals. Differences in the anthers of these species were first observed by Bartlett (see p. 18), but they have been overlooked by subsequent workers. Dioscorea floridana grows in moist thickets and swamps, as well as in moist to dry woods and hammocks in South Carolina (Berkeley, Charleston, Dor- chester, Orangeburg, and Williamsburg counties), Georgia (Baker, Chatham, Clarke, Dougherty, Jenkins, Lowndes, Sumter, and Tattnall counties), and Florida (Alachua, Calhoun, Columbia, Duval, Hernando, Hillsborough, Jack- son, Levy, and Santa Rosa counties). It has been mapped from Bamberg and Beaufort counties, South Carolina (Ahles, 1968), but we have not seen any material from either. No chromosome counts are available for either Dioscorea floridana or D. villosa. The few counts for D. guaternata are inconsistent, and more studies are needed before any meaningful conclusions can be reached. Dioscorea bulbifera L. (Helmia bulbifera (L.) Kunth; see Coursey (1967) for ten additional synonyms), aerial or potato yam, 2” = 36?, 40, 54?, 60, 70, 80, 100, is the only edible yam believed to be native to both Asia and Africa (Prain & Burkill, 1936). Introduced to the New World during the slave trade, it is a widespread weed in Florida, where it grows in disturbed woods and thickets (Wunderlin), but is uncommon in Mississippi. Edible and poisonous forms are known, but the Florida populations produce bulbils that remain inedible and nauseating even after repeated washing and boiling (Ward). Prain & Burkill (1934) recognized ten varieties primarily on the basis of the shape and taste of both bubils and tubers. It is doubtful, however, that these entities are sig- nificant taxonomically. Tubers of D. bulbifera, if produced, are very small, bitter, and usually hard. The bulbils, on the other hand, may weigh up to 2 kg. The species is easily distinguished from the other introduced dioscoreas that grow in the Southeastern States by its sinistrorse, terete, wingless stems; large, alternate, broadly cordate, acuminate leaves with rounded lobes; large, smooth, subspherical bulbils; and auriculate petiolar bases. The sectional disposition of Dioscorea bulbifera is problematic. If the species is accepted as the lectotype of the genus, as is done here, it should be assigned to sect. DioscorEA. Accordingly, sect. OPSOPHYTON Uline, to which D. bulbifera is assigned by nearly all students of the Dioscoreaceae, should be reduced to 12 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 URE |. Dioscorea. a-l, D. quaternata: a, rhizome, x 2—note scars left by annual aerial stems; b, portion of mature plant with fruits, x '2; c, flowers and buds from staminate plant, x 6; d, staminate flower from above, x 6; e, abaxial view of anther, x 40; f, lateral view of anther, x 40; g, tip of carpellate inflorescence with buds and 1 1989] AL-SHEHBAZ & SCHUBERT, DIOSCOREACEAE 73 a synonym of sect. DioscorEA. Knuth (1924) recognized 26 species and three subsections in sect. OpsoPpHYTON and placed D. bulbifera in subsect. Ewopso- phyton, which also included three other species believed to be of hybrid origin. Prain & Burkill (1936), on the other hand, defined the section to include only D. bulbifera and its ee hybrids with three members of sect. ENANTIOPHYLLUM Ulin The two remaining species of Dioscorea that are naturalized in the South- eastern States are D. alata L., white yam, greater yam, 2n = 20, 30, 40, 50, 60, 70, 80, and D. oppositifolia L. (D. batatas Dene.; D. opposita Thunb., see Taxon 24: 173. 1975), Chinese yam, cinnamon vine, 2n = 138, 140, 142, 144. They belong to sect. ENANTIOPHYLLUM Uline (tubers nonpoisonous, stems dex- trorse, leaves alternate or opposite, filaments as long as the anthers, seeds winged all around). This section, the largest in Dioscorea, includes about 120 species restricted to Africa and Asia, none of which is common to both con- tinents (Prain & Burkill, 1939). Dioscorea alata, a native of southeastern Asia not known to grow in the wild state (Coursey, 1967), is the most widely cultivated edible species in the genus. It has escaped from cultivation in several counties in Florida (Ward; Wunderlin) but is not naturalized in the other Southeastern States. Dioscorea alata is highly variable, and many cultivars with different ploidy levels are grown. It resembles D. bulbifera in the production of bulbils and in having auriculate petiolar bases but differs in having dextrorse, quadrangular, broadly winged stems. Dioscorea oppositifolia, which is naturalized in North and South Carolina, Georgia, Tennessee, Alabama, and Arkansas, is native to China and is widely cultivated in eastern Asia for its edible tubers. It is grown in the United States primarily as an ornamental (Bailey ef a/.) and has escaped from cultivation in many parts of the country. It is a weed of thickets, alluvial woods, waste places, flood plains, roadsides, limestone outcrops, slopes, and fence rows. According to Purseglove, D. oppositifolia (as D. opposita) was cultivated experimentally in Europe during the mid-nineteenth century as a substitute for potatoes (Sola- num tuberosum L.) that were threatened by blight. Dioscorea oppositifolia 1s most likely a 14-ploid based on ten, and the aneuploid counts above are probably approximations of 2” = 140. It is readily distinguished from the other bulbil-producing species by its terete, dextrorse, wingless stems, nonauriculate petiolar bases, and somewhat hastate, usually opposite leaves. Hooker’s statement (p. 288) that “the species of Dioscorea are in a state of indescribable confusion” may be justified because of the inadequacy of material for many tropical species. It is often difficult, if not impossible, to identify specimens that lack well-developed staminate flowers or mature fruits. Because of dioecism, there are many cases in which the carpellate and staminate plants of a given species were described as two species assigned to different sections. opened flower, x 6; h, carpellate flower from above, showing 3 styles and 6 staminodia, from above, x 10—note stylar rudiments and 6 inwardly curved filaments; 0, abaxial view of anther, x 40. 74 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 A case in point is D. hondurensis Knuth: carpellate material was described in sect. TRIANGULARES Knuth, while staminate material of the same species was described as D. tabascana Matuda in sect. MACROGYNODIUM Uline. According to Schubert (1966), D. hondurensis probably belongs to the South American sect. SARCANTHA Uline. The reproductive biology of Dioscorea has been poorly studied. Although all species are dioecious, monoecism is known in at least 17 species (Burkill, 1960; Hawley). It 1s not known, however, if the carpellate flowers of monoecious plants are fertile. The flowers of Dioscorea are generally small, and their di- ameter in the great majority of species is 2-4 mm. The smallest flowers (ca. | mm wide) are found in the Madagascan D. nako Perr. (Burkill, 1960), while the largest (tepals more than 3 cm long) belong to the Mexican D. insignis Morton & Schubert (see Schubert & Morton). Because of their reduced size and dull colors (usually dirty white, creamy, greenish, or brownish), the flowers of Dioscorea were believed to be wind pollinated. However, this is highly unlikely because the pollen is glutinous. Furthermore, Coursey (1967) suggested that the flowers are sweetly scented and are likely to be pollinated by night-flying insects. Sadik & Okereke observed thrips of the genus Larothrips carrying pollen from the staminate to the car- pellate flowers of D. rotundata Poiret. The flowers of D. composita Hemsley and D. floribunda Martens & Gal. are said to be inefficiently wind pollinated in Puerto Rico (Martin, Cabanillas, & Ortiz). Chromosome numbers have been reported for about 90 species (ca. 9.5 percent) of Dioscorea. Apparently all of the Old World species are either dip- loids or polyploids based on ten, while most of the New World ones are polyploids based on nine (Martin & Ortiz, 1963a). However, only about three and 17 percent, respectively, of the New and Old World species have been surveyed, and more counts are needed to confirm these cytological observa- tions. Aneuploidy has been well documented in Dioscorea alata, D. bulbifera, D. cayenensis Lam., D. dumetorum (Kunth) Pax, and D. oppositifolia. All of these species are among the principal food yams, and it is not unusual to encounter variations in chromosome numbers of crop plants that have been propagated vegetatively for hundreds or perhaps thousands of generations. A continuous ploidy series (2-8x) is known in D. alata (Bolkhovskikh et al.), and higher ploidy counts (10-14x) based on ten are known in at least 12 species. The lowest chromosome number (2 = 20) in the genus characterizes most species of sect. STENOPHORA, the most primitive group in Dioscorea (Chin et al.; Pei et al.), while the highest (2n = 140) has been reported for D. cavenensis, D. oppositifolia, and D. pentaphylla L. Onwueme’s conclusion (p. 3) that ‘the highest chromosome numbers and the smallest chromosome sizes occur in the more tropical Dioscorea spp., while the smallest numbers and largest sizes occur in the more temperate species”’ needs further study. It is evident, however, that polyploidy has played a major role in the evolution of Dioscorea (Ra- machandran, 1962). Except for a few aberrant cases, species of Dioscorea are always dioecious, and within a given population staminate plants usually outnumber the car- 1989] AL-SHEHBAZ & SCHUBERT, DIOSCOREACEAE he pellate ones by three to five times. The mechanism of sex determination in Dioscorea is controversial. Nakajima, Ramachandran (1962), and B. W. Smith have all suggested that the staminate plant is heterogametic and that the “male- determining” genes are borne on the Y chromosome. On the other hand, Bhat & Bindroo concluded that in D. de/toidea Wall. staminate plants have homo- morphic chromosomes and carpellate ones have heteromorphic. Furthermore, Jensen could not find any evidence of sex chromosomes in D. quaternata, and Martin & Ortiz (1963a) and V. R. Rao & Murty reached similar conclusions for four Central American species. The presence of an extra chromosome in D. alata and D. brachybotrya Poeppig (as D. reticulata C. Gay) was considered to be responsible for “male” expression in an XO sex-determining mechanism (B. W. Smith). Because of their high numbers and small sizes, the chromosomes of Dioscorea are very difficult to distinguish morphologically. As suggested by Martin (1966), however, the cytological evidence supports the staminate plant as heteroga- metic, and therefore an XO sex-determining mechanism is highly unlikely. It is possible that the staminate plants are produced by different mechanisms. Sex ratios of progeny from a carpellate parent vary according to the staminate plant, and such ratios are constant among the progeny of a staminate parent. On the basis of these facts, Martin (1966) concluded that in tetraploid taxa the staminate plants have XX YY or XXXY genotypes. Natural interspecific hybridization is very rare in Dioscorea. The origin of one such hybrid (D. composita x D. floribunda) has been supported by chem- ical, cytological, and morphological data (Martin & Cabanillas, 1963). Artificial crosses among some of the New World species (D. composita, D. floribunda, D. Friedrichsthalii Knuth, and D. spiculiflora Hemsley) are successful in all possible combinations (Martin & Cabanillas, 1966; V. R. Rao et al.; V. R. Rao & Murty). The reduction of hybrid fertility is caused by some meiotic abnor- malities (univalent and multivalent formations), high percentage of abnormal pollen, low crossability of hybrids, and poor seed set. Crosses between the Old and New World species, as well as between D. alata and D. deltoidea, have been unsuccessful (V. R. Rao et a/.). Alkaloid production in the tubers is apparently restricted to some of the Old World species of Dioscorea that belong to sects. LASIOPHYTON Uline, PAR- AMECOCARPA Prain & Burkill, and TestupinarIA (Salisb.) Prain & Burkill. The principal compounds are dioscorine and its stereo isomer dihydrodioscorine. These highly toxic alkaloids cause general paralysis of the central nervous system. The African D. Dregeana (Kunth) Durand & Schinz and D. dumetorum and the Asiatic D. hispida Dennst., which are rich in alkaloids, are used locally to poison animals (see below). No attempts have been made to determine the chemotaxonomic significance of these compounds in the genus. A few Asiatic species accumulate in their tubers relatively high concentrations of tannins that are economically important. Dioscorea cirrhosa Lour. (= D. rhipogonoides Oliver) contains between six and 13 percent tannins per dry weight and is widely used in Taiwan and Indochina for tanning leather and for fishing cordage and nets (Burkill, 1966). More than 20 steroidal sapogenins have been isolated from some 70 species 76 JOURNAL OF THE ARNOLD ARBORETUM [vVoL. 70 of Dioscorea. Diosgenin is found in all the saponin-containing species. Its content in the tropical dioscoreas is generally much higher than in those of other regions (Takeda). The distribution of sapogenins in Dioscorea supports certain sectional groupings and is therefore of some chemotaxonomic value (Akahori, 1965a; Kadkade et al.; Takeda). For example, sect. ENANTIOPHYLLUM, which contains the most important food crops of the genus, lacks the steroidal sapogenins. These compounds are also lacking in a few sections that produce bulbils and have sinistrorse stems. On the other hand, sections with alternate leaves, sinistrorse stems, and no bulbils are often rich in sapogenins. Takeda concluded that the chemical data tend to support Prain & Burkill’s (1936, 1939) sectional classification of the eastern Asiatic Dioscorea. However, the chemistry of the genus is far from being adequately studied, and the great majority of the New World sections have not been surveyed. Using polyacrylamide-gel electrophoresis of tuber proteins, Ikediobi & Ig- boanusi have found significant differences in the locations and intensities of certain bands in five species. They were also able to distinguish among nine cultivars of Dioscorea rotate Several authors (e.g., RaSper & Coursey) observed that differences in the viscosity, size, and shape of the tuber starch grains are useful in separating several edible species of Dioscorea. Only a few species, however, have been surveyed for tuber proteins and starch grains. Ayensu (1972, 1973) observed that stems of the Old World species have two distinct phloem units below the innermost pair of metaxylem vessels, whereas those of the New World species have only one. This anatomical differentiation between the two groups of Dioscorea, however, requires further study to assess its phylogenetic value. On the basis of stem anatomy, Ayensu (1972) concluded that the anatomical data are very useful taxonomically at the intersectional level but have little value within sections. He proposed several alterations to the sectional classi- fication of Knuth (1924), including the reduction of many sections to synonymy and the elevation of a few subsections to sectional rank. In general, stem anatomy supports the sectional subdivision of the genus as accepted by Prain & Burkill (1936, 1939) and Matuda (1953b) but not by Knuth (1924). Extrafloral nectaries of certain species (e.g., Dioscorea rotundata) are embed- ded in the leaf blade and open on the lower leaf surface. Each gland consists of small cells with dense cytoplasmic contents and is surrounded by a layer of cells with apparently little or no cytoplasm. The secretions of these nectaries contain fructose, glucose, and sucrose, as well as ninhydrin and traces of gal- actose (Grout & Williams). Leaf glands of the West African D. macroura Harms form enlarged pockets within the mesophyll and are connected to the upper leaf surface by slits or ducts. The pockets are lined with multicellular glandular “trichomes” that secrete mucopolysaccharide, on which nitrogen-fixing bac- teria grow (Orr). As shown by Behnke, however, these bacteria are parasitic, and in response to their presence the plant produces an abundance of simple trichomes that fill the pockets and cause the bacteria to disintegrate. In the African Dioscorea praehensilis Bentham and the Asiatic D. esculenta (Lour.) Burkill and D. piscatorum Prain & Burkill, roots produced on the tuber 1989] AL-SHEHBAZ & SCHUBERT, DIOSCOREACEAE 77 surface are modified into thorns or spines. Such modified roots apparently evolved independently several times within the genus. Tubers of Dioscorea originate primarily from the seedling hypocotyl follow- ing the initiation of a cambial zone around the vascular tissue. The intensive activity of the cambium produces the storage parenchyma. These tubers differ from stem tubers of other genera in lacking both scale leaves, buds, or eyes that identify nodal positions and a terminal bud that marks the growing point of the tuber. Furthermore, most tubers of Dioscorea exhibit positive geotro- pism. The corky outer portion of the tuber is derived from successive cork cambia each originating beneath another. The ground tissue, which is derived from the vascular cambium, forms the bulk of the tuber and consists of thick-walled parenchyma densely packed with starch grains. The vascular bundles are col- lateral and have xylem comprising tracheids and parenchyma but no vessels. Some species of Dioscorea develop enormous tubers: up to 365 kg in D. elephantipes (L’Hér.) Engler (Purseglove) and up to 110 kg and 3.5 m long in D. alata (Haynes & Coursey). Species of Dioscorea can be propagated by tubers, bulbils, vine cuttings, tissue cultures, or seeds. A leaf cutting with the complete pulvinus dipped in “hormone powder” is capable of producing new plants (Blunden, Hardman, & Trease). Bulbil production, which is reduced by the increase of day length (Allard), promotes dispersal, particularly in habitats that are subjected to floods or rain washes. Because of their occurrence in many species of unrelated sec- tions, bulbils probably evolved independently a few times within Dioscorea. Murty & Purnima, who studied 11 species, suggested that bulbils develop as composite structures from the diffused growth of accessory buds that arise on the abaxial side of the axillary bud. Ayensu (1972) and Burkill (1960) define the bulbil as a modified branch of the aerial stem. They stated that it resembles the tuber anatomically. Little is known about the ecology of Dioscorea. Although the majority of species are tropical, many rhizomatous taxa grow in temperate areas. In general, most species require more than | m of annual rainfall and temperatures higher than 25°C (Purseglove). The effect of photoperiodism has not been studied adequately. According to Purseglove, long days apparently favor the devel- opment of aerial parts, whereas short days promote the growth of tubers. Food crops of the genus Dioscorea, which are known as yams in most parts of the world but are often confused with the sweet potatoes (Ipomoea Batatas (L.) Lam., Convolvulaceae) in the United States, provide the staple foodstuff for millions of people in many subtropical and tropical countries, particularly in West Africa (Coursey, 1967). The dietary position of yams in the tropics, however, has declined substantially due to socioeconomic reasons, introduction of better crops, and higher costs and greater labor needed for the cultivation and preparation of yams for food. The so-called “‘yam zone,” which extends from the Cameroun Mountains west into central Ivory Coast, is the area where most of the world’s yams are grown. In fact, Nigeria alone produces nearly half of the world’s crop (Ayensu & Coursey; Coursey, 1967). Yams are eaten 78 JOURNAL OF THE ARNOLD ARBORETUM [vVoL. 70 baked, boiled, fried, mashed, pounded, or roasted. They are also used to prepare chips, flakes, and flour Although about 50 species of Dioscorea are collected from the wild or cul- tivated as food crops, only ten are considered to be of major importance. These are associated with three independent centers of domestication. In southern China and southeastern Asia D. alata, D. bulbifera, D. esculenta, D. japonica Thunb., D. oppositifolia, and D. pentaphylla were first domesticated, while in the Caribbean region D. trifida L. and in the West African forest belt D. cay- enensis, D. dumetorum, and D. rotundata were first brought into cultivation. It has been suggested that the domestication of Dioscorea probably took place ca. 10,000 s.p. in Asia and ca. 11,000 b.p. in Africa (Coursey, 1976a). The West African Dioscorea cayenensis and D. rotundata were described from plants cultivated in Jamaica and French Guiana, respectively. Both species were introduced to the New World during the slave trade (Ayensu & Coursey). There is some controversy whether D. rotundata, the most important African yam that is not known to grow in the wild, 1s a distinct species or a subspecies of the earlier-published D. cayenensis. It has been suggested that D. rotundata probably evolved from hybridization between D. cayenensis and D. praehen- silis, although the second parent may have been either D. abyssinica Hochst. or D. togoensis Knuth (Akoroda & Chheda; Coursey, 1976a). The medicinal value of various species of Dioscorea has increased dramat- ically following Marker’s (see Tyler ef a/.) pioneering research on the conversion of diosgenin to precursors of steroidal drugs. These drugs include the sex hor- mones androgen, estrogen, and progestogen, as well as oral contraceptives and anti-inflammatory compounds such as the systemic corticosteroids and topical hormones. Diosgenin is commercially extracted mainly from the tubers of the Mexican D. composita, D. floribunda, and D. spiculiflora, the South African D. sylvatica Eklon, and the Himalayan D. de/toidea and D. Prazeri Prain & Burkill. Sapogenin content in the tubers of the Mexican species above can be as high as 10-15 percent of the dry weight. Although the use of Dioscorea tubers in traditional medicine is likely based on superstition and magic rather than on actual physiological effects, some of the sapogenins and other steroidal com- ponents may have certain medicinal properties. In fact, preparations from the tubers of various species have been prescribed to cure colic, dysentery, ulcers, syphilis, sore throat, swellings, pulmonary complaints, diarrhea, hemorrhoids, boils, tumors, corns, diabetes, cuts, superficial lesions, and hysteria (Hartwell; Karnick, 1969; Perry; Sastri; Watt & Breyer-Brandwyk). Tubers of numerous species of Dioscorea are highly toxic. The poisonous substances are alkaloids, sapogenins, or tannins. Because of their hemolytic properties, sapogenins are highly toxic if introduced into the bloodstream. Therefore, the tubers have been widely used as a source of poison for fishing, hunting, or criminal purposes. Fish poisons are obtained from D. deltoidea in India, D. hispida in Java, D. piscatorum in Malaya, D. Poilanei Prain & Burkill in Vietnam, D. Prazeri in Sikkim, D. sansibarensis Pax in tropical East Africa, D. tokoro Makino in China, D. composita in Mexico, and D. bulbifera, D. Dregeana, and D. rupicola Kunth in various parts of Africa. The placement of 1989] AL-SHEHBAZ & SCHUBERT, DIOSCOREACEAE 72 grated tubers into a stream can stupefy fish at a considerable distance. Arrow or dart poisons are prepared in Malaysia and Sikkim by mixing the juice of Antiaria toxicaria Lesch (Moraceae Link) with that of the tubers of D. hispida, and in Africa by using the tuber extracts of D. dumetorum or D. sansibarensis with extracts of Strophanthus DC. (Apocynaceae Juss.). These poisons are employed to kill tigers and monkeys. In India tigers are hunted by placing pounded tubers of D. hispida in carcasses (Karnick, 1969). Because of their high saponin content, tubers of various species, particularly Dioscorea deltoidea and D. Prazeri, are used in Burma, southwestern China (Yunnan), and northern India as soap for washing hair, silk, or wool. Decoctions of the tubers are also used in these countries to kill lice in human hair and clothing. In Malaysia tubers of D. piscatorum are used in the preparation of an insecticidal powder and are employed to destroy rice parasites (Coursey, 1967), while in Tanzania the tubers of D. dumetorum are used as a remedy for schistosomiasis (Watt & Breyer-Brandwijk). A paste prepared from the bulbils of D. bulbifera is said to cure scorpion stings and snake bites (Karnick, 1969). Bulbils of the last species are also used as a fish bait in Kashmir (Sastri). Other minor uses of Dioscorea include extracting starch from D. alata, dis- tilling alcohol from certain species (Sastri), brewing beer from D. dumetorum (Corkill; Irvine), substituting D. po/ygonoides Humb. & Bonpl. (as D. /utea G. F. W. Meyer) for coffee (Uphof), tanning leather and making fishing nets (see above), dyeing clothes and cotton threads with tubers of C. cirrhosa (Karnick, 1969; Uphof), coloring ice cream with phenolic pigments extracted from D. alata (Ingram & Greenwood-Barton), feeding livestock (Bailey et a/.), and utilizing stems as cordage (Burkill, 1966). A drink prepared by the Meskwakis from the rhizomes of D. villosa is said to relieve the pain of childbirth (Lewis & Elvin-Lewis). The juice of a certain Dioscorea is used in Tanganyika for tattooing (Watt & Breyer-Brandwijk). Several species of Dioscorea are grown as outdoor twining ornamentals, and forms of D. alata and D. oppositifolia with variegated foliage or colored younger parts are highly desirable (Hawley). Dioscorea elephantipes, elephant’s foot, is occasionally grown in Europe and North America as an indoor plant for its large, curious, above-ground tubers with corky bark. Yams have played a central role in the social and religious life of people in the tropics, particularly in West Africa. Cultural, social, magic, and religious practices associated with almost every aspect of yam planting, harvest, and storage are especially evident in Nigeria and neighboring countries (Ayensu & Coursey; Coursey, 1967, 1972; Purseglove). The cultivation of Dioscorea pen- taphylla has been intimately associated with priesthood and has therefore been considered sacred in various parts of Malaysia and Indonesia (Coursey, 1967). The poisonous properties of many species form the basis of several magical beliefs in West Africa. Many of the highly toxic species (e.g., D. dumetorum and D. hispida) are used during famine for food. The tubers are peeled, sliced, boiled, pounded, and placed in running water for a few days to remove the toxins before they are consumed (Irvine; Karnick, 1969). Different tribes, how- ever, follow different procedures to detoxify the tubers. 80 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 REFERENCES: The literature on Dioscorea is too voluminous to be presented fully here. The interested reader should consult Lawani & Odubanjo and the indexes of the Bibliography of Ag- riculture for additional leads. The aaa listed below were selected from a file con- taining more than 3500 entries. All the family references contain pertinent data on Dioscorea except the following that deal with other genera of the Dioscoreaceae: AYENSU (1966, 1968); BURKILL (1937, 1939, 1944, 1946); CLARKE & JONES; CONRAN; GRAHAM; KALE & PAI; MULLER; PERRIER DE LA BATHIE; RAO; REICHE; ROGERS; SCHLITTLER; TAN; and TOMLINSON & AYENSU. Family references cited in the text of Dioscorea are: AYENSU (1972); BOLKHOVSKIKH et al.; BuRKILL (1960); HARTWELL; HOOKER; KNUTH (1924, 1930); and RAMACHANDRAN (1962). ABRAHAM, K., S. G. Nair, M. T. SREEKUMARI, & M. UNNIKRISHNAN. Seed set and seedling variation in greater yam (Dioscorea alata L.). Euphytica 35: 337-343. 1986. AbROL, B. K., L. O. KApoor, & I. C. CHopra. Pharmacognostic study of the rhizome of Dioscorea deltoidea Wall. Pl. Med. 10: 335-340. 1962. [Anatomy.] Au es, H. E. New combinations for some vascular plants of southeastern United States. Jour. Elisha Mitchell Sci. Soc. 80: 172, 173. 1964. [D. floridana and D. hirticaulis reduced to varieties of D. villosa.] Dioscoreaceae. Pp. 317, 318 in A. E. RApForp, H. E. AHLEs, & C. R. of the vascular flora of the Carolinas. Chapel Hill, North Carolina. en [D. aEpos terol (as D. batatas), D. villosa; D. floridana and D. hirticaulis recognized as varieties of D. villosa, maps, figure. ] AKAHoRrI, A. Studies on the steroidal components of domestic plants— XLIV. Steroidal sapogenins contained in Japanese Dioscorea sp. Phytochemistry 4: 97-106. 1965a. [Survey of 12 species; — between sapogenin contents and taxonomic group- ing; key to the species of Jap Morphological and aes | seine differences between D[/oscorea] tokoro and D. tenuipes. (In Japanese.) Acta Phytotax. Geobot. 21: 149-152. 1965b. Axkoropa, M. O. Floral biology in relation to hand pollination of white yam. Euphytica 32: 831-838. 1983. [D. rotundata; low fruit set caused by inefficient pollination; sex ratio. ] . CHHEDA. Agro-botanical and species relationships of Guinea yams. Trop. Agr. Tanidad 60: 242-248. 1983. [D. cayenensis and D. rotundata as distinct species; see MARTIN & RHODES, 1978.] ALBANS, J. W. Diosgenin and the new synthesis of cortisone. Discovery 17: 122, 123. 1956 ALEXANDER ,J. The ip aa of yams: a multi-d roach. Pp. 229-234 nD. Broruwet & E Hiaains, eds., Science in sae eae Rene ed. New York so Washington. 1969, & D. G. Coursey. The origins of yam cultivation. Pp. 405-425 in P. J. Ucko G. W. DIMBLEBY, eds., The domestication and exploitation of plants and animals. Chicago. 1969. [Centers of cultivation, cultivated species, archaeology.] ALLARD, H. A. Some behaviors of the yams (Dioscorea) of the family Fores: Castanea 10: 8- 13. 1945. [Habit, twining of stem, effects of day length on the Ax-Rals, A. H., A. Myers, & L. WATSON. The isolation and properties of oxalate crystals from plants. Ann. Bot. II. 35: 1213-1218. p/. 1971. [D. alata.] | W. Notes on the flora of Tennessee: Dioscorea. Rhodora 36: 344-346. 934. [D. auaternat D. villosa.) ae HIBALD, E. E. A. The genus Dioscorea in the Cape Province west of East London. Jour. S. Afr. Bot. 33: 1-46. 1967. [Eight species; a new section (Perennia), mor- phology, distribution, key, illustrations. 1989] AL-SHEHBAZ & SCHUBERT, DIOSCOREACEAE 81 AYALA, F. Four new species of Dioscorea from Amazonian Peru. Ann. Missouri Bot. Gard. 68: 125-131. 1981. [D. Claytonii, D. Revillae, D. Schunkei, D. tamshiya- cuensis. | os nomina nova para ea americanas de Dioscorea. Phytologia 55: 296. 1984. [D. Ravenii, D. Schuber AYENSU, E. S. Analysis of ae vasoulanty in stems of Dioscorea composita. Jour. Arnold Arb. 51: 228-240. 1970. [Comparison with the palm Rhapis.] . Comments on Old and New World dioscoreas of commercial importance. Publ. Espec. Inst. Nac. Invest. Forest. México 8: 75-81. 1973. [Anatomical differences between the Old and New World species. ] . Coursey. Guinea yams. The botany, ethnobotany, use and possible future of yams i in West Africa. Econ. Bot. 26: 301-318. 1972. [A review; cultivation, origin of the major yams, food values, prospects; emphasis on D. cayenensis and D. rotundata. ] BarLey, L. H., E. Z. BArLEy, & BAILEY HorTORIUM STAFF. Hortus third. xiv + 1290 pp. New York and London. 1976. [Dioscorea, 387, 388; yam, 1178. BAKER, E. A., J. T. MARTIN, & A. P. WiLson. The distribution of diosgenin in Dioscorea spp. Ann. Appl. Biol. 58: 203-211. pi. Hee [D. deltoidea, D. sylvatica; concentra- tions of diosgenin in the various plant par BAKER, J. G. Dioscoreaceae. Jn: W. T. oe Dyer, ed., Fl. Tropical Africa 7: 414— 421. 1898. [Dioscorea; 20 species. ] BaAQquar, S. R. Chromosome behaviour in Nigerian siey (Dioscorea). Genetica 54: 1- 9. 1980. [Nine species; chromosome counts, polyplo BaRRAU, J. Les ignames alimentaires des iles du Pacific ue Sud. Jour. Agr. Trop. Bot. Appl. 3: 385-401. 1956. [D. alata, D. bulbifera, D. esculenta, D. pentaphylla, tax- onomy, distribution, cultivation Barroso, G. M., E. F. GUIMARAES, & D. Sucre. Margarethia, uma seccéo nova do género Dioscorea L., com a espécie D. margarethia Barroso, Guimaraes et Sucre. (English summary.) Loefgrenia 49: 1-7. 1970. [Section has divided leaves and six staminal filaments connate into a thick column.] ,— Dioscorea pseudomacrocapsa Barroso, Guimaraes et Sucre, uma espécie nova de Dioscoreaceae da flérula do estado da Guanabara. (English summary.) Revista Brasil. Biol. 31: 309-312. 1971. [Species assigned to the new sect. Spinosa.] etal. Flora da Guanabara. Familia: Dioscoreaceae. (English summary.) Sellowia ci 9-256. 1974. [Twenty-five species; piaeen axonomy, anatomy, pollen, starch; ect. Hyperocarpa raised to a genus; 213 figures.] cree H. H. The source of the dam ees with a consideration of the Dios- coreae found in the United States. U. 8S. Dep. Agr. Bur. Pl. Industry Bull. 189: 1- 29. 1910. [Five species; D. floridana and D. hirticaulis, spp. nov.; distributions, medicinal properties, maps. Barua, A. K., D. CHAKRAVARTI, & R. N. CHAKRAVARTI. Steroid sapogenins from Indian Dioscorea plants. Part II. Jour. Indian Chem. Soc. 33: 799-803. 1956. [Survey of 11 species. Part I, ibid. 31: 173-178. 1954, deals with eight other species. ] BEHNKE, H.-D. Plant faehoriices Weir: and | terminology, taxo- nomic Seen: and aspects of trichome-bacteria interaction in leaf tips of Dios- corea. Pp. 1-21 . Ropricuez, P. L. HEALY, & I. MEHTA, eds., Biology and chemistry of ore, trichomes. New York and London. 1984. [D. macroura, ultra- structure of De tips, 6-19, figs. 6-21.] U Ariastomoses: in internode of Dioscorea: their frequency, structure, and Cee Pecks Pflanzenphysiol. 66: 82-92. 1971. Buat, B. K., & B. B. BinpRoo. Sex chromosomes in Dioscorea deltoidea Wall. Cytologia 45: 739-742. 1980. [Homomorphic sex chromosomes in staminate plants, hetero- morphic in carpellate ones.] 82 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 BLAKE, S. F. Notes on the Clayton herbarium. Rhodora 20: 48-54. 1918. [D. villosa, 48, 49; indicates that Bartlett’s choice of D. paniculata Michx. as a replacement for the Linnaean D. villosa was based on insufficient grounds; a new combination; see BARTLETT. ] BLUNDEN, G., C. J. BRicGs, & R. HARDMAN. Steroidal constituents of the aerial parts of Dioscorea and Tamus species. Phytochemistry 7: 453-458. 1968. [Nine species; sapogenin contents of leaves are six to 40 times ica than those of stems & R. RDMAN. Thin-layer chromatogra of Dioscorea sapo . Jour. Chromatography 15: 273-276. 1964. [D. cas (as D. belizensis), D. as ee D. villosa.) ,& F. J. Hinp. The eee glans and anatomy of Dioscorea sylvatica Eckl. from Natal and the Transvaal. Bot. Jour. Linn. Soc. 64: 431-446 1971. [Leaf and stem anatomy; a new ae : & G. E. TREASE. Some observations on the propagation of Dioscorea belizensis Lundell and other steroid-yielding yams. Pl. Med. 14: 84-89. 1966. [D. hondurensis, leaf cuttings.] Bourret, D. Etude ethnobotanique des — alimentaires: ignames de Nou- velle-Calédonie. 135 pp. Thesis, Fac. Sci. 1973.* Braun, E. L. The vascular flora of Ohio. an - The Monocotyledoneae. vill + 464 pp. [Columbus,] Ohio. 1967. [D. oppositifolia (as D. batatas), D. quaternata, D. villosa, 390-392.] BriIpGE, J. Nematodes of yams. Pp. 253-264 in J. MiEGE & S. N. Lvonaa, eds., Yams. Oxford. 1982. [A review.] Britton, N. L., & A. BRown. An illustrated flora of the northern United States, Canada and the British possessions. ed. 2. Vol. 1. xxix + 680 pp. New York. 1913. [Dios- coreaceae, 535.] Brown, M. L., & R. G. Brown. Herbaceous plants of Maryland. xlvi + 1125 pp. + 32 ‘unnumbered color pls. College Park, Maryland. 1984. [D. hirticaulis, D. opposi- tifolia (as D. batatas), D. quaternata, D. villosa, 357-359; color photograph of D. oppositifolia.] ILL ; correct botanic names for the white and the yellow Guinea yams. Gard. Bull. Straits Settl. 2: 438-441. 3 p/s. 1921. [D. cayenensis, D. rotundata.] A list of oriental vernacular names of the genus Dioscorea. Ibid. 3: 121-244. ee 1-5, 8, 9. 1924, [Distribution maps of generic sections, sources and origins of vernacular names, names within various geographic regions, origin and distribution of the most oe cultivated species; alphabetic list of several hundred common names, the areas in which they are used, and the scientific names of plants to which they are applied. i . The phyllotaxy of Dioscorea glauca Muhl. Jour. Bot. London 74: 89-101. 1936. [D. quaternata (as D. glauca), vasculature of the nodes and internodes. ] —. The contact of the Portuguese with African food-plants which gave words such as ‘“‘yam”’ to European languages. Proc. Linn. Soc. London 150: 84-95. 1938. [The English word yam, the French igname, the Spanish name, and the Portuguese ynhame are most likely derived from the African niam (Mande) or enyame (Temne Two notes on dioscoreas in the Congo: (1) the acarodomatia of D. minutiflora Engl. and D. smilacifolia DeWild., and (2) twining in both directions in D. Baya DeWild. /bid. 151: 57-61. 1939a. [Domatia occupied by mites, found in four African species of sect. Enantiophyllum, are shallow pits covered by inrolled leaf margin and are located on the lower leaf surface near the base; stems of all species of sect. Enantiophyllum except D. Baya are dextrorse.] otes on the genus Dioscorea in the Belgian Congo. Bull. Jard. Bot. Bruxelles 15: 345-392. 1939b. [Twenty-three species in six sections; detailed descriptions of sections; notes on distribution, origin, and cultivation of several species; D. bulbi- fera.] 1989] AL-SHEHBAZ & SCHUBERT, DIOSCOREACEAE 83 . Slight zygomorphy in Dioscorea sylvatica Ecklon. Jour. Bot. London 78: 100, 101. 1940. . A plea for a description from life of the African Dioscorea minutiflora Engl. Proc. Linn. Soc. London 159: 77-81. 1947. [Key to the types of tubers and rhizomes of Dioscorea and Tamus.] The rise and decline of the greater yam in the service of man. Adv. Sci. London Ws 443-448. 1951. Testudinaria as a section of the genus Dioscorea. Jour. S. Afr. Bot. 18: 191. 1952. [Section endemic to eastern South Africa; includes D. ee. is hemicrypta, and D. sylvatica; taxonomic history, key, distributions. A dictionary of the economic products of the Malay Peninsula. ed. 2. Vol. 1. Frontisp. + xiv + 1240 pp. Kuala Lumpur. 1966. [Dioscorea, 824-838; economic importance of 18 species, including D. alata, D. bulbifera, and D. oppositifolia (as D. opposita).] & R. tuM. Dioscorea piscatorum or tuba-ubi, a fish poison. Gard. Bull. Straits Settl. 3: "360. 1924. [Test of tuber poison.] PERRIER DELA BATHIE. Dioscoreaceae. /n; H. HUMBERT, ed., Fl. Madagascar Comores 44: ae 1950. [Dioscorea; 33 species in 12 sections.] units in the taxonomy of the Madagascan Dioscoreaceae. Not. Syst. Paris 14: (32 137. 1951. [Dioscorea; four species, new taxa; four new sections without descriptions; sectional relationships and differences.] BUTTERFIELD, H. M. Dioscorea and Testudinaria. Cact. Succ. Jour. 43: 117. 1971. CARTONI-CRETTON, N. Etude sur la bulbille de Dioscorea batatas. (English summary ) Saussurea 9: 1-21. 1978. [D. oppositifolia (as D. batatas); formation and germination of bulbil buds, anatomy.] CASTILLON, L. Las Dioscoredceas argentinas. Bol. Mus. Hist. Nat. Univ. Nac. Tucuman 11: 1-41. 1927. [Dioscorea; 32 species in 15 sections; see HAUMAN. CHAN, K. C. Steroidal sapogenins from PDioscorea in Malaysia. Planter 49: 146-151 Loa. CHANG, M.-C., Z.-J. Wu, H.-C. Coin, & C.-T. TinG. Comparative anatomy of Chinese Dioscorea and its meaning in sectional eee (In Chinese; English summary.) Bull. Nanjing Bot. Gard. (Mem. Sun Tan Sen) 1982: 1-8. pls. 1-5. 1982. [Compar- ative stem anatomy, fruit and seed morphology; 35 species of five sections ana A. Contribution a 7 ee de quelques espéces africaines du genre Pastore us. Hist. Nat. Paris, II. 8: 520-551. 1936. [Notes on 22 species in eight eee ell new Pee Notes sur deux ignames. Sees me edee Lamk. et D. minutiflora Engler.) i Revue Bot. Appl. Agr. Trop. 18: 33-3 8. [Apparently all species of sect. En- antiophyllum with opposite upper nue . have spiny stem bases and edible tubers. Nouvelles recherches sur les ignames cultivées. /bid. 26: 26-31. 1946. [Origin and cultivation of the major African, American, and Asiatic yams.] Un igname d’Afrique employé dans les empoisonnements criminels. /bid. 27: 56, 57. 1947. [D. sansibarensis, toxic effects.] e quelques Dioscorea d’Afrique équatoriale toxiques dont plusieurs variétes sont alimentaires. bid. 32: 14-19. 1952. [D. anthropophagorum, D. bulbifera, D. latifolia, D. sansibarensis.] Cun, H.-C., M.-C. CHANG, P.-P. Linc, C.-T. TinG, & F.-P. Dou. A cytotaxonomic study on Chinese Dioscorea L.—the chromosome numbers and their relation to the origin and evolution of the genus. (In Chinese; English summary.) Acta Phytotax. Sinica 23: 11-18. 1985. [Chromosome numbers of 23 taxa of five sections CLEWELL, A. F. Guide to the vascular plants of the Florida Panhandle. [viii] + 605 pp. — 84 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 Tallahassee, Florida. 1985. [D. alata, D. bulbifera, D. floridana, D. quaternata, 103, 104 ConzaTti, C. Flora taxonémica mexicana. Vol. 2. 220 pp. Mexico. 1947. [Dioscorea- ceae, 206-210; 26 species of Dioscorea, key, brief descriptions. ]} CopeLAND, E. B. Growth phenomena of Dioscorea. Philip. Jour. Sci. Bot. 11: 227-241. 1916. [Stem nutation in darkness and in light.] CorkiLL, N. L. The poisonous wild cluster yam, Dioscorea dumetorum Pax, as a famine food in the Anglo-Egyptian Sudan. Ann. Trop. Med. Parasitol. 42: 278-287. pls. 9 10. 1948, CorreLL, D. S., & M. C. JoHNston. Manual of the vascular plants of Texas. Frontisp. + xv + 1881 pp. Renner, Texas. 1970. [Dioscoreaceae, 424, 425; Dioscorea qua- iernata, D. villosa.] B. G. SCHUBERT, H.S. Gentry, & W. O. HAw.ey. The search for plant precursors of cortisone. Econ. Bot. 9: 305-375. 1955. [Dioscorea, 327-336, 372, 373, figs. 3, #, 16-23, 28-38, 48-63, map 2; fig. 38 shows some of the leaf variation in the genus. | COURSEY, = G. The role of yams in West African food economies. World Crops 17(2): 74-82. 1965. —, rae An account of the nature, origins, cultivation and utilisation of the useful members of the Dioscoreaceae. 230 pp. London. 1967. [A comprehensive account of the history, nae a. agriculture, pests, chemical composition, social and cultural importance, and toxic and pharmacological constituents of var- 10us species of Dioscorea.] The civilizations of the yam: interrelationships of man and yams in Africa and the Indo-Pacific region. Archaeol. Phys. Anthropol. Oceania 7: 215-233. 1972. [Uses, domestication. | he origins and domestication of yams in Africa. Pp. 187-212 in M. L. Arnott, , Gastronomy: the anthropology ene and food habits. Paris. 1975. [Domes- ae in Africa, the Americas, and Asi Yan Dios ‘orea spp. eras Pp. 70-74 in N. W. Stmmonps, ed., eee of crop plants. London and New York. 1976a. [Cytology, ms and recent history, prospects, relationships of the cultivated yams, role of hybridization, eco- nomic importance ——.. The origins and oe cue of yams in Africa. Pp. 383-408 in J. R. HARLA J.M. J. De Wet, & A. B. L. STEMLER, eds., Origins of African plant nee 1976b.* & C. K. Coursey. The new yam festivals of West Africa. Anthropos 66: 444- 484. 1971. Cox, D. K., A. HERNANDEZ Corzo, E. MAtuba, & J. G. GONZALEZ DuRAN. Estudio de las dioscoreas mexicanas I. Dioscorea se tas sisi Bol. Soc. Bot. Méx. 22: 12-27. 1958. EP Soup OOn, eHOmOsOMe nun Der nee Sie J Crete, P. A propos de I’ 2 L. Bull. Soc. Bot. France 100: 306, 307. 1953. CUSHING, A. M. Monograph on Dioscorea villosa and dioseoune: their physiological effects with their use in disease. 52 pp. Detroit. 1869. Dace, E. On the origin, development, and morphological nature of the aerial tubers in Dioscorea sativa L. Ann. Bot. 15: 491-501. pl. 26. 1901. [D. bulbifera.] Deam, C. C. Flora of Indiana. 1236 pp. Indianapolis. 1940. [D. glauca, D. hirticaulis, D. quaternata, D. villosa, 330-332. Decras, L., ed. L’igname. Inst. Natl. Rech. Agron. Vol. 2. 296 pp. Paris. 1980. [Inter- national symposium; 29 papers on the cultivation, physiology, pests, and repro- uction of various cultivated species of Dioscorea. ———.. La reproduction végétative de ’igname: données fondamentales et applications 1989] AL-SHEHBAZ & SCHUBERT, DIOSCOREACEAE 85 récentes. (English summary.) Pp. 60-87 in J. MIEGE & S. N. Lyonaa, eds., Yams. Oxford. 1982 , R. ARNouin, A. Porrout, & C. SUARD. Quelques aspects de la biologie des ignames (Dioscorea spp.) I.—Les ignames et leur culture. (English summary.) Ann. Amél. Pl. 27: 1-23. 1977. [D. alata, D. bulbifera, D. cayenensis, D. trifida.] DEYouna, J. E., ed. Yam cultivation in the Trust Seekeee Anthropol. Working Pap. pp. Guam. 1959. [Cultivation in Palau, Ponape, a Bieonr: R. Etude morpho-botanique des ignames Dione Gee eb Dioscorea cayenensis cultivées au Nord-Bénin. Agron. Trop. 32: 225-241. Duncan, W. H., & J. T. Kartesz. Vascular flora of Georgia. An noted checklist. ix + 143 + 3 unnumbered index pp. Athens, Georgia. 1981. [D. floridana, D. hirticaulis, D. oppositifolia (as D. batatas), D. quaternata, D. villosa, 43; see JONES CoILe.] Exunpayo, C. A. Stomatal development in Dioscorea and Elaeis guineensis. Trans. Missouri Acad. Sci. 6: 6-11. 1972. [D. alata, D. cayenensis, D. dumeitorum, D. EssAp, S. Variation géographique de nombres chromosomiques de base et polyploidie dans le genre Dioscorea a propos du dénombrement des espéces fransversa Brown, pilosiuscula Bert. et trifida L. ae summary.) Agronomie 4: 611-617. 1984.* FERNALD, M. L. Gray’s manual of botany. ed. 8. lxiv + 1632 pp. New York and other cities. 1950. [D. hirticaulis, D. oppositifolia (as D. batatas), D. quaternata, D. villosa, 451, 452.] Forsytu, C., & J. VAN STADEN. pes aus of Dioscorea bulbifera stem nodes in culture. Jour. Pl. ne 115: 79-83. 1984. GIRAL, F. Las dioscorea la industria ae. Revista Soc. Cuba. Bot. 14: 12- 26. 1957. [Historical ae of phytochemical studies; names and formulas of isolated steroidal sapogenins. | GIRARD, F. Les gens de l’igname. Les Buang de la vallée du Snake, District de Morobé, Nouvelle-Guinée. Jour. Agr. Trop. Bot. Appl. 14: 287-338. pls. 1-8. 1967. [Detailed anthropological account of yam cultivation for food and rituals; no scientific names given. Gieason, H. A. The new Britton and Brown illustrated flora of the northeastern United States and adjacent Canada. Vol. 1. Ixxv + 482 pp. Lancaster, Pennsylvania. 1952. [Dioscoreaceae, 439, Dioscorea hirticaulis, D. oppositifolia (as D. batatas), D. quaternata, D. villos GOEBEL, K. Mopbion che und biologische Bemerkungen. 16. Die Knollen der Dios- coreen und die Wurzeltrager der Selaginellen, Organe, welche zwischen Wurzeln und Sprossen stehen. Flora 95: 167-212. 1905. [Tuber formation and eae, Gomez-Pompa, A. Notas botanicas sobre algunas dioscéreas de importancia farmacév tica. (English summary.) Ciencia 21: 221-229. 1962. [D. composita, D. Fen D. mexicana, D. spiculiflora, descriptions, ecology, distribution; history of their discovery as sources for precursors of steroidal hormones. ] Goopinc, H. J. West Indian Dioscorea alata cultivars. ae Agr. Trinidad 37: 11-30. 1960. [Sixteen cultivars; descriptions, variation in leaf and tuber, key.] Grout, B. W. W., & A. WittiAMs. Extrafloral nectaries of Dioscorea = iaie Por. their structure and secretions. Ann. Bot. II. 46: 255-258. pis. 1, 2. 1980. ieee of leaf-blade nectaries; sugar content; light and transmission electron microscopy. ] HANSEN, B. Studies in the flora of Thailand 42. Dioscoreaceae. Dansk. Bot. Arkiv. 23: 459-463. 1968. [Dioscorea; 14 species, distribution, notes. Harrison, I. T., M. VALAsco, & C. Dyerassi. Chiapagenin and isochiapagenin. Two new steroidal sapogenins from Dioscorea chiapasensis. Jour. Organ. Chem. 26: 155- 158. 1961. HASHIMOTO, T., K. HASEGAWA, & A. KAWARADA. Batatasins: new dormancy-inducing 86 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 ae of yam bulbils. Planta 108: 369-374. 1972. [D. oppositifolia (as D. ba- atas), occurrence of three inhibitors of bulbil sprouting, reversal of inhibition by nee temperature stratification.] HauMan, L. Les Dioscoréacées de l’Argentine. Anal. woes he Hist. Nat. Buenos Aires 7: 441-513. 1916. [Dioscorea; 19 species; see CAST Haw ey, W. O. Dioscoreas as ornamental foliage nen ‘Natl Hort. Mag. 35: 23-29. 1956. Haynes, P. H., & D. G. Coursey. Gigantism in the yam. Trop. Sci. 11: 93-96. 3. 1969, fibers of D. alata reach up to 3.5 m in length and 110 kg in weight, i. of D. elephantipes up to 350 kg.] ENRY, G. Studies on botanical and agronomic characteristics in cush- epee eHoscanee trifida L. f.). Ph.D. dissertation, McGill Univ., Montreal, Can a 67. HerK Lots, G. A. C. Vegetables in South-East Asia. xii + 525 p ae York. 1972. [Diosc orea, 416-428 .] Hircucock, A. S., & M. L. GREEN. Standard-species of Linnaean genera of Phanero- gamae (1753-54). Pp. 110-199 in International Botanical Congress, Cambridge (England), 1930. Proposals by British botanists. 1929. [D. bulbifera as the lectotype of Dioscorea, 192.] Hopce, W. H. The elusive elephant’s foot. Nat. Hist. 62: 324-329. 1953. [D. elephan- tipes.] Hout, T. ners ae of North America. 97. Dioscorea villosa L. Merck’s Rep. 22: 311-315. 1912 HUuSSAIN, 4 M. ail use of Dioscorea deltoidea Wall. Pakistan Jour. Forestry 3: 91, 92. 1953.* Ikeprosl, C, O., & L. C. IGBoanusr. Identification of yam (Dioscorea spp.) species and cultivars by use of electrophoretic patterns of soluble tuber proteins. Biotropica 15: 65-67. 1983. [D. alata, D. cayenensis, D. dumetorum, D. esculenta, D. rotundata (nine cultivars). ] INGRAM, J. S., & L. H. GREENWoop-BARTON. The cultivation of yams for food. Trop. Sci. 4: 82-86. 1962. [Cultivation, economic importance IrvinE, F. R. Supplementary and emergency food plants of West Africa. Econ. Bot. 6: 23-40. 1952. [Dioscorea, 25-27.] JENSEN, H. W. Meiosis in several species of dioecious Monocotyledoneae I. The pos sibility of sex-chromosomes. Cytologia, ae Jub. Vol.: 96-103. 1937. [D. quater- nata, n = 7: no evidence of sex chromosom Jones, S. B., JR., & N.C. Corte. The distribution of the vascular flora of Georgia. 230 pp. Athens, Georgia. 1988. [County eae maps of D. floridana, D. opposi- tifolia (as D. batatas), and D. villosa, 28, 29.] KADKADE, P. G., C. Rotz, & J. D. Dwyer. Steroidal sapogenins of the tubers of Dios- coreaceae: a chemotaxonomic study. Lloydia 39: 416-419. 1976. [Dioscorea, seven known and six undetermined species; distribution of 13 sapogenins. ] Karnick, C. R. Dioscorea (yams)—the food of slaves, with potentials for newer drugs. Quart. Jour. Crude Drug Res. 9: 1372-1391. 1969. [Review of taxonomy, anatomy, embryology, physiology, ethnobotany, phytochemistry, pharmacology, and econom- ic importance .N ects regarding the “nodal plexus” of some Dioscorea species of India. Ibid. 10: 1607-1625. 1970. [D. alata, D. bulbifera, D. sativa, D. Wallichii; nodal Phytochemical investigations of some Dioscorea species and varieties found in India. /bid. 11: 1761-1773. 1971. [Fifty-one taxa screened for alkaloids, glycosides, sage gre saponins, sterols, tannins, proteins, and phenolic compounds.] effects of photoperiod on steroidal sapogenins and other constituents in oe deltoidea Wall. Ann. Bot. II. 36: 605-610. p/. 1972. an and phenolic components increased 3 in leaves and tubers during growth.] 1989] AL-SHEHBAZ & SCHUBERT, DIOSCOREACEAE 87 Kaut, B., & E. J. Stasa. Dioscorea tissue cultures. I. Biosynthesis and isolation of diosgenin from Dioscorea deltoidea callus and suspension cells. Lloydia 31: 171- 179. 1968. [Undifferentiated cells of tissue culture produce significant amounts of diosgenin; a related paper in ibid. 32: 347-359. 1969. Knutu, R. Dioscoreaceae novae. I. Repert. Sp. Nov. 21: 77-81. 1925. [Part II, ibid. 22: 344-347. 1925: parts INI-V, ibid. 28: 81-88. 1930; part VI, ibid. 30: 158-161. 1932; part VII, ibid. 36: 125-128. 1934; part VIII, ibid. 38: 117-121. 1935.] Kocn, W. Las dioscoreas de Guatemala. Ceiba 12(1): 58-60. 1966. [Key to 14 species based on tuber characters; diosgenin contents. ] . BRuuN. Uber die Morphologie der Speicherorgane einiger mittelameri- ieumicenee Dioscoreen. Flora 152: 670-678. 1962. [D. chiapasensis, D. composita, D. floribunda, e spiculiflora.] Kreic, M. B. Green medicine. The search for plants that heal. 462 pp. Chicago, New York, and San Francisco. 1964. [Historical account of the isolation of steroidal hormones from some Mexican species of Dioscorea, 255-292. Lawanl, S. M., & M. O. OpuBANJo. A bibliography of yams and the genus Dioscorea. vi + 192 pp. Ibadan, Nigeria. 1976. [Excellent compilation of 1562 annotated bibliographies through 1975; bibliographies classified into 11 sections, including botany (anatomy, morphology, histology, physiology, pee cee breeding. See and phytochemistry: indexes to authors and subjec Lawton, J. R.S. A key to the Dioscorea species in Nigeria. West es Sci. Assoc. Jour. LELonG, M. G. Annotated list of vascular plants in Mobile, Alabama. Sida 7: 118-146. 1977. [D. villosa, 128.] Lewis, W. H., & M. P. F. E-vin-Lewis. Medical botany. Plants affecting man’s health. xv + 515 pp. New York, London, Sydney, and Toronto. 1977. [Dioscorea, 6, 19, 93, 95, 285, 318, 319, 322, 387.] Li, D.-G., & Y.-C. RUAN. Diosgenin and yamogenin from four species of Dioscorea L. in Yunnan. (In Chinese; English summary.) Acta Bot. Yunnanica 2: 476-479. 1980. [D. Collettii, D. panthaica, D. parviflora.) Lina, P.-P., Z.-J. Wu, & H.-C. Cu1n. Some observations on the epidermis of Chinese Dioscores and its taxonomic significance. (In Chinese; English summary.) Bull. Nan- jing Bot. Gard. (Mem. Sun Yat Sen) 1982: 9-16. pis. 1-5. 1982. [Thirty-eight species of five sections.] moe: & T.-C. HUANG. On the Taiwan species of Dioscorea. Bot. Bull. Acad. Sinica, s. 3: 133-149. 1962. [Fifteen ee ae distributions, key 7 Dioscoreaceae. Fl. Taiwan 5: 99-109. 1978. meee ‘14 species. |} MACckRIDE, J. F. Dioscorea. Fl. Peru. Publ. "Field Mus. Bot. 13(1): 690-707. 1936. [Forty- nine species. MacRoserts, D. T. The vascular plants of Louisiana. An annotated checklist and bibliography of the vascular plants reported to grow without cultivation in Louisiana. Bull. Mus. Life Sci. Louisiana State Univ. Shreveport 6: 1-165. 1984. [D. quaternata, D. villosa, 35; distinctness of the two species questioned. ] MartTIN, F. W. Sex ratio and sex determination in a ee Hered. 57: 95-99. 1966. [Sex en ere controlled by X-Y chromosom Species of Dioscorea containing sapogenin. Econ. Bot. 23: 373-379. 1969, [Com- pilation of sapogenin contents in about 125 species.] —. Potencialidades para mejorar los fames que contienen sapogeninas. Publ. Espec. Inst. Nac. Invest. Forest. Méx. 8: 89-110. 1973. [D. alata, D. composita, D. flori- bunda : opical yams and their potential. Part 1. Dioscorea eee U.S. Dep. Agr. Agr. Handb. 457. 18 pp. 1974a. [Botany, cytology, cultivation.] Tropical yams and their potential. Part 2. Dioscorea buena Ibid. 466. 20 pp. 1974b. [Origin, distribution, morphology, cytology, culture 88 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Tropical yams and their potential. Part 3. Dioscorea Ibid. 495. 40 pp. 1976. [Origin, distribution, morphology, cytology, culture E. CABANILLAS. A wild hybrid of sapogenin- sheatine ee species. Bull. Torrey Bot. Club ae 232-237. 1963. ae composita x D. floribun & hybrids of some sapogenin-bearing Dioscorea species. Am. Jour. Bot. 53: 350- ae 1966. [Artificial hybridization among | 5 species of Dioscorea and one of Rajania in 60 different combinations. ] ,& S. ee Natural pollination, hand pollination and crossability of me Mexican species of Dioscorea. Trop. Agr. Trinidad 40: 135-141. 1963. [D ie dD. ee D. spiculiflora, fruit set, cross fertility. ] ——— & L. Decras. Tropical yams and their potential. Part 5. Dioscorea trifida. U Dep. Agr. Agr. Handb. 522. 26 pp. 1978a. [History and origin, distribution, mor- phology, eg oe number, cultivation. & ical yams and their potential. Part 6. Minor cultivated Dioscorea species. [bid. 538. 3 pp. 1978b. [D. dumetorum, D. hispida, D. nummularia, D. pentaphylla.] L, & H. J. CRuzapo. Dioscorea Friedrichsthalii, another sapogenin- bearing species. Turrialba 13: 159-163. 1963. [Description, sapogenin content, chro- mosome number, interspecific crosses. ] .H. Gaskins. Cultivation of the sapogenin- cana ne ee species. U.S. . Res. Rep. 103. 19 pp. 8. [D. composita, D. floribunda, D. Priednichsthale D. spiculiflora, propagation, ae production, breed- ing. ] & S. Ortiz. Chromosome numbers and behavior in some species of Dioscorea. Cytologia 28: 96-101. 1963a. [Counts for 13 species, summary of previous counts; a and Old World species are based on nine and ten, respectively. . Origin and anatomy of tubers of Dioscorea floribunda and D. spicu- ina Bot Bot. ae ah 416-421. 1963b chromosome numbers in some Dioscorea species. ae 31: an 1966. icone for 11 determined and four undetermined spec & A.M. RHopes. Correlations among greater yam (Dioscorea alata L) eer Trop. Agr. Trinidad 50: 183-192. 1973. [Correlations among 100 characters of 30 cultivars. ] & Intraspecific classification of Dioscorea alata. Ibid. 54: 1-13. 1977. & —— relationship of Dioscorea cayenensis and D. rotundata. Ibid. 55: 193-206. 1978. [Numerical analysis of 75 characters supports the recognition of one species complex. ] S. SADIK. Tropical yams ae their er iaue Part 4. Dioscorea rotundata and Dioscorea cayenensis. U. S. Dep. Agr. Agr. Handb. 502. 36 .L. TELEK, & R. M. eee Yellow ae of Dioscorea bulbifera. Jour. Agr. Food Chem. 22: 335-337. 1974. [Bulbil color due to saponifiable esters of xantho- phylls such as lutein, neoxanthin, violaxanthin, zeaxanthin, auroxanthin, and cryp- Matupa, E. Nue vas dioscoreas de México y Guatemala. Anal. Inst. Biol. Méx. 24: 55- 61. 1953a. [_D. Deamii, D. igualamontana, D. spiculoides, D. tancitarensis, D. tubi- perianthia, spp. nov. Las dioscoreas de México. bid. 279-390. 1953b. [Treatment of 63 species; descriptions distributions, keys, 58 figures. s dioscoreas del Estado de México. Toluca, Mexico. 61 pp. 1955. [Twenty- seven species; descriptions, synonymy, distributions in Mexico, key, illustrations. ] Dioscoreas nuevas de México. Bol. Soc. Bot. Méx. a 1-8. 1957. [D. alboho- losericea, he Coxtl, D. esc MINEISS. Dd. ee spp. nov.] plantas de Méx nal. Inst. Biol. Méx. 39: 143-155. 1961. [Nana- repenta ae ana, gen. et sp. nov., a a oO 1989] AL-SHEHBAZ & SCHUBERT, DIOSCOREACEAE 89 . Nueva Nanarepenta de Guerrero. (English summary.) Cact. Sucul. Méx. 19: 70, 71. 1974. [N. guerrerensis, sp. nov. McC ure, F. A. A note on a Chinese vegetable dye. Lingnaam Agr. Rev. 4: 31-37. pi. 5. 1927. [Brown dye from the tubers of D. rhipogonoides, preparation, dyeing silk.] Miéce, J. Contribution a l’étude esystemnaudue des Dioscorea d’Afrique occidentale. Thése ee Nat. 266 pp. 1952. hénoménes de nutation chez les Dioscorea. Ann. Sci. Univ. Besancon, II. Bot. 12: 63-70. 1958. [D. bulbifera, D. dumetorum, and D. esculenta are dextrorse, while D. alata and D. minutiflora are sinistrorse; physiology of nutation L’appui de la pal dans la distinction des espéces africaines de Dioscoree. Webbia 19: 841 845, pls. Ge 66. 1965. [Pollen of 31 species of eight sections. ] —. Dioscoreaceae. Jn: J. HUTCHINSON, J. M. DAvzieL, & F. N. Hepper, eds., Fl. West Trop. Africa, ed. 2. 3(1): 144-154. 1968. [Dioscorea; 20 species, evs to sta- minate and carpellate plants . La densité de nervation foliaire, caractére possible d’identification et de sépa- ration spécifiques chez les Dioscoréacées. Ann. Sci. Univ. Besancon, III. Bot. 12: 91-95. pl. 1972 ques caractéres discriminatoires entre les taxons sae ae Dioscorea oe (English summary.) Pp. 185-196 in J. MiEGE & S. N. Lyo eds. . Oxford. 1982a. [Comparison of wild and cultivated forms; aoe cera a eee me numbers, bulbil morphology, scanning-electron microscopy plexe Dioscorea cayenensis—D. rotundata. (English summary.) Pp. 197-231 in J. Mice & S. N. Lyonaa, eds., ibid. 1982b. ah al aero of tuber and seed proteins, close affinities and grouping among cultivars. ] . Note sur les espéces Dioscorea cayenensis Lamk. et D. rotundata Poir. Pp. 367- 375 inJ. MieGE & S. LYonGa, eds., ibid. 1982c. [English translation, without figures or bibliography, pp. 377-383: evidence pee the reduction of D. rotundata to a subordinate of the earlier-published D. cayen Dioscoreaceae of Ethiopia. Symb. Bot. Upsal. 26(2): 157-168. 1986. [Eleven species in five sections; preliminary distribution maps of seven taxa . LYonGA, eds. Yams. xviii + 411 pp. Oxford. 1982. [Twenty- five papers (12 in English and 13 in French); five parts dealing with plant improvement, ag- ronomic studies and taxonomy, plant protection, post-harvest biology and tech- nology, and economics of production and mechanization; discussions at the end of each part; peer English and French subject indexes. Mice, M.-N., & J. MizGe. Recherches taxonomiques et biologiques sur la famille de Dioscoréacées: I. Les protéines des graines de quatre espéces de Dioscorea d’ Afrique occidentale: méthodologie et résultats systématiques. (English and German sum- maries.) Arch. Sci. 24: 177-205. 1971. [Electrophoresis of seed proteins supports the placement of four species in different sections.] MILNE-REDHEAD, E. Dioscoreaceae. /1: R. M. POLHILL, ed., Fl. Trop. East Africa. 26 pp. 1975. [Dioscorea, 19 species.] MOou.ensrock, R. H. The illustrated flora of Illinois. Flowering plants, lilies to orchids. xii + 288 pp. Carbondale and Edwardsville, Illinois. 1970. [D. quaternata, D. villosa, 146-151. Mourer,C. Plant life of Alabama. xii + 921 pp. Montgomery, Alabama. 1901. (Reprinted from Contr. U. S. Natl. Herb. 6. 1901.) [D. villosa, 449. Morton, C. V. Notes on Dioscorea, with special reference to the species of the Yucatan Peninsula. Carnegie Inst. Publ. 461: 239-253. 1936. [Key and treatment of ten species of the Yucatan Peninsula, notes on 15 other species; five new species and three new sections. ] Dioscoreaceae. Jn: R. E. Woopson, Jr., & R. W. ScHERy, eds., Fl. Panama. 90 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Ann. Missouri Bot. Gard. 32: 26-33. 1945. [Dioscorea; 11 species, descriptions, distributions, key, illustrations. Morton, J. F. Major medicinal plants— botany, culture and uses. Frontisp. + xix + 431 pp. Springfield, Illinois. 1977. [Dioscoreaceae, 73-83; D. composita, D. flori- bunda, and related species; descriptions, origins, constituents, cultivation, medicinal uses, toxicity. ] Murty, Y.S., & PurNimA. Morphology, anatomy and development of bulbil in some dioscoreas. Proc. Indian Acad. Sci. Pl. Sci. 92: 443-449. 1983. [Eleven species. ] Nakajima, G. Cytological studies in some dioecious plants. Cytologia, Fujii Jub. Vol.: 282-292. 1937. [D. gracillima, D. tokoro.] Ninan, C. A., S. ABRAHAM, & P. G. PILLAr. On the cultivar taxonomy of Dioscorea alata. New Bot. 2: 151-153. 1975. [Three cultivar groups of tubers based on presence of spines and pigment Nyoxu, E. The propagation of as (Dioscorea spp.) by vine cuttings. Jour. W. Afr. Sci. Assoc. 8: 29-32. pls. 1, 2. 1963. Onwt, J. Flora of Japan. eer + 1x + 1067 pp. Washington, D. C. 1965. [Dios- coreaceae, 313-315; ten species of Dioscorea recognized. ] ONWUEME, I. C. The tropical tuber crops: yams, cassava, sweet potato, and cocoyams. xiv + 234 pp. Chichester, New York, Brisbane, and Toronto. 1978. [Dioscorea, 3- 106; ten chapters dealing with origin, classification, botany, propagation, cultivation, storage, diseases, economic importance, and prospects. OrnburF, R., ed. Index to plant chromosome numbers for 1966. Regnum Veg. 55: 1- 126. 1968. [Dioscorea, 39.] Orr, M. Y. The leaf glands of Dioscorea macroura Harms. Notes Bot. Gard. Edinb. 14: 57-72. pls. 197, 198. 1923. [Nitrogen-fixing bacteria growing in secretions of leaf glands. ] Ozo, O. N., J.C. CayaiLi, & D. G. Coursey. Phenolics of five Phytochemistry 23: 329-331. 1984. [D. alata, D. ulbiferd, D. capensis Dp. aie metorum, D. Lies [Per, C., & C.-T. Tinc.] Studies on Chinese Dioscorea sect. Stenophora Pr. et Burk. and we pene numbers. (In Chinese; English summary.) Acta Phytotax. ees 14(1): 65-72. 1976. [D. banzhuana, D. biformifolia, spp. nov.; chromosome number of 16 species. ] , H.-C. Cun, P. Su, S.-Y. TANG, & H.-C. CHANG. A preliminary sys- tematic study of Dioscorea L. sect. Stenophora Uline. (In Chinese; English summary.) Acta Phytotax. Sinica 17(3): 61-72. 1979. [Section considered primitive; chromo- some numbers, palynology.] PELLEGRIN, F. Quelques remarques sur les Dioscoréacées du Paraguay. Bull. Soc. Bot. Geneve 10: 383-388. 1918. [Species of ten sections.] Perry, L. M. Medicinal plants of East and Southeast Asia: attributed properties and uses. 1X + 620 pp. Cambridge, Massachusetts. 1980. [Dioscorea, 125-127.] PRrAIN, D., & I. H. BuRKILL. A synopsis of the dioscoreas of the Old World, Africa excluded, with descriptions of new species and of varieties. Jour. Proc. Asiat. Soc. Bengal, II. 10: 5-41. 1914. Leer of 107 species in eight sections. & “Dioscorea sativa.” Bull. Misc. Inf. Kew 1919: 339-375. 1919. [Critical evaluation ot the Linnaean species (Sp. Pl. 2: 1032-1034. 1753), particularly D. Sativa. ee lagnoses specierum novarum generis Dioscoreae. Ibid. 1925: 58-66. 1925. [Nineteen new species; 27 nomina resorben The genus Dioscorea in Siam. Ibid. 1927: 225-247. 1927. [Thirty-two species; dchion distributions, maps.] Dioscorea: section Stenocorea. Ibid. 1931: 88— a a [Five species endemic to southeastern Asia; D. sumatrana, sp. nov.; key, m —. Dioscoreaceae. Jn: H. Lecomte, H. HUMBERT, & °. "GAGNEPANN, eds., la 1989] AL-SHEHBAZ & SCHUBERT, DIOSCOREACEAE 9] Fl. Gén. Indo-Chine 6: 698-745. 1934. [Fifty-one species in nine sections; descrip- tions, keys. & . An account of the genus Dioscorea in the East. Part I. ae species which twine to the left. Ann. Bot. Gard. Calcutta 14: 1-11, 1, u, 1-210, 1 pls. I- 85 (bound separately). 1936. [A car aed eae of 76 ae Hctailed descriptions, di eae economic importa D. bulbifera, 111-132. & ccount of the genus eee in the East. es II. The species which twine to the ane with addenda to Part I, and a summary. /bid. 14:1, ii, 211- 528, i-xx. pls. 86-150 (bound separately). 1938 [1939]. [Sect. Bea Solan: 211- 415; D. alata, 302-342: D. oppositifolia (as D. opposita), 243-256; addenda (pp. 416-528) includes 35 distribution tables, as well as a summary on morphology, phylogeny, and classification.] PuURNIMA, & A. RIVASTAVA. Leaf epidermal studies in Dioscorea Linn. Acta Bot. Indica 13: ae 180. 1985. [Eighteen species PURSEGLOVE, J. W. Tropical crops. Monocotyledons 1. x + 334 pp. London. 1972. [Dioscorea, 97-117; brief description and importance of several species, botany, chemical composition, cultivation, diseases, production; key to the principal edible pecies. ] QuiGcLey, F. R. Diosgenin in West as Dioscorea plants. Pl. Med. 33: 414, 415. 1978. [Contents in tubers of ten specie a R.S. A chromosome survey sPindics dioscoreas. Proc. Indian Acad. Sci 48: 59-63. pls. 5-7. 1958. [Chromosome counts for nine species; photographs oF se and tubers. ] Studies on the genus Dioscorea L. Bull. Bot. Surv. India 2: 379-386. 1960. [Morphology and chromosome numbers of 13 species Rao, A.N., & A. S. TAN. Shoot apex and bulbil development in Dioscorea sansibarensis Pax. Bot. Jour. Linn. Soc. 72: 285-298. pls. 1-3. 1976. Rao, P. S., & R. M. Bert. Dioscorea starches. Sci. Cult. 20: 387-398. 1955.* Rao, V. R., R. K. Bam, & G. S. RANDHAWA. Interspecific hybridization in the genus Dioscorea. Ann. Bot. I. 37: 395-401. 1973. [D. alata, D. composita, D. deltoidea, D. floribunda, D. Friedrichsthalii; artificial hybridization, fertility and sterility re- lationships. ] .R. Murty. Meiotic studies in species and hybrids in medicinal yams. Cur Sci. Bangalore 44: 24, 25. 1975. [D. composita, D. floribunda, D. Friedrichsthalii and their artificial hybrids. ] Boat V., & D. G. Coursey. Properties of starches of some West African yams. Jour. Food Agr. 18: 240-244. 1967. [D. alata, D. dumetorum, D. esculenta, D. Se granule size, viscosity. ] Ruopes, A. M., & F. W. Martin. Multivariate studies ot vanahons) in yams (Dioscorea alata L.). Am. Soc. Hort. Sci. Jour. 97: 685-688. 1972. RicketTT, H. W. ie flowers ae: United States. Vol. 2. The Southeastern States. Part 1. x + 322 pp. New York. 1967. [D. floridana, D. hirticaulis, D. quaternata (as distinct ee, glauca), D. ee 80, 81.] Ripv.e, J. M. Dioscorides on pharmacy and medicine. Frontisp. + xxvii + 298 pp. Austin, Texas. 1985. Rizzint, C. T., & A. DE Mattos-FiLHo. Trés espécies novas de Minas Gerais e Bahia Brasil. (English summary.) Revista Brasil. Biol. 46: 317-321. 1986. [D. hasiclavi- caulis (sp. nov.) wath enormicus inflated, spiny stem base. ] ROMEIKE a AS angios perms Bot. Not. 131: 85-96. 1978. [Dioscorea, 88, 91, 92; synthesis of tropane alkaloids originated independently within several evolutionary lines of angiosperms. ] Row ey, G. Testudinaria—the elephant’s foot. Natl. Cact. Succ. Jour. 8: 46-50. 1953. Coe ae recognized as a genus with six species; cultivation, key.] SADIK, S., . U. Oxereke. Flowering, pollen grain germination, ‘Guiting! seed ger- a2 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 mination and seedling oe of white yam, Dioscorea rotundata Poir. Ann. Bot. IL. 39: 597-604. pls. 1 75. SANTAPAU, H. The genus eee in Bombay State. Jour. Bombay Nat. Hist. Soc. 49: 624-636. pis. 1-3. 1951. [Nine species; descriptions, SEABMMONS, key. ] SARWAR, M. Growth and development of D nd cross fertility with Dioscorea composita Hemsl. Ang. Bot. 45: 157-161. 1971.* SAsTRI, B. N., ed. The wealth of India. Vol. 3. Frontisp. + xx + 236 + xxx pp. + 22 pls. New Delhi. 1952. [Dioscorea, 67-76, pl. 9; description, economic importance, and chemical composition of 11 species; D. alata, D. bulbifera.] ScHRETER, A. I., et a/. Nomenclature and distribution of Dioscorea from the Soviet Far East and the Sree - raw materials supplied by this plant. (In Russian.) Rastitel Resursy 1: 307-397. SCHUBERT, B. G. ze in : opto I: a collection from British Honduras. Jour. Arnold Arb. 47: 147-159. 1966. [D. Gaumeri, D. nse D. Sunda sp. nov. . Aspects of taxonomy in the genus Dioscorea. Publ. Espec. Inst. Nac. Investig. Forest. Méx. 8: 29-42. 1973. [Genera of Dioscoreaceae, evaluation and oe from Dioscorea, subgenera of Dioscorea, eee eee the genus; 5 figur & C. V. Morton. Studies in Dioscorea, Il; an unusual new ae from Mexico. Proc. Biol. Soc. Wash. 84: 445-448. 1971. [D. ng sp. n SEALE, C. C. Estudios con Dioscorea en el sur de la Florida. (English summary.) Pro Caribbean Reg. Am. Soc. Hort. Sci. 5: 56-58. 1961. [Cultivation of D. oe and D. Wena for sapogenins. ] SEIDEMANN, J. Mikroskopische Untersuchung verschiedener Dioscorea-Starken. Starke 16: — 19 — ne L., & B. V. GopaL. Some observations on the diversity of stomata and es in six species of Dioscorea. Ann. Bot. IL. 36: 997-1004. 1972. [D. alata, . oe D. bulbifera, D. hispida. D. oppositifolia, D. Wallichii.] SHAH, J. i Coenocytic vessel elements in Dioscorea alata. Nature 197: 1125. 1963. JNNIKRISHNAN, & K. V. POULOSE. Vessel members i in the stem of Dioscorea ae L. Canad. Jour. Bot. 45: 155-167. 1967. [Differences between vessels of nodes and internodes. | SHARMA, A. K., & D. N. De. Polyploidy in Dioscorea. Genetica 28: 112-120. 1956. [D. alata, 2n = 30, 40, 50, 70.] SHARMA, O. P. Anatomy, origin and development of the rhizome of Dioscorea deltoidea Wallich. Proc. Indian Acad. Sci. B. 84: 50-55. pls. 1-4. 1976a. ———. Ontogeny and morphology of bulbil and its seedling tuber in Dioscorea melano- phyma. Phytomorphology 28: 192-197. 1978. [Related papers on D. glabra in ibid. 24: 297-305. 1975, and 26: 87-96. 1976b. Ontogeny and morphology of the tuber of Dioscorea floribunda Mart. et Gal. Proc. Indian Acad. Sci. (Pl. Sci.) 89: 47-55. 1980. [Tubers result from the intensive unilateral activity of the primary thickening meristem in the hypocotyl.] Ontogeny of seedling rhizome in Dioscorea Prazeri. Phytomorphology 36: 229- 234. 1986. [Rhizome dev Ses eae from the first plumular internode, sympodial growth of rhizomes, origin of tuber SHRuUM, J. E., P. K. SopERHOLM, & J. R. HAUN, Preliminary studies on Dioscorea growth in Florida. Proc. Florida State Hort. Soc. 71: 315-317 8.* Sippia1, M. A. Dioscoreaceae. Fl. West Pakistan - 1-6. 1973. [Dioscorea, four species. ] SMALL, J. K. Manual of the southeastern flora. xxii + 1554 pp. New York. [Dioscorea (under Tamaceae), 324, 325; D. oe D. glauca, D. hirticaulis, D. quaternata, D. villosa.] SmitH, A. I. A guide to wildflowers of the Mid-South. vii + 281 pp. Memphis, Tennessee. 1979. [D. villosa in Georgia and Alabama, 78: color photograph.] SMITH, B. W. Notes on the cytology and distribution of the Dioscoreaceae. Bull. Torrey 1989] AL-SHEHBAZ & SCHUBERT, DIOSCOREACEAE 93 Bot. Club 64: 189-197. 1937. [Dioscorea; chromosome numbers of 11 species; base number for i Ce is ten: sex chromosomes present. SmituH, E. B. An and annotated list of the vascular plants of Arkansas. iv + 592 pp. eee an 1978. an ie (as D. batatas), D. villosa (including D. hirticaulis and D. quaternata), 421.] SMITH, P. M. The development of the ae and seedling of Dioscorea villosa. Bull. Torrey Bot. ee 43: 545-558. pls. 31-34. 1916. SODERHOLM, P. K., et a/. Yield trials of steroid-producing Dioscorea on Florida’s Ev- erglades peat ls Econ. Bot. 22: 80- ae rae [D. composita, D. ee Sosa, V., B. G. SCHUBERT, & A. GOMEz-Pompa. Dioscoreaceae. Fl. Veracruz 53: 1-46. 1987. [Dioscorea, 18 species; descriptions, Astabutibas habitats, illustrations, ae SPARSHOTT, E. N. Observations on the formation, development, and structure of the tuber of Testudinaria elephantipes, and on the origin of the vegetative shoot. Jour. Linn. Soc. Bot. 49: 593-610. p/. 32. 1935. [D. elephantipes.] STANDLEY, P. C., & J. A. STEYERMARK. ee eae. Fl. Guatemala. Fieldiana Bot. 24(3): 145-159. 1952. [Two introduced and 22 indigenous species of Dioscorea.] STEYERMARK, J. A. Flora of ea Ixxxill + 1725 pp. Ames, Iowa. [1963.] [D. quaternata, D. villosa, 454, 4 STRAUSBAUGH, P. D., & E. L. ees Flora of West Virginia. ed. 2. xl + 1079 pp. Grantsville, West Virginia. 1978. ee 290.2513 descriptions and illustrations of D. oppositifolia (as D. batatas), D. quaternata, and D. villosa.] Su, P. Pollen morphology of Dioscorea in China. (In Chinese; English summary.) Acta Phytotax. Sinica 25: 357-365. pls. 1-4. 1987. [Light and scanning-electron micros- copy of pollen of 33 species in five sections.] TAKAMI, W. On the formation of calcium oxalate crystals in the young leaf of Dioscorea japonica and Zingiber Miyoga. (In Japanese, English summary.) Bot. Mag. Tokyo 79: 350-354. 1966. [Formation of crystals promoted by adding glucose to culture medium and retarded by supplying oxygen, glutamic acid, or methionine.] TAKEDA, K. The steroidal sapogenins of the Dioscoreaceae. Pp. 287-333 in L. REINHOLD & Y. Liwscuitz, eds., Progress in phytochemistry. Vol. 3. London, New York, Sydney, and Toronto. 1972. [Structure, blosynthess, and distribution of steroidal sapogenins in Dioscorea, chemotaxonom TAKEUCHI, Y. Embryo sac development in Dinseaes Japonica Thunb. and D. guinque- loba Thunb. Acta Phytotax. Geobot. 25: 57-60. 1972. [Ovules anatropous, bitegmic, crassinucellate; embryo sac of the Sy sae type. ] ; o, & A. AKAHORI. Chromosome numbers of some Japanese Dioscorea species. hee Phytotax. Geobot. 24: 168 173. 1970. [Eight species. ] TAN, A. S., .N. RAo. Studies on the developmental anatomy of te a Sansi- harens Pax (Dioscoreaceae). Bot. Jour. Linn. Soc. 69: 211-227. pls. 1, 2 74 TATTERSFIELD, F., J. ARTIN, & F. N. Howes. Some fish-poison plants i their insecticidal properties. Bull. Misc. Inf. Kew 1940: 169-180. 1940. [D. piscatorum, 173.] TEICHMAN UND LOGISCHEN, I. VON, & P. J. ROBBERTSE. Die anatomie en ultrastruktuur van die ekstraflorale nektarkliere van Dioscorea sylvatica Eckl. en die samestelling van die nektar. (English summary.) Jour. S. Afr. Bot. 45: 63-77. 1979, N DER SCHIFF, & P. J. ROBBERTSE. The genus Dioscorea L. in South Aiea, Boissiera 24a: 215-224. 1975. [Distributions aa habitats of 17 species; schematic representations eee characters. TELFORD, I. R. H. Dioscoreaceae. Jn; A. S. GeorGe, ed., Fl. Australia 46: 196-202. 1986. [Dioscorea, five species maps, p. 21 io TELLEz, O., & B. G. ScHuBERT. Una nueva especie del género Dioscorea (Dioscoreaceae) del Estado de eee MCCS Ann. Missouri Bot. Gard. 74: 539-541. 1987. [D. Matudae, sp. no 94 JOURNAL OF THE ARNOLD ARBORETUM [vVoL. 70 Tina, C.-T., & M.-C. CHANG. New taxa of Dioscorea from China. (In Chinese and Latin.) Acta ee Sinica 20: 205-209. 1982. & . Linc. Dioscoreaceae. /n: C. Per & C.-T. Tina, eds., Fl. Reipubl. Pop. Sinicae 16(1): 54-120. 1985. [Dioscorea, 49 species in six sections; keys, de- scriptions, illustrations. ] Tycer, V. E., L. R. Brapy, & J. E. Rea Pharmacognosy. ed. 7. x + 537 pp. Philadelphia. 1976. [Dioscorea, 91, 216, 217. Une, E. B. Dioscoreae mexicanae et et americanae. Bot. Jahrb. 22: 421-432. 1896. [Enumeration and descriptions of 37 species. ——. Dioscoreaceae. Jn: A. ENGLER & K. PRANTL, Nat. Pflanzenfam. Nachtrage zu I]. 5: 80-87. 1897 : ia nbohanie: A new genus, and other new Dioscoreaceae. Publ. Field Mus. Bot. 1: 413-418. pi 22. 1899. onan of four seeds in each locule of the fruit is the key difference between the new genus and Dioscorea, see SCHUBERT, 1973.] Unprg, A. S., & P. I. AKUBUE. ie evaluation of Dioscorea dumetorum tuber used in traditional antidiabetic therapy. Jour. Ethnopharmacol. 15: 133-144. 1986. [Experimentation on mice and rabbits suggests the possibility of obtaining tuber mee that could act in conditions of severe lack of insulin.] Upnor, J. C. T. Dictionary of economic plants. ed. 2. 591 pp. New York and Codicote, England. 1968. [Dioscorea, 180-182; economic importance of 36 species.] Waitt, A. W. Yams, Dioscorea species. Field Crops Abstr. 16: 145-157. 1963.* WALENS, H. A., S. SeRotTA, & M. E. WALL. Steroidal sapogenins xxx1. Gentrogenin and Se aa new sapogenins from D. spiculiflora. Jour. Am. Chem. Soc. 77: 5196 5197, WALKER, . ee Dioscorea’s du Gabon. Revue Bot. Appl. Agr. Trop. 32: 191-193. 1952. [Enumeration of edible and toxic species, medicinal importance.] Wa tL, M. E., C. S. Fenske, H. E. KENNEY, J. J. WILLAMAN, D. S. CorreLt, B. G. SCHUBERT, & H. S. GENTRY. Steroidal sapogenins XLIII. Survey of plants for steroidal sapogenins and other constituents. Jour. Am. Pharm. Assoc. 46: 653-684. 1957. [Dioscorea, 663-670, 683, 684. Related papers in ibid. 43: 1-7, 503-505. 1954; 44: 438-440, 1955; — 695-722. 1959. _ J. W. Garvin, J. J. WILLAMAN, Q. Jones, & B. G. SCHUBERT. Steroidal sapogenins LX. Survey of plants for ene eo a other constituents. Jour. Pharm. Sci. 50: 1001-1034. 1961. [Dioscorea, 1002- Warb, D. B. Keys to the flora of Florida—5, Bee Phytologia 38: 151-154. 1977. [D. alata, D. bulbifera, D. floridana, D. quaternata, D. oppositifolia (as D. batatas) and D. villosa excluded from the flora of Florida.] Watt, J. M., & M. G. BREYER-BRANDWIJK. The medicinal and poisonous plants of southern and eastern Africa. ed. 2. x11 + 1457 pp. Edinburgh and London. 1962. a 383-387.] Wel, H., C. G, & K. Xu. The application of cluster analysis on the species level for the at Bon ‘orea L. sect. Stenophora Uline in China. (In Chinese; English sum- mary.) Acta Bot. Yunnanica 5: 231-237. 1983. [Numerical analysis of morpholog- ical, cytological, palynological, and chemical characters supports the recognition of 18 species in sect. ala WickHaM, L. D., L. A. Witson, & H. C. PAssAm. Tuber germination and early growth in four edible Die ae es. Ann. Bot. II. 47: 87-95. 1981. [D. alata, D. esculenta, D. Lat D. trifida, origin of roots, shoots, and organs of renewed growth.] WILLAMAN, J. J., . FENSKE, & D. S. CorreELL. Occurrence of alkaloids in Dioscorea. Science 118: 329. 330. 1953. [Alkaloids absent in the New World taxa and present in some of the Old World ones — & H.-L. Li. Alkaloid- bearing plants and their contained alkaloids, 1957-1968. “Bievdia 33(3A, supplement). vil + 286 pp. 1970. [Dioscorea, 87.] 1989] AL-SHEHBAZ & SCHUBERT, DIOSCOREACEAE O53 .G. ScHusert. Alkaloid-bearing plants and their contained alkaloids. U. S. Dep. Agr. Tech. Bull. 1234. 287 pp. 1961. [Dioscorea, Tamus, 81. Wixus, J. C. A dictionary of the flowering plants and ferns. ed. 8. (Revised by H ArrY SHAW.) xxil + 1245 + Ixvi pp. Cambridge, England. 1973. Dee oi 369.] WILson, J. E. Progress in a breeding of ae Dioscorea spp. Pp. 17-22 in J. MIEGE & S. N. Lyonaa, eds., Yams. Oxford. WONDERLIN: R. P. Guide - the vascular ae “OE central Florida. 472 pp. Tampa and other cities. 1982. [D. alata, D. bulbifera, D. floridana, 141. XIFREDA, C. C. Estudios en Dioscoreaceae. I. (English summary.) Bol. Soc. Argent. Bot. 20: 317-327. 1982a. [D. Haumanii, sp. nov.; D. furcata recorded for Argentina; D. ceratandra. | Estudios en Dioscoreaceae. II. (English summary.) /bid. 21: 273- i 1982b. [Seven species of Dioscorea; descriptions, distributions, and illustratio Estudios en Dioscoreaceae. III. (English summary.) /bid. 22: 337-344, 1983. [D. Hasslerana, D. pilcomayensis.] . Estudios en Dioscoreaceae. IV. (English eae Kurtziana 17: 119-123. 1984. [D. Demourae, D. spectabilis, new to Argentina.] . Una nueva especie de Dioscorea L. ioe ‘ed a (English sum- mary.) Darwiniana 26: 371, 372. 1985. [D. Hunzikeri, sp Yakuwa, T., T. HARADA, N. Kasat, H. ARAKI, & I. OKUYAMA. eects on the botanical characteristics of genus Dioscorea. 1. Observation of the flower, fruit and seed of female plants of Chinese yams (Dioscorea opposita Thunb. cv. Nagaimo). (In Jap- anese; English summary.) Mem. Fac. Agr. Hokkaido Univ. 12: 271-280. 1981. TUCKER, COMMELINACEAE a} THE GENERA OF COMMELINACEAE IN THE SOUTHEASTERN UNITED STATES! GORDON C. TUCKER? COMMELINACEAE R. Brown, Prodr. 268. 1810, ‘“Commelineae,”” nom. cons. (SPIDERWORT FAMILY) Perennial or annual, mucilaginous herbs [twiners]. Roots fibrous or thick- ened. Stems soft. Leaves several, cauline or basal; sheaths closed; blades flat to V-shaped, the margins often with short prickles; vernation involute or su- pervolute. ee eae 1 to several, variously thyrsiform or cymose, terminal or axillary, sometimes subtended by spathes. Pedicels slender. Flowers perfect [imperfect], Boece eae a or zygomorphic, unscented [strongly or weakly € et if present, 2—4; filaments glabrous or hairy; anthers narrowly oblong, dorsifixed, tetrasporangiate, introrse [extrorse or poricidal]; pollen ovoid, monosulcate [trisulcate], binucleate [trinucleate]. Ovary syncarpous, [2- or] 3-locular; ovules 'Prepared for the Generic Fl f the Sout! United States, a long-term project made possible by grants from the National Science Foundation and al this wae SUBpOLEG: ey BSR-8415367 (Norton G. Miller, principal investigator), under which this d, and BSR-8415769 (Carroll E. Wood, Jr., principal investigator). This treatment, 128th i in the § series, ae the format established in the first paper r (Jou r. Arnol d Arb. 39: 296-346. 1958) and continued to the present. plants of this area, with information about extraregional members of a family or genus in brackets References that I have not verified are marked with an asterisk. I have continued to enjoy working with Norton Miller and Carroll Wood on the Generic Flora project, and I thank them for their ae and advice. Ihsan A. Al-Shehbaz and Barbara Nimblett have helped frequently. Stephen A. Spongberg and Elizabeth B. Schmidt improved the final manu- script with their editorial precision and expertise. Thanks are extended to the staffs of the New York following ee ie pee sent eee or a ided access to collections a hospitality pn my visits: A, ALU, CON . NEBC, NHA, NYS, UNA Irene re Bea ne lint en of Callisia tee Carroll Wood’s direction from material that R. J. Eaton collected in Citrus County, Florida, and that Wood cultivated at the Arnold Arboretum. Dorothy H. Marsh prepared the illustration of Commelina under R. B. Channell’s direction from material Mrs. Marsh collected in Cambridge, Massachusetts, and Wood collected in Ayer, Massa- chusetts. Contribution number 590 from the New York State Science Servic ?Biological Survey, New York State Museum, The State Education mete Albany, New York 230. © President and Fellows of Harvard College, 1989. Journal of the Arnold Arboretum 70: 97-130. January, 1989. 98 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 1-3 per locule, orthotropous. Fruit a loculicidal capsule [berry or nutlet]. Seeds small, circular to oblong, flattened laterally [winged], often reticulately pitted [arillate]; endosperm abundant; embryo small, broadly capitate, beneath a disclike callosity on seed coat. TyPpE GENUS: Commelina L. A medium-sized family of 40-50 genera and 500-600 species. The largest genus is Commelina, with about 230 species. The family is pantropic and warm temperate, with centers of diversity in Mesoamerica, tropical Africa, and south- ern India. There 1s general agreement that the family most closely related to the Commelinaceae 1s the Mayacaceae (see Thieret, Tomlinson). No consensus exists, however, concerning suprageneric classification in the family: although two tribes have commonly been recognized, their circumscriptions have varied. The system of Faden (1985) is followed here. The Tradescantieae C. B. Clarke are characterized by stomata with four subsidiary cells (except two in Callisia sect. CUTHBERTIA (Small) Hunt); inflorescences consisting of basally fused pairs of cincinni or thyrses of cincinni subtended by large to small leaflike or modified spathes; generally actinomorphic and most often self-incompatible flowers; a gynoecium of three locules, each with two ovules; and six stamens, all fertile. This chiefly New World tribe is represented in the United States by three genera, all with species present in the Southeast: 7Tradescantia L. (including Zebrina Schnizl., Setcreasea K. Schum. & Sydow, and Rhoeo Walp.), Callisia Loefl. (including Cuthbertia Small and Tradescantella Small), and Gibasis Raf. Representatives of tribe Commelineae have stomata with six subsidiary cells (Tomlinson, 1969). Flowers are often zygomorphic and are generally self-com- patible. There are usually only three stamens and three (four) or no staminodia. The Commelineae are most diverse in the Old World tropics. Two genera, Commelina and Murdannia Royle, are represented in the United States, both in the Southeast. Clarke (1881) recognized three tribes: the Pollieae C. B. Clarke (fruits in- dehiscent), the Tradescantieae (stamens 5 or 6, staminodia lacking; fruits cap- sular), and the Commelineae (stamens 2 or 3, staminodia lacking or 1-4; fruits capsular). Briickner (1930) followed Clarke but treated the last two tribes as subfamilies (with the Pollieae included in the Tradescantioideae Briickner). Pichon, in an apparent frenzy of taxonomic splitting, recognized ten tribes and accorded familial status to the Australian genus Cartonema R. Br. Brenan (1966) recognized 15 groups to which he assigned no rank, although several of them correspond to tribes recognized by Pichon. Brenan’s insightful (but in- formal) classification pointed out morphological similarities among groups of genera; these have later been confirmed by anatomical or cytological studies. Hutchinson recognized no suprageneric ranks in the family, noting that rela- tionships of genera were poorly understood and largely defined by “Linnean” characteristics such as stamen number. Woodson emphasized the importance of inflorescence characters for classification. He noted that some genera placed by Briickner (1930) in the supposedly “‘zygomorphic” subfam. Commelinoi- deae actually had actinomorphic flowers. Faden (1985) recognized two subfam- ilies, the Cartonematoideae* and the Commelinoideae, the latter with two 1989] TUCKER, COMMELINACEAE Se tribes, the Tradescantieae and the Commelineae, a classification followed here. Recent diverse studies (discussed below) have provided eclectic information on the distribution of characters in the family. So far, these data have not been brought together and used for a synthetic classification based on evolutionary relationships. In the present account, broad generic concepts are followed. This seems prudent both in view of the difficulty in defining genera and because of the diverse circumscriptions given to many of them (e.g., Callisia and Tradescan- tia). The generic classification of the Tradescantieae following Hunt is accepted here. Thus, Tradescantia is defined by its conspicuous, paired spathes and its fused cincinni. It is amply distinguished from Tripogandra Raf., which has zygomorphic flowers and dimorphic pollen (Handlos, 1975; Mattson). The various satellite genera near Tradescantia, such as Rhoeo and Zebrina, are treated as sections, as this seems a better indication that they differ less from sect. TRADESCANTIA than does Tripogandra. Callisia includes Phyodina Raf. and Cuthbertia. The genus as so circumscribed is defined by its reduced bracts and solitary cymes. Gibasis presents comparatively little problem in circum- scription. It has paired (but unfused) cymes, with bracteoles and flowers ar- ranged distichously on a conspicuous axis. The American species included in Aneilema R. Br. by recent authors (e.g., Fernald; Radford er al.) are placed here under Murdannia (q.v.). Various workers agree that Cartonema is the most anomalous genus in the family, differing from all other Commelinaceae in its spiciform raceme, its lack of glandular microhairs and of raphides, and its poorly developed embryo and embryostega. Tomlinson (1969) has brought to light some additional anatom- ical differences (lack of transverse leaf veins and stem collenchyma, presence of radiate chlorenchyma in leaves) that support the unique status of Cartonema. Triceratella Brenan, a genus endemic to southern Africa, has similarities both to Cartonema and to other genera of the family. Like Cartonema, Tri- ceratella lacks glandular microhairs. However, it has raphide canals in the leaves, although the canals are adjacent to the veins, while in the remainder of the family they are not. Tomlinson (1969) suggested (but did not validly publish) subfamilial status for Triceratella. Vessels are present in the vascular tissue of roots, stems, and leaves in all genera of the family that have been investigated except Cartonema, where they are present only in the roots (Tomlinson, 1969). The vessels have simple perforation plates. Sieve-tube plastids are of the “monocotyledonous type,” in which there are several to many cuneate proteinaceous crystalloids. Although silica bodies are generally absent in the Commelinaceae, specialized epidermal cells containing silica are known from Callisia (including Hadrodemas Moore), Coleotrype C. B. Clarke, Forrestia A. Rich., Gibasis, and Tripogandra. Stomatal types have been surveyed by Tomlinson (1969). Both hexacytic and tetracytic stomata are known. (In the Mayacaceae they are paracytic.) Se dled hai ae R. Faden ex G. Tucker, stat. nov., based on Cartonemataceae Pichon, Not. Syst. Paris 12: . 1946. 100 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 The family has received much attention from cytologists; the chromosomes are generally large and the meiotic material easy to work with. Chromosome numbers have been published for all but six genera, and numbers for one third of the species are in the literature. The family shows a wide range of variation in base number and other features of the karyotype. Base numbers vary from four to 20, including nearly every value between. The chromosomes of most karyotypes are generally alike in size. Species of the genus Stanfieldiella Brenan are a notable exception: their chromosomes are small, but there is typically one in the complement that is twice as long as the others. The species of Stanfieldiella have the smallest chromosomes (2-3 um), and Tradescantia vir- siniana L. and its allies the largest (about 12-14 um). Most genera have asym- metric karyotypes—1.e., most of the chromosomes have an acrocentric or telo- centric kinetochore. In this family a genus is usually characterized by a single base chromosome number and a particular karyotype. The lowest base chromosome number recorded (4, 5) is for species of Gibasis and the highest (20) for Palisota Rei- chenb. Each of the tribes (groups used by Brenan) shows considerable chro- mosomal diversity, which in some cases argues against their naturalness. Polyploidy has been detected in several genera. Usually the taxa are tetra- ploids, although higher levels have been reported. Polyploidy is also known within species. Chromosomal change is a common feature of speciation in the family. Closely related species often differ in chromosome number or in karyo- type. The family is characterized by the presence of various flavonols and glu- coflavones, often in quantity. Syringic acid and cyanidin are common. Sulfated phenolic acids have been detected in species of Coleotrype, Commelina, Gibasis, Tradescantia, and Tripogandra. Quercetin 3-glycoside is also widespread in the family but is most common in the genera of the Commelineae (Del Pero de Martinez). Steroidal saponins are known only from Cyanotis D. Don. Some 30 species have been examined for cyanogenesis, which is apparently rare (reported from Commelina and Tinantia Scheidw.). In T. erecta Schlecht. the cyanogenic glucoside is taxiphyllin (Tjon Sie Fat). The report of cyanogenesis in Commelina benghalensis L. by Juliano is considered doubtful (Tjon Sie Fat). Calcium-oxalate raphides in raphide canals are common; they are generally found in the members of the Commelinaceae but are absent in the Cartone- matoideae and the Mayacaceae. Commelinin, a magnesium-containing me- tallo-anthocyanin, gives the blue color to the petals of Commelina species (Goto ef al.; Hayashi). Zantovska Stirton & Harborne investigated anthocyanins in 28 species rep- resenting seven genera (as recognized herein following Hunt, 1975; 1980; 1986a, b, d). All species were found to contain acylated glucosides, a rare class of anthocyanins not reported from any other monocotyledonous family. In Cy- anotis, Dichorisandra Mikan, Gibasis, Tradescantia (including representatives of the segregate genera Campelia L. C. Rich., Rhoeo, and Zebrina), and Tri- pogandra, the most abundant anthocyanin was cyanidin 3,7,3'-triglucoside. In Commelina (seven species tested), however, the most abundant was p-cou- maroyl-delphinidin 3,5-glucoside (commelinin). Thus there are two different 1989] TUCKER, COMMELINACEAE 101 anthocyanins responsible for blue petal color in the family. It would be en- lightening to learn whether other genera purportedly related to Commelina (e.g., Murdannia, Aneilema) also contain commelinin. It would also be inter- esting to identify the compounds responsible for the yellow petal color of such species as C. africana L Cleistogamous flowers are produced by certain species of Commelina, Mur- dannia (both Commelineae), and 7inantia (Tradescantieae). There are six fertile stamens per flower in more than half the genera of the family (including the Cartonematoideae). Numerous genera have only three stamens, these alternating with or opposing in an adaxial-abaxial situation two to four staminodia. There is a single functional stamen in some species of Callisia, Murdannia, and Pseudoparis H. Perr. The filaments are hairy in many genera such as Cartonema, Cyanotis, Tinantia, and Tradescantia (including sect. RHOEO), and the filaments are fused in Cochliostema Lem. Anthers are tetrasporangiate (vs. bisporangiate in the Mayacaceae) and generally introrse. A few taxa are characterized by extrorse dehiscence, but Dichorisandra and Porandra Mikan (and members of the Mayacaceae) have poricidal anthers. The anther walls have girdle-type thickenings, and the tapetum is plasmodial. The pollen grains are always binucleate (seven genera studied by Owens) except in Floscopa Lour., where they are reported to be trinucleate (Davis). Poole & Hunt surveyed pollen morphology with scanning electron microscopy. All grains are monosulcate, except those of one species of Tinantia, which are trisulcate. Pollen dimensions are variable within and among genera. Four main types of ornamentation are encountered in the Tradescantieae and five in the Com- melineae. The tectum tends to be insulate in species with actinomorphic flowers and spinulate, verrucate, or tuberculate in those with zygomorphic ones; in some species of Callisia and of Gibasis it is finely reticulate. All species of Tradescantia (including Rhoeo and Zebrina), the remaining species of Gibasis, and Callisia sects. HADRODEMAS (Moore) Hunt and CUTHBERTIA are char- acterized by the ““7radescantia-type”’ tectum (1.¢., insulate, the insulae forming a cerebroid pattern). Species of Commelina and Aneilema have a perforate tectum bearing spinules; those of Murdannia have tubercles. Species of 7ripogandra are characterized by dimorphic pollen (Handlos) produced by two kinds of stamens: monosulcate, spheroidal, and fertile by the three anthers on short filaments; anasulcate, oblong, and infertile by the three on long filaments (Mattson). Both kinds of pollen have a granular-verrucate tectum interspersed with micropores (Poole & Hunt). Placentation is axile in the Commelinaceae (vs. parietal in the Mayacaceae). The ovules of most genera are orthotropous, although they are hemianatropous in some and anatropous in Cartonema (orthotropous in the Mayacaceae). A parietal cell is generally present in the family but is known to be lacking in Commelina and Cyanotis. Megagametophyte development is of the Polygonum type in most genera but of the Allium type in 7radescantia. The ovary is syncarpous, and the style 1s three-branched. Aneilema, Callisia, Commelina, and Tradescantia have dry stigmas, while Cyanotis, Gibasis, Thyrsanthemum Pichon, and We/denia Schultes f. have wet ones (Heslop-Harrison & Shivanna). Endosperm formation 1s nuclear (several genera reported). Embryogeny con- 102 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 forms to the Asterad type. Endosperm 1s abundant and starchy. The mature embryo is capitate, broad, and achlorophyllous. Fruits are capsules in most genera, except in Athyrocarpus Schlecht., Pollia Thunb. (both nutlets), and Palisota Reichenb. (berries). A few genera (Amis- chotolype Hassk., Dichorisandra, Porandra) have arillate seeds. Pollination biology is little known, considering the size of the family. Faden (1983) studied isolating mechanisms in Anei/ema. At a site in lowland Kenya where five species were growing in a small area, no hybrids were found. Some species pairs were separated by time of flower opening (morning vs. early afternoon), others by soil preference (well-drained sandy soils vs. waterlogged ones). Bees of several genera visited the flowers. The size of the flowers deter- mined the bee that could gather pollen, thereby enforcing pollinator constancy. In certain species the lateral stamens were covered by the vernate petals, while in others the lateral and medial stamens were exposed simultaneously. Faden hypothesized that species in which the lateral stamens are covered longer would have a greater likelihood of outcrossing, since bees could gather less pollen from a flower with one or two anthers exposed than from one with three and would thus visit more flowers in a collecting trip. Numerous species from several genera are worthy horticultural subjects. Several species of Tradescantia are cultivated in gardens in the United States and southern Canada. Trailing species of Commelina are used as ground covers. The most commonly cultivated taxa are 7. virginiana and its hybrids with 7. subaspera Ker or T. ohiensis Raf. (7. x Andersoniana,” nom. inval.; see Hunt, 1984), all easily grown in well-drained garden soils. Tradescantia Zebrina Bosse (Zebrina pendula Schnizl.), wandering Jew (a name also applied to 7. flumi- nensis Vell. and T. albiflora Kunth), is commonly grown as an indoor hanging- basket plant. The common cultivar is ‘Purpusii’. All the above species are readily propagated by division or from seed. Some species of the African genus Palisota are occasionally grown indoors. Murdannia nudiflora (L.) Brenan and several species of Commelina are se- rious weeds in numerous warm-temperate and tropical countries worldwide. Commelina species, especially C. benghalensis, are persistent and not easily controlled by herbicides (Wilson). The tuberous roots of certain western African species of Anei/lema contain abundant starch and are eaten (Augier), while those of Commelina coelestis Willd. (Sturtevant), C. graminifolia HBK., and C. tuberosa L. have been a source of food in Mexico. The young stems and leaves of C. communis L. (Sturtevant; Gibbons & Tucker), C. diffusa Burman f., and (according to Tana- ka), C. benghalensis are edible as a cooked vegetable. The tender young shoots of Tradescantia occidentalis were eaten uncooked by Indians of the Great Plains (Yanovsky); the leaves and new shoots of 7. virginiana L. can be similarly used (Fernald & Kinsey). The fresh or dried leaves of 7radescantia Zanonia (L.) Sw. (Hegnauer) and Commelina communis (Preston) have been employed as a treatment to stop the flow of blood from surface wounds. A Mexican species (perhaps C. virginica) has similar effects and has also been reported to cure cancer of the stomach 1989] TUCKER, COMMELINACEAE 103 (Herrera). Commelina virginica and T. elongata G. F. W. Meyer have been employed as diuretics in Latin America. The medically active principle in these species has apparently never been isolated. REFERENCES: Auaier, J. Les glucides de quelques monocotylédones de l’ordre des Farinosae. Bull. Soc. Bot. France 103: 251- ie 6. BarLey, L. H., E. Z. BAILEY, & Sta F BaiLey Hortorium. Hortus third. xiv + 1290 pp. New York. 1976. Ceres 301.] BRENAN, J. P. M. Tricerate/la, a new genus of Commelinaceae from Rhodesia. Kirkia 1: 14-19. 1961. . The classification of the Commelinaceae. Jour. Linn. Soc. Bot. 59: 349-370. 1966. [Numerous illustrations. ] BRUCKNER, G. Beitraége zur Anatomie, Morphologie, und Systematik der Commelina- ceae. Bot. Jahrb. 61(Beibl. 137): 1-70. p/s. vii. 1926. . Zur speziellen Systematik der Commelinaceae. Notizbl. Bot. Gart. Berlin 10: 55-61. 1927. —. Commelinaceae. Jn: A. ENGLER, ed., Nat. Pflanzenfam. ed. 2. 15a: 159-181. 930. Calus, J. The medicinal spiderworts of India. Jour. Bombay Nat. Hist. Soc. 39: 361- 365.1937, CELARIER, R. P. Cytology of the Tradescantieae (Commelinaceae). Bull. Torrey Bot. Club 82: 30-38. 1955. CnHANT, S. R. Commelinaceae. Pp. 279-281 in V. H. HEywoon, ed., Flowering plants of the world. New York. 1978. CHEADLE, V. I., & H. Kosakr. Occurrence and specialization of vessels in Commelinales. Phytomorphology 30: 98-117. 1980. CLARKE, C. B. Commelinaceae. Jn: A. & C. DE CANDOLLE, eds., Monogr. Phanerog. 3: 113-324. 1881. CLEWELL, A. F. Guide to the vascular plants of the Florida Panhandle. viii + 605 pp. 72.] Davis, G. L. Systematic embryology of the angiosperms. vil + 528 pp. New York, London, and Sydney. 1966. [Commelinaceae, 87, Det PERO DE MarTiNnEz, M.A. L Los flavanoides de las Commelinaceae y su interpretacion con otras Sea Monocotiledoneas. (Abstract.) Resumenes Jornadas Argent. Bot. 18: 72, 73. Fapen, R. B. A Beate study of the genus Anei/ema R. Br. (Commelinaceae). xil + 626 pp. Unpubl. Ph.D. dissertation, Washington Univ., St. Louis. 1975. —. Pollia Thunb. (Commelinaceae): | a Oaees record from the New World. Ann. Missouri Bot. Gard. 65: 676-680. . Isolating mechanisms among five espa species of Aneilema R. Br. (Com- melinaceae) in Kenya. Bothalia 14: 1002. 1983. Commelinaceae. Pp. 381-387 in R. M. T. on , H. T. CLirForp, & P. F. Yeo. The families of the sSpnenoiae Berlin, Heidelberg, New York, and Tokyo. 1985. Comparative vegetativ d ductive Ses of forest and non-forest Com- melinaceae. (Abstract.) Am. Jour. Bot. 73: 763. 1986. & D.R. Hunt. Reunion gee Se, ae with Commelina (Commelinaceae). Ann. Missouri Bot. Gard. 74: 121, 122. — & Y. Supa. Cytotaxonomy of Commelinaceae: ea numbers of some African and Asiatic species. Bot. Jour. Linn. Soc. 81: 301-325. 1980. 104 JOURNAL OF THE ARNOLD ARBORETUM [VoOL. 70 FERNALD, M. L. Gray’s manual of botany. ed. 8. Ixiv + 1632 pp. New York. 1950. [Commelinaceae, 392-395. . Kinsey. Edible wild plants of eastern North America. (Revised by R. C. ROLtms, 1958.) xi1 + 452 pp. New York and Evanston, Illinois. 1943. [Com- melinaceae, 122-124. Gippons, E. T., & G. C. Tucker. Euell Gibbons’ handbook of edible wild plants. [xi +] 319 pp. bleeee Beach, Virginia. 1979. Goto, T., T. Ho o, & S. TAKAsE. A proposed structure of commelinin, a sky-bl anthocyanin Cees obtained from the flower petals of Commelina. Tetrahedron Lett. 31: 2905-2908. 1979. GUuERVIN, C., & C. LE Cog. Karyology of Commelinaceae; application to 7 prob- lems relative to their evolution. Rev. Cytol. Biol. Veg. 29: 267-328. 6. Hamann, U. Character stability and relationships of feta a ee to the system of monocotyledons. Willdenowia 2: 639-768. HANDLOos, W. L. Cytological investigations of some bet ae from Mexico. Bai- leya 17: 6-33. 1970. . The taxonomy of 7ripogandra (Commelinaceae). Rhodora 77: 213-333. 1975. HAYASHI, K. Further experiment on the structure of commelinin—an approach to the problem of flower color variation. (Abstract.) XI Int. Bot. Congr. Abstr. 87. 1969. HEGNAUER, R. Chemotaxonomie der Pflanzen. Band 2. Monocotyledoneae. 540 pp. Basel and Stuttgart. 1963. [Commelinaceae, 117-121. Herp, Y., & D. J. BEADLE. The site oe mechanism in Tradescantia pallida. Ann. Bot. II. 45: 251-256. 1980. HERRERA, A. Yerba de pollo. Am. ras Pharm. 69: 290-294. 1897. [Mexican species of Commelina.] pane ae ARRISON, Y., & K. R. SHIVANNA. The receptive surface of the angiosperm a. Ann. Bot. Il. 41: 1233-1258. 1977 Te eee M. H. Notes on ae rad of Commelina communis L. Contr. Bot. Lab. Univ. Penn. 3: 275-287. Hom, L., J. V. PANcHO, J. P. Aes. & D. L. PLuckNetr. A geographical atlas of world weeds. (Arabic, Chinese, English, French, German, Hindi, Indonesian, Japanese, Russian, and Spanish introductions.) xliv + 391 pp. New York and other cities. 1979. , K. S. Morphological study of Commelinaceae in Taiwan. |. Inflorescence and oe (In Chinese.) Quart. Jour. Chin. Forest. 5(4): 87-95. mn [Illustrations.] Hunt, D. R. The reunion of Setcreasea and Separotheca with Tradescantia. American Commelinaceae: I. Kew Bull. 30: 443-458. 1975a. [Includes distribution map.] otes on Gibasis. American Commelinaceae: I]. /bid. 709-714. 1975b. Further notes on Gibasis. American Commelinaceae: V. bid. 33: 145, 146. 1978. . Sections and series in Tradescantia. American Commelinaceae: IX. Ibid. 35: 437-442. 1980. [Broad generic concept.] . New taxa in Commelinaceae. Jour. nate Soc. Bot. 83: 137-139. 1981. ; a ey eee Pp. 25-31 inS.M TERS ef al., eds., The European garden flora. II. Monocotyledons (part II). oe England, and other cities. 1984. (Callisia, Cochliostema, Commelina, Cyanotis, Dichorisandra, Geogenanthus, Gi- basis, sens Tinantia, Tradescantia, Tripogandra, Weldenia. sion of Gibasis Rafin. Kew Bull. 41: 107-129. 1986a. [Eleven species; stration of G. geniculata, G. pellucida.] npelia, Rhoeo, and eee united with 7radescantia. American Comme- nace XIII. Ibid. 41: 401-405. 6b. ee and a new eB ees Tradescantia. American Commelinaceae: as Ibid. 406. 986c. [7. exaltata (sect. MANDONIA), 2” = 12, from Oaxaca, Mexic : iseidatied of Callisia Loefl. American Commelinaceae: XV. Lbid. 407-412. 1986d. 1989] TUCKER, COMMELINACEAE 105 HutTcuHINnson, J. The families of flowering plants. ed. 3. xix + 968 pp. Oxford. 1973. [Commelinaceae, 692-700; Cartonemataceae, 700, 701. Jones, K., & C. JopLinc. Chromosomes and c lassification of Commelinaceae. Jour. Linn. "Soc. Bot. 65: 129-162. 1972. JuutANno, J. B. Additional cyanophoric plants in the Makiling region. Philipp. Agr. 11: 231, 232. 1923. Lewis, W. H. Meiotic chromosomes in African Commelinaceae. Sida 1: 274-293. 1964. Lona, R. W., & O. Laxeta. A flora of tropical Florida. xvii + 962 pp. Coral Gables, Florida. 1971. [Commelinaceae, 269-274. Martin, A.C. The comparative internal morphology of seeds. Am. Midl. Nat. 36: 513- 660. 1946. [Commelinaceae, 533, 535.] MattHews, J. F. A paper chromatographic study of North and South American species of the genus 7radescantia. Bot. Gaz. 127: 74-78. 1966. [Supports removal of Cuth- bertia MATISON, O. The development of dimorphic pollen in 7ripogandra (Commelinaceae). Pp. 163-183 in I. K. FerGuson & J. MuLLeER, eds., The evolutionary significance of the exine. London and New York. 1976. Matupba, E. Nuevas Commelinaceae de México y Guatemala. Anal. Inst. Biol. Méx. 26: 59-77. 1955. [New species of Callisia, Commelina, Setcreasea, and Trades- cantia. | . Las Commelinaceas mexicanas. [bid. 26: 303-432. 1956. [Keys, descriptions, illustrations. ] Mayer, A. M., & A. PoLJAKOFF-MaAyBER. The germination of seeds. ed. 3. ix + 211 pp. Oxford and other cities. 1982. [Tradescantia virginica (sic), 9; illustration of seedling showing hypogeal germination. ] Mirsukurt, Y., & K. KoBayasui. Cytological studies in Commelinaceae. IV. Chro- mosome numbers and karyotypes of some species. (In Japanese; English summary.) Kromosomo 51/52: 1734-1739. 1962 Morton, J. K. The Commelinaceae of West Africa: a biosystematic survey. Jour. Linn. Soc. Bot. 60: 167-221. Murthy, K. L. Gametogenesis and embryogeny in some Commelinaceae. Jour. Indian Bot. Soc. 17: 101-116. 1938. Owens, S. J. Self-incompatibility in the Commelinaceae. Ann. Bot. II. 47: 567-581. 1981. N. J. HorsFieLp. A light and electron microscopic study of stigmas in Aneilema and Commelina species (Commelinaceae). Protoplasma 112: 26-36. 1982. [Illus- trations. ] & F.M. Kimmins. Stigma morphology in Sr enn iat Ann. Bot. I. 47: 771-- 783. 1981. [Callisia, Commelina, Gibasis, Tradescant Parks, M. Embryo sac development and ae ae in Conaeinditie Pringlei. Bull. Torrey Bot. Club 62: 91-104. pis. 7, ¢ 35. Picuon, M. Sur les Commelinacées. Not. oe Paris 12: 217-242. 1946. [Keys to tribes and Cate detailed discussion of generic limits; three new genera from New World.] re M.M., & D.R. Hunt. Pollen morphology and the taxonomy of the Comme- inaceae: an | sh American Commelinaceae: VIII. Kew Bull. 34: 639-- 660. 1980. [Many Toceee ns.] ee G. K. A species of Commelina Am. Jour. Pharm. 70: 321-335. 1898. [Prob- bly C. communis; illustrat ene A.E., H. E. AHLEs, EC. R. Bett. Manual of the vascular flora of the Carolinas. Ixi + 1183, pp. Chapel Hill, North Carolina. 1968. [Commelinaceae, 268-271. RAGHAVAN, R. S., & S. R. Rotia. Cytological observations on Indian species of Com- melinaceae. Curr. Sci. Bangalore 30: 310, 1. Rao, R. S., R. S. RAGHAVAN, & R. V. MMATHY. Biosystematic studies on Indian Commelinaceae— oo pattern and evolutionary trends. Bull. Bot. Surv. India 12: 242-254. 1970. 106 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 ROHWEDER, O. Die Farinosae in der Vegetation von El Salvador. Abh. Geb. Auslandsk. 61: 98-178. 1956. Beitrage zur Blitenmorphologie und -anatomie der Commelinaceen mit An- sane as zur Begrenzung und Gliederung der Familie. Ber. Schweiz. Bot. Ges. 79: 9, 199-220, Row .ey,J.R. i fine structure of the pollen wall in the Commelinaceae. Grana Palynol. 2: 3-31. 1959. A.O. DAHL. The aperture of the pollen grain in Commelinantia. Pollen Spores 4: 221-232. 1962. SHARMA, A. K., & A. SHARMA. Further investigations on cytology of members of Com- melinaceae with special ta to the role of polyploidy and the origin of ecotypes. Jour. Genet. 56: 63-84. STEBBINS, G. L., & S. K. ae Developmental studies of cell differentiation in the epidermis of monocotyledons. I. Allium, Rhoeo, and Commelina. Develop. Biol. 2: 409-426. 1960. STURTEVANT, E. L. Sturtevant’s edible plants of the world. (U. P. Heprick, ed.) vii + pp. Albany, New York. 1919. TANAKA, T. Tanaka’s cyclopedia of edible plants of the world. [viii +] 924 pp. Tokyo. 976. THARP, B. C. igs a new genus of the Commelinaceae. Bull. Torrey Bot. Club 49: 269-275. p 1922. . ise eir Cia an evaluation of its generic status. Ibid. 83: 107-112. 1956 THIERET, J. i. The Mayacaceae in the southeastern United States. Jour. Arnold Arb. 56: 248-255. 1975. [Also useful at species level.] Tiwari, S. C., & B. E. S. GUNNING. Cytoskeleton cell surface and the development of invasive plasmodial tapetum in Tradescantia virginiana. Protoplasma 133: 89-99. 1986a. — & —. An ultrastructural cytochemical and immunofluorescence study of post-meiotic devciooment of plasmodial tapetum in 7Tradescantia virginiana and its relevance to the pathway of sporopollenin secretion. bid. Pie a . 1986b. & Colchicine inhibits plasmodium formation and dis ts pathways of sporopollenin secretion in the anther tapetum of 7radescantia ae Ibid. 115- 128. 1986c. TJON Ste Fat, L. A. Contribution to the knowledge of cyanogenesis in angiosperms. 8th communication. Cyanogenesis in Commelinaceae. Lloydia 41: 571-573. 1978. a P. B. eee data in the classification of the Commelinaceae. Jour. nn. Soc. Bot. 59: 371- 6. : CommlnalesZingberales Vol. 3 in C. R. MeTcALre, ed., Anatomy of the monocotyledons. Oxfor TROLL, W. Cochliostema oe aie Lem.; structure and habits with comparative outlooks on other Commelinaceae. (In German.) Beitr. Biol. Pflanzen 36: 325-389. 1961. WALKER, N. E. Cuticular comparison of genera cine eed x + 129 pp. Unpubl. Ph.D. dissertation, Univ. Missouri, Columbia. 1974. WILSON, A. K. Commelinaceae— a review of the fe eed biology, and control of the important weeds belonging to this family. Trop. Pest Managem. 27: 405-418. 1981. [Commelina benghalensis, C. diffusa, Murdannia nudiflora, an important paper. Woopson, R. E. Commentary on the North American genera of Commelinaceae. Ann. Missouri a Gard. 29: 141-154. 1942. WUNDERLIN, R. P. Guide to the vascular plants of central Florida. [vi +] 472 pp. Tampa and other . 1982. [Commelinaceae, 128-131.] 1989] TUCKER, COMMELINACEAE 107 YANovsky, E. Food plants of the North American Indians. 87 pp. Washington, D. C. 1936. ZANTOVSKA STIRTON, J., & J. B. HARBORNE. Two distinctive anthocyanin patterns in the Commelinaceae. i ea Syst. Ecol. 8: 285-287. 1980. KEY TO THE GENERA OF COMMELINACEAE IN THE SOUTHEASTERN UNITED STATES General characters: fibrous-rooted or tuberous perennials (annuals) of open or disturbed places in full sun or light shade; stems and leaves soft, exuding mucilage when bruised; leaves basal or cauline, th : closed, the blades linear, lanceolate, or broadly ellipsoid; inflorescences terminal or axillary and terminal, variously thyrsiform or cymose; flowers trimerous, pe erfect, actinomorphic or zygomorphic, sepals somewhat persistent, petals de- liquescent, staminodia sometimes present; fruits more or less dehiscent 2- or 3-locular capsules; seeds oblong, ellipsoid, or triangular, often operculate, the seed coat pitted or roughened. A. Fertile stamens 6; filaments hairy (rarely glabrous). B. Inflorescences subtended by conspicuous, paired, leaflike spathes; anther con- fectives trapezoidal Or thangular, <4 424414 40sas Geers 1. Tradescantia. B. Inflorescences subtended pease aa paired or solitary, scarious or scalelike bract(s); anther connectives rectangu Inflorescences dense, branches a epeuate or absent; cymes solitary, with- OULCONSPICGUOUS: AXIS: 22eaeu vy a eas ee ne ete ee eee 2. Callisia. C. Inflorescences open, branches conspicuous; cymes paired, with evident axis bearing 2-ranked orbiculate bracteoles. ..............-.+-++-- 3. Gibasi Fertile stamens 3, staminodia 3; filaments elabrou D. Spathes inconspicuous, scalelike; petals equal; eae alternating with stam1- nodia 4. Mu > rdannia. D. eae conspicuous, leaflike; petals unequal; stamens below staminodia. ..... 5. Commelina. SUPPLEMENTAL KEY EMPHASIZING VEGETATIVE FEATURES SPECIALLY FOR USE WITH DRIED SPECIMENS? Ae Spathes paired: terminating SCI « 4.0 «saci asada Haaw e's 6 Rea eien 1. Tradescantia. A. Spathes single, terminal or axillar Inflorescences unbranched, closely enclosed by spathe. a Spathes.composed of sheath Only 2.4 keane yan ieee ene 2. Callisia C. Spathes with oblong to pear blades, sheathless or nearly so. ......... es Md iets ead ate ond Baca ie als See Gon amelina. B. Inflorescences branched (or if unbranched, the individual flowers pedicellate, not closely eee by spathe). Leav Scaeyiminetnc BASALLY, 230-214 eden Seeks the eae 3. Gibasis. D. poe symmetrical. Ee Stems: Creeping. cs .0% vi oon 0b oh bee eee ee de ea 2. Callisia. E. Stems ascendent or erect. _ Leaves linear; stamens 6, staminodia lacking. ......... 2. Callisia. F. Leaves lanceolate; stamens 3, staminodia 3. ....... 4. Murdannia. ‘Provided b floral ch t listinguishi are not well preserved in dried specimens. 108 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 Tribe TRADESCANTIEAE Meisner ex C. B. Clarke in DC. Monogr. Phanerog. 3: 120. 1881. 1. Tradescantia Linnaeus, Sp. Pl. 1: 288. 1753: Gen. Pl. ed. 5. 139. 1754. Perennial herbs. Roots fibrous, fleshy, sometimes tuberous. Stems branched or simple, erect to ascendent [creeping], sometimes much shorter than the leaf blades. Leaves several, both cauline and basal [basal only]; sheaths closed; blades broadly lanceolate to linear, sometimes slightly constricted just above sheath, apex acute to broadly rounded, membranaceous or succulent, narrowly to broadly canaliculate to flat, glabrous or pubescent [villous]. Inflorescences | to several, terminal [and/or axillary], cymose, forming tight umbels, sub- tended by 2 opposed, linear to broadly lanceolate, oblong, or suborbiculate spathes (sometimes the uppermost internode so short that the uppermost leaf appears to be a third spathe). Pedicels slender, erect or spreading, reflexed after anthesis, sometimes colored like petals. Flowers several, perfect, actinomorphic [slightly zygomorphic]. Sepals 3, equal [the uppermost larger], oblong, broadly rounded to acute. Petals 3, ovate [obovate], not clawed [clawed], distinct or sympetalous (fused along edges to each other) [or fused to edges of filaments]. Stamens 6, identical, all fertile; filaments densely hirsute, trichomes monili- form; anthers ellipsoid, the connective broad, trapezoidal to triangular, the anther sacs reniform, dehiscent longitudinally. Ovary ovoid to obovoid, gla- brous or pubescent, 3-locular; ovules (1 or) 2 or 3 per locule, orthotropous. Fruit a capsule, more or less dehiscent [fleshy]. Seeds oblong to ovoid, laterally compressed, funicular scar linear, 2 to nearly equal the length of the seed. Base chromosome number 6. (Including Cymbispatha Pichon, Setcreasea K. Schum. & Sydow, Rhoeo Walpers, Zebrina Schnizl.; excluding Cuthbertia Small, Tra- descantella Small.) Tyre species: T. virginiana L., the only species included by Linnaeus. (Named for John Tradescant, Jr., 1608-1662, British naturalist and gardener.)— SPIDERWORT. A genus of about 60 species of the tropical and warm-temperate regions of the New World. Seventeen species occur in the Southeast, 13 of them native. This circumscription follows Hunt (1975a, 1980, 1986b), who recognized 12 sections, five of which are represented in our area. Cuthbertia is included in Callisia following Hunt. The largest section, TRADESCANTIA, was divided by Hunt (1980) into four series. All the southeastern species of this section are in ser. TRADESCANTIA (ser. Virginianae Hunt, nom. superfl.), the members of which are perennials with short rhizomes, linear (leaflike) inflorescence spathes, and glandular-pu- bescent or glabrous pedicels and calyces. All 17 species of ser. TRADESCANTIA are native to the United States, and 13 of these occur in the Southeast. The remainder are present in the Great Plains and the Southwest. The taxonomy of this series 1s rather well known, thanks to monographs by Anderson & Woodson and MacRoberts and the cytotax- onomic studies by Anderson & Sax and Anderson (1954). All are diploids or tetraploids with a similar karyotype (six or 12 large metacentric chromosomes), 1989] TUCKER, COMMELINACEAE 109 and all are self-incompatible (Anderson & Sax; Anderson & Woodson; Owens). An account of the southeastern species of ser. TRADESCANTIA follows, then discussion of the three remaining series of sect. TRADESCANTIA Tradescantia subaspera Ker-Gawl. (T. pilosa Lehm.) is the tallest (30-100 cm) species of the series. The stems are conspicuously flexuose, and the broad leaf blades are lanceolate (not linear as in the other species of the series) and constricted to a subpetiolate base. It ranges from West Virginia to northern Illinois south to northwestern Florida, southern Alabama (Lelong), and south- ern Louisiana. The eastern populations have been distinguished as var. mon- tana (Shuttlew. ex Small & Vail) Anderson & Woodson on the basis of their narrower leaves and sessile cymes. MacRoberts questioned the significance of the varieties because populations from Louisiana were intermediate between var. montana and the typical variety. Tradescantia gigantea Rose (2n = 12; stems 30-60 cm tall; spathes saccate, velutinous; petals broadly obovate) is endemic to sandy soils of central and eastern Texas and three parishes of western Louisiana (MacRoberts Tradescantia virginiana L., the type species (“perhaps the most hain cteristic feature ... 1s its turgid or inflated, delicately foliaceous, eglandular pubescent calyx” Een deicon & Woodson, 1935, p. 65), has long been cultivated both in the United States and in temperate regions worldwide. Its original range in North America can only be surmised. It is now known from southwestern Connecticut, southern Michigan, and western Wisconsin south to north-central North Carolina, northwestern Georgia, central Tennessee, and southeastern Missouri. Reports from the Gulf Coastal Plain are apparently based on mis- identifications of 7. hirsuticaulis (MacRoberts). Many cultivated plants iden- tified as 7. virginiana (especially outside North America) are actually ‘“T. x Andersoniana,” the hybrids of T. virginiana and T. ohiensis. Tradescantia Ernestiana Anderson & Woodson (basal and juvenile leaves linear, narrower than mature leaves and spathes: leaf blade broader than sheath, lacking basal constriction; roots long, slender, not fleshy), an Ozarkian endemic, occurs in southwestern Missouri, western Arkansas, and eastern Oklahoma. It is Intermediate between 7. virginiana and the Texas endemic 7. edwardsiana Anderson & Woodson. Tradescantia ohiensis Raf. (T. canaliculata Raf., T. reflexa Raf.) (roots slen- der, plants glaucous, upper leaves as long as the lower, the blades narrower than sheath, sepals glabrous) 1s a wide-ranging species found from Massachu- setts, Michigan, Minnesota, and Nebraska south to southern Florida and south- ern Texas. Tradescantia paludosa Anderson & Woodson (roots slender, fleshy; upper leaves shorter than lower; base of leaf blade not constricted; foliage not glau- cous) is closely related to 7. ohiensis. Its range is much more restricted (southern Alabama, southeastern Arkansas, eastern Louisiana, and perhaps also Missis- sipp1). MacRoberts treated this as 7. ohiensis var. paludosa (Anderson & Wood- son) MacRoberts Tradescantia hirsutiflora Bush (stems spreading-pubescent; leaves broad, stiff, roots glabrous, not thickened; sepals glandular-pubescent or not), the most poorly understood species of ser. TRADESCANTIA (Anderson & Woodson), is 110 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 70 closely related to 7. virginiana. It occurs from Georgia and western Florida to eastern Oklahoma and central Texas. Tradescantia longipes Anderson & Woodson (2n = 24; stems 2-10 cm tall, leaves and spathes pilose, hairs eglandular), was described from southern Mis- souri and has recently been reported from northern Arkansas (Montgomery County; see Timme & Faden). It is limited to acidic soils in pine-oak forests. A close relative, 7. Tharpii Anderson & Woodson, occurs ae on calcareous soils derived from limestone or chert, in southwestern Missour Tradescantia Reverchonii Bush, a close relative of 7. i aire can be distinguished by its densely tomentose roots. Such hairy roots are unique in the genus. The species is rare (southern Arkansas, Louisiana (Natchitoches and addo parishes), and westward to central Texas). Tradescantia ozarkana Anderson & Woodson (leaf blades abruptly con- stricted just above the sheath, veinlets obvious in dried specimens, capsules 6-8 mm long) is perhaps not distinct from 7. edwardsiana of central Texas. It is a rare species of western Arkansas, southwestern Missouri, and eastern Okla- homa. Tradescantia occidentalis (Britton) Smyth (7. virginiana var. occidentalis Britton) is a species mainly of the Great Plains (Minnesota, Arkansas, and Louisiana west to southeastern Montana, southern Utah, and southern Ari- zona). The plants are glabrous, with the spathes leaflike, the sepals having some glandular hairs, and the hilum about half as long as the seed. Tradescantia occidentalis var. melanthera MacRoberts is distinguished only by its brown anther connectives (other North American species have yellow ones). Its more restricted range includes southern Arkansas, eastern Oklahoma, western Lou- isiana, and eastern Texas (MacRoberts). Tradescantia roseolens Small is close to 7. occidentalis, from which 1t differs in its longer hilum (nearly as long as the seed) and its allopatric distribution (western South Carolina and northern Georgia to southern Florida). The species grows in sandy soil in open woods. Its fresh flowers smell like tea roses (An- derson & Woodson). Tradescantia hirsuticaulis Small (sepals petaloid, pinkish to purplish) has a peculiar distribution: Georgia, Alabama, Arkansas, and eastern Oklahoma. MacRoberts considered it a dubious species. The three remaining series of sect. TRADESCANTIA Comprise species of Mexico and the southwestern United States. Series SILLAMONTANAE Hunt (plants pe- rennial, shortly rhizomatous, nontuberiferous, lanate) is restricted to north- eastern Mexico and consists of two species, 7. si/lamontana Matuda (7. pexata H. E. Moore), 2” = 24, and 7. Rozynskii Matuda, 2n = 12, 24, both of which are self-incompatible (Owens). Series TUBEROSAE Hunt comprises about ten species of tuberous perennials of southern Arizona, southern New Mexico, and Mexico. All are self-incom- patible. There are diploid, tetraploid, and hexaploid species (base number 6). Members of series ORCHIDOPHYLLAE Hunt are subacaulescent tuberiferous perennials with glandular pedicels and calyces. There are two species, 7ra- descantia orchidophylla Hemsley and 7. Mirandae Matuda, both 2n = 24 and both of western Mexico. 7radescantia Mirandae is known to be self-incom- patible. 1989] TUCKER, COMMELINACEAE 11] The five remaining sections of Tradescantia are widely distributed in the warm regions of the New World, although none is represented by indigenous species in the Southeast. The single species of sect. SEPAROTHECA (Waterfall) Hunt, 7radescantia pyg- maea Hunt (2n = 12), is a tuberiferous perennial of the mountains of north- western Mexico (Waterfall; Hunt, 1975a). Its anthers have broad connectives, and the edges of the filaments are adnate to the edges of the petals, forming a tube at least 1 mm long. The plants are self-incompatible. Section MANDONIA Hunt (Mandonia Hasskarl, non Wedd.) comprises about ten species of tuberiferous perennials of Mexico, Guatemala, Bolivia, Brazil, and northern Argentina. There are both self-compatible and self-incompatible species. Tradescantia crassifolia Cav. (2n = 12, 24; self-incompatible) 1s an important weed in Mexico (Holm ef a/.). In sect. PARASETCREASEA Hunt the petals and filaments of the three antesep- alous stamens are fused along their edges to form a short tube (Hunt, 1980). There is but one species, 7. Andrieuxii, 2n = 12, 24, which 1s native to Mexico from Oaxaca to Chihuahua. Both self-compatible and self-incompatible plants have been reported (Owens). Plants of sect. SETCREASEA (K. Schum. & Sydow) Hunt (Setcreasea K. Schum. & Sydow) are rhizomatous. The petals are fused basally to different heights in the five species, which occur in the southwestern United States and Mexico: Tradescantia brevifolia (Torrey) Rose, 2n = 24, 7. Buckleyi (Rose) I. M. John- ston, 2n = 24, T. hirta Hunt, 2n = 24, 7. leiandra Torrey, 2n = 12, and T. pallida (Rose) ae (Setcreasea pallida Rose), 2n = 12, 18, 24. All are self- incompatible. Tradescantia pallida has recently been reported as an adventive in Pinellas County, Florida (Wunderlin ef a/.). Section CyMBISPATHA (Pichon) Hunt (Cymbispatha Pichon) differs from sect. TRADESCANTIA in its base chromosome number (x = 7 vs. 6) and in having various acrocentric karyotypes derived by Robertsonian fusions. The upper- most sepal is larger than the others. This section is intermediate between sects. ZEBRINA (discussed below) and TRADESCANTIA (Hunt, 1980). It includes about seven species that occur from the mountains of western Mexico south to Brazil and Bolivia. There are both self-compatible and self-incompatible species, and T. commelinoides Schultes f., 2m = 14, 22, 23, 30, includes both self-compatible and self-incompatible plants. Section ZEBRINA (Schnizl.) Hunt (Zebrina Schnizl.) comprises two species of southern Mexico and northern Central America, Tradescantia Schippii Hunt and 7. Zebrina Bosse (Z. pendula Schnizl.), 2n = 22, 23, 24, 41, 47. The corolla is sympetalous to various degrees, and the hilum is punctiform. The karyotype is variable, a mixture of meta-, subtelo-, and telocentric chromosomes. The plants are self-compatible. Zebrina has sometimes been accorded generic status, but sympetaly varies in degree within the two species and has arisen elsewhere in the genus (e.g., sect. SETCREASEA). Tradescantia Zebrina is reported as a rare escape from cultivation in central Florida (Wunderlin). Section RHoeo (Hance) Hunt (Rhoeo Hance) includes only Tradescantia discolor L’Hér. (T. spathacea Sw., Rhoeo spathacea (Sw.) Hance), boat-lily, Moses-in-the-bulrushes, three-men-in-a-boat, 2” = 12, which is native to the Yucatan peninsula and Belize and naturalized in southern Florida. The plants 112 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 are succulent and bromeliadlike in habit. The spathes are broadly ovate, not linear as in sect. TRADESCANTIA, and there is only one ovule per locule. The karyotype, similar to that of the species of sect. TRADESCANTIA (Hunt, 1986b), consists of small, meta- to submetacentric chromosomes. Many translocations have occurred because the chromatids form rings or chains meiotic at metaphase (Lin & Paddock). Some plants are self-compatible, while others are not. The boat lily is widely cultivated in tropical regions and in greenhouses. Three of the remaining four sections are monotypic. Section CAMPELIA (L. C. Rich.) Hunt (Campelia L. C. Rich.) has one species, 7radescantia Zanonia (L.) Sw., 2” = 16. Its fleshy capsules and sepals are an adaptation to dispersal by birds. The axillary inflorescences perforate the subtending sheaths (Faden, 1985). The karyotype comprises 2 meta-, 2 submeta-, and 12 acrocentric chro- mosomes. The plants are self-incompatible. In the one species of sect. CoHOLOMIA Hunt, Tradescantia guatemalensis C. B. Clarke ex J. D. Smith, coholom, 2 = 42, the spathes subtending the inflo- rescence are connate basally, forming a short tube that encloses the buds. The species is endemic to southern Mexico and northern Central America. The plants are self-incompatible. Section CorINNA Hunt also includes a single species, 7radescantia soco- nuscana Matuda (Campelia Standleyi Steyerm.), 2n = 26, of southern Mexico and Guatemala. Its karyotype consists of small chromosomes, of which s1x are metacentric, 16 acrocentric, and four telocentric. Plants of sect. AUSTROTRADESCANTIA Hunt are procumbent and leafy stemmed; the spathes are conspicuously broader than the stem leaves. The chromosomes are very small for the genus, the diploid number ranging from 40 to 110. The four to six species are found in southeastern Brazil, Uruguay, Paraguay, and northern Argentina. Tradescantia fluminensis Vell., 2n = 40, 67 (Owens), a native of southeastern Brazil, is adventive in southwestern Georgia, central and western Florida, southern Alabama, Louisiana, and California. In New Zealand, where it is also naturalized, it forms dense, matlike stands in lowland forests. These populations exclude most tree seedlings and inhibit the repro- duction of the native woody plants. Apparently all reproduction of 7. flumi- nensis in New Zealand is vegetative. A node with as little as | cm of stem attached is able to sprout and form a new plant. Thus, trampling stands of T. fluminensis fragments the plants and enhances their reproduction (Kelly & Skipworth). Tradescantia virginiana, T. ohiensis, and T. paludosa, or hybrids between them, are regularly used to monitor radiation and other environmental con- taminants. Clones heterozygous for flower color are exposed to the prospective hazard. Radiation or chemical mutagens cause somatic mutations, resulting in production of pink (recessive) rather than blue (dominant) flowers. When a cell of a stamen hair is affected, the change can be readily spotted under low-power 16 clones of Tradescantia (pure species, hybrids, diploids, triploids, and tet- raploids). The sensitivity to radiation damage was 50-60 times greater in the 1989] TUCKER, COMMELINACEAE 113 most than in the least sensitive clone. The differences in sensitivity of the taxa showed no apparent taxonomic correlation but were believed to result from differences in the genetic repair system The Tradescantia micronucleus- test. (Ma, 1981) makes use of the highly synchronous prophase and tetrad stages in microsporogenesis in clones of 7. paludosa. Radiation or chemical mutagens can break chromatids, and acentric fragments are formed. At the tetrad stage these fragments form micronuclei that are easily spotted and tallied from aceto-carmine squashes. The frequency of micronuclei is indicative of the level of radiation or the mutagenicity of the chemical in question. This highly sensitive test is often used to monitor the purity of drinking-water supplies (Ma et al.). REFERENCES: Under family references see BRENAN (1966); BRUCKNER (1926, 1927, 1930); CELARIER; CHANT; CLARKE; CLEWELL; FADEN (1985); FERNALD; FERNALD & Kinsey; GUERVIN & Le Cog; HANDLOs (1970); HERD & BEETLE; HouM et al.; Hunt (1975a; 1980; 1981; 1984; 1986b, c); Jones & JopLinc; LonNG & LAKELA; MATTHEWS; MATuDA (1955, one MAYER & POLIAKOFF-MAYBER; OwENS; Owens & KIMMINS; PICHON; POOLE & Hunt; ROWLEY; STURTEVANT; eas & GUNNING; TOMLINSON (1966, 1969); Woopson; ea and YANOVSK ALLEN, N. S., G. B. Witson, & S. PowELL. Comparative effects of colchicine and sodium nucleate on somatic chromosomes of Allium and Tradescantia. Jour. Hered. 41: 159-163. 1950. ANpeRSON, E. Binary variation in Tradescantia bracteata. Ann. Missouri Bot. Gard. 28: 147-163. 1941. [A species of the upper Mississippi drainage. ] field survey of chromosome numbers in the species of Tradescantia closely allied oe oa plan virginiana. Ibid. 41: 305-327. 1954. ——. AX. tological analysis of self-sterility in Tradescantia. Bot. Gaz. 95: 50962 1934. filueeatione _E. Woopson. The species of Tradescantia indigenous to the United States. Ge Arnold Arb. 9: 1-132. pls. 1-72. 1935. Baum, J. W., & C. H. NAUMAN. Influence of strong magnetic rae on genetic end points in Tradescantia tetrads and stamen hairs. Environm. Mutagen. 6: 49-58. 1984. BisHop, C. J. Differential x-ray sensitivity of Tradescantia fa ee during the mitotic cycle. Genetics 35: 175-187. 1950. Influence of polyploidy on the x-ray sensitivity of cells. Canad. Jour. Bot. 30: 139- 146. 1952. & L. J. McGowan. The role of the vegetative nucleus in pollen tube growth and the division of the generative nucleus in Tradescantia paludosa. Am. Jour. Bot. 40: 658, 659. 1953. Bouck, M. R. Tradescantia bracteata Small and T. occidentalis (Britt.) Smyth in the Great Plains. (Abstract.) Bot. Soc. Am. Misc. Ser. Publ. 160: 63. 1981 Bora, K. C. Delayed effects of chromosome breakage by x-rays in Tradescantia brac- teata. Jour. Genet. 52: 140-151. 1954. _ The influence of temperature on the production of chromosome aberrations by fractional x-ray doses in the microspores of Tradescantia bracteata. Ibid. 53: 29- Busn, B. F. Two new Texas tradescantias. Ann. Rep. Missouri Bot. Gard. 16: 100, 101. 1905 CELARIER, R. P. Cytology of the Tradescantieae. Bull. Torrey Bot. Club 82: 30-38. 1955. 114 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 . Desynapsis in Tradescantia. Cytologia 20: 69-83. 1955 A new species of 7radescantia (Commelinaceae) from south Texas. Field Lab. 24: 5-9. 1956. CLARKE, E., & M. W. STEER. Cytoplasmic structure of germinated and ungerminated pollen grains of Tradescantia virginiana. Caryologia 36: 299-306. 1983. CONGER, A. ae Tes studies with 7radescantia at nuclear test detonations. Am. Nat. 88: 215-224. CouoT-GASTELIER, J., D. LAF AY, & P. LouGuet. Comparative ultrastructural study of opened and closed fiat in ey virginiana. Canad. Jour. Bot. 62: 1505-1512. 1984. Dean, D. S. Distribution of tetraploid and diploid Tradescantia ohiensis in Michigan and adjoining areas. Am. Midl. Nat. 61: 204-209. 1959. DryAnovskA, O. A. Cytochemical and radiographic studies of meiosis and microspo- tbe nese ie Tradescantia paludosa. Compt. Rend. Acad. Sci. Bulg. 34: 1169-1172. 1981.* & H. Trrsu. Electron microscopic investigations of microsporogenesis and the pollen, grain in 7radescantia paludosa. Compt. Rend. Acad. Sci. Bulg. 34: 257-260. 1981. EMMERLING-THOmpson, M., & M. M. NAwrocky. Evidence of gametic mutation for flower color in Tradescantia subacaulis. Environ. Exp. Bot. 22: 403-408. 1982. FERNALD, M. L. Two of Rafinesque’s species of Tradescantia. Rhodora 46: 310, 311. 1944 Fourcroy, M. Nouvel exemple de racine bistélique: 7radescantia virginica. Bull. Soc. Bot. France 103: 590-595. 1956. arque au sujet de la systématique de trois Tradescantia. Ibid. 104: 605— 608. 1957. [“T. reflexa et T. virginica, souvent confondus par les auteurs, oo une structure radiculaire différente qui vient a l’appui de leur distinction.”’] a a O. M. A dwarf Tradescantia in southwestern Abbeville County, South Car- na. Castanea 22: 120, 121. 1957. [T. brevicaulis.] Soho. H. Karyological 8 in Commelinaceae. IV. Karyotypes of genus Trad- escantia € species in affinitive genera. (In Japanese; English summary.) Kromosomo 85/86: ee 2727. 1971. GALE, R. M. O WENS. The occurrence of an epicuticular layer on the abaxial .O surface of petals in ae ae spp. (Commelinaceae). Kew Bull. 37: 515-520. 1983. Gites, N. Chromosome behavior at meiosis in triploid Tradescantia hybrids. Bull. Torrey Bot. a 68: 207-221. 1941. GUNCKEL, J. E., I. B. Morrow, A. H. SpARROw, & E. CHRISTENSEN. Variations in the floral Ss of normal a nee plants of Tradescantia paludosa. Bull. Torrey Bot. Club 80: 445-456. Haque, A. Effect of infra-red rays on ae ee breakage in Tradescantia. Cytologia 20: 229-236. 1955. Heynowicz, Z. Interpretation und Tatsache in Bezug auf die Entwicklungsmechanik der Meas ie bei Tradescantia fluminensis var. albostriata. Acta Soc. Bot. Polon. 26: 1 255, 256 HikGewower: . A. Longitudinal axis thickenings in whole mount spreads of synap- tonemal oe from Tradescantia ohiensis var. paludosa. Chromosoma 90: 285-288. 1984. ENZEL. Whole-mount spread preparations of ay chromosomes of Tradescantia. (Abstract.) Am. Jour. Bot. 71(5, part 2): 104. 1984. Horie, K. The behavior of the petals in the fading of the flowers oc een Protoplasma 53: 377-386. 1961. Hunt, D. R. singe Mirandae, Commelinaceae. Curtis’s Bot. Mag. 178: ¢. 675. 3 unnumbered pp. 197 1989] TUCKER, COMMELINACEAE a is ICHIKAWA, S. In-situ monitoring with 77 pe around nuclear power plants. En- viron. Health Perspect. 37: 145-164. 1981. . Spontaneous somatic mutation eae in the stamen hairs of 14 different Tradescantia clones heterozygous for flower color. Environ. Exper. Bot. 24: 259- 266. 1984. _C. S. TAKAHASHI, & C. IsHi-NAGASHIMA. Somatic mutation frequency in the stamen hairs of Tradescantia KU-7 and KU-9 clones exposed to low level gamma- rays. Jap. Jour. Genet. 56: 409-424. 1981. Jones, K., A. Kenton, & D. R. Hunt. Contributions to the cytotaxonomy of the Commelinaceae. C ene evolution in 7radescantia section Cymbispatha. Jour. Linn. Soc. Bot. 83: 157-188. Kamizyo, A., & N. TANAKA. eae. on the generative nuclear division, 4. Effect of sucrose in artificial medium on the generative division in Tradescantia paludosa. Cytologia 47: 207-218. 1982. KEHLHOFENER, J. L., & J. DieTRICH. Synaptonemal complex and a new type of nuclear poly-complex in 3 higher plants: Paeonia ae Paeonia Delavayi, and Tra descantia paludosa. Chromosoma 88: 164-170. Keity, D., & J. P. SkipwortuH. 7Tradescantia Pes Bee in a manawatu New Zealand forest. New Zealand Jour. Bot. 22: 393-398. 1984 Kexso, J. M., & J. S. Turner. Protoplasmic streaming in Tradescantia. |. The effects of indoleacetic acid and other growth-promoting substances on streaming. Austral. Jour. Biol. Sci. 8: 19-35. 1955. LAFLeur, G. J., A. E. Gross, & J. P. MAsCARENHAS. Optimization of culture conditions for the eee sperm iG in pollen tubes of Tradescantia paludosa. Gamete Res. 4: 35-4 LELONG, M. G. ates monocots of Mobile and Baldwin counties, Alabama. Sida 13: 101-113. 1988. [7. fluminensis.] Levin, G. G. Age-related changes in the resistance to light and moisture deficiencies in Tradescantia fluminensis (oe leaves and stems. (In Russian.) Bot. Zhur. eae ad 71: 334-341. Lima DE FariA, A. Structure, caee na delimitation of the kinetochore in Trades- cantia. ee 41: 209-226. 1955. LING Koss _F. Pappock. Ring-position and frequency of adjacent distribution of meiotic chromosomes in Rhoeo spathacea. Am. Jour. Bot. 60: 685-690. 1 Ma, T. H. Thin-layer lactose agar for pollen tube culture of Tradescantia to enhance planar distribution of chromosomes. Stain Tech. 42: 285-291. 1967. [T. paludosa.]| Tradescantia paludosa micro-nucleus bioassay and pollen tube chromatid ab- erration test for in-situ monitoring and mutagen screening. Environ. Health Perspect. 37: 85-90. 1981. , V. A. Anperson, M. M. Ha _R. E. Neas, & T. S. Lee. Mutagenicity of danlane water detected by the 7 aa nicionmicte test. Canad. Jour. Genet. Cytol. 27: 143-150. 1985. MacRoserts, D. T. Notes on Tradescantia (Commelinaceae). V. Tradescantia of Lou- isiana. Bull. Mus. Life Sci. Louisiana State Univ. Shreveport 4. 15 pp. 1980. [Seven species; keys, descriptions; distribution maps for Louisiana only. ] MARIMUTHU, K. M., L. A. SCHAIRER, A. H. SPARROW, & M. M. NAwrocky. Effects o space flight aera II) ai radiation on female gametophyte development in Tradescantia. Am. Jour. Bot. 59: 359-366. 1972. Martin, E. S., & H. ee. ae stomatal movements in Tradescantia vir- giniana. New Phytol. 70: 923-928. 197 MatTTHEws, B. Called flower ae day, an blooms through three seasons. New Zealand Gard. 2: 258, 259. a Mitsukurl, Y., & K. SHIRATO. em studies on Commelinaceae. [II]. The study 116 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 of Tradescantia having the chromosome number 2n = 18. (In Japanese; English abstract.) Jap. Jour. Genet. 32: 251. 1957. , T. Tovopa, & K. Nomura. Karyotype analysis in Tradescantia. (In Japanese: English abstract.). Jap. Jour. Genet. 31: 304. 1956. Moore, H. E. Campelia Zanonia in cultivation. Baileya 7: 42-44. 1959. . Tradescantia pexata (Commelinaceae), a new species. Ibid. 8: 99-101. 1960. NAKAMURA, Y., & Y. OHMOMOo. Transfer of gaseous iodine to Tradescantia. Jour. pera Res 25: 251-259. 1984. Owens, S. J., & S. McGratu. Self-incompatibility and the pollen-stigma interaction in Se ohiensis. Protoplasma 121: 209-213. 1984. PaAnbiITA, T. K., & P. N. MEHRA. oe of tetraploid Tradescantia canaliculata. Nucleus (Calcutta) 26: 90-93, 1983.* PARCHMAN, L. G. The morphogenesis of the stamen hairs of 7radescantia paludosa. Diss. Abstr. 25(12, part 1): 6915. 1965. Puza, V. Die Amitose bei Tradescantia virginica L. Protoplasma 59: 589-593. 1965. ReCHEL, E. A., & W. F. CAMPBELL. The effects of radiation from uranium mill tailings on Tradescantia. Great Basin Nat. 38: 456-462. 1978. Rivey, H. P. oo hybridization in a natural population of Tradescantia. Ge- netics 24: 753-769. 1939. Rivigre, S. Une fausse le de la racine de 7radescantia virginica. Bull. Soc. Bot. France 103: 596-598. 1956 Rose, J. N. Three new oe a Tradescantia from the United States. Contr. U. S. Natl. Herb. . 204-206. 18 SAKAI WapaA, A. Ult ee h I studi the synaptonemal complex of ae reflexa. Cytoogi 48: a 868. 1983. Sassen, M.M.A.,& A.M. C. Wocters-Arts. Cell wall texture and cortical microtubules in growing staminal hairs eae virginiana. Acta Bot. Neerl. 35: 351-360. 1986. Sax, K., & H. W. EpMonps. Development of the male gametophyte in Tradescantia. Bot. Gaz. 95: 156-163. 1933. ScumitT, F. O., & G. T. JOHNSON. Optical and chemical studies on the granules in microspores of Tradescantia. Ann. Missouri Bot. Gard. 25: 455-466. 1938. Scott, L. I. & J. H. Priesttey. Leaf and stem anatomy of Tradescantia fluminensis Vel. Jour. Linn. Soc. Bot. 47: 1-28. 1925. SkirM, G. W. Bivalent pairing in an induced tetraploid of Tradescantia. Genetics 27: 635-640. 1942. SLAVIN, A. D., & J. A. NreUWLAND. A new form of Tradescantia reflexa. Am. Midl. Nat. 11: 600. 1929. SMALL, J. K. Tradescantia longifolia. Addisonia 17: 17, 18. pl. 553. 1932. STEARN, W. T. The boat-lily (RAoeo spathacea). Baileya 5: 195-198. 1957. STEFFENSEN, D. Irregularities of chromosome division in Tradescantia grown on low sulfate. sty Cell Res. 6: 554-556. 1954. [T. paludosa.] Stinson, J. R., & J. P. MAscARENHAS. Fate of messenger RNA in the mature pollen grain i Tr adesc antia aes during germination and pollen tube growth. Pp. 510, S11 in D. L. Mutcany er al., eds., Biotechnology and ecology of pollen. Secaucus, New Jersey, and Berlin, 1986. Swanson, C. P. The distribution of inversions in 7radescantia. Genetics 25: 438-465. 1940. Secondary association of fragment chromosomes in generative nucleus of Trad- escantia rand its bearing on their origin. Bot. Gaz. 105: 108-112. 1943, Taser, R. A., & M. E. StrronG. Vesicular-arbuscular mycorrhizae in roots and xylem of Tradescantia virginiana. Mycologia 74: 152-156. 1982. & W. A. Taser. Evidence for ectomycorrhizal fungus-mediated nutrient transfer between Pinus and Tradescantia. Forest Sci. 30: 892-896. 1984.* 1989] TUCKER, COMMELINACEAE 117 THiERET, J. W. Additions to the flora of Louisiana. Castanea 28: 169, 170. 1963. [7. subaspera, ae TimME, S. L.,& R. B. FADEN. Ti fa longipes (C li )in the southeastern .SLA, Cae. a5 83-85. 1984. pee and Arkansas. ] ToMmos, A. D., E. STEUDLE, U. ZIMMERMANN, & E.-D. ScHULZE. Water relations of leaf epidermal cells Tradescantia virginiana. P|. Physiol. 68: 1135-1143. 1981. UNDERBRINK, A. G., R. C. SPARROW, & A. H. SPARROW. Relations between phenotypic aberrations and loss of en integrity in Tradescantia stamen hairs. Radiat. Bot. 11: 473-481. 197 WATERFALL, U. T. ae a new genus (Commelinaceae) from Mexico. Rhodora 61: 136-139. 1959. WHALLON, J. Scanning electron microscopy of meiotic chromosomes of Tradescantia gigantea. Cytobios 38: 137-148. 1983 WUNDERLIN, R. P., B. F. HANSEN, & D. W. HALL. The vascular flora of central Florida: taxonomic and nomenclatural changes, additional taxa. II. Sida 13: 83-91. 1988. 2. Callisia Loefling, Iter Hisp. 305. 1758. Perennial [annual] herbs. Rhizomes present or lacking. Stems erect or de- cumbent and rooting at the nodes. Leaves several to many, cauline, spiraled or distichous; sheaths shorter than blades, the apex ciliate; blades oblong- lanceolate to linear. Inflorescences axillary or terminal, cymose; spathe solitary, with or without leaflike blade; bracteoles smaller than leaves, generally scarious. Pedicels elongate or very short, glandular-pubescent and/or villous. Sepals 3, equal, broadly elliptic to lanceolate, the apex rounded or acute, free or connate for the basal 3, the outer surface with many short glandular hairs. Petals (2 or) 3, orbiculate to narrowly oblong, free, subequal, glabrous. Stamens [1, 3, or] 6, all fertile; filaments slender, glabrous or hairy; anthers broadly ovate, the connective rectangular, about as wide as the anther locule. Ovaries oblong- ellipsoid, 3-locular, each locule with 2 ovules; style shorter than to 2—3 times longer than ovary; stigma club shaped, rough (sometimes shortly plumose). Fruit a capsule, [2- or] 3-locular, each locule containing [1 or] 2 seed(s). Seeds ovoid, dorsiventrally flattened, the surface rugose. Base chromosome numbers 6, 7, 8. (Including Aploleia Raf., Cuthbertia Small, Leptorhoeo Hemsley, Phyo- dina Raf., Tradescantella Small.) Types species: C. repens L. (Name from Greek kallos, pretty.) — BASKET-PLANT. A genus of about 20 species of the tropical and subtropical regions of the New World. The circumscription of the genus adopted here follows that of Hunt. Four of the six sections are represented in the United States, with three of these in our area. Like its close relatives Tradescantia and Tripogandra, Callisia has fused pairs of cymes. It differs from Tradescantia in having single (vs. paired) inflorescence spathes and glabrous (vs. hairy) filaments; from 77i- pogandra, in lacking dimorphic pollen (Mattson) and in having actinomorphic (vs. zygomorphic) flowers. Ecologically, species of Callisia generally grow in drier situations than those of either Tradescantia or Tripogandra. Section CUTHBERTIA (Small) Hunt (Cuthbertia Small) consists of three species, all endemic to the Coastal Plain and Piedmont from Virginia to Florida. They are erect plants with linear leaves and both lateral and terminal inflorescences 118 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 subtended by inconspicuous spathes. The base chromosome number is 6, and the karyotype comprises 2 metacentric and 4 subtelocentric chromosomes. Although these species have been included in Tradescantia by some workers, they differ from that genus in morphology, anatomy (Tomlinson, 1969), cy- tology (Giles, 1942, 1943), and chemistry (Matthews). The removal of Cuth- bertia from Tradescantia has gained general acceptance. Rohweder included the species in Phyodina, which in his classification comprised sects. CUTHBERTIA and Leproca..isiA. The three species have all been treated as varieties of Cuthbertia rosea (= Callisia rosea) by some workers (e.g., Fernald). Callisia rosea (Vent.) Hunt (7radescantia rosea Vent.; Cuthbertia rosea (Vent.) Small) (stems approximate, leaf blades as broad as or broader than sheaths), the most wide ranging of the three species, occurs from central North Carolina to southeastern Georgia and northeastern Florida. Callisia graminea (Small) G. Tucker (stems densely caespitose, roots gla- brous, inflorescence equaling or surpassed by leaf blades, spathes somewhat foliaceous) occurs on the Coastal Plain from southeastern Virginia to southern Florida. (It has also been reported from southeastern Oklahoma (Cleveland and Oklahoma counties) by Stemen & Myers, but I have seen no specimens to confirm this.) There are diploid, tetraploid, and hexaploid populations (Giles, 1942, 1943). The diploids are endemic to the sandhills region along the western edge of the Coastal Plain in southern North Carolina, while the tetraploids occupy habitats on the Coastal Plain from southeastern Virginia to southern Florida. Hexaploids are rare and have only been found from sites in South Carolina and Florida. The tetraploids averaged 25 percent larger in vegetative and floral parts than diploids. Lakela described a white-flowered form from central Florida, C. graminea f. leucantha (Lakela) G. Tucker. It produces veg- etative plantlets from the axils of the upper leaves, but there are no other reports of vivipary in the genus Callisia ornata (Small) G. Packer (stems approximate, roots woolly, inflo- rescence surpassing leaf blades, spathes scarious) is endemic to palm or pine scrub 1n central Florida. Section LepTocaLLisiA Bentham & Hooker (Ap/loleia Raf., Leiandra Raf., Phyodina Raf. in part, Leptorhoeo Hemsley, Tradescantella Small) includes six species of Florida, Texas, Mexico, and Central America. The base chromosome number is 7, and the stigmas are subcapitellate to penicilliform. One species, Callisia cordifolia (Sw.) Anderson & Woodson (Tradescantia floridana S. Wat- son, Tradescantella floridana (S. Watson) Small), grows in central and southern Florida. It is a procumbent plant that forms small mats on shaded rock or moss. The inflorescence is terminal, and the flowers pedicellate. The Neotrop- ical C. multiflora (Martens & Gal.) Standley (Aploleia multiflora (Martens & ‘Three new combinations are needed for the inclusion of Cuthbertia in Callisia: Cc epee graminea (Small) G. Tucker, based on Cuthbertia graminea Small, Fl. Southeast. U. S. 237, 1328. 1903. Cc nee cial : ae (Lakela) G. Tucker, based on Cuthbertia graminea f. leucantha Lakela, Sida 5 Callisia ornata eels G. Tucker, based on Cuthbertia ornata Small, Man. Southeast. Fl. 259. 1933 1989] TUCKER, COMMELINACEAE 119 Figure 1. Callisia sect. LePTOCALLISIA. a-h, C. cordifolia: a, stem with inflorescence, x 1: b, inflorescence with buds and 3 opened capsules, x 6; c, flower, x 15; d, stamen, x 25: e, diagrammatic longitudinal section of gynoecium showing 2 orthotropous ovules in 1 locule, x 12; f, open loculicidal capsule with marcescent petals and persistent sepals, x 8: g, seed, hilum below center, x 25; h, seed from opposite side, operculum at center, x 50. Gal.) H. E. Moore), 2” = 28, is self-incompatible, while pai cordifolia is unknown in this regard. The flowers of C. multiflora are unsce Section BRACHYPHYLLA Hunt (Phyodina Raf., in part) ace ine species of Texas and northern Mexico. The plants are procumbent perennials with short, succulent leaves, subsessile terminal inflorescences, six stamens per flow- er, barbed filaments, and capitate stigmas. Ca/lisia navicularis (Ortg.) Hunt (P. navicularis (Ortg.) Rohw.), 2n = 32, and C. micrantha (Torrey) Hunt (7’ra- descantia micrantha Torrey, P. micrantha (Torrey) Hunt), 2n = 24, are both self-incompatible (Owens). Section HADRODEMAS (Moore) Hunt contains only Callisia Warszawicziana (Kunth & Bouché) Hunt, 2” = 16, of Guatemala, a self-incompatible species. 120 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 The plants are thick-stemmed perennials with broadly lanceolate leaves, and their purple sepals are persistent (see Moore, 1962, for illustration). Section Laura Hunt contains just Ca/lisia Laui (Hunt) Hunt (Phyodina Laui Hunt), 2 = 12, which is known only from southeastern Oaxaca, Mexico. The plants are tuberous, the inflorescences terminal and pedunculate, and the flow- ers pedicellate. The species is self-incompatible. Section CALLIsIA (Hapalanthus Jacq., nom. superfl., Spironema Lindley non Raf., Rectanthera Degener) comprises some ten Neotropical species with the base chromosome number six. The small, often bladeless petals and the bushy, exserted stigmas may be an adaptation to wind pollination, but this has not been demonstrated. One species, Callisia repens L. (including C. insignis C. B. Clarke), basket-plant, 2n = 24, occurs in the Florida panhandle (Franklin County) and in peninsular Florida from Pinellas and Broward counties, and southward through the West Indies and Mexico to Brazil and Peru. It is similar in habit to C. cordifolia but has both axillary and terminal inflorescences of subsessile flowers partly enclosed in a bladeless sheath. The flowers of C. repens are unscented. Ca/lisia repens is self-compatible, while the three other species of this section for which information is available—C. Gentlei Matuda (C. elegans E. J. Alexander ex H. E. Moore; C. Macdougallii Miranda) of northern Central America, C. fragrans (Lindley) Woodson of southeastern Mexico, and C. soconuscensis Matuda of Guerrero, Mexico, all 2n = 12 (Owens)—are all self-incompatible. Several species of Callisia are cultivated as creeping ground covers in frost- free areas. Callisia fragrans and similar procumbent species are sometimes grown indoors as hanging-basket plants. REFERENCES: Under family references see FADEN (1985); FERNALD; HunT (1984, 1986d); MATTHEWS; MATTSON; Owens; ROHWEDER; and TOMLINSON (1969). ANDERSON, L. C. Noteworthy plants a North Florida. III. Sida 13: 93-100. 1988. [Callisia repens in Franklin Count Gites, N. H., Jr. Autopolyploidy and seeiaphied distribution in Cuthbertia graminea Small. res Jour. Bot. 29: 637-645. 1942. . Studies of natural populations a Cuthbertia graminea in the Carolinas. Jour. mee Mitchell Sci. Soc. 59: 73-80. 1943. AKELA, O. Field observations of See (Commelinaceae) with description of a new form. Sida 5: 26-32. 1972. [Ilustrations; key to species of sect. CUTHBERTIA. ] Moore, H. E. The genus Aploleia (Commelinaceae). Baileya 9: 13-19. 1961. [Illustra- tions of C. monandra, C. multiflora.] Hadrodemas, a new genus of Commelinaceae. Jhid. 10: 130-136. 1962. [Illus- trations.] STEMEN, T. R., & W.S. Myers. Oklahoma flora. xxxi + 706 pp. Oklahoma City. 1937, 3. Gibasis Rafinesque, Fl. Telluriensis 2: 16. 1837. rawling, weak-stemmed [erect] perennials [annuals]. Roots fibrous [tu- Sp beriferous]. Stems much branched, rooting at the nodes. Leaves several: sheaths ciliate apically; blades oblong-lanceolate to broadly or narrowly ovate. Inflo- 1989] TUCKER, COMMELINACEAE 121 rescences several, terminal or axillary and terminal, cymose, 2[-8] per stipe; cincinnal axis short, thick; cincinnal bracts distichous, broadly rounded, sub- orbiculate; pedicels short. Flowers perfect, actinomorphic. Sepals 3, oblong- ovate, more or less keeled. Petals 3, oblong [orbiculate]. Stamens 3; filaments bearded, the hairs moniliform; anthers orbiculate, dehiscing laterally; connec- tive rectangular, broader than long. Ovary 3-locular, ovules 2 per locule; style glabrous; stigma subtly 3-lobed, capitellate. Fruit a capsule. Seeds ovoid [ob- long-ovoid], the hilum elongate-punctiform [linear]. Base chromosome number 6. Type species: Tradescantia pulchella HBK. (= G. pulchella (HBK.) Raf.), the only species included by Rafinesque. (Name from Latin gibba, gibbous, and basis, base, referring to the “gibbous base,” presumably of the asymmetric leaves.)— BRIDAL-VEIL. A genus of 11 species of Mexico and, to a lesser extent, the Caribbean region and tropical South America. Gibasis is characterized by features of the inflo- rescence, which is made up of paired cymes borne on stipes. The paired cymes are not fused, as in the closely related Tradescantia. The cyme axis in Gibasis bears two rows of broad scalelike bracteoles, each subtending a single flower. In his revision of the genus, Hunt (1986a) recognized two sections. Section Gisasis has tuberiferous roots, oblong-lanceolate to broadly ovate leaves, and pink flowers. The plants are self-incompatible. The chromosomes are large, and the base numbers are 4, 5, or 6. Sulfated phenolic acids are absent (Del Pero de Martinez). There are eight species, mostly restricted to the uplands of Mexico: G. chihuahuensis (Standley) Rohw.; G. consobrina Hunt, 2n = 20; G. Karwinskyana (Schultes f.) Rohw., 2n = 10, 20, 30; G. linearis (Bentham) Rohw., 2” = 10, 12, 20, 22; G. Matudae Hunt, 2n = 10; G. pulchella, 2n = 10, 15; G. triflora (Martens & Gal.) Hunt, 2” = 10; and G. venustula (Kunth) Hunt, 2n = 12. Cyme axes in G. pulchella are reflexed 180° and fused to the upper stipe, so the first flowers to open seem to be those most distal in the inflorescence. Plants of sect. HETEROBASIS Hunt have nontuberiferous roots, linear to ovate- lanceolate leaves, and white, self-compatible flowers. The chromosomes are small, and the base number is 8 (Owens, 1977). Sulfated phenolic acids are present. There are three species. Gibasis oaxacana Hunt, 2n = 12, is endemic to Oaxaca, Mexico. Gibasis pellucida, 2n = 10, 16, with leaves and stems glabrescent, and filaments hairy throughout, is widespread in the lowlands of southern Mexico (dubiously reported from Cuba and the Dominican Republic); it has been collected as a weed in southern Florida (Dade County, Hansen & Sauleda 10732, usF, and Sarasota County, Waldo s.n., 4 May 1971, nys). Gibasis geniculata, 2n = 48, 52, with leaves and stems villous and filaments hairy basally, is the only wide-ranging species, occurring from Haiti and south- ern Mexico south to northern Argentina. It has been reported from central Florida as an escape (Wunderlin), but I have seen no specimens. Hunt (1986a) did not mention the genus as being represented in the United States. Leaf anatomy of Gibasis geniculata, G. Matudae, and G. Schiedeana (Kunth) Hunt (= G. pellucida) has been investigated by Stant. She found three kinds of hairs: short, 2-celled, hook or prickle hairs; glandular, 3-celled ones; and 122 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 long, 4- to 6-celled trichomes. All three types were present in various amounts in the three species, but differences were apparently not taxonomically useful. Gibasis pellucida, Tahitian bridal veil, is widely cultivated in the United States and Mexico. It has often been confused with G. geniculata (Hunt, 1986a). REFERENCES: Under family references see ares (1966); BRUCKNER (1926, 1927, 1930); CLARKE; Det PERO DE MARTINEZ; FADEN (1985); Hunt (1975b, 1978, 1984, 1986a); JONES & JOPLING; OWENS (1981); Sees & KimMins; PooLte & HUNT; ROHWEDER; TOMLINSON (1966, 1969); Woopson; and WUNDERLIN. BRANDHAM, P. E., & S. BHATTARAI. The effect of B chromosome number on chiasma frequency within and See aa of Gibasis linearis (Commelinaceae). Chromosoma 64: 343-348. 19 Hunt, D. R. Gibasis oaxacana. ee Bot. Mag. 179: ¢. 624. 1972. . Gibasis a eine Ibid. t. 636. 1973. [= G. pellucida. ] Jones, K. Chromosome evolution by Robertsonian translocation in Gibasis (Comme- linaceac). saan 45: 353-368. 1974. _S. BHATTARAI, & D. R. HUNT. Contributions to the cytotaxonomy of the Com- melinaceae. Gibasis linearis and its allies. Jour. Linn. Soc. Bot. 83: 129-162. 1981. APES, & . Hunt. Contributions to the cytotaxonomy of the Comme- linaceae. II. Further observations on Gibasis geniculata and its allies. Jour. Linn. Soc. Bot. 71: 145-166. 1975 KENTON, A. Giemsa C- banding i in Gibasis (Commelinaceae). Chromosoma 65: 309- 324. 1978. . Chromosome evolution in the Gibasis linearis alliance (Commelinaceae) I. The Robertsonian differentiation of G. venustula and G. speciosa. Ibid. 84: 291-304. 1981. —. Chromosome evolution in the Gibasis linearis group (Commelinaceae) II. Rob- ertsonian differentiation and preferential pairing revealed in species and F, hybrids of the Gibasis linearis group (Commelinaceae). Pl. Syst. Evol. 144: 221-240. 1984a. h osome evolution! in ne Gibasis linearis group (Commelinaceae) III]. DNA variatio and speciation in G. venustula and G. heterophylla (2n = 13. Chromosoma 99; 1-8. 1984b. Owens, S. J. Incompatibility studies in the genus Gibasis section Heterobasis D. R. Hunt (Commelinaceae). Heredity 39: 365-371. 1977 REYNOLDS, T. An analysis of enzyme multiple form variation in relation to cytology and fea in Gibasis Schiedeana I. Peroxidases. Kew Bull. 34: 83-111. 1979. & I. L. Gitpert. An analysis of enzyme multiple form variation in relation to cytology and morphology in Gibasis Schiedeana (Commelinaceae) II. Acid phos- phatases. [bid. 36: 575-589. 1981. Strant, M. Y. Scanning electron microscopy of silica bodies and other epidermal features in Gibasis (Tradescantia) leaf. Jour. Linn. Soc. Bot. 66: 233-243. pls. 1-5. 197 Tribe COMMELINEAE 4. Murdannia Royle, Illus. Bot. Himalayan Mtns. 403. 1840, nom. cons. Perennial or annual herbs. Roots fibrous [sometimes uniformly thickened throughout]. Stems smooth. Leaves several, cauline [basal only, or both basal and cauline]; sheaths closed; petioles lacking; blades [broadly] narrowly lan- ceolate, the base rounded [cordate], the margins and apex with papillae or 1989] TUCKER, COMMELINACEAE 123 prickles. Inflorescences axillary and terminal [terminal only], thyrsiform, open. Pedicels erect to ascending. Flowers | to many, perfect, actinomorphic. Sepals 3, equal. Petals 3, equal, oblong, not clawed. Stamens 2 or 3, antesepalous: filaments glabrous or bearded; staminodia [lacking, 1, 2, or] 3, antepetalous; sterile anthers 3-lobed. Ovary trilocular, glabrous or glandular-pubescent; ovules uni- or biseriate, 2 (or 3) per locule; stigma capitate. Fruit a capsule. Seeds broadly oblong, hilum punctate to linear. Base chromosome number 10. Type SPECIES: Commelina scapiflora Roxb. (M. scapiflora (Roxb.) Royle = M. edulis (Stokes) Faden). (Named for Murdan Aly, plant collector and keeper of the herbarium at Saharanpur, India, in the mid-nineteenth century.) A genus of 40-45 species of tropical and warm-temperate regions of the Old and New worlds. The genus 1s most diverse in southern Asia, where some 30 species occur (Faden, 1975). There are seven species in the New World, four of them South American endemics. Murdannia has long been confused with Anei/lema R. Br., but Faden (1975) pointed out that the confusion seems to have been from the publication of numerous species of Murdannia under Aneilema rather than from close rela- tionship. The two genera are distinguished on the basis of androecial structure, but since floral features are poorly preserved in pressed specimens, this char- acter has often been overlooked. In Murdannia the stamens alternate with the staminodia, while in Anei/ema they are clustered below the staminodia (see Faden, 1975, for illustration). The closest relative of Murdannia is the mono- typic African genus Anthericopsis Engler. The alternating arrangement of stamens and staminodia that characterizes Murdannia is also found in Palisota Reichenb. (in this genus the stamens are antepetalous and the staminodia antesepalous, the reverse of the situation in Murdannia), Pseudoparis (antepetalous staminodia, but petiolate leaves), and Anthericopsis (3-locular dehiscent capsules as in Murdannia, but sessile sub- opposite inflorescences, larger chromosomes, and base chromosome num- ber 7). Murdannia nudiflora (L.) Briickner is widespread in Latin America and extends northward to Florida and southeastern Georgia; it is also known from Louisiana (reported from coastal North and South Carolina by Radford and colleagues; no specimens seen). The plants are caespitose and do not root at the nodes. The longest pedicels about equal the capsule, which is usually 1.25 times longer than wide. When the fruit is mature, the sepals are hyaline. The seeds are 1.2-1.5 mm long. Murdannia spirata (L.) Brenan, an Asian species, is adventive in southern Florida (specimens seen from Charlotte, Glades, Lee, Hendry, and Collier counties). Although this species and M. nudiflora are similar, M. spirata has persistent bracteoles on the inflorescence branches and the longest pedicels two to three times longer than the capsules. It was first reported in the New World from Collier County, Florida, in 1968 (Lakela). Murdannia Keisak (Hassk.) Hand.-Mazz. (Aneilema Keisak Hassk.), 2n = 32, 40, is naturalized in the area from eastern Maryland to Florida westward to Louisiana, and it has recently been reported from central and eastern Ten- nessee and southeastern Kentucky. The plants are single stemmed and often 124 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 root at the lower nodes, the pedicels are several times longer than the capsules, the sepals are green and develop sessile reddish glands in fruit, and the seeds are about 3 mm long. The karyotype of Murdannia Keisak comprises five metacentric, |2 submetacentric, and three subtelocentric pairs of chromosomes (Fujishima). Fernald thought this species to be native to eastern North America, but Shinners pointed out that its rapid spread since the 1930's indicated the weedy behavior of an adventive. In western Australia Murdannia nudiflora produces cleistogamous flowers during rainy weather and chasmogamous ones when it is sunny (Keighery). Chasmogamous flowers are self pollinated, with the anthers dehiscing directly onto the receptive stigmas. No pollinators were observed visiting the open flowers. Murdannia nudiflora is self-compatible (Owens), as is M. simplex (Vahl) Brenan. Murdannia Keisak is a significant weed in Japan. Although it has spread rapidly in the Southeast in the last 50 years, it does not appear to have become a weed here. REFERENCES: Under family references see BAILEY et a/.; FADEN (1975, 1985); FERNALD; Owens; and RADFORD @f a C.M. Murdannia Keisak (Hassk.) Hand.-Mazz. (Commelinaceae), Bothriochloa ‘hs ‘brida (Gould) Gould (Poaceae), and Scutellaria racemosa Pers. (Lamiaceae) new to ape Sida 10: 189, 190. 1983. Darwin, S. P., E. G. SUNDELL, & A. S. BRADBURN. Noteworthy vascular plants from Louisiana, U. S. A. Sida 9: 70-75. 1981. FuyIsHImMa, H. Kary aimee noes Commelinaceae, II. Karyotypes of Pollia japonica and Aneilema Keisak. B . Mag. Tokyo 83: 21-26. 1970. KEIGHERY, G. J. Facu ee Eien in Murdannia nudiflora (Commelinaceae). Flora 172: 327, 328. 1982. [Illustrations.] LAKELA, O. Occurrence of Murdannia spirata in Florida. Rhodora 70: 571-573. 1968. SHINNERS, L. H. Aneilema (Commelinaceae) in the United States. Sida 1: 100, 101. 1962. [M. ee (as Aneilema Keisak), M. nudiflora (as A. midiflorum).) Sippce, W. S., & R. H. WHEELER. On the presence of three vascular plants, Melothria pendula, Carex eee and Aneilema Keisack [sic], in Maryland. Chesapeake Sci. 15: 173, 174. 1974. [M. Keisak in Charles and Wicomico counties Wuite, P. S. New and noteworthy plants from the Great Smoky Mountains National Park, North Carolina and Tennessee. Castanea 47: 78-83. 1982. [M. Keisak, 79, from Blount County, Tennessee. ] WorrorD, B. E., D. H. Wess, & W. M. DENNis. State records and other recent note- worthy collections of Tennessee plants II. Castanea 42: 190-193. 1977. [M. Keisak from Roane County, Tennessee. ] 5. Commelina Linnaeus, Sp. Pl. 1: 40. 1753; Gen. Pl. ed. 5. 25. 1754. Annual or perennial herbs. Rhizomes lacking; roots fibrous. Stems simple or slightly branched [acaulescent], glabrous. Leaves several, cauline; sheaths closed; blades broadly to narrowly lanceolate, flat or nearly so. Inflorescences axillary [scapose], each subtended by an orbiculate to oblong, acuminate to 1989] TUCKER, COMMELINACEAE 125 acute, strongly conduplicate spathe, its margins free or fused basally. Flowers several, perfect, zygomorphic. Sepals 3, ovate, equal. Petals 3, unequal, clawed, the upper 2 broadly ovate to suborbiculate, the lower 1 oblong. Stamens 3, dimorphic; staminodia 3. Ovary 2- or 3-locular; ovules | or 2 per locule. Fruit a capsule. Seeds rectangular to oblong. Base chromosome numbers 11, 14, 15. LECTOTYPE SPECIES: C. ee L. (Britton & Brown, Illus. Fl. No. U.S. & Canada. ed. 2. 1: 457. 1913). (Named for Caspar Commelijn [also spelled Commelin], 1667/1668- ical and Jan Commelijn, 1629-1692, Dutch bota- nists.°— DAYFLOWER. Commelina, with some 250 species, is the largest genus in the family. The center of diversity for the genus is southern Asia, but there are also numerous representatives in tropical Africa and a lesser number in the Neotropics. There are five native species in the United States (four in the Southeast) and four introduced ones, all of which occur in the Southeast. Commelina virginica L. (2n = 30, 60: werent leaves broadly lanceolate; inflorescences terminal, | to several together; spathes oblong, the margins fused for the basal 3-2), occurs from Maryland, southern West Virginia, Kentucky, [linois, and Missouri south to southern Florida and southern Texas; it is also found in the Neotropics. The apex of the leaf sheath bears reddish, uniseriate, multicellular trichomes 2-5 mm long, a feature that distinguishes this species from other native species of the genus in North America. Commelina erecta L. (2n = 60; perennial; inflorescences single, terminal and axillary; spathes orbiculate, the basal 4 of the margins fused), occurs from New Jersey to northern Illinois, southern Wisconsin, and Nebraska south to Florida, Texas, and Arizona. Commelina erecta is cytologically diverse in West Africa (Morton, 196 Commelina Haniel Delile (leaves linear; spathes long-attenuate, the margins free) occurs from western Texas, eastern Colorado, and Arizona south to El Salvador. Its long-attenuate spathes distinguish it from other New World species. The southern Asian C. appendiculata C. B. Clarke, which also has long-attenuate spathes, appears to be closely related. Commelina gigas Small (perennial; weak stemmed, clambering; spathes ob- long, the margins free) is endemic to the Lake Okeechobee area in Florida (Wunderlin). The plants have a few white trichomes 1-2 mm long at the apex of the leaf sheaths. Commelina communis L. (2n = 48, 90; annual; leaves 15-40 mm wide; spathes oblong (orbiculate when spread out and pressed flat), acute, the margins free) occurs spontaneously in an area from southwestern Maine, southern New a 10, central Michigan, southern Minnesota, and Nebraska south to Georgia, northern Mississippi, and northeastern Texas. Pennell believed it to be native &*Commelina has flowers with three petals. two of which are showy, while the third is not con- spicuous: fom the two Bene called Commelin: for the third died pe accomplishing anything in botany.” Linnaeus, Crit. Bot. (transl. A. Hort), 62, 63. 1737. The third Commelin was Caspar, , 1700-1734, son of uae Commelin. “Actually Commelina was ee by Plumier, who was a not aware of the existence of Caspar, Jr.” (Wijnands, pp. 10, 11). 126 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 to eastern North America, while Brashier thought it was introduced from Asia. It is occasionally cultivated in the eastern United States (Bailey er a/.). There are two varieties: var. communis (petals light blue, anthers entirely yellow) occupies the entire range of the species in North America; var. /udens (Miq.) C. B. Clarke (C. communis subsp. /udens (Miq.) Penn.; petals intense violet, anther connective with purplish brown eye spot) occurs from New England to Kentucky southward to Virginia. Both varieties are also present in eastern Asia. The smut Ustilago commelinae (Komarov) Zundel (Tilletia commelinae Ko- maroy) infects var. /udens, but not var. communis, even when the two varieties are found growing together.’ Ustilago commelinae is also known to parasitize C. communis (variety unknown) in China, Japan, and the eastern Soviet Union. The restriction of the smut to one of the two varieties of C. communis 1n eastern North America suggests a biochemical difference between the varieties, em- phasizing their taxonomic significance Commelina diffusa Burman f. (C. caroliniana Walter) (2n = 28, 30, 56, 60, 72: annual: leaves 9-15 mm wide: spathes oblong when flattened, acuminate, the margins free) is a pantropic and warm-temperate species found in North America from Maryland, southern Ohio, northern Illinois, and eastern Kansas south to southern Florida and southern Texas. It has prostrate stems and erect flowering shoots. Japanese populations are 2 = 72 (Fujishima, 1984). The west African C. diffiisa var. aquatica (Morton) Morton, 2n = 28, grows in shallow water in seasonally inundated grasslands Commelina nigritiana Bentham var. gambiae (C. B. Clarke) Brenan is native to West Africa and a recent adventive in central Florida (Manatee Co., Wun- derlin 5721, usr, Polk Co., Wunderlin & Beckner 9096, USF). Its flowers are peach colored; the spathe is fused for the basal half along the margin and bears long white trichomes on the abaxial surface. Commelina benghalensis L., 2n = 22, 28, 44, 56, is an important weed in the croplands of tropical Africa and Asia. A recent adventive in the Southeast, it was first noted in southeastern Georgia in 1966 (Duncan) and has since been reported from central and western Florida. It has broadly ellipsoid, distinctly petiolate leaves with reddish trichomes at the apex of the sheath. The spathe margins are fused for the basal third of their length. The plants bear both chasmogamous and cleistogamous flowers. Seeds from the former weigh about one seventh as much as those from the latter (see Wilson for illustrations of both kinds of seeds). Dormant seeds that are able to germinate far below ground, easy vegetative reproduction from fragments, and resistance to weeding are features accounting for the spread and abundance this species. Mechanical cultivation breaks up the plants. Fragments of stems can root in wet weather, reestablishing the population; they are able to sprout even if buried beneath 2 cm of soil. Seeds are dormant when freshly shed. After disruption of the seed coat by scarification or pricking, 40-70 percent of treated seeds germinate in 20 days (Budd et ale, The larger seeds from cleistogamous flowers are able to germinate if covered on smut parasites kindly provided by Clark T. Rogerson (New York Botanical Garden), ee on unpublished observations by him and the late Stanley J. Smith (New York State Museum). are ANNE LEIS OT rts —_ ESE PT £ SAD = FiGureE 2. Commelina communis var. communis: a, stem with flowers and fruits, x 1; b, flower, showing 2 sepals, 2 conspicuous petals, reduced petal eae fertile stamens, and 3 staminodia, x 3; c, 1 of the 2 isomorphic fertile stamens, x 10; d, 2 views of anisomorphic fertile men, x 10; €, staminodium with rudimentary anthers, x 10: f, staminodium, completely sterile, x 10; g, diagrammatic transverse cross se of ovary at poe showing 3 ee G fertile, each with | ovule; 1 sterile), x 15: h, infructescence math 2 capsules, | side of folded spathe removed, x 1; 1, detail of same, x 3; j, seed, showing operculum, x 20. AVAOVNITAWWOD “WAMONL [6861 EG 128 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 by less than 14 cm of soil; the smaller seeds from open flowers, only if covered by less than 7 cm. In a natural population, 88 percent of the seedlings came from large seeds. Another introduced species in the Southeast is Commelina Forskalaei Vahl, known only from southern Florida (Dade County, D. S. Correll & Popenoe 51166, usF, and H. B. Correll, Popenoe, & Sanders 54190, usF). It is distin- guished from other species in our area by its undulate leaf margins. The spathes are orbiculate, and the margins are fused for the basal third. Like C. ben- ghalensis, it also produces cleistogamous flowers on short subterranean shoots. In West Africa, where it is native, some plants have both perfect and staminate flowers (Morton, 1967). Morton (1967) has provided much information on the biosystematics of the genus in West Africa. Commelina ascendens Morton, 2n = 28, is a scrambling herb with the leaves and stems covered with uncinate hairs. The main stems are procumbent, and the erect branches bear inflorescences. Commelina zam- besiaca C. B. Clarke, 2n = ca. 56, has leaf blades that are asymmetric at the base. Commelina africana L., 2n = 28 (30), has yellow flowers. As a general- ization, the polyploid West African taxa of Commelina occur at higher ele- vations and in moister habitats than the diploids. Barnes, in an excellent account of floral biology of the southern Indian species, documented a diversity of phenological types in the genus. The opening and pollination of flowers in a cyme affect the flower type and maturation of remaining flower buds. In general, if early flowers of a cyme are pollinated and set seed, the remaining flower buds develop stamens only. If the buds or earlier flowers are removed, the later buds may develop ovaries and stamens. There is a developmental interdependence among the flowers of an inflorescence that probably is mediated by as-yet-undetermined plant growth substances. Barnes also described anthesis and self-pollination. After the flower has been open for a few hours, the filaments coil, bringing the anthers into contact with the stigma and effecting self-pollination, as in Commelina clavata C. B. Clarke and C. attenuata Koenig, 2n = 48. In C. hirsuta C. B. Clarke, the filaments coil, but not enough to bring the anthers into contact with the stigmas. Neither style nor filaments coil in C. indehiscens E. Barnes, and no selfing occurs. Selfing occurs before coiling in C. tricolor E. Barnes, where the anthers are very close to the stigma at anthesis. In C. Kurzii C. B. Clarke and C. undulata R. Br., both the style and the filaments coil two to three times, bringing about effective selfing. In C. Forskalaei the filament of the median stamen bends slightly, while the style coils strongly, bringing it into contact with the stigma and effecting self-pollination. Owens tested 13 species of Commelina (including C. benghalensis, C. diffusa, and C. erecta from the Southeast) and found that all were self-compatible. Faden has investigated the floral biology of Commelina erecta in eastern Texas. As in the Indian species, both perfect and staminate flowers are produced in the same cyme. Perfect flowers are produced first, and if seed set occurs, later flowers are staminate. Little selfing occurs. 1989] TUCKER, COMMELINACEAE 129 REFERENCES: Under family references see BAILEY ef a/.; FADEN & HUNT; Morton (1967); OWENS; RADFORD et al.; WILSON; and WUNDERLIN. BarAnov, A. I. A contribution to the knowledge of the life history of Commelina communis. Quart. Jour. Taiwan Mus. 17: 81-113. 1964. [Illustrations.] BARNES, E. Some observations on South Indian commelinas; two new species of Com- melina from South India. Jour. Bombay Nat. Hist. Soc. 46: 70-89. 1946. [Illustra- tions. ] Bates, J.C. Seed germination and enearee of the seedling of Commelina virginica. Trans. Kansas Acad. Sci. 42: 109-112. 1939. BERGER, C. A., A. L. LAFLEur, & E. R. Witkus. Mitosis in Commelina communis L. Jour. Hered. 43: 243-247. 1942. BRASHIER, C. K. A revision of the genus Commelina (Plum.) L. in the U. S. A. Bull. Torrey Bot. Club 93: 1-19. 1966. [Four species; C. gigas omitted.] Bubp, G. D., P. E. L. THomas, & J. C. S. ALLISON. Vegetative regeneration, depth of germination, and seed ae in Commelina benghalensis L. Rhodesian Jour. Agr. Res. 17: 151-153. DuncaAN, W. H. Oe an a species new to the United States. Brittonia 19; 282. 1967. FApEN, R. B. Reproductive biology of Commelina erecta L. (Commelinaceae) in eastern Texas. (Abstract.) Am. Jour. Bot. 72: 953. 1985. Fassett, N. C. Another driftless area endemic hy ommelina erecta var. Greenei, var. nov.). Bull. Torrey Bot. Club 70: 398, 399. 1943. FusISHIMA, H. Karyological studies on Comme oe eae. VI. Chromosome numbers in Commelina communis complex and their geographical distribution in Japan. Kro- mosomo 21/22: 605-610. 1981. . VIL. Karyotypes of Commelina auriculata and C. diffusa from Japan. Jour. Fac. Educ. Tottori Univ. 4. 1984. Fukumoto, K. Exam ination of chromosome numbers and karyotypes in the genus Commelina. I. Intraspecific variation of the chromosome numbers in Commelina communis L. Kromosomo, II. 13: 350-354. 1979 Hsu, K. S. Seed morphology and seed germination of genus Commelina in Taiwan. Quart. Jour. Chin. Forest. 5: 55-60. 1971. Lewis, W. H., H. L. Striptine, & R. G. Ross. Chromosome numbers for some angio- sperms of the southern United States and Mexico. Rhodora 64: 147-161. MacKeever, F.C. A variegated foliage form of Commelina. Rhodora 63: 88-90. 1961. MAHESHWARI, S. C., & B. BALDEV. A contribution to the ae ae and embryology of Commelina Forskalaei Vahl. Phytomorphology 8: 277-298. & AHESHWARI. Floral dimorphism in Commelina eee L. and C. benghalensis L. Phytomorphology 5: 413-422. 1955 INGH. A preliminary note on the morpholoss of the aerial and underground flowers of Commelina benghalensis Linn. Curr. Sci. Bangalore 3: 158-160. 1934. MAL LocH, K. R., & R. FENTON. Reversible effects ae on Commelina communis L. New Phytol. 88: 249-254. 1981. [Stomatal closur McLain, D. D. The cytology and taxonomy of Ustilago ee Diss. Abstr. 21(5): 1960. MItTsuI, S., K. HAYASHI, & S. HATToRI. Crystallization and properties of commelinin, a blue a ee from Commelina. (In Japanese.) Bot. Mag. Tokyo 72: 325-333. p) a ———. Further studies on commelinin, a crystalline blue metallo- 130 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 anthocyanin from the flowers of Commelina, studies on anthocyanins XX XI. Proc. Japan Acad. 35: 169-174. 1959b. Morton, J. K. Cytotaxonomic studies on the Gold ee species of the genus Com- me lina Linn. Jour. Linn. Soc. Bot. 55: 507-531. PENNELL, F. W. The eae als (Plumier) L. in the United States. Bull. Torrey Bot. Club 43: 96-111 6. [Four species. ] SAXE, ae “ Suc and ae eee study of the coordinated reactions of individual Co tomata (Commelinaceae). Am. Jour. Bot. 66: 1044-1052. 1979. SmitH, R. J., Jk. Competition of spreading dayflower (Commelina diffusa) with rice (Oryza sativa). Weed Sci. 32: 116-119. 1984 Souéces, R. Embryogénie des Commelinacées; développement de l’embryon chez le Commelina communis L. Compt. Rend. Acad. Sci. Paris 246(1 4): 2082-2086. 1958. Upuor, J. C. T. Vergleichende Bliitenmorphologie und bliitenbiologische Studien an Commelina virginica L. Ber. Deutsch. Bot. Ges. 52: 173-180. 1934 Trocuain, J. Sur la biologie de deux Commelinacées (Commelina For. ae Vahl et C. benghalensis Linn.). Compt. Rend. Acad. Sci. Colon. 194: 743-7 1932. WALKER, S. R., & J. P. EvENSON. Biology of Commelina pe eae: in oe Queensland, Australia. 1. Growth, development, and seed production. Weed Res. 25: 239-244. 1985; 2. Seed dormancy, germination, and ae Ibid. 245-250. WuNANDS, D. O. The botany of the Commelins. viii + 232 pp. 64 pls. Rotterdam. 1983 1989] KRAL, ERIOCAULACEAE 131 THE GENERA OF ERIOCAULACEAE IN THE SOUTHEASTERN UNITED STATES! ROBERT KRAL? ERIOCAULACEAE Desvaux, Ann. Sci. Nat. XIII. 39: 45. 1828, “Eriocau- loneae,” nom. cons. (PIPEWORT FAMILY) {1 Annual or perennial often dioecious), diminutive or robust, stemless or caulescent, usually rosulate; scapose aguatie or wetland herbs, mostly of acidic waters and substrates. Rootstock a rhizome or caudiciform, the roots fibrous, spongy, or spongy-septate (diaphrammed), with simple or paired root hairs. Stems short or elongate (in the latter often spreading-as- cending), repent or prostrate, the axis sympodial or monopodial, sometimes branching. Leaves mostly polystichous, usually in rosettes, sometimes in looser spirals in the elongate-stemmed species, in rosettes spaced along an elongate slender axis, or in involucral rosettes in compound-scaped cake mostly graminiform, linear-lingulate, parallel nerved, the bases usually , open, clasping, the stomata paracytic. Scape(s) simple or compound, Boe ae om or toward rosette center, | to many, each arising axillary to a rosette leaf, enclosed at base by a tubular, distally open, spathelike sheath, the scape axis terete, usually twisted, mostly few-to-many-costate, terminating in a single, 'Prepared for the Generic Flora of the Southeastern United States, a long-term project made possible by grants from the National Science Foundation and currently supported by BSR-8415769 (C. E. Wood, Jr., principal investigator), under which this research was done, and BSR-8415637 (N. G. Miller, principal investigator). This account, the 126th in the series, follows the format established in the first paper (Jour. Arnold Arb. 39: 296-346. 1958) and continued to the present. The area and ith references that were not immediatly available to me, and to Barhara Nimblett and Ihsan The illustration of Syngonanthus was drawn for the Generic Flora in 1975 or = ee made Kenneth R. Robertson. The liquid-preserved specimens used were collected by R mber Hillsborough County, Florida; herbarium material by Kral and Kral (6777; Gu) in a ae Cou mt Florida. Living plants aeoee by George K. Rogers from Columbus County, North Carolina, were used to check various deta 2Department of General ee Herbarium, Box 1705, Station B, Vanderbilt University, Nash- ville, Tennessee 37235 © President and Fellows of Harvard College, 1989. Journal of the Arnold Arboretum 70: 131-142. nue 1989. 132 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 chaffy, involucrate, sometimes proliferative, buttonlike head, the involucral bracts usually spirally imbricate in few to many gradate series, the receptacle smooth or variously hairy or chaffy. Flowers (florets) imperfect, regular (acti- y ophore). Petals lacking, 2, or 3, evident as lobes diverging from a stipelike base (androphore or gynophore) or merely a part of a lobeless column. Stamens 2- 4(-6), when 2 or 3 alternating with sepals, often unequal; filaments epipetalous or appearing to arise from the apex of an androphore when petal lobes are reduced; anthers introrse or versatile, 2- or 4-sporangiate, I- or 2-locular, dehiscing longitudinally; pollen monosulcate, trinucleate. Staminodia present in some 2- or 3-merous flowers, then glandlike or peglike and sometimes fimbriate or variously cleft, rarely with a stipitate blade. Gynoecium syncar- pous, 2- or 3-carpellate; style terminal, 2- or 3-branched, the branches simple or rebranched apically, in some species the stylar apex producing appendages alternate with style branches; ovary superior, raised on a gynophore, 3-locular, the ovules | in each locule, ventral-apical, pendulous, orthotropous, 2-integ- umented. Fruit a thin-walled, loculicidal capsule; seeds small, variously ribbed or sculptured, the embryo small, apical, the endosperm copious, mealy-starchy, with compound starch grains. Type GENUS: Eriocaulon L A large, mainly pantropic family of about 13 genera and at least 1200 species, with a distinct majority in the Neotropics and in the genera Paepalanthus Mart. (about 500 species), Eriocaulon L. (about 400), and Syngonanthus Ruhland (about 200). Under the assumption that reduction in ovule number and uni- formly imperfect flowers constitute evolutionary advancements, a present and popular view is to place the family as derived from the Commelinales, but distinct from that order (cf. Cronquist; Hutchinson; Takhtajan, 1980; Thorne, 1983 In the New World only five genera—namely (in order of number of species) Eriocaulon, Paepalanthus, Lachnocaulon Kunth, Syngonanthus, and Tonina ublet—are known to occur north of South America; just three of these are found in the United States, with only Eriocau/on being represented outside the Southeast. Of the three, Eriocau/on is the sole representative of the subfamily Eriocauloideae, while Lachnocaulon and Syngonanthus are treated under the other subfamily, Paepalanthoideae Ruhland. The family is mainly distinguished by its combination of rosulate and sheathed- scapose habit; vessels with simple or scalariform perforations in all organs; leaves with polystichous arrangement and open, clasping bases (true wetland or aquatic species often with lacunar tissue evident); vascular bundles of stem closed and bicyclic; cuneate proteinaceous inclusions in sieve-tube plastids; root hairs on most roots; monoecious [or dioecious] habit; flowers and invo- lucral bracts and bractlets chaffy, the flowers crowded in tight spirals in in- volucrate heads; perianth 2- or 3-merous, |- or 2-cyclic, regular or irregular, usually valvate in bud and often with mealy pubescence, the petals distinct or variously connate, at base incorporated into an androphore or a gynophore; androecium of 2 or 4, or 3 or 6 stamens, often in the trimerous flowers with 1989] KRAL, ERIOCAULACEAE 133 3 staminodia or glands at tube apex; gynoecium variously raised on the gyn- ophore, 2- or 3-carpellate, the ovary 2- or 3-locular, each locule with a single ventral-apical, pendulous ovule; style branches 2 or 3, sometimes rebranching at stigma level [in many species producing appendages from the apex of the stylar column]; and fruit capsular, thin walled, loculicidal, the seeds small, with copious mealy endosperm and a tiny lateral-apical embryo. Trichomes in most species are many and varied on all or nearly all organs. These may constitute at once the most neglected and the most useful taxonomic characters. REFERENCES: Arser, A. Leaves of the Farinosae. Bot. Gaz. 74: 80-94. 1922. [Eriocaulaceae, 84.] onocotyledons, a morphological ae Frontisp. + 258 pp. Cambridge, En- gland. 1925. [Eriocaulaceae, 88 (fig. 66), 211.] BEAL, E. O. A manual of marsh and aquatic vascular plants of North Carolina. N. Carolina Agr. Exper. Sta. Tech. Bull. 247. iv + 291 pp. 1977. [Eriocaulaceae, 145- 147.] CHAPMAN, A. W. Flora of the southern oe States. ed. 3. xxxix + 655 pp. Cambridge, Massachusetts. 1897. [Eriocaulaceae, 52 see CorRELL, D. S., & M. C. Jonnston. Manual of the vascular plants of Texas. x1 + 1881 pp. Renner, Texas. 1970. ee 352-354. Cronautst, A. The evolution and classification of flowering plants. vii + 396 pp. Boston. 1968. Cera Eriocaulaceae, 335, 336; phylogeny. . An integrated system of classification of flowering plants. xviii + 1262 pp. New York. 1981. [Eriocaulales, 1116-1118.] Fassett, N. C. A manual of aquatic plants. | + 382 pp. York, Pennsylvania. 1940. grea ga 169. FERNALD, M. Gray’s manual of botany. ed. 8. Ixiv + 1632 pp. New York and other cities. 5 [Eriocaulaceae, 290-292; Eriocaulon, Lachnocautlon.] G EASON, H. A. The new Britton and Brown illustrated flora of the northeastern United States and adjacent Canada. Vol. |. Ixxv + 482 pp. New York. 1963. [Eriocaulaceae; Eriocaulon, Lachnocaulon, 372-374 Goprrey, R. K., & J. Wooten. Aquatic and wetland plants of southeastern United States. Monocotyledons. x + 712 pp. Athens, Georgia. 1979. [Eriocaulon, Lach- nocaulon, Sy ngonanthus, 503-530: illustrated. ] HEGNAUER, Chemotaxonomie der Pflanzen. Band 2. Monocotyledoneae. 540 pp. Basel and Stuttgart 1963. [Eriocaulaceae, 152-154.] HELLouist, C. B., & G. E. Crow. Aquatic vascular plants of New England. Part 5. Araceae, pee ee Xyridaceae, Eriocaulaceae and Pontederiaceae. New Hamp- Hieronymus, G. Eriocaulaceae. Jn; A. ENGLER & K. PRANTL, eds., Nat. Pflanzenfam. 88 Hotcukiss, N. Common marsh, underwater and floating-leaved plants of the United States and Canada. Dover Press ed. 124 pp. New York. 1970. [Enocaulaceae, Er- iocaulon, 25; illustrated.] HurtTcuinson, J. The families of flowering plants. ed. 3. xvi + 968 pp. a 1973. [Eriocaulaceae, 710-712; very useful for synoptic presentations of ge KoeRNICKE, F. Eriocaulaceae. /n: C. F. P. von Martius, FI. Brasil. 3(1): 374-508. 1863. [The beginnings of subfamilial and generic classifications in the family. ] KRAL, R. Eriocaulaceae of continental North America north of Mexico. Sida 2: 285- 134 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 332. 1966. [A general discussion of the family, plus descriptions of Eriocaulon, Lachnocaulon, and Syngonanthus. eae all species described and ue ys : Lawrence, G. H. M. Taxonomy of vascular plants. xiit + 823 pp. New aa [Eriocaulaceae, 404, 405; still one ic best descriptions of the fam Martin, A.C. The comparative internal morphology of seeds. Am. Midl. nie 36: 513- 660. 1946. [Eriocaulaceae, 533.] Martius, C. F. P. von. Die Eriocauleae, als selbstandige Pflanzen- ers aufgestellt und erldutert. Nova Acta Acad. Leop.-Carol. 17(1). 72 pp. 5 pls. Mo.penkg, H. N. Eriocaulaceae. Jn; N. Am. Fl. 19: 17-50. 1937. [A pele treat- ment of paces ene Paepalanthus, Syngonanthus, Tonina, and the basis for most subsequent studies on the plants in North America. Additional sites on nile EAocaulscese 1 Phytologia 1: 309-336. 1939a; II, ibid. 343- 364. 1939b; V, ibid. 3: 382-400. 1950a; VI, ibid. 468-472. 195la; VU, ibid. 491-502. 195 1b; VIII, ibid. 4: 134-152. 1952a; IX, ibid. 200-208. 1952b; xX, ibid. 295-304. 1952c; XI, ibid. 311-341. 1953a; XII, Bull. Jard. Bot. Bruxelles, B. 27: 115-141. 1957; XXX VIII, Phytologia 24: 333-358. 1972a; XLI, ibid. 25: 121-166. 1973a; XLII, ibid. 245-248. 1973b; XLII, ibid. 26: 15-47. 1973c; XLIV, ibid. 136- 147. 1973d; XLVI, ibid. 226-262. 1973e, XLVII, ibid. 455-484. 1973f. Novelties in the Eriocaulaceae and Verbenaceae. /bid. 2: 6-32. 1941; II, Bull. Torrey Bot. Club 77: 389-405. 1950b . Supplementary notes on the Eriocaulaceae, a = Verbenaceae of Texas. I. Phytologia 2: 123-128. 1944; II, ibid. 152-168. . The known geographic distribution of the members ar he Eriocaulaceae. 64 pp. Ann Arbor, Michigan. The known geographic distribution of the members of the Eriocaulaceae, to- “7 with a check list of scientific names proposed in the group. New York. 1946c. 1946d; Suppl. 2, Revista Sai Bot. 8: 165-174. 1950c; Suppl. 4, Phytologia 3: 374-382. 1950d; Suppl. 5, ibid. 448-468. 1951c; Suppl. 8, ibid. 4: 65-88. 1952d; Suppl. : a 119-124. 1952e; Suppl. 11, ibid. 450-452. 1953b; Suppl. 12, ibid. 5: 2951. ad notes on the Eriocaulaceae, Avicenniaceae and Verbenaceae of Texas I. Wrightia 1: 220-246. 1948. . Plants collected in Ecuador by W. H. Camp. Eriocaulaceae, Verbenaceae. Mem. N. Y. Bot. Gard. 9: 175-177. 1955. . A résumé of the Verbenaceae, Avicenniaceae, Stilbaceae, Symphoremaceae, and Eriocaulaceae of the world as to valid taxa, geographic distribution and syn- onymy. 495 pp. Yonkers, New York. 1959. [Followed . a 2-volume treatise, Yon- kers, 1960a, and another, Supplementum III, 45 pp., . Notes on new and noteworthy plants XXV. Lae 7: 119-123. 1960b. fifth summary of Verbenaceae, Avicenniaceae, Stilbaceae, pie uenimena Symphoremaceae, Nyctanthaceae and Eriocaulaceae of the world as to valid taxa, geographic distribution and synonymy, suppl. 1. /bid. 23: 413-438. 19736: Suppl. 2, ibid. 25: 225-245. 1973g; Suppl. 3, bid. 26: 356-377. 1973h. —. Seven undescribed subspecific taxa in — and Verbenaceae. /bid. 23: 453, ee 1972c. [Includes new species in Eriocau A sixth summary of the Verbenaceae, Avicenniaceae, Stilbaceae, Chloanthaceae, ace Nyctanthaceae, and Eriocaulaceae of the world as valid taxa, geographic distribution, and synonymy. Suppl. |. /bid. 50: 233-270. 1980. Morona, T. Notes on the North American species of Eriocaulaceae. Ba Torrey Bot. Club 18: 351-362. 1891. PouLsen, V. A. Anatomiske Studier over Eriocaulaceerne. Vidensk. Medd. Naturh. For. Kjobenhavn, IV. 10: 221-386. 1888. 1989] KRAL, ERIOCAULACEAE 135 —. Bemaerkninger om Tonina fluviatilis Aubl. Bot. Tidssk. 18: 279-292. pls. 20, RaApForD, A. E., H. E. AHLES, & C. R. BELL. Manual of the vascular flora of the Carolinas. Ixi + 1183 pp. Chapel Hill, North Caroline: 1968. [Eriocaulaceae, 265-267; illus- trated. Ramaswamy, S. N., B. G. L. Swamy, & D. A. GovinpappaA. From zygote to seedling in Eriocaulon robusto- Fe ninmum Ruhl. (Eriocaulaceae). Beitr. Biol. Pflanzen 55: 179-188. 1981. [Developmental anatomy; illustrated.] Rickett, H. W. Wildflowers of the United States. Vol. 2. The Southeastern States, part I. New York. 1966. [Eriocaulon, Lachnocaulon, Syngonanthus, 135, pl. 27.] RUHLAND, W. Eriocaulaceae. Pp. 482-494 in I. URBA Bee a Antillanae. Vol. 1. Berlin. 1900. [Eriocaulon, Paepalanthus, Sy eee ne Tonin —. Eriocaulaceae. Jn: A. ENGLER, ed., Pflanzenr. IV. 30(Heft 13). 294 pp. 1903. . Zur geographischen Verbreitung der Eriocaulaceen. Bot. Jahrb. 50: 363-374, 1914. iocaulaceae. Jn: A. ENGLER, ed., Nat. Pflanzenfam. ed. 2. 15a: 39-57. 1930. [Synoptic presentation of subfamilies and genera; illustrations. ] SMALL, J. K. Flora of the southeastern United States. xi1 + 1370 pp. Lancaster, Penn- sylvania. 1903. [Eriocaulaceae, 234-236; Eriocaulon lineare Small, 236. anual of the southeastern flora. xxii + 1554 pp. New York. 1933. (Reprinted Sel Hill, North Carolina. 1953.) eabeuieoe 355 258.] SOLEREDER, H., & F. J. MEYER. Eriocaulaceae. Systematische Anatomie der Monoko- tyledonen. Vol. 4, Pp. 50-70. Berlin. 1929. STAUDERMANN, W. Die Haare der Monokotylen. ae ae 8: 105-184. 1924. [Erio- caulaceae, 125, 126, pl. 9, figs. 8-11; pl. 10, figs. STEUDEL, E. G. Eriocauloneae. Synopsis plantarum pre ee Part 2. Pp. 267-283. See 1855. [Eriocaulon, Syngonanthus.] TAKHTAJAN, A. Die Evolution der Angiospermen. 344 pp. Jena. 1959. [Superorder Farinosae, order Eriocaulales, 273. . Outline of the classification of flowering a [Magnoliophyta]. Bot. Rev. 46: 225-359. 1980. [Superorder V. Commelinana ae NI, G. Contribution to the pollen manson of Eriocaulaceae. Pollen Spores 7: 181-191. 1965. TuHorng, R. F. A phylogenetic classification of the Angiospermae. Evol. Biol. 9: 35- 106. 1976. [Superorder Commeliniflorae, order Commelinales, suborder Eriocau- lineae. ] Proposed new realignments in the angiosperms. Nordic Jour. Bot. 3: 85-117. 1983. [Superorder Commeliniflorae, order Juncineae, suborder Eriocaulineae.] TomLINson, P. B. Notes on the anatomy of Aphyllanthes ae ei and comparison with secon, ay Linn. Soc. Bot. ~ ae ae caulaceae. 46-192 in C. R. CAL ee Anatomy of the mono- coiled, Vol. 3. rie 1969. [The ae ae treatise on the anatomy of the fam UPHOF, 1 "C T. The floral behavior of some Eriocaulaceae. Am. Jour. Bot. 14: 4-48. 1927 WARD, D. B. Checklist of the vascular flora of Florida, Part J. Univ. Florida Agr. Exper. Sta. Tech. Bull. 726. 1968. [Eriocaulon, Lachnocaulon, Syngonanthus, 47, 48.] Key TO THE GENERA OF ERIOCAULACEAE IN THE SOUTHEASTERN UNITED STATES A. Lacunar tissue evident at leaf bases; larger roots thickened, septate, appearing artic- ulated and unbranched; perianth parts in 2’s [3’s in the Asian adventive Eriocaulon cinereum], each petal adaxially glandular; stamens (3 or) 4(-6), the anthers 2-locular, 136 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 staminal cua apically with 2 or 3 glands at center; carpels 2, style branches unappendaged, 2 [carpels 3, style branches 3 in many western, tropical American, or Asian ar path p hole sie oan Ge ee tee eae ee ees 1. Eriocaulon. . Lacunar tissue not evident at leaf bases; larger roots either fibrous and evidently branched or thickened and fleshy (and then appearing neither septate nor branched): perianth parts in 3’s, the petals if present eglandular; stamens 2 or 3, the anthers 1- or 2-locular, apex of staminal column usually with 3 appendaged or unappendaged ag carpels (2 or) 3. ots dark, slender, fibrous, evidently branched; scapes smooth or hairy, the ae neither swollen nor glandular apically; expanded inflorescence globose or short-cylindrical, the outermost involucral bracts reflexed and obscured by the rest of the inflorescence; trichomes of perianth tips clavate; staminal filaments ae to rim of an androphore, the anthers unilocular; style branched with bifid Se, water aie es gate ane cme oe ea Shea aes Lachnocaulon. ce ot thickened, spongy, appearing unbranched; scapes hairy, at least some of the hairs swollen at base, often glandular distally; expanded inflorescence hemispheric, outer bracts not obscured by the rest of the inflorescence; trichomes of perianth tapering, acute, not clavate; staminal filaments separating from corolla toward apex of tube, the anthers bilocular; style branches undivided. ......... af Ses cecais oie TAB a ta if cap b iad tee Ty alk ids austaian ay atid atk tne aacdear ee 3. Syngonanthus. > ow Subfamily ERIOCAULOIDEAE [Ruhland] 1. Eriocaulon Linnaeus, Sp. Pl. 1: 87. 1753: Gen. Pl. ed. 5. 38. 1754. Rosulate, scapose, often tufted herbs, the roots evidently cross partitioned, the stems short or elongate, sometimes sparingly branched. Leaves linear- lingulate, often attenuate, gradually or abruptly flaring toward a clasping base, there often noticeably lacunar and pale, greener and less distinctly lacunar distally, the chlorenchymatous part of leaves increasing with extent and du- ration of emergence. Inflorescences scapose, the usually fluted scape encased at base by a tubular, fluted, distally open, often bifid or trifid sheath and terminating in a single hemispheric to globose or short-cylindrical, involucrate head, all parts chaffy and imbricate in several to very many tight spirals. Involucral bracts in gradate spirals, hardly (if at all) longer than flowers, usually broad, green, gray-green, or paler, at least the margins translucent; receptacle trichomiferous, chaffy, or smooth. Sepals 2, scarious, usually equal, fused to a stipelike base, pale or colored, often translucent, navicular, frequently covered toward tips with farinose or translucent, multicellular trichomes. Petals mostly 2, fused at least toward bases, equal or unequal, similar to sepals but generally narrower, shorter, separated from them by a pronounced tubular stalk (partly petal and, in staminate flowers, partly stamen) and distinguishable from them by a dark gland on adaxial distal surface of each; outer and sometimes inner surfaces or margins frequently with multicellular, translucent or opaque tri- chomes; distal end of staminate perianth tube concave, sometimes with 2 or 3 dark glands at its center like those on corolla lobes. (Sepals and petals 3 [in the Californian FE. microcephalum and] in the Asian adventive E. cinereum.) Stamens 3 or 4 to 6, 2 or 3 opposite the petals, the anthers introrse, dorsifixed and usually versatile, bilocular, tetrasporangiate, well exserted from the inflo- rescence at anthesis. Carpels 2, the ovary on a gynophore, 2-locular, the style — 1989] KRAL, ERIOCAULACEAE la? 2-branched (carpels 3, locules 3 in E. microcephalum, E. cinereum). Pericarp thin, often velumlike, the style persisting. Seeds ovoid, ellipsoid, or broadly fusiform, variously lined, ridged, or papillate. LECTOTYPE SPECIES: E. Ree gulare L.; see Britton & Brown, Illus. Fl. No. U. S. & Canada, ed. 2. 1: 454. 1913. (Name from Greek erion, wool, and caulos, stalk, from the i at the base of the scape in the original species, fide Arthur Stanley Pease in Fernald (1950).)— PIPEWoRT, BUTTON-RODS, HAT-PINS (Small). About 400 species, mostly pantropic, with a few representatives extending to temperate or boreal-transitional climates in both Asia and eastern North America, and only one (Eriocaulon aquaticum (Hill) Druce) in northern Europe (the Hebrides, Scotland). Of the ten species reported for the United States and Canada, one (Friocaulon cinereum R. Br.) is Asian, now sporadically established in rice-growing parts of California and Louisiana; another (£. microcephalum HBK.,) is disjunct from the montane Neotropics and has been collected only once from southern Cal- ifornia; and two (£. aquaticum and E. Parkeri B. L. Robinson) are principally of Pleistocene-glaciated latitudes in the eastern or central United States and Canada, each with extensions southward along the Appalachians or the coast into the southeastern United States. The remainder either center in the Coastal Plain of the Southeast with a few extensions into contiguous areas or are endemic to it; of these, only E. decangulare extends to (or 1s extended from) the Neotropics; not surprisingly, this is perhaps the most ecologically ample of our species. Of the three southeastern genera in the Eriocaulaceae, Eriocau/on has the largest number of truly wetland or aquatic species and the most anatomical characters associated with an aquatic habit. As to soil or soil-water reactions, all species except Eriocaulon Parkeri, usually in tidally influenced, circumneutral or even brackish sites, and £. Rav- enelii Chapman, which favors wetlands transitional to brackish marsh, are found in acidic wet sites, such as pine flatwoods, low savannas, acid pond shores and shallows, sphagnous seeps and bogs, and seeps or pools on or around arenaceous rock. Of these, EF. decangulare appears to be the most drought tolerant and is commonest in the drier transitions to wetlands. Harold N. Moldenke, the world authority on the pee treated 43 species for the North American Flora (1937); a clear majority a atively narrow endemics, some still unknown beyond the type or a ee collections: Since trichome characters and the relative dimensions of accrescent parts tend to vary considerably within well-known seek it is quite possible that a bio- systematic overview will be more conservativ Of the three genera that occur in the ie only Eriocaulon has nectar- iferous glands. Insect visitors are rare. It could be assumed that, for our species, the most common agent of pollination 1s wind. REFERENCES: Under family references see KRAL and MOLDENKE (1937). ERLANDSON, S. The chromosome numbers of three Eriocau/on species. Ark. Bot. 30B(2). pp. 1942. [Literature cited; illustrations. 138 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 FERNALD, M. L. Misinterpretation of Atlantic Coastal Plain species. Rhodora 44: 245. 1942. [Eriocaulon septangulare in Georgia?] Fyson, P. F. The Indian species of Eriocaulon. Jour. wa Bot. 2: 133-150. 1921. [Comments on differences in number of perianth parts in same species. Hare, C. L. The structure and development of ae septangulare With. Jour. Linn. Soc. Bot. 53: 422-448. 1950. Hout, P. Eriocaulon decangulare L., an anatomical study. Bot. Gaz. 31: 17-37. 1901. Hupson, W. Nasmythia. Flora Anglica. ed. 2. Pp. 414, 415. London. 1778. [Nasmythia articulata Hudson (Friocaulon articulata Hudson; Eriocaulon aquaticum (Hill.) Druce).] Hutten, E. The amphi-Atlantic plants and their phytogeographical connections. Sv. Vet-akad. Handl. IV. 7(1): 1-340. 1958. [Eriocaulon septangulare, 206, map 188.] KOERNICKE, O. Eriocaulaceae. Linnaea 27: 567-596. 1854. [Contains stir ae for cats texanum Koern., Lachnocaulon Beyrichianum Sporl., L. glabrum Koern., L. digynum Koe KRAL, R. A oe on some rare, threatened, or endangered, forest-related vascular plants of the South. Vol. |. Isoetaceae one: Euphorbiaceae. U.S. Dep. Agr. Forest Serv. So. Region Tech. Publ. R8-TP2. x + 718 pp. Atlanta, Georgia. 983, [Erio- caulon pee ene 171-174; pene ecology, flowering times, description, management. ] MaeKAwa, F. Eriocaulon heleocharioides may belong to the same section with American E. era (In Japanese.) Jour. Jap. Bot. 26: 116. 1951. Movpenkg, H. N. Notes on new w and a eer plants VI. Phytologia 3: 35-46. 1948; XII, ibid. 406-448. 1951; XVI, ibid. 4: 285-295. 1953. Eriocaulaceae. Inst. Paran. Bot. Catal. Gén. 10: 1, 2. 1956. [Eriocaulon, 441 species. | RAMASWaMy, S. N., & G. D. AREKAL. rer of Eriocaulon setaceum (Eriocau- laceae). Plant Syst. Evol. 138: 175-188. 19 Rosrinson, B. L. A hitherto undescribed ne from New Jersey. Rhodora 5: 175, 176. 1903. LE. Parkeri.] Royven, P. vAN. Eriocaulaceae. Nova Guinea, IJ. 10: 21-44. 1959. w species in Eriocaulon. Blumea 10: 126-135. 1960. RUHLAND, W. Kritische aie der afrikanischen Arten der Gattung Eriocaulon L. Bot. Jahrb. 27: 65-85. SATAKE, Y. On what I eee a Dr. F. Maekawa’s opinion, “Eriocaulon heleocharioides may belong to the same section with American EF. septangulare.” (In Japanese.) Jour. Jap. Bot. 26: 221. 1951. . Miscellany on £riocaulon. |. (In Japanese.) [bid. 27: 264-268. 1952. [Latin and English diagnoses of Friocaulon.] SmitH, R. W. The floral development and embryogeny of Eriocaulon septangulare. Bot. Gaz. 49: 281-289. pis. 19, 20. 1910. SoLtomon, R. The anatomy : and root of Eriocaulon septangulare. Jour. Indian Bot. Soc. 10: 139-144. Subfamily PAEPALANTHOIDEAE Ruhland 2. Lachnocaulon Kunth, Enum. Pl. 3: 497. 1841. Annual or perennial, rosulate, scapose, often tufted herbs, the roots branched, slender, fibrous, with paired root hairs; perennation by lateral offshoots or short ascending rhizomes. Rosette leaves linear, acute or attenuate, in very flat spirals toward the apex of a short or elongate, sparingly branched stem, the bases 1989] KRAL, ERIOCAULACEAE 139 clasping, crowded, none of the leaf evidently lacunar, often hairy with slender, multicellular, translucent hairs; old leaf bases persisting as scales on stems. Inflorescences with scape base in a tubular, costate, apically open sheath, this hairy like leaves; scapes usually several, from axils of central rosette leaves, slender, twisted, variously costate, often with trichomes, terminating in a small (1 cm broad or narrower), white, gray, gray-brown, or brown, globose to ovoid or short-cylindric, involucrate head, each comprising an outer involucre of few to many spirally imbricate series of bracts and an inner bracteate, spirally imbricate mass of bracteate, chaffy, imperfect flowers, both staminate and carpellate. Receptacle of head covered by pale or yellowish, multicellular, fili- form trichomes and by the floral bracts (1 or 2 per flower), these mostly about the size of involucral bracts, often navicular or strongly convex and with a few translucent or opaque, usually clavate, multicellular trichomes at the tip. Sepals 3, nearly distinct, similar to receptacular bracts, smooth or sparingly clavate- hairy. Petals lacking or reduced to hairs or small scales. Stamens 2 or 3, elevated ona tubular androphore at least as long; filaments linear; anthers oblong-linear, bisporangiate, unilocular, yellowish or pale, dorsifixed, introrse or versatile, exserted from heads at anthesis; staminodia (2? or) 3, lance-ovoid or peglike, often with fimbriate appendages. Gynoecium on a short, usually comose gyno- phore, 3-carpellate, the ovary 3-locular, 3-ovulate (2-merous in L. digynum); styles (2- or) 3-branched, the stylar column appendaged at apex, with appendages similar to those of staminate flowers. Fruit and seeds like those of Eriocaulon. Type species: L. Michauxti Kunth (Eriocaulon villosum Michx.) = L. anceps (Walter) Morong. (Name from Greek, /achnos, wool, and chaulos, stem, in reference to the long, soft, upwardly pointed hairs on the scapes of the type species.) — HAIRY PIPEWORT, BOG BUTTONS Seven species confined to the southeastern United States and Cuba. This small genus evidently centers in the Coastal Plain of the Southeastern States, with two species (Lachnocaulon cubense Ruhl. and L. Ekmannii Ruhl.) en- demic to Cuba. Our Lachnocaulon and Syngonanthus species inhabit ecologically drier sites than most species of Eriocau/on within the Coastal Plain, the physiographic province in which all the species occur. The optimum habitat is an acid, sandy, or sandy-peat-loam soil, usually one with a gley horizon and therefore saturated much of the year, but definitely one in which water does not stand except during the wettest periods. Soil and soil-water reaction are distinctly acid. Sphagnum is the most commonly associated bryophyte. All species appear to be best adapted to moist ecotones around acid wet areas, with the most diminutive species, L. digvnum Koern., tolerating the wettest—and L. anceps and L. Bey- richianum Sporl. extending into the least moist—sites. Root systems within the genus are shallow, fine, and diffuse; the roots are eseptate and with paired, elongate and delicate hairs, providing a good adaptation to drier situations. Both staminate and carpellate flowers develop peglike, often cleft or fringed appendages apically, in the former case on the androphore and alternating with the filament bases and in the latter alternating with the style branches. In the carpellate flowers, petals are reduced to three clumps of trichomes arising from . 70 a! \ . \ NY oN 4 \ i Hh A LHe _ Figure 1. Syngonanthus. a—p, S. flavidulus: a, flowering plant, leaves decumbent from the basal rosette, showing sheaths at bases of peduncles, x '2; b, sheath enclosing base of peduncle, x 12; c, flowering head at beginning of anthesis, bracts of head a anthers of staminate flowers visible, x 12; d, detail of peduncle just below head, showing indumentum of divergent glandular trichomes and adpressed, simple ones, x 24; e, uniseriate, septate, glandular trichome, x 100; f, portion of head in vertical section, anthers of staminate flower and styles and petals of carpellate flower projecting above bracts, 2 bracts in section at extreme right, x 24; g, staminate flower, the calyx lobes connate below, pubescent above, the petals united, vertically folded around the filaments, carpellate flower, free petals (see “‘l’’) ees Laan pubescent | calyx lobes, tips ‘of 3 stigmas protruding between petals, x 4 bands, = 6; 1, fruit partly enclosed by ener nase toate lobes removed), showing seed through opening in | locule of capsule, x 50; m, fruit, perianth removed, styles 1989] KRAL, ERIOCAULACEAE 141 the gynophore above the sepals. Stamens provide yet another distinguishing feature, the anthers being bisporangiate but one-locular at anthesis. REFERENCES: Under family references see KRAL and MoLpDENKE (1937); under references for Erio- caulon, see KRAL and MOLDENKE (1956) 3. Syngonanthus Ruhland in Urban, Symbol. Antill. 1: 487. 1900. Low [to moderately tall, slender to coarse, solitary to] densely caespitose, [annual or] perennial, monoecious [or dioecious], rosulate herbs, the stems short [to very elongate], the primary roots mostly thickened-porous, pale, esep- tate, the branch roots much more slender, angular, with paired root hairs. Stems short [to elongate, sometimes branched]. Leaves mostly slenderly linear, smooth or variously hairy, often recurved, mostly firm and not evidently la- cunar. Scape(s) solitary or many, from axils of inner rosette leaves, each en- closed at base by a tubular, open-ended, costate sheath, slender, terete and smooth or variously costate and hairy, erect or declinate, simple and arising directly from a basal rosette [or 1 or more times compound, each secondary or subsequént whorl subtended by an involucre or “rosette” of bracteal leaves comparable to but progressively smaller than those of the basal rosette, or scapes proliferative and substoloniferous, glabrous to sparsely or copiously pubescent, usually pilose and with sharp- or gland-tipped hairs often mixed with shorter, glandular ones]. Heads globose, urceolate, turbinate, or hemi- spheric, with many involucral bracts, these pale, papery-chaffy, imbricate in several tight, flat spirals, often gradate, shorter or longer than florets. Receptacle of heads pilose to subglabrate, the hairs elongate, sharp tipped, mostly silvery. Flowers trimerous and nearly regular. Sepals mostly chaffy and pale, distinct to near base, there comprising the outer part of the column (androphore, gyn- ophore), smooth or with long hairs comparable to those of receptacle. Petals subequal; in the staminate flowers much more connate, the androphore a pro- nounced stipe, at the apex flaring into a cuplike, funnelform or clavate, concave- apexed corolla, the rim of which is often 3-toothed; in the carpellate with 3 blades departing from the gynophore and connate or connivent above at the level of the ovary or higher. Stamens 3, the filaments adnate to corolla at apex of androphore, alternating with 3 central glands at its concave center and opposite the lobes of its flaring rim (corolla limb), exserting the anthers beyond them: anthers 2- to 4-sporangiate, bilocular, the locules basally divergent. Ovary trilobed, tricarpellate, the gynophore shorter than the androphore of staminate flowers, the style elongate [or short, clavate], at apex with 3 elongate-linear, papillose stigmatic branches [their bases alternating with stipitate-bladed or connate below and above but free in middle, seed visible through slit in capsule, x 50: n, fruit in diagrammatic vertical section, seed pendulous from apical, axile placenta, endosperm evenly stippled, embryo unshaded, = 50; 0, seed, funiculus above, micropvle below, x 100; p, same in vertical section, endosperm evenly stippled, seed coat and embryo unshaded, x 100. 142 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 simple peglike, bearded appendages, or these lacking.] Capsule and seeds as in Eriocaulon, LECTOTYPE SPECIES: Eriocaulon umbellatum Lamarck (= Syngo- nanthus umbellatus (Lam.) Ruhland). (Name from Greek syngonos, joined to- gether, and anthos, flower, from the connate petals of the carpellate flowers.) About 200 species in Africa, the West Indies, and Central and South America, the earlier-described ones placed in various sections of Paepalanthus, subse- quently transferred to Syngonanthus by Ruhland (1900). The center for these is the planalto of Brazil. In the Neotropics the genus becomes progressively less well represented northward. In North America north of Mexico, there is only S. flavidulus (Michx.) Ruhland (Paepalanthus flavidulus Kunth), shoe- buttons, bantam buttons, a caespitose plant with rhizomes slender, ascending; rosette leaves narrowly linear, acute, strongly recurved, often with a scattering of pustular-based, uniseriate, multicellular hairs; scapes numerous, pilose-glan- dular, wiry, erect or ascending, to 30 cm long; heads pale whitish yellow, chaffy, hemispheric, ca. 5-10 mm broad; receptacles densely translucent-pilose; an- drophores and gynophores hairy at the base: staminate corollas broadly fun- nelform, with 3 broadly triangular lobes; and carpellate corollas with 3 narrow lobes connivent over the ovary, with the slender stylar apex unappendaged. Syngonanthus flavidulus 1s one of the commoner associates of most Lach- nocaulon species, usually in sandy-peaty bog ecotones, flat, exposed sandy washes and seeps, and savannas and flatwoods in the lower Coastal Plain from eastern North Carolina, south through Florida, and west in the Gulf Coastal Plain to northwestern Florida and contiguous Alabama. While this species agrees with many of the tropical ones in the character of its androphore (which appears as a long, pedicellike stipe) and in its cuplike or funnellike corolla with a triglandular center, it differs from most in having staminate flowers with densely pubescent sepal tips, carpellate flowers with a hairier gynophore and petals connate only at the base, and an unappendaged stylar apex. Interpretation of the evolution of species in this now very large genus is best left to those tropical botanists who work within its centers of diversity, particularly Brazil. Many of the tropical species of Syngonanthus and Paepalanthus are both large and showy; the ones with larger and papery heads produce effects much like the “everlastings” (e.g., Helichrysum Gaertner, Anaphalis DC.) and are commercially exploited for dried plant decorations. REFERENCES: Under family references see FERNALD (1950), KOERNICKE, KRAL, MOLDENKE (1937), RUHLAND (1900, 1903), and SMALL; under references for Eriocaulon, see MOLDENKE (1948, 1951 FERNALD, a L. Additions to and subtractions from the flora of Virginia. Rhodora 49: 128. 1947. [Syngonanthus | flavidulus subtracted. Hess, H. sree einige afrikanische Arten der Gattungen Friocau/on und ee aus dem Botanischen Museum der Universitat Ziirich. Bull. Soc. Bot. Suis 263-271. | Mo.penkgE, H. N. Four novelties from Brazil. Phytologia 24: 498, 499. 1972. [Paepa- lanthus, Syngonanthus.| 1989] WHITEFOORD, DOMINICA 143 RECENT PLANT COLLECTIONS FROM DOMINICA C. WHITEFOORD! Recent collections from the Commonwealth of Dominica, Lesser Antilles, are reviewed, and attention is drawn to first records of 40 phanerogams and six ferns. Second collections are also listed where they are considered note- worthy. Dominica is a volcanic island about 28 miles long and 15 miles wide. It is the third largest of the Lesser Antilles and has the highest rainfall. Its moun- tainous terrain has so far protected its magnificent forests from the intensive agricultural development that has cleared the other, less-rugged islands of the group. An account of the ecology and botanical exploration of the island can be found in Hodge (1954). Between 1888 and 1889 Ramage collected plants in Dominica, and the British Museum (Natural History) has 485 of his excellent specimens. The Museum’s Dominican holdings were later increased by duplicates from Hodge, collected between 1937 and 1940, and from Howard, collected in 1950. Other botanists, notably the Bredin-Archbold-Smithsonian collectors of the 1960’s and 1970’s, have subsequently visited the island, but few of their duplicates have reached the British Museum. Since Ramage and Hodge made their col- lections, there have been major changes in Dominica. The hurricanes of 1979 and 1980 devastated both the natural vegetation and the agricultural industry, and much of the resulting financial aid went toward building roads, making more of the island accessible for farming and logging. Large areas of rainforest known to Ramage are now under cultivation, and the pressure to plant more and more bananas is ever increasing. I have visited Dominica on four occasions: 28 August-19 October 1983 (collection numbers 3400-4126), 17 September-7 November 1984 (4/30- 4706), 11 January-—16 February 1986 (51730-5528), and 18 March--12 April 1987 (5550-5859). About 1600 numbers were collected, the first set being deposited at the British Museum (Natural History) (BM), with incomplete sets at the Roseau Botanic Garden, the Smithsonian Institution (us), the Arnold Arboretum (A), and the Missouri Botanical Garden (Mo). The following notes draw attention to those collections that are first records for the island, or are otherwise of interest. The families are listed systematically in the Bm and Kew herbarium order, which is modified from Bentham & Hooker (1862-1883). ‘Department of Botany, British Museum (Natural History), Cromwell Road, London SW7 SBD, England. © President and Fellows of Harvard College, 1989 Journal of the Arnold Arboretum 70: 143-152. January, 1989. 144 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 SPERMATOPHYTA NYMPHAEACEAE Nymphaea rubra Roxb. (First record; cultivated) 5509, growing in the pool at Canefield, where it had presumably been planted for ornament. Since there is a photograph of this striking plant on the tourist brochure, it is strange that it has not been previously recorded. MELIACEAE Trichilia septentrionalis C. DC. (Second record) 4329, growing in forest at the edge of the ravine at Syndicate; sterile. Fruiting material (535/) was later collected from the same tree. It was first collected by Ramage at Laudat in 1888. This species occurs in mainland Central America from Costa Rica to Panama, and in northern South America. OLACACEAE Schoepfia schreberi Gmelin (Second record) 5207, with very young fruit, growing in dry scrub vegetation near Dublanc, and 5252, in similar scrub on the Cabrits. The previous record of this plant was based on an Imray specimen collected sometime between 1832 and 1875. Widespread in the West Indies, it occurs in both Guadeloupe and Martinique (Fournet, 1978). CELASTRACEAE Celastrus racemosus (Reisseck) Loes. (First record) 3869, in fruit, growing in montane thicket at Freshwater Lake. It is distributed throughout Central America. Celastrus grenadensis Urban, from Grenada, is considered a synonym. RHAMNACEAE Gouania lupuloides (L.) Urban (Second record) 4666, scrambling over thickets at Petite Coulibrie. Fruiting material (5202) was later collected near Dublanc. Although common in the West Indies, this plant had not been recorded from Dominica since Imray’s collection (Nicolson, unpubl. ms.), sometime after 1837. SAPINDACEAE Allophylus racemosus Sw. (First record) 4440, growing in dry woodland between Fond Hunte and Fondchére. It occurs generally in Central America and the West Indies but apparently has not been collected in Dominica before. 1989] WHITEFOORD, DOMINICA 145 LEGUMINOSAE Acacia tortuosa (L.) Willd. (First record) 5686, in fruit, growing beside the road near St. Joseph. It occurs in Florida, the West Indies, and South America. Anadenanthera peregrina (L.) Speg. var. peregrina (First record; cultivated or naturalized) 5833, in full flower, and 5834, with older flowers and a few pods. Both were growing near the bottom of the road to Springfield Estate. Native to northern Brazil, Guyana, Colombia, and Venezuela and probably naturalized in some of the West Indies, it is often planted for ornament. Calopogonium mucunoides Desv. (Second record; naturalized) 5399, trailing among grasses beside the road near La Plaine. It was collected in Dominica once before, by Fairchild in 1932. It is native to South America and has become naturalized in the West Indies. Cassia javanica L. (First record; cultivated) 5856, growing in Roseau Botanic Garden; sterile. It is native to Indonesia but is cultivated in the West Indies for its large, attractive, pink and white flowers. Indigofera hartwegii Rydb. (First record) 5550, growing beside the path to Scotts Head. It is a Mexican species and new to the West Indies. Inga vera Willd. subsp. vera (First record) 5634, bearing very young pods, growing at the edge of a banana field near Syndicate. Although it is native from Mexico to northern South America and the Greater Antilles, it is new to the Lesser Antilles. Mimosa arenosa (Willd.) Poiret var. arenosa (First record; naturalized) 5360, in flower, growing beside the road near Coulibistri. A specimen bearing pods (5822) was later collected from the waste ground below the road near Mahaut. Native to Venezuela and naturalized in Mexico, it has been collected recently in Martinique and Puerto Rico (Nicolson, pers. comm.). Mucuna pruriens (L.) DC. f. cochinchinensis (Lour.) Backer (Second record) 5562, growing in dry roadside scrub near Dublanc. Stehlé collected it on Guadeloupe and reported it as introduced from Dominica (Nicolson, unpubl. ms.), but this may be the first collection from this island. 146 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Piscidia carthagenensis Jacq. (First record) 5710, in flower, growing high up on a cliff near Coulibistri. Fournet (1978) described it as endemic to Guadeloupe. Rhynchosia phaseoloides (Sw.) DC. (Second record) 5704, growing at Milton. Widely distributed throughout the Caribbean and South America, it was first collected in Dominica by Imray. HALORAGACEAE Myriophyllum aquaticum (Vell.) Verdc. (First record) 3745, a pondweed from Canefield (sterile), and 55/2 (fertile). This is a South American species naturalized in North America, Southeast Asia, and Australia. It is the first record of the family for the Lesser Antilles. LYTHRACEAE Cuphea hyssopifolia Kunth (Second record; cultivated) 5457, growing at Baiac. It is native to Central America. This plant was attributed to Dominica by Grisebach (1860), based on an Imray specimen that may have been misidentified (Nicolson, unpubl. ms.). Lagerstroemia speciosa (L.) Pers. (First record; cultivated?) 5828, in fruit, growing beside the road near Hillcrest Estate, east of Ports- mouth. Occurring from India to south China and from Indonesia to northern Australia, it is often planted for ornament in the tropics. BEGONIACEAE Begonia heracleifolia Schldl. & Cham. (First record; cultivated) 5459, growing at Baiac. It is native to southern Mexico, Guatemala, and Belize. RUBIACEAE Hedyotis callitrichoides (Griseb.) W. Lewis (Second record) 5374, growing out of concrete steps, at La Plaine. This is a West Indian species and occurs on Guadeloupe, but it has not yet been recorded from Martinique. Ernst made the first collection in Dominica, about 1964. Hedyotis commutata Schultes (First record) 4589, growing at the edge of a cultivated field beside the road between Trafalgar and Trafalgar Falls. Originally from tropical Africa and Madagascar, it has become naturalized in the West Indies. 1989] WHITEFOORD, DOMINICA 147 Vangueria madagascariensis Gmelin (Second record; cultivated) 3747, bearing a few flowers and young fruits, growing beside the road above Canefield. It is also cultivated at Baiac. The previous record was based on a tentatively identified specimen collected by Ernst (Nicolson, unpubl. ms.). This plant is grown for its fruits, which are sold in the market and are known locally as “tamarins des Indes.” COMPOSITAE Chromolaena integrifolia (Bertero ex Sprengel) King & Robinson (First record) 5816, growing on rough ground below the path east of Capucin. It is known from Guadeloupe and Montserrat but not from Martinique. Pectis elongata Kunth var. floribunda Kiel (First record) 4039, growing thickly in a large patch beside the path between the swamp and the sea, East Cabrit. It is no longer to be found there because of the new road to Fort Shirley, but a few more plants were seen on the roadside between Petite Baie and the Picard Estate. It is native to Central America and the West Indies. In Jamaica the popular name for this plant 1s ““verbena grass” (Adams, 1972): ERICACEAE Gaultheria domingensis Urban (First record) 4212, with flowers and fruits, growing in boggy soil in the Valley of Deso- lation, and 5482, from the same locality. First described from Hispaniola, it has also been recorded from Guadeloupe and Martinique (Fournet, 1978). MYRSINACEAE Ardisia crenata Sims (First record; cultivated) 5464, planted at Baiac. It is native to Japan and northern India. ASCLEPIADACEAE Cynanchum parviflorum Sw. (Second record) 5298, hanging from a roadside thicket near Portsmouth. This specimen has exceptionally large, pale yellow flowers. It is the first collection from Dominica since Imray’s (Nicolson, unpubl. ms.), although the species is generally common in the West Indies. 148 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 CONVOLVULACEAE Ipomoea obscura (L.) Ker Gawler (First record) 5505, twining among roadside vegetation near Massacre. It is an Asiatic species but has been recorded from Marie Galante and Guadeloupe (Fournet, 1978) SOLANACEAE Solanum triste Jacq. (First record) 3769, growing near the Belvédére Estate, Délices. Native to coastal Vene- zuela and Trinidad, it is fairly common on Martinique but is not yet recorded from Guadeloupe. SCROPHULARIACEAE Stemodia verticillata (Miller) Hassler (Second record) 5573, growing on a path at Syndicate. It was previously collected by Hodge in 1938 (Nicolson, unpubl. ms.). Widespread in the Neotropics, it is often found as a weed of cultivation. GESNERIACEAE Chrysothemis pulchella (Don ex Sims) Decne. (First record; cultivated?) 3638, growing at the edge of a banana field at Syndicate, but perhaps planted for ornament. Native to Panama and Trinidad, it is sometimes cultivated. Gloxinia perennis (L.) Druce (First record; cultivated) 5458, growing at Baiac. It occurs naturally from Colombia to Peru and is well known in cultivation. ACANTHACEAE Asystasia gangetica (L.) T. Anderson (First record; cultivated) 4218, with mauve flowers, and 42/9, with greenish yellow ones, both growing near the house at Baiac. This is an Indian plant, sometimes cultivated, and is naturalized on the Malay Peninsula and in tropical Africa. Justicia betonica L. (First record; cultivated) 4225, growing at Baiac. It is native to India. LABIATAE Plectranthus verticillatus (L. f.) Druce (First record: cultivated) 4222, flowering, trailing over a tree stump at Baiac. It 1s South African in origin. 1989] WHITEFOORD, DOMINICA 149 AMARANTHACEAE Celosia argentea L. (First record) 4080, growing in swampy ground near the sea, East Cabrit. This is a common tropical plant and has been collected in both Guadeloupe and Martinique. EUPHORBIACEAE Breynia disticha J. R. & G. Forster (First record; naturalized) 5192, growing at the edge of scrub woodland near the beach at Douglas Bay. Originally from the Pacific islands, it is often found naturalized in the tropics and is commonly known as “snow-bush.”’ Sebastiania hexaptera Urban (First record) 3649, growing in hurricane-damaged forest beside the Picard gorge, Morne Diablotin. Fruiting specimens, 4333 and 5356, were later collected at the same locality. It is reported as rare on Guadeloupe and Martinique (Howard, 1974). ORCHIDACEAE Elleanthus dussii Cogn. (First record) 3862, in fruit, growing beside the road from Laudat to Freshwater Lake. It is endemic to the Lesser Antilles and has been recorded from Guadeloupe and Martinique (Howard, 1974). Pleurothallis wilsonii Lindley (First record) 5676, in fruit, growing in Milton on the trunk of Syzygium jambos (L.) Alston. It is distributed throughout the Greater Antilles and has been recorded from Guadeloupe. Spiranthes adnata (Sw.) Bentham (First record) 5638, growing at the edge of the Picard gorge. It is widely distributed in Central America and the West Indies and has been collected in several of the Lesser Antilles. HELICONIACEAE Heliconia psittacorum L. f. (First record; naturalized) 3762, growing in Délices on a slope below a path. It is native to Central and South America, commonly cultivated in the West Indies, and often found as an escape. 150 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 DIOSCOREACEAE Dioscorea bulbifera L. (First record; naturalized) 3903, fruiting, found in the forest above Fond Pie, and 3941, a sterile spec- imen bearing aerial tubers. This is an Old World species, but it 1s widely cultivated as a food plant, often escaping and becoming naturalized. AMARYLLIDACEAE Zephyranthes rosea Lindley (First record; naturalized) 4670, growing in a deserted garden near Morne Rouge. It is Central American and occurs generally in the West Indies. HyYDROCHARITACEAE Hydrilla verticillata (L. f.) Royle (First record) 5508, growing in the pool at Canefield. This plant is distributed throughout the warmer parts of the Old World and has now become a nuisance in Florida. CYPERACEAE Cyperus iria L. (First record) 3744, growing in waterlogged pasture at Canefield. It occurs in most tropical parts of the Old World, Australia, and North America. In the West Indies it has been recorded only from Cuba. Eleocharis cellulosa Torrey (First record) 5283, growing in the Cabrit swamp. It 1s widespread in Central America but new to the Lesser Antilles. Fimbristylis littoralis Gaudich. (First record) 4474, growing beside the track from Syndicate to Trou Cochon, and 5271, from the Cabrit swamp. This is an Old World sedge that has been recorded from St. Lucia and Grenada. Kyllinga nemoralis (J. R. & G. Forster) Dandy ex Hutch. (First record) 3726, growing near the sea at Pointe Mulatre. It was also found at Fort Shirley and on waste ground near the bridge at Goodwill. Originally from the Pacific islands, it has been recorded from Jamaica and Trinidad. Scleria distans Poiret (First record) 4526, growing beside the road above Salisbury. It is known in Jamaica, Cuba, and the Dominican Republic but is new to the Lesser Antilles. 1989] WHITEFOORD, DOMINICA 15] GRAMINEAE Pennisetum purpureum Schum. (First record) 5365, growing on a bank at La Plaine. It is native to tropical Africa but has become naturalized in the West Indies and tropical America, where it has been introduced as a forage crop. PTERIDOPHYTA POLYPODIACEAE? Acrostichum aureum L. (First record) 4054, growing in large clumps in the Cabrit swamp. It is found generally in the tropics, including the other Lesser Antillean islands. Elaphoglossum smithii (Baker) Christ (First record) 4552 and 4554, epiphytic on buttresses on Morne Diablotin. This is endemic to the Lesser Antilles and has been recorded from St. Vincent and Grenada. Polypodium aureum L. var. areolatum (Humb. & Bonpl. ex Willd.) Baker (First record) 3609, epiphytic on citrus, Syndicate. It occurs in the Greater Antilles, con- tinental tropical America, the Galapagos Islands, and Guadeloupe. Polypodium plumula Humb. & Bonpl. ex Willd. (First record) 3842, ona fallen slab of bark at Baiac. It is distributed in Florida, the Greater Antilles, continental tropical America, and the Lesser Antilles (Saba, Guade- loupe, and Martinique). Pteris grandifolia L. (First record) 3937, ona buttress, in forest near Syndicate. It occurs in the Greater Antilles, the Lesser Antilles (Guadeloupe, Grenada), Trinidad, and continental tropical America from Mexico to Peru. Thelypteris reptans (J. F. Gmelin) Morton var. tenera (Fée) Proctor (First record) 4073, growing on a wall, Fort Shirley. It may occur in Puerto Rico and has been recorded from St. Kitts, Nevis, and Guadeloupe in the Lesser Antilles. *The following ferns were not recorded from Dominica in Howard’s Flora of the Lesser Antilles (Vol. 2, 1977) and so are listed as first records. 152 JOURNAL OF THE ARNOLD ARBORETUM ACKNOWLEDGMENTS I should like to thank D. H. Nicolson for the use of his unpublished manu- script, Flora of Dominica, and for his help. I am grateful to the Forestry Division, Commonwealth of Dominica, and to the Botany Department, British Museum (Natural History), for their support. I would also like to thank the many other people who have aided me, including C. D. Adams and M. L. de Rico Arce, for help with identifications; A. O. Chater and other colleagues, for critical advice; and B. U. Fawssett and S. Davie, for assistance in preparing the manuscript. I am especially grateful to P. G. H. Evans for allowing me to join his Dominica Multiple Land Use Project. LITERATURE CITED ADAMS, C. zs 1972. Flowering plants of Jamaica. University of the West Indies, Mona, Jamai BENTHAM, G. & J.D. Hooker. 1862-1883. Genera plantarum. Vols. 1-3. Reeve and o., London FourRneT, J. 1978. Flore illustrée des phanérogames ae oe et de Martinique. Institut national de la Recherche agronomique, GRISEBACH, A. H.R. 1859-1860. Flora of the British West Indian Islands. Lovell Reeve and Co., London Hopace, W. H. 1954, Flora of Dominica, B.W.I., part 1. Lloydia 17: 1-238. Howarp, R. A., ed. 1974-1979. Flora of the Lesser Antilles. Vols. 1-3. Arnold Ar- boretum of Harvard University, Jamaica Plain, Massachusetts. Nicotson, D. H. Flora of Dominica. Unpublished manuscript. Postal Service STATEMENT OF OWNERSHIP, , MANAGEMENT AND CIRCULATION 1A. Title of Publication 2. Date of Filing ocurnal of Arnold Arboretum O;O | O44 | 216 | 215 10-4-88 3. Frequency of Issue 3A. No. of Issues Published 38. Annually Quarterly $50.00 4 City, County, +4 Code) (Not printers 1041 New Hampshire St., Lawrence, (Douglas), KS 66044 5 (Not printer) | | Sam | 6. Full Names and Complete Mailing Address of Publisher, Editor | Publisher a and Complete sep Address) | Arnold Arboretum of Harvard University, 1041 New Hampshire St., Lawrence, KS | 66044 l Editor (Nan |S. A, Spongberg, 22 Divinity Avenue, Cambridge, MA 02138 2 yee B. By Sehmidt, 22 Divinity Avenue, Cambridge, MA 02138 Owner (/f owned by a corporation, its name nd add ress must be d and ai | percent or more ae total amount of stock. If not or ips unincorporated firm and iso immediately thereunder the -s and addresses of stockholders stat owning or holdin: wned by a raises the names and addresses of the individual owners must be given. If owned by a pa e its name and pee as well as that of each individual must be given. If the publication is published by a nonprofit organization, its address must be stated.) (Item must be completed.) Full Name i pS re ee | Arnold Arboretum of Harvard University, 10d] New Hampshire, Lawrence, KS ‘4 | [ a | 8. Known Bondholders. Mortgagees, g 1 Percent or More of Total Amount of Bonds, Mortgages or Other | Securities (If there are none, so state) | ull Name eres | ON Peete pam Ee 9. For Comp! The purpose, function, p 9 ere (1) (2) ie) Has Changed During (If changed l XJ Preceding 12 Months Preceding 12 Months change with this statement.) | 10 Extent and Nature of Circulation Wivaracatnio Actual No. Copies of Single Issue (See instructions on reverse side) Preceding 12 Months Published Nearest to Filing Date — - A. Total No. Copies (Net Press Run) res 750 750 8. Paid and/or Requested Circulation 1. Sales through dealers and carriers, street vendors and counter sales | i 2. Mail Subscription | (Paid and/or requested) 547 571 (Sum or 1OBI and 10B2) 547 571 | OD. Free Distribution by Mail, Carrier or Other Means Samples, Complimentary, and Other Free Copies 12) 12 E - and D | eee a 559 Sig Ses eee F. Copies Not Distributed 1. Office use. left over, unaccounted, spoiled after printing 19 1 162 2. Return from News Agents G TOTAL (Sum of E. Fl and 2 4) W 750 2 : Pee tr Ba Signature a , Publisher, Business Manager, or Owner certify that t Y = : Leary a CigebdK J Litt PS Form 3526, Dec 1987 (See instructions on reverse) Journal of the Arnold Arboretum January, 1989 CONTENTS OF VOLUME 70, NUMBER | Taxonomic Studies in the Piperaceae—I. The Pedicellate Pipers of Mexico and Central America (Piper Subg. Arctottonia). ALLAN J. BORNSTEIN ee ere ee ee Ae eee ee eee eee 1-55 The Dioscoreaceae in the Southeastern United States. IHSAN A. AL-SHEHBAZ AND BERNICE G. SCHUBERT ............ 57-95 The Genera of Commelinaceae in the Southeastern United States. SORDON C.. LUGKRER: 2445-6 405:555 053 See bees sack ooeaeen 97-130 The Genera of Eriocaulaceae in the Southeastern United States. ROBERT KRAL ........ 0.00 ccc eee e ceed bebe bb evnns 131-142 Recent Plant Collections from Dominica. . WHITEPOORD: coc ogc code dad sae ane bia bene sd iieehe ieee 143-152 Volume 69, Number 4, including pages 289-450, was issued 21 October 1988. JOURNAL OF THE ARNOLD ARBORETUM HARVARD UNIVERSITY VOLUME 70 NUMBER 2 ISSN 0004-2625 Journal of the Arnold Arboretum The Journal of the Arnold Arboretum (ISSN 0004-2625) is published quarterly in January, April, July, and October for $70.00 per year, plus $10.00 or $15.00 postage for addresses outside of the United States, by the Arnold Arboretum of Harvard University. It is printed and distributed by Allen Press, Inc., 1041 New Hampshire Street, Lawrence, Kansas 66044. Second-class postage paid at Lawrence, Kansas. POSTMASTER: send address changes to Journal of the Arnold Arboretum, Y% Allen Press, Inc., P.O. Box 368, Lawrence. Kansas 6604 Subscriptions and remittances should be sent to Journal of the Arnold Arboretum, 1041 New Hampshire Street, Lawrence, Kansas 66044, U.S.A. Claims will not be accepted after six months from the date of issue. Volumes I-51, reprinted, and some back numbers of volumes 52-56 are available from the Kraus Reprint Corporation, Route 100, Millwood, New York 10546 U.S.A EDITORIAL COMMITTEE S. A. Spongberg, Editor E. B. Schmidt, Managing Editor P. S. Ashton K. S. Bawa P. F. Stevens C. E. Wood, Jr. Printed at Allen Press, Inc., Lawrence, Kansas COVER: The stylized design appearing on the Journal and the offprints was drawn by Karen Stoutsenberger. THIS PUBLICATION IS PRINTED ON ACID-FREE PAPER. JOURNAL OF THE ARNOLD ARBORETUM VOLUME 70 APRIL 1989 NUMBER 2 THE BIOSYSTEMATICS AND EVOLUTION OF THE POLYALTHIA HYPOLEUCA COMPLEX (ANNONACEAE) OF MALESIA, I. SYSTEMATIC TREATMENT STEVEN H. RoGSTAD! A component of tropical lowland rain-forest flora diversity is the presence of series of sympatric, este related species. The three major hypotheses priate objects of investigation when examining them. To that end, a classifi- cation of the Polyalthia hypoleuca complex (Annonaceae) is provided to form the foundation for a set of forthcoming articles exploring which of the three hypotheses best applies to a complex. After presentation of evidence that the complex is monophyletic, the results of uni-, bi-, and multivariate statistical analyses ne character dau taken from herbarium specimens are reviewed and shown to ort th of six species in the complex, with members ene sympatrically in various combinations throughout Malesia. One new species, P. ovalifolia, is described. The hypothesis of coexistence that best applies to the complex is briefly noted, with supporting details to be given in subsequent articles. If the traveller notices a particular species and wishes to find more like it, he may often turn his eyes in vain in every direction. Trees of varied forms, di- mensions, and colours are around him, but he rarely sees any one of them repeated. Time after time he goes tow ard a tree which looks like the one he seeks, but a closer examination proves it to be distinct Alfred Russel Wallace (1878, p. 65) As the species of the same genus have usually, though by no means invariably, much similarity in habits and constitution, and always in structure, the struggle ‘Department of Organismic and Evolutionary Biology, Harvard University, Cambridge, Massa- chusetts 02138. Current address: Biology Department, Box 1137, Washington University, St. Louis, Missouri 63130 © President and Fellows of Harvard College, 1989. Journal of the Arnold Arboretum 70: 153-246. April, 1989. 154 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 will generally be more severe between them, if they come into competition with each other, than between species of distinct genera. .. . One species of charlock has been known to supplant another species, and so in other cases. We ca dimly see why the competition should be most severe Seale allied forms, which fill nearly the same place in the economy of natt Charles Darwin (1859, p. 76) INTRODUCTION Although the prodigious organismic diversity of tropical lowland rain forests (TLRFs) is increasingly being catalogued, the biological principles generating and maintaining it remain enigmatic. This study has been conducted as part of an attempt to illuminate one aspect of the biology of tropical lowland com- munities that contributes to such diversity: the occurrence of series of mor- phologically very similar yet consistently distinct entities, or species, in ap- parent sympatry. In this Introduction, I first review the relevant observations and explanatory theories of previous researchers who have contemplated such series of species. Then I summarize them as a set of competing hypotheses. Finally, I outline the research methods I have used to test these hypotheses with one such series of entities. As will be demonstrated, a clear understanding of the systematics of these species complexes is extremely important to the examination of the competing hypotheses forwarded to explain their existence. I therefore include a systematic treatment of a complex of Malesian TLRF trees, the species of the Polyalthia hypoleuca complex (Annonaceae), that will serve as the foundation for a series of forthcoming articles investigating the outlined hypotheses with regard to this species constellation. Each of these articles will be related to and will reflect upon the data and conclusions pre- sented in all of the others. For example, the results to be reported in subsequent papers on the distributional ecology (Rogstad, in prep. a) and floral biology (Rogstad, in prep. b) of the species included in the complex both are based on and support the systematic conclusions reached primarily on morphological grounds here. THE PROBLEM OF SERIES OF CLOSELY RELATED, SYMPATRIC SPECIES One often-noted component of the high species diversity of TLRFs 1s the presence of series of morphologically very similar, sympatric entities (e.g., see Klopfer & MacArthur, 1961; Fedorov, 1966; Ashton, 1969; Richards, 1969; Van Steenis, 1969; Hubbell, 1979; Leigh, 1982; Whitmore, 1984). These en- tities are usually recognized at the species level, and because of their morpho- logical similarity, they are presumed to be closely related species. I will give a more precise definition of these species complexes below. That these series pose a problem for community theory can be traced back at least to the work of Lotka, Volterra, and Gause (Hardin, 1960; Armstrong & McGehee, 1980). Gause demonstrated experimentally that the predictions of the Lotka-Volterra equations were fulfilled in competition experiments be- tween two closely related Paramecium species. This principle is usually referred 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 155 to as Gause’s Principle of Competitive Exclusion: two or more species that occupy the same niche cannot coexist in the same habitat. While theoreticians (e.g., Hubbell, 1979; Huston, 1979; Armstrong & McGehee, 1980; Pickett, 1980; Agren & Fagerstrom, 1983) have argued that this principle probably is not valid for nonequilibrium environments, a large body of empirical evidence (see, for example, MacArthur, 1957; Hutchinson, 1959: Mayr, 1963, 1970; Schoener, 1974, 1983; Harper, 1977; Jones, 1978; Werner, 1979; Aarssen, 1983, 1985: Armbruster, 1985; Pulliam, 1985), based primarily on geographic distributions or niche shifts of closely related species, suggests that it generally holds, at least in temperate environments. The many studies on character displacement (e.g., Grant, 1972) are also germane, because they indicate that even species without broad overlap in many niche charac- teristics may compete intensively with regard to the overlapping areas. For temperate trees, the prevailing belief has been summed up by Fedorov (1966, p. 1): “*... taxonomically close ... species of plants, do not as a rule occupy the same area, but are most usually geographically isolated from one another. . . the isolation of closely allied species may be not only geographical, but ecological as well, as when such species are found in different habitats within one area.” According to V. Grant (1963), this principle becomes more evident as the size of the organisms under consideration increases, due to more intense competition for scarce resources among larger organisms. He cited several temperate examples of this concept, noting that while some herbs may have series of sympatric, closely related species, this condition is rare for trees. Two often-noted, seeming exceptions (members of the genera Quercus and Pinus) show, upon detailed examination, clear geographic or ecological sepa- ration of closely related species (also see Stebbins, 1950; Reich & Hinckley, 1980). There may be a few groups of temperate trees that do not conform to Gause’s Exclusion Principle, but these should generally be regarded as the anomalies that highlight the rule. In contrast, forests of the humid tropics often appear to deviate significantly from Gause’s principle. For example, in Ashton’s (1977, 1984) surveys of five two-hectare (ha) plots in TLRF at Pasoh Forest Reserve (central peninsular Malaysia), 484 species were identified (DBH > 10 cm); approximately 40 percent of the 191 genera had three or more congeners, and [5 percent of the genera had six or more congeners. Some genera had many species, with Eugenia topping the list at 28. Viewed in another manner, if (as Ashton extrapolates) 700 species can be found on 50 ha at Pasoh, more than 300 (43%) will be growing sympatrically with at least five congeners. Interestingly, Pasoh is by no means the most species rich of the forests that have been surveyed in Malesia (Ashton, 1984), so even more striking examples probably exist. Sympatric ensembles of closely related species also exist in the richest TLRFs of South America, such as the Yanomamo region of Amazonian Peru, and in Africa, where Cola (Sterculiaceae) and Diospyros (Ebenaceae) are examples of genera exhibiting this pattern (A. Gentry, pers. comm.). Thus, while the intensity of this phenomenon may vary from region to region, it appears to be a general feature of TLRF communities. As noted above, the occurrence of sympatric, closely related species is not 156 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 cause to discard the Competitive Exclusion Principle if such species are found to have diverged in at least one niche characteristic. However, it will be seen below that some authors have argued, on both theoretical and empirical grounds, that series of closely related species may coexist in forests of the humid tropics without such niche divergence. Are there fundamental differences in temperate and tropical forests in the forces guiding evolution and community organiza- tion? DEFINITIONS Greater precision can be given to the above concepts, and to the development ofa means to investigate them, by defining a few of the terms more rigorously. The term “sympatry” has a long and varied history of usage (e.g., see Mayr, 1970). Here I use it to mean the concurrent, regular occurrence of different taxa within potential gene exchange distance of one another. This rather broad definition has been chosen over more restrictive ones that may overlap with niche definitions (e.g., ecological or temporal sympatry). The concept of the niche has also been extensively debated (e.g., see dis- cussion in Aarssen (1983)). I use Whittaker’s (1967, p. 210) definition: “‘the position of the species in the community, its particular way of relating to other species, environment and space within the community, and seasonal and diur- nal time.”’ A niche difference between two tree species exists if they interact with other species differently (e.g., pollinators or herbivores), if they differ in some aspect of the environments and space they occupy (e.g., understory vs. canopy; differences in substrate requirements), or if their seasonal or daily rhythms of activity differ temporally (e.g., flowers functional at different times of the year or day). It is also not always clear what is meant by the various authors cited above when they use concepts such as ““morphologically very similar,” “closely re- lated,” ‘‘taxonomically allied,” or ““congeneric.”’ All of these expressions refer to groups of species whose members are assumed to have a close phylogenetic relationship. However, they are relative, depending upon context, and they can be (and have been) applied to different types of groups. Several discussions of the types of groups that can be constructed exist in the literature (e.g., see Wiley, 1981), but for present purposes I will follow Eldredge and Cracraft (1980) in reducing the possibilities to two main classes, monophyletic and nonmono- phyletic groups. As will be seen below, these groups differ as to their precision in delimiting assemblages of species with immediate phylogenetic proximity. After a brief explanation of the differences in the construction of these two types of groups, the importance of the type chosen to examine the phenomenon of series of “closely related”’ sympatric species will be outlined. I will establish that monophyletic groups are preferred over nonmonophyletic ones in inves- tigating this phenomenon. A comparison of these two types of groups is presented in FiGureE |. In this discussion, it is assumed that the phylogenetic relationships among the taxa (designated by letters) as drawn reflect the “true” history of the groups. Monophyletic groups, as defined by Hennig (1966), comprise all and only 1989] ROGSTAD, POLYALTHIA HYPOLEUCA Lo? TAXON 1 TAXON @2 SK Ficure |. Monophyletic and nonmonophyletic groups (see text for explanation). those taxa descended from a common ancestor (the “strict”? monophyletic groups of Stevens (1986)). Examples include such terminal groups as W and X (note that these are sister taxa); W, X, Y, and Z; or all of taxon 1, in which case node A may be considered as the common ancestor. Monophyletic groups are usually recognized by the presence of a unique character or suite of char- acters (synapomorphies) that is shared among all the members of the group and that presumably arose only once in the ancestral lineage before its diver- sification. Nonmonophyletic groups do not include all and only the species descended from a single ancestor but rather exclude at least one of the terminal taxa. For example, nonmonophyletic groups would include, for taxon 1, ones made up of only V, W, and Y, or of V, W, X, and Z (often referred to as “paraphyletic” groups). They can also be constructed of taxa with even more distant phylogenetic relationships such as only taxa V, Y, and M (they are then frequently termed “‘polyphyletic’’?). While monophyletic assemblages are by definition composed of taxa with the most immediate phylogenetic relation- ships (the most closely related contemporary species), nonmonophyletic groups exclude some of these taxa. THE NEED FOR MONOPHYLETIC GROUPS Of these two types of groups, small monophyletic ones will always be most informative in examining whether or not series of “closely related” species live in sympatry in TLRFs. Although series of apparently sympatric, congeneric species have been identified at several tropical forest sites (e.g., see Ashton, 1977; Wheelwright, 1985), a phylogenetic perspective of these genera is required to demonstrate that any particular congeners occurring at one site can be considered to be members of a small monophyletic group. For example, the genus under investigation here, Polya/thia Blume, comprises up to 150 species, nine of which have been found at Pasoh Forest Reserve, Peninsular Malaysia (P. S. Ashton, pers. comm.). If, upon analysis of the systematics of the entire genus, no two of these nine species can be placed together in a small mono- phyletic group, and all nine are best described as phylogenetically distant from one another within the genus, then making a case that a series of “closely related,” sympatric species of Polyalthia occurs at Pasoh becomes more diffi- cult. In studying this interesting ecological phenomenon, then, a clear under- standing of the systematics and biogeography of the group is crucial. 158 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 If, as is widely believed (e.g., see Mayr, 1970) allopatric speciation 1s more prevalent than other modes of speciation, it might be expected that sister species, especially recently derived ones, will rarely occur in sympatry. The term ‘‘monophyly” has nothing to do with age, of course, yet when most or only some of the members of a small monophyletic group are found sympatri- cally, their coexistence is still of interest as they are perhaps the most closely related entities of the group that can be found growing together in nature. For several reasons, either small monophyletic groups or those including only all the species of a small monophyletic group that occur in sympatry locally are also preferred over nonmonophyletic groups when investigating the origin and maintenance of series of ‘“‘closely related” sympatric species with regard to competitive exclusion. First, if nonmonophyletic groups are chosen to examine “‘closely related” species 1n sympatry, it 1s possible to miss one or more interactions between sympatric species of closer phylogenetic relationship than those studied. As an example, consider a forest where all of the congeneric species of taxon | in Figure | are known to occur in sympatry. Ifan investigator chooses to study only species V and X and finds that they have differentiated niches, the implications of this finding would have to be revised if, later, no niche differences could be detected among species V, W, Y, and Z. Alternatively, an investigator might find an extensive and seemingly random pattern of over- lap with regard to some niche gradient measured for all of the species of FIGURE 1 but may discover well-differentiated patterns when comparing only the mem- bers within taxon | or taxon 2 nother approach to understanding why small monophyletic groups are preferred in addressing the problem of series of closely related, sympatric species derives from a consideration of how phylogenetic distance may be related to niche distance. It 1s thought often to be the case that, ceteris paribus, the closer the degree of phylogenetic relationship, the greater the degree of morphological similarity. It follows that monophyletic groups are often (al- though not always) those with the greatest degree of morphological similarity among member taxa. If structure is related to function, then members of mono- phyletic groups should have the most similar niches. It is also often argued (e.g., see Grant, 1963; Wheelwright, 1985; Glazier, 1987) that since most closely related species have the same ecological and physiological heritage via their common ancestral lineage, such species will usually have relatively similar niches (for a contrasting view, see the discussion of the adaptive speciation model, below). This principle has been invoked, at least implicitly, in many of the experimental studies purporting to investigate competitive exclusion, beginning with Gause’s experiments with two Paramecium species. Studying the niche characteristics of the members of monophyletic groups ensures that the most closely related entities at a site are under consideration and therefore provides the most rigorous test of whether or not sympatric closely related species show niche differentiation. Finally, if one is interested not only in present patterns (for example, patterns of niche differences of contemporary species) but also in evolutionary processes, monophyletic groups are again the more informative type. To use another illustration from Ficure |, if sympatric species U and Z are found to have 1989] ROGSTAD, POLYALTHIA HYPOLEUCA [59 only slight niche differences, this might be interpreted as indicating a minimal degree of niche divergence from the state of their common ancestor at node A. However, if niche characteristics of V, W, X, and Y are subsequently found to diverge greatly in a similar fashion from U, then it is most parsimonious to interpret the slight differences between U and Z as convergence. In conclusion, monophyletic groups are the preferred groups for interpreting the biological significance of series of sympatric, closely related species. Such groups are composed of the most phylogenetically recent products of speciation, and thus their members are always the most closely related species that can be investigated when considering any lineage. Only the study of monophyletic groups gives a minimal estimate of the total extent of niche divergence of a lineage radiating from a common ancestor and may yield information about this divergence in relation to speciation and the total pattern of the geographic distribution of the descendant species. As speciation is the wellspring of di- versity, an understanding of sympatry-allopatry patterns in relation to niche divergence within small eae bs ee will be informative as to the origin and maintenance of TLRF dive Unfortunately, most, if not all, of - eae purporting to investigate sym- patric, closely related species were carried out at a time when such consider- ations concerning monophyletic groups were not appreciated. For example, several studies have examined flowering phenology and have arrived at a range of conclusions for different groups. However, what appear to be either random or highly organized patterns of flowering may require reinterpretation in re- lation to new information concerning the monophyly, phylogeny, and bio- geography of the taxa involved (or a more inclusive set of taxa), if monophyletic groups have not been used in such studies. With this caveat in mind, it 1s important to note that, as will be shown in the systematic treatment below, congeneric groups are by no means necessarily monophyletic ones. PREVIOUS OBSERVATIONS AND HYPOTHESES Hypothesis I. Polymorphic Entities One explanation of monophyletic series of sympatric species in TLRFs is that the phenomenon has been largely misinterpreted: many or all of the mem- bers of such series are more appropriately combined into single polymorphic species (e.g., see Levin, 1979). To establish, on morphological grounds, that this is the case for any group of entities, at least one of three fundamental patterns of morphological variation must be demonstrated for the group. The first consists of a few or many characters that are extremely variable on a local scale. The second has been termed “kaleidoscopic” character distributions (Cullen, 1968). In this case, the possible character states for a number of characters exhibit several or all of the possible permutations among taxa. While Cullen used this term to indicate patterns of characters among species, it is also applicable at other levels. The third pattern, clinal variation (e.g., see Mayr, 1970), can lead to problems of classification when two taxa from distant sites are recognized as distinct until morphologically intermediate specimens are found. However, even if local species can be thus “‘explained away,” local 160 JOURNAL OF THE ARNOLD ARBORETUM [vVoL. 70 differentiation may be biologically significant despite a broader geographic pattern of kaleidoscopic or clinal variation within a taxon. Hence, Leenhouts’s (1967) reduction of approximately 255 species of A//ophylus (Sapindaceae) to one, Jacobs’s (1962) similar reduction of species in Pometia (Sapindaceae), and descriptions of variable taxa such as Diospyros (Ebenaceae) species by White (1962), Ficus deltoidea Jack (Moraceae) by Corner (1970), Drimys piperita Hooker f. (Winteraceae) by Vink (1970), Licania ie ota ge js by Prance (1972), Calophyllum blancoi Planchon & Tri and C. c Hooker f. (Guttiferae) by Stevens (1980), Calamus is are by Drans- field (1981), and Chisocheton (Meliaceae) species by Mabberley (1979), among others (see list in Leenhouts, 1967), are difficult to interpret. Some may indeed be examples of taxa in which a high degree of polymorphism leads to problems of definition of morphological species. In this context species delimitations based upon morphological criteria may require revision as new data from allied studies (e.g., chemosystematics, re- productive biology, or ecology) accumulate, since such new information may not support the decisions based upon morphological patterns alone (Raven, 1976; Levin, 1979; Stevens, 1980). Collections for the great majority of tropical taxa are sparse, and as areas become more thoroughly studied, any one of the above possibilities might alter taxonomic decisions about groups previously recognized as species rich. It is important to note here that hybridization and hence hybrid swarms are thought to be rare in TLRFs (Ashton, 1969, 1984; Burger, 1980). On the other hand, if it is determined that facultative apomixis occurs to a significant degree in TLRF communities (Kaur et a/., 1977; Jong, 1980; Kaur, 1980), then poly- morphic apomictic assemblages are to be expected (V. Grant, 1971). A high degree of polymorphism may also be expected for species in which inbreeding is the rule. On theoretical grounds the latter has been proposed as the case for trees of the humid tropics (Corner, 1954; Baker, 1959; Fedorov, 1966; Van Steenis, 1969; but see Chapter 4 in Rogstad, 1986). Hypothesis II. Groups of Discrete, Closely Related Species Do Occur in Sympatry Various investigators (references below) have accepted that a number of these complexes comprise discrete, sympatric species, and two further hypotheses have been forwarded to explain their existence. Hypotuesis ITA. SPECIES HAVE OVERLAPPING NICHES. It has been suggested that members of such series may have similar or identical niche characteristics; sympatry here contradicts the competitive exclusion principle, or at least the observations on temperate forest communities noted above. This line of thought has developed from empirical observations, as well as from theoretical con- siderations of evolutionary and community-organizing forces occurring in TLRF communities. Fedorov (1966, 1976), Poore (1968), Richards (1969), Van Stee- nis (1969, 1976, 1981), Baker (1970), Wong and Whitmore (1970), Hubbell (1979), Ashton (1979, 1984), Gan ef a/. (1981), Tanner (1982), Yap (1982), and Wheelwright (1985) have all discussed examples where no evidence could 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 161 be found of niche divergence or niche specialization among sympatric groups of TLRF tree species (also see Klopfer & MacArthur (1961) and Terborgh (1985a) for interesting studies of TLRF birds). It has been proposed that in many cases where series of very similar species are sympatric in TLRFs, the distinguishing morphological characters are not associated with differences in niche parameters, implying that these series arise not by differential adaptation but rather by selection-neutral processes. Noting that correlates of the high species diversity of these forests are a low density of individuals and, therefore, a small effective population size for most species, Fedorov (1966) suggested that these conditions favor speciation by random drift. His proposal stems, in part, from the work of Wright (1931), who pre- sented a population genetics model in which random fixation of changes in gene frequencies (drift) is likely in organisms with small effective population sizes. If gene flow among small populations is consistently low, speciation by random drift results (King & Jukes, 1969; Maynard Smith, 1970; Stern & Roche, 1974; Bernstein ef a/., 1985; Slatkin, 1985). Hubbell (1979) and Ng (1983) have demonstrated that a large percentage of the species found locally in TLRFs exhibit extremely low population numbers and patchy distributions. For example, Ng examined Poore’s (1968) data and determined that of the 377 species identified on 23 ha of peninsular Malaysian TLRF, 307 (81%) had ten or fewer mature individuals and 143 (38%) were represented by only one (see Lovejoy (1975) and Terborgh (1985a) for parallel results with TLRF bird species). Such low population densities and patchy distributions of species increase the chances for disruption of gene flow by population isolation, leading to heightened localized inbreeding, and therefore also increase the chances for random drift. For an even more radical yet sup- portive view, see Barton and Charlesworth (1984). In this connection, several authors have indicated that since tropical forests are less seasonal than temperate ones, the environmental cues triggering in- traspecifically synchronized flowering in temperate forests are largely absent. Flowering within local populations of some species of TLRFs may thus be asynchronous (Holttum, 1953; Koriba, 1958; Fedorov, 1966; Wong, 1983). The occurrence of noncoordinated flowering within species with small popu- lation sizes would act to inhibit gene exchange even further and would promote random drift. The few investigations documenting levels of gene flow for TLRF tree pop- ulations (Whiffin, 1978; Chan, 1980; Gan et al., 1981) give evidence that gene exchange is very local and occurs at low levels. In addition, Lovejoy (1975) and Terborgh (1985a) have shown that a large proportion of the birds found in TLRFs have low population densities and very low vagility. Similar patterns have been found for TLRF canopy insect species (Elton, 1975). Gene dispersal in TLRF tree populations will be commensurate with the density and the degree of vagility of their pollen and seed vectors. Lewis’s (1966) suggestion that small populations with a high degree of in- breeding are more susceptible to novel chromosome rearrangements that po- tentially lead to speciation is worthy of consideration with respect to TLRF tree population structure (also see Bush, 1981). It is important to note that 162 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 speciation via such changes no longer necessarily requires the troublesome bottleneck of Goldschmidt’s (1940) hopeful monsters but rather may occur through genetic phenomena in small populations (Dover & Flavell, 1982; Rose & Doolittle, 1983; Fitch & Atchley, 1985; Wallace, 1985; Krieber & Rose, 1986). Several authors (e.g., Fischer, 1960; Baker, 1970; Stebbins, 1974; Van Stee- nis, 1978) have argued that tropical environments are more stable than tem- perate biomes with respect to the degree and predictability of the fluctuations of physical parameters and are therefore more neutral in selection of variation by physical extremes. In other words, TLRFs are benign environments where a greater degree of marginal or nonselected variation persists. Many of the characters distinguishing among similar species are apparently not adaptive (e.g., see Ashton, 1979) but have arisen through genetic discontinuities origi- nating in stochastic processes (drift) and are functionally of little consequence (Van Steenis, 1969, 1976, 1978). A contrasting view of the effects of relaxed abiotic selection in TLRFs is presented below. Another theoretical approach that predicts little chance for niche differen- tiation within complexes of similar species (or among anything but very general guilds of species, for that matter) derives from community-level considerations. If TLRF communities are nonequilibrium communities (Connell, 1978; Hub- bell, 1979: Acevedo, 1980, 1981; Chesson & Warner, 1981; Abugov, 1982; Wright & Hubbell, 1983) the composition of any local biota may be more a product of historical chance events than of the relative competitive interactions of the species involved. The taxonomic assemblage at any locale is constantly and randomly changing, a phenomenon that Hubbell (1979) has termed “com- munity drift.” Several authors (e.g., Aubréville, 1938, 1971; Grubb, 1977; Strong, 1977; Ewel, 1980; Lang & Knight, 1983; Brokaw, 1985; Lieberman et al., 1985) have noted the importance to all TLRF tree species of regeneration at tree-fall sites. If successful reproduction is largely dependent on having progeny in the appropriate stage of development present in unpredictable forest gaps, then forest structure has a proportionately large stochastic component (Connell, 1978; Burger, 1980; Buckley, 1983; Connell et a/., 1984; Comins & Noble, 1985). Since most TLRF species have extremely low population den- sities, if the community-drift hypothesis holds most TLRF species will be randomly associated through time. Competitive interactions between or among species, as well as resource availability, will be completely unpredictable, and therefore persistent directional selection will be rare. Such stochastic compe- tition and distribution of resources may lead to selection for generalists (Went, 1973: Hubbell, 1979; Burger, 1980; Buckley, 1983; Ashton, 1984) and thus act upon a significant portion of TLRF species by generalizing their niches rather than driving niche divergence. The consequences of nondirectional selection can also be explored in more formal terms. Kimura and Weiss (1964) and Kimura and Ohta (1971) have shown that in diploid populations, mutations become fixed predominantly by random drift when the quantity 4N.s (N, represents the effective population size, and s denotes the coefficient of selection) is sufficiently smaller than | (for example, Hartl and colleagues (1985) suggest when 4N.s < 0.1). In tropical 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 163 humid forests the unpredictability of both one’s competitors and the matrix of available resources over long periods of time may give vastly varying and often conflicting values of s for most alleles, thus reducing the value of s to close to 0 (selection nondirectional). As s approaches 0, 4N.s becomes < 0.1, and this is exacerbated by the small population sizes found in TLRF trees. Lande (1976) has presented a model that suggests phenotypic evolution may be subject to drift even in large effective populations where selection is weak. A few empirical studies (Malecot, 1959; Selander, 1970; Carson et al., 1982) have shown that when directional selection 1s weak or nonexistent, character variability may increase or drift. The link between the theories of community drift and the random drift of species in TLRF communities (and the implications of this link for niche divergence of sister species descendent from a TLRF tree ancestor) 1s perhaps best established by considering the most widely accepted model of speciation, an allopatric one. Under this model, two lineages from a common ancestor first become geographically isolated. Due to the low population sizes common to TLRF trees, random sampling of the once-common genetic pool and ac- cumulation of different mutations in the two isolated lineages lead to genetic divergence of the daughter populations. Gottlieb (1984, 1985) has pointed out that many of the morphological differences distinguishing species are under simple genetic control involving a low number of alleles, a situation facilitating rapid fixation of morphological differences by random drift. Since competitors and resource distribution are unpredictable, no directional selection takes place. Divergence is due to genetic sampling errors only, and morphological differ- ences need not have any pronounced functional significance (Van Steenis, 1978, 1981; King, 1984; Davis & Gilmartin, 1985), especially if selection for gen- eralists prevails. If the two sister lineages subsequently achieve sympatry, it is highly unlikely, due to the extremely low population numbers and densities of species in TLRF tree communities, that the two lineages will ever compete over resources in any predictable manner; therefore, the forces driving com- petitive exclusion and niche divergence (or character displacement) envisioned as active in temperate communities will be extremely weak in TLRF com- munities. Daughter populations that have diverged morphologically and ge- netically (to a degree that they have become genetically incompatible) may coexist indefinitely in TLRF communities without having diverged with regard to realized niche characteristics. Community drift and random drift thus may interact to generate series of closely related, sympatric species with broad niche overlap in these biomes. HypotuHesis IIB. SPECIES HAVE DIVERGENT NICHES. An alternative hypothesis has also been forwarded to describe the sympatric occurrence of series of closely related entities (e.g., species), granted that it can be demonstrated that distinct entities exist. A number of authors have argued that similar species coexist in TLRFs by occupying narrowly defined, distinct niches, thereby not engaging in competition along at least one niche gradient. Wallace (1878, p. 66) noted that, “In the equatorial zone, there is no struggle against climate.’ The forces of natural selection arise from biological interactions rather than physical con- 164 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 straints, leading “‘. . . to the filling up of every place in nature with some specially adapted forms.” Dobzhansky (1950) and Williams (1964) have reemphasized that biological competition is the most important factor in natural selection in the tropics (as opposed to physical environmental factors in temperate forests), and that this has led to more specialization for narrow niches and, in turn, a greater diversity of spec Dobzhansky (1950) i pointed out that TLRFs have a greater variety of niches than other terrestrial forest habitats due to their greater structural com- plexity, a notion that has been further reinforced by Corner (1954), Ricklefs (1977), and Terborgh (1985b). If this is true, the higher structural heterogeneity of TLRF communities would increase the possibilities for the coexistence of similar, closely related species, each with a distinct niche, thereby permitting greater species packing. Ashton (1969) has proposed that such series of species may be the result of adaptive speciation (see Ringo ef al., 1985, for a summary of the model). Each species in such a series may actually have subtly different niche characteristics with respect to at least one resource gradient. Such niche differentiation may take place with allopatric speciation or occur through divergent selection driven by competition between sister species once they come into sympatry again. Studies of TLRF organisms reporting empirical evidence of similar, closely related species each having at least one unique niche characteristic and growing sympatrically include those by Snow (1965), Diamond (1973), Burger (1974, 1980, 1981), Vandermeer and colleagues (1974), Stiles (1975), Gentry (1976, 1982), Chan and Appanah (1980), Janzen (1980), Haber and Frankie (1982), Fleming (1985; although significant overlap was found for all parameters ex- amined), Martin (1985), and Moulton (1985). Burger (1974, 1981), Stevens (1980), and Ashton (1984) have argued that in many cases, sister species grow allopatrically in the tropics. Such patterns, however, need careful scrutiny, as they may reflect more about how taxonomists recognize taxa than about anything of biological significance (Stevens, 1980) and may therefore falsely support Gause’s conclusions. A METHOD OF EXAMINING THE COMPETING CONCEPTS Most of the theorizing and supporting evidence reviewed in the paragraphs above are based on preliminary, descriptive, or anecdotal information. The monophyly of the groups discussed has not been established, and quantitative studies of character-state variation within and among putative entities to es- tabish their delimitation are lacking. Further, attempts to demonstrate the sympatry of distinct entities of advocated monophyletic groups, and then to distinguish whether the entities of such a complex are in fact ecologically equivalent or rather have differentiated with respect to one or more niche characteristics, have been neglected. Obviously, a clear understanding of the taxonomy of the members of any group intended for such investigations is crucial. Attempting to compare niche characteristics of sympatric entities whose definitions as species are not well understood will most likely be misleading. The following systematic treatment 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 165 of the Polyalthia hypoleuca complex is the foundation for a series of articles in which I will address all of the concerns noted in the previous paragraph, thereby enabling a choice of which of the above three hypotheses is most applicable with regard to this group of taxonomically allied Malesian tree species. To examine whether or not the niches of the species in this complex were segregated in any way, I employed methods commonly used to define niche characteristics for temperate plants, recognizing from the outset that this ap- proach might yield only negative evidence. That is, the failure of these methods to detect niche differentiation would not eliminate the possibility that niche differentiation does exist between entities for one or more uninvestigated fea- tures. Nevertheless, the finding that two morphologically similar species have similar or identical niche characteristics as defined by techniques that have successfully detected niche differentiation in temperate plants is theoretically interesting and prepares the ground for alternative or more detailed future investigations. However, as will be seen in the subsequent associated articles, such negative evidence was not a problem in this case. In fact, information concerning the niche characteristics of the species of this complex serves to support my systematic conclusions and must thus be consulted if one is inter- ested in understanding the sum of the evidence speaking to the systematic decisions, based primarily on morphometric analyses, presented here. Only one group was examined in this research. A number of them will have to be investigated with regard to these hypotheses before we can do much more than speculate about the causal forces generating the patterns. Information on these issues is central to understanding the origin and maintenance of TLRF diversity. A forest structured predominantly under hypothesis I is a very dif- ferent community than one composed of groups conforming strictly to hy- pothesis IIB. Of course, a range of possible combinations between these two extremes is probable locally (will generalizations be possible at more synoptic levels?), and a more accurate description of the distribution curve for the realized possibilities will yield deeper insight into evolutionary processes in TLRF biomes, as well as have implications for forest utilization and manage- ment practices. SYSTEMATICS OF THE POLYALTHIA HYPOLEUCA SPECIES COMPLEX After examining several groups of potential candidates for this investigation, I chose the Polyalthia hypoleuca complex. I was able to determine several characters suggesting that it was a monophyletic group, and 1n preliminary field work I found that some of the species do grow in sympatry in statistically tractable sample sizes. The specific goals of this systematic treatment are to demonstrate that this complex of species is indeed monophyletic, to apply various morphometric techniques to support the conclusion that discrete enti ] species) exist within this monophyletic group, and to present a formal classification for the group. 166 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 THE GENUS POLYALTHIA AND THE PLACEMENT OF THE P. HYPOLEUCA COMPLEX Polyalthia comprises approximately 100 to 150 species and is therefore one of the larger genera of the Annonaceae. Members of the genus are generally restricted to tropical latitudes and lower altitudes. They are distributed in humid regions of Africa and Madagascar, and from India and Sri Lanka through Southeast Asia and Malesia to Fiji and the associated islands. Sinclair (1955), the last author to revise a significant portion of the genus, placed it in tribe Unoneae, which he defined primarily on the basis of a single character: petals that are valvate in the flower buds. This agrees with the placement of the genus by other systematists, including Bentham and Hooker (1862), Fries (1959; he further subdivided the tribe into informal groups, Polyalthia being placed in the Polyalthia group), and Hutchinson (1964; Polyalthia is placed in Group A of subtribe Xylopineae). In contrast, Walker (1971) included Polyalthia in his tribe Uvarieae, based upon considerations of pollen morphology. However, all of these classification schemes need further investigation. For example, as is shown in FiGure 2, within the Polyalthia hypoleuca complex, which as I shall demonstrate below is clearly monophyletic, some members (e.g., P. discolor Diels) have petals that are distally (but not basally) imbricate in the floral buds, while at least one (P. glauca (Hassk.) Mueller) apparently has only valvate petals. Imbricate bud petals have been noted for other species in Polyalthia and in other genera usually placed in the Unoneae (Okada & Ueda, 1984; pers. obs.), suggesting that the distribution and the taxonomic importance of this character are in need of review. Further examinations of petal aestivation should include observations on degree of petal overlap, and whether the petals overlap distally, basally (including insertion), or both. Indication that previous treatments are in need of further scrutiny also de- rives from Walker’s (1971) observations that two species of Polyalthia differ in pollen characters from the other congeners examined. These characters, as well as several other nonpollen characters (discussed below), unite these two species with four not examined by Walker, and the mutual possession of these characters constitutes the evidence for the monophyly of the P. hypoleuca complex (the pollen data will be treated in detail in Rogstad & Le Thomas, in prep.). The pollen morphology of these species 1s different from that of the other Polyalthia species examined to date, with the few important exceptions discussed below It is difficult to reconcile the pollen characters of the Polyalthia hypoleuca complex with those of Walker’s Uvarieae, to which other members of Poly- althia conform (see also Le Thomas, 1988). The distally imbricate floral bud petals and the distinguishing pollen type within the complex render its place- ment both within the genus and within the subfamilial classification proposals noted above problematic. Thus, the clearly monophyletic P. hypoleuca complex may be the sister group to some taxon in another tribe, or it may be derived from within Polyalthia. Until the phylogenetic relationships of the complex are more clearly understood, assigning new generic or sectional status to this constellation of species may create future nomenclatural problems. Thus, for 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 167 URE 2. Petal aestivation in buds from Polyalthia hypoleuca complex. A, B, P. glauca es ie 939): A, outer petals not substantially overlapping: B, inner petals not sub- stantially overlapping (outer ones pulled back). C, D, P. discolor (Rogstad 814): C, outer petals slightly overlapping; inner petals significantly overlapping (lower outer one removed). Scale bars = | m the purposes of this investigation, the most prudent approach is to maintain the group within Polyalthia until the needed reclassification is attempted. The implications of the above taxonomic problems when using congeneric species to investigate the problem of closely related, sympatric species are explored below. Sinclair (1955, and references therein) followed most previous researchers in dividing the genus into two sections, Eu-Polyalthia Blume (or, by later convention, Polyalthia), defined as those species with two or more ovules per carpel, and Monoon Miq., comprising those with only one. Sinclair further suggested that distinct species groups existed within each section but did not explicitly propose character suites by which they could be distinguished. Ex- amining these groups as candidates for a monophyletic group with sympatric species, I determined from herbarium and field work that one complex of species, the P. hypoleuca complex, was especially suitable in that a unique suite of characters is shared by all the members of the group and various members were found growing sympatrically at different field sites. My final concept of this complex, although related to Sinclair’s Group | of sect. Monoon (1955; he included only P. sumatrana (Miq.) Kurz, P. hypoleuca Hooker f. & Thomson, P. glauca, P. longifolia (Sonn.) Thwaites, and P. parkinsonii Hutch.), is quite different from his as to both the limits of the group and the circumscriptions 168 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Figure 3. Leaves. A-D, species of Polyalthia hypoleuca complex (note weak ondary venation and uniformly white eure surfaces): A, P. ovalifolia (left; Rogstad 843) and P. hypoleuca (right; Rogstad s.n., 19 Dec. 1983), scale bar = 5 cm; B, sympatric at — Forest Reserve, ae Malaysia (P. hypoleuca above, P. SU- uniformly distributed papillae, scale bar = 0.5 mm; D, P. multinervis (Rogstad 813), uniformly distributed papillae, scale bar = 0.5 mm. E, F, abaxial surfaces representative of Polyalthia species not in P. hypoleuca complex (note absence of dense, uniformly distributed papillae): E, P. aus (Nur 18585), F, P. insignis (Hooker f.) Airy-Shaw (SAN A 4325), scale bars = 0.5 n of the species within it. Nevertheless, Sinclair was probably segregating his Group | by implicitly recognizing some of the characters by which I explicitly delimit the P. hypoleuca complex here THe LIMITS OF THE POLYALTHIA HYPOLEUCA COMPLEX I include in the Polyalthia hypoleuca complex only those species with all of the following seven characters: densely and uniformly distributed papillae cov- 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 169 eRe NG Fier é ~ gealione, ¢ ECE —« Ficure 4. Features of leaf anatomy shared by all members of Polyalthia eae complex. A, B, epidermal peels showing Pau anticlinal walls: A, P.s frana (Rogstad 508); B, P. discolor (Rogs 14). ae C, P. glauca (Rogstad ne D, P. ae ease 508); E, P. multinervis eee 813). Scale bars = 30 u ering the abaxial surface of leaves; extremely thin secondary veins differing little in diameter from the tertiary ones; secondary veins usually not forming a strong or relatively straight intramarginal vein, and relatively more numerous ape Cane together than in most taxa of the family; fundamentally white bark; monosulcate, “boat-shaped”’ pollen; pollen-wall architecture including a psi- ate, Seca perforated tectum, regular columellae, and a bipartite, foliated basal layer (Rogstad & Le Thomas, in prep.); and spiniform endosperm rumi- nations. The most easily recognized of these characters is the densely and uniformly distributed papillae on the abaxial surface of the leaves (FIGURES 3, 4C—E). These papillae, which are lacking in all other species of Polyalthia (see, for example, FIGURE 5), cannot be seen readily with a 10x hand lens, and so the use of a dissecting microscope (at least 25 x) is recommended. It is probably the dense distribution of these papillae that colors the lower leaf surfaces of all species in this group white (often mixed with pale green, brown, gray, or gold). While several genera of this family (e.g., the Dasymaschalon Dalla Torre & 170 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 FiGure 5. Leaves from Polyalthia species not in P. igen complex (note strong secondary venation): A, P. longifolia (Rogstad 960), scale bar = 2 cm; B, P. cinnamomea (Rogstad 936), scale bar = 2 cm; C, P. cauliflora (Rogstad 951), scale bar = 1 cm; D, P. rumphii (above; Rogstad 922) and P. jenkinsii (below: Rogstad 956), scale bar = 3 cm. Harms—Desmos Lour. complex; some species of Friesodielsia Steenis, Popowia Endl., and Yylopia L.) have at least a few species with white abaxial leaf surfaces, most lack papillae, so this color must be caused in the latter genera by some other feature of the abaxial surface, for example, the nature of the waxy cuticle. As far as I am aware, with the possible exceptions discussed here and below, leaf papillae of similar distributional uniformity, density, and structure are not found regularly in any other taxa of the Annonaceae. Roth (1981; pers. comm.) has indicated that within the family, the papillae most similar to those of the Polyalthia hypoleuca complex are found in Onychopetalum R. E. Fries, Bo- cageopsis R. E. Fries, Richella A. Gray, Rollinia St. Hil., Ruizodendron R. E. Fries, and Woodiella Merr. I have examined the holdings of these taxa (see APPENDIXES |, 2) at A, GH, MO, and us, and in none have I seen leaf papillae identical in form or distribution to those found in the P. Aypoleuca complex. Further, all of these taxa are quite different in floral and fruit morphology from members of the complex; only Woodiella is also an Old World genus. In a survey of 38 Neotropical genera, Van Setten and Koek-Noorman (1986) re- ported papillae for at least some species in Bocageopsis, tas S. Moore, Onychopetalum, Ruizodendron, Annona L., and Rollinia. However, a com parison of their description of the papillae and eae 4 and 9 with pci 4 here demonstrates differences in papilla structure and distribution between members of the P. hypoleuca complex and all of the genera they listed. Still, 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 171 further investigations of the distribution and taxonomic importance of abaxial leaf papillae in the genera noted by Roth, as well as by Van Setten and Koek- oorman, are needed. While none of the members of these genera has all seven of the characters defining the P. hypoleuca complex, and all share with their congeners characters not found in the complex, some of these species are worthy of consideration in the search for possible outgroups to the complex. Of all the characters setting the Polyalthia hypoleuca complex apart, those relating to leaf venation are the most difficult to describe because occasionally only subtle differences exist between members of the complex and other species. Rarely, no differences are readily apparent. Thus, these characters are best examined in direct comparisons between specimens. First, the secondary veins are very fine and differ little in diameter from the tertiary ones, while in most Annonaceae they are much broader. In the P. hypoleuca complex, veins of both degrees are only very slightly raised above the leaf surface and are often difficult to see. Second, the secondary veins do not usually form a strong or relatively straight intramarginal vein, and they are relatively more numerous and closer together than in most taxa of the family. The vein characters of the P. hypoleuca complex can be compared with those of other representative species of the genus in Ficures 3 and 5. The bark of species of the Polyalthia hypoleuca complex is a rare type in the family in that it is fundamentally white (FiGuRE 6), although especially on the young twigs and branches it may be tinted with shades of yellow or red in some species. I have been unable to find identical bark coloration in any other species of the genus, or even within the family (e.g., see Sinclair (1955); pers. obs.), with the following exceptions. White bark has been noted for Pseudox- andra cuspidata Maas (although brown bark also occurs in this species) and in Oxandra leucodermis (Spruce) Warm., apparently the only two known Neo- tropical species with this characteristic (Maas et al., 1986). Another character uniting the species of the Polyalthia hypoleuca complex was first noted by Walker (1971). In his survey of annonaceous pollen types, he placed Polyalthia in his tribe Uvarieae, which consists of those taxa with solitary, inaperturate, radiosymmetric pollen. He noted, however, that of the 22 species examined, P. glauca and P. hypoleuca stood apart in having sulcate, “‘hoat-shaped” pollen. All of the taxa in the P. hypoleuca complex have pollen that is indistinguishable from that of these two species (Rogstad & Le Thomas, in prep.). I have also examined the pollen of four other Malesian species of Polyalthia (i.e., P. lateriflora (Blume) King, P. obliqua Hooker f. & Thomson, P. sclerophylla Hooker f. & Thomson, P. socia Craib; Rogstad 931, 935, 930, and 958, respectively) not treated by Walker; none of these species has pollen of the type found in the P. hypoleuca complex (see also Le Thomas, 1988). Walker (1971) found that sulcate, boat-shaped pollen was extremely rare in the Old World, occurring only in the genus Enantia Oliver of Africa and the two species of Polyalthia noted above. Walker did find such pollen in ten Neotropical genera, and these along with Enantia constitute his tribe Malmeae. Le Thomas (1981) has also found this pollen shape in a few other African genera and, more importantly here, in at least seven species of Polyalthia from 172 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 vs 4 PSOE FiGurE 6. A-C, bole characters in Polyalthia hypoleuca complex: A, P. hypoleuca (79 cm girth at breast height), fissured bark lacking hoop marks; B, P. discolor (123 cm girth at breast height), smooth bark with horizontal hoop marks; C, P. glauca (104 cm 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 173 Africa and Madagascar. It should be noted that Gottsberger and Silberbauer- Gottsberger (1984) have cautioned that the apparent pollen shape of some Annonaceae may be an artifact of preparation. Another feature that indicates a close genealogical relationship among the species here included in the Polyalthia hypoleuca complex is the architecture of the pollen wall (Rogstad & Le Thomas, in prep.). The five species of the complex for which pollen-wall examinations have been possible have a psilate, moderately perforated tectum, with regular columellae and a bipartite, foliated basal layer. While pollen appropriate for TEM observation of this feature has not yet become available for the sixth member of the complex (P. ovalifolia S. H. Rogstad), observations by compound microscope indicate that it agrees in external tectal construction with that from other members here incuded in the complex. This type of pollen architecture appears to be rare in the family; Enantia chlorantha Oliver (Le Thomas, 1981) and Ephedranthus amazonicus R. E. Fries (Waha, 1985), both placed by Walker in his tribe Malmeae, are the only examples thus far known from outside of the genus. The majority of the numerous other Annonaceae examined have a different surface architecture (Walker, 1971; Le Thomas, 1981; Waha, 1985; pers. obs.). Within Polyalthia, Le Thomas (1981) has described a pollen-wall architecture very similar to that of the species of the P. hypoleuca complex for P. capuronii Cav. & Keraudren, P. emarginata Diels, P. heteropetala Diels, and P. oligosperma (Danguy) Diels, all from Madagascar. A final character that unites the members of the Polyalthia hypoleuca complex concerns the nature of the endosperm ruminations. All Annonaceae have en- dosperm that is divided by “ruminations as transverse folds of the tegmen or, also, of the testa or middle integument” (Corner, 1976, p. 68). In cross section the seeds have ruminations that can be a) very regular and platelike with a central ‘“‘cross” of endosperm (Corner, 1949, fig. 124A; FiGuRE 7C), b) irregular as though the regular plates just mentioned are broken into pieces of various sizes and distributions (Corner, 1949, fig. 14D, E; Figure 7D); or c) numerous, very fine, and needlelike (spiniform; Ficure 8A, B). The endosperm rumina- tions of most of the annonaceous species that have been described in the literature are type a or b (e.g., Corner, 1949, 1976; Periasamy & Swamy, 1961; Rao, 1975, 1979, 1982), while those of all of the species here included in the P. hypoleuca complex are spiniform. Outside of the genus Polyalthia, spiniform ruminations have been found in Bocageopsis, Cremastosperma R. E. Fries (not all species), Enantia, Onychopetalum R. E. Fries, Oxandra A. Rich. (not all species), Piptostigma Oliver, Polyceratocarpus Engler & Diels, Popowia EndL., Pseudoxandra R. E. Fries, Pseuduvaria Miq., Richella A. Gray, and Unonopsis R. E. Fries (see APPENDIX 1). girth at breast height), mature tree, flaking white bark with red bole beneath (note hoop _ D, P. glauca, distal portion of emergent pneumatophore (note annular scars indicating rhythmic extension growth). 174 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 URE 7. Seedling, fruits, aie seeds: A, Polyalthia hypoleuca seedling, mature leaves Bis ae: in order of development; B, P. discolor (Rogstad 814), mature fruits, micro- pylar region of seed exposed nner ventral orientation of seed (arrow indicates URE 8. FIG spe, with spiniform endosperm ruminations, Be: aril scar (arrow), and circumferential ridge; B, P. sumatrana (Rogstad 508), sections at various levels, showing spiniform endosperm ruminations throughout; C, P. discolor Ceo 814), regular, lozenge shaped, circumferentially grooved, lacking cet unced aril scar at micropylar region (arrow); D, P. glauca (left; Rogstad 939), P. hypoleuca (center, Bra 912), P. sumatrana (right; ’ Rogstad 508). I have also examined species from over 24 additional genera of Annonaceae (including 34 species of Polyalthia, and all genera now known to share at least one other character from the suite defining the Polyalthia hypoleuca complex) for this character (see APPENDIX 1). Seeds with exclusively needlelike rumi- nations were lacking in these other genera but were found in P. oliveri Engler & Diels, P. stuhlmannii (Engler) Verdc., P. suaveolens Engler & Diels, P. su- berosa (Roxb.) Thwaites, P. capuronii, and P. oligosperma. Of these species, only the latter two have as many as three of the seven characters I use to delimit the complex, a finding discussed further below. Although wider surveys of this seed character are needed, the above evidence indicates that the spiniform endosperm ruminations found in all of the species of the P. hypoleuca complex can be included in the suite of characters distinguishing this group. “stigmatic crest” remnant on oe ioe ree = | cm; C, P. longifolia (Rogstad 960), cross section of mature seed, scale = ; D, P. jenkinsii (Rogstad 956), near- mature carpels and seed, scale bar = fy mm 176 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Another character deserves mention as a potential member of the suite defining the Polyalthia hypoleuca complex. A survey of the leaf anatomy of several species of the genus has revealed that all species of the complex have an adaxial epidermis in which all or most cells have straight, thickened anti- clinal walls (Figure 4A, B). In this survey 34 other species from the genus were examined for this character (see APPENDIX 2); it was lacking in all but two of them. These results contrast with the findings of Van Setten and Koek- Noorman (1986), who found similar-looking adaxial epidermis in several taxa, which they did not list. These authors suggested that the anticlinal walls appear thickened due to “thickened cuticular ledges following the anticlinal . . .”’ walls (p. 22; see figs. 6, 7). With the species of the P. hypoleuca complex, it is not clear whether thickened ledges or thickened anticlinal walls give rise to this effect (see Figure 4C). While this character is not common in Polyalthia, its distribution in the rest of the family is unclear, so it cannot yet be included in the list of characters defining the complex. Interestingly, the only other two species of Polyalthia found to have an adaxial epidermis similar to that seen in all members of the P. hypoleuca complex are from Africa, and one of these, P. oligosperma (Dangy) Diels, is a leading potential candidate as an outgroup species for the complex. In summary, a unique suite of seven characters is shared by all and only the members I am placing in the Polyalthia hypoleuca complex. Although these characters are uncommmonly found in other taxa of the family, each usually occurs in relative isolation from the other characters. A detailed analysis of the distribution of these characters outside of the P. hypoleuca complex has been conducted (Rogstad & Le Thomas, in prep.), and only an outline of the resulting conclusions regarding the monophyly and taxonomic position of the complex is given here. Within Polyalthia, only P. capuronii and P. oligosperma (both from Madagascar) have as many as three of the seven characters, making them the best congeneric choices as possible sister lineages to the P. hypoleuca complex. Outside of the genus, it is interesting that the seven characters find their densest distribution in a group of eleven genera designated by Walker (1971) as tribe Malmeae. Of these, one African and four Neotropical genera have species possibly possessing three of the seven characters, while only the Neotropical genera Oxandra and Pseudoxandra have species perhaps possess- ing as many as five, making these latter species the most likely choices as sister taxa to the complex. Note that all of the species having more than two of the seven characters are from Africa/Madagascar or the Neotropics, giving strength to the notion that at the least, the members of the P. hypoleuca complex are the only Asian-Oceanic members of a larger monophyletic group with other members occurring in other areas. Further, all of the taxa noted above as having at least one of the seven characters share with their congeners characters not found in the P. hypoleuca complex (Rogstad & Le Thomas, in prep.). The latter finding, together with the fact that the character suite discussed above is found only in all members of the P. hypoleuca complex, leads to the conclusion that, by parsimony, the P. hypoleuca complex either is a monophyletic group (Rog- stad & Le Thomas, in prep.) or is composed of all the Malesian members of 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 177 a larger, small monophyletic group with complementary members in Africa or the Neotropics. TAXA EXCLUDED FROM THE POLYALTHIA HYPOLEUCA COMPLEX One group of species, treated by Sinclair (1955) as members of the genus Melodorum Lour., 1s problematic because they bear abaxial leaf papillae most similar to those of the Polyalthia hypoleuca complex. In his survey of leaf papillae, Roth (1981; pers. comm.) did not see a close similarity between the papillae in Melodorum and those of the P. hypoleuca complex. The placement of Melodorum 1s contested (e.g., see Okada & Ueda, 1984), and various authors have described some of the relevant taxa as species in Polyalthia (including P. aberrans Maingay ex Hooker f. & Thomson, P. aberrans Pierre ex Finet & Gagnep., P. affinis Teysm. & Binnend., P. diospyrifolia Pierre ex Finet & Ga- gnep., and P. stamensis Boerl.). While this is not the place to discuss in detail the proper placement of Melodorum, in my opinion these species differ from the genus Polyalthia be- cause they are climbing shrubs, the sepals are connate, the petals are thick and coriaceous, with the inner ones distinctly smaller than the outer and not spread- ing as in Polyalthia (see Sinclair, 1955), and the chromosome number (2n = 16) has not yet been found in Polyalthia (Okada & Ueda, 1984; Rogstad, unpubl. data). Further, although the papillae of some of the specimens of Melodorum are similar to those found in the P. hypoleuca complex (sheets that are Clearly Melodorum according to the characters listed below are occasionally annotated as P. glauca), they are less dense and are only sometimes present. Additionally, the species of Me/odorum (as denoted by Sinclair, 1955, for example) differ from all the species I assign to the P. hypoleuca complex in having very distinct leaf venation with an intramarginal vein well removed from the leaf margin, deeply grooved stigmas, globose pollen (Walker, 1971), and platelike endosperm ruminations. In light of these considerations, I con- clude that little justification exists for including the taxa of Sinclair’s Melodorum in Polyalthia or the P. hypoleuca complex, and that any similarity between their abaxial leaf papillae is of dubious value in trying to establish a close phylogenetic relationship between them. Polyalthia longifolia must also be considered for inclusion in the complex. Although Sinclair (1955) placed this species in his Group 1, several lines of evidence argue against it. As noted above, Sinclair did not explain criteria for group inclusion, so I cannot address his reasons for the placement of P. /on- gifolia. However, the species has not one of the seven character states listed above that I am using to define the P. hypoleuca complex (e.g., see FIGURES SA and 7C, and Rao (1979)). Further, the staminate and carpellate portions of the torus in P. /ongifolia are pubescent, a character altogether lacking in any of the members of the P. Aypoleuca complex. I would include P. longifolia in Sinclair’s Group 6 because it has a pubescent torus, midribs deeply grooved longitudinally on the abaxial surface of dried leaves, platelike endosperm rumi- nations with a “cross” of endosperm clearly visible, the outer seed-coat surface 178 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 a 1 marked with parallel fin iat ing pery to the circumferential ring of the seed, and the first above-ground growth of the germinating shoot deriving from extension of the first internode above the cotyledons rather than from the hypocotyl (the latter being the case for all members of the P. hypoleuca complex). Species that other authors recognize but that I have reduced to synonymy under other species within the Polyalthia hypoleuca complex are discussed below. SPECIES CONCEPT While it is not my purpose here to discuss the nature of species, it is important that I provide the working definition I have employed to define the species of the Polyalthia hypoleuca complex. My species decisions have been made pri- marily on a morphological basis: are there clear discontinuities by which species can be delimited? I have searched for such discontinuities by recording data on numerous characters from both herbarium and living specimens and then analyzing these data in several different ways, as outlined in the methods section below. Clear discontinuities were found in character-state distributions, and these form the basis for my species concepts. For two cases where these dis- continuities are relatively slight, and in fact for all of the species discussed, the decisions reached here will be supported in later articles detailing the com- parative autecologies of these species. Thus, my species are defined not only on morphological considerations, but also on an assessment of whether or not they are likely to be isolated genetically or ecologically in nature. The reader interested in weighing the sum of all the evidence bearing on the taxonomic decisions presented here may wish to consult the subsequent articles. Ulti- mately, the species concept I have used in this treatment is perhaps most fully in accord with the “evolutionary” species concept of Simpson (1961, p. 153): “ta lineage (an ancestral-descendant sequence of populations) evolving sepa- rately from others and with its own unitary evolutionary role and tendencies.” This is not to say that the evolutionary species concept is the preferred one for all plant species, but after considering all of the information bearing on species in the P. hypoleuca complex, I believe it to be the most applicable here. METHODS Morphometric analyses were used to search for evidence of morphological species. Character observations and measurements were taken both from her- barium specimens at A, BK, BKF, BO, F, GH, K, KEP, KLU, L, LAE, MO, P, PNG, SAN, SAR, SING, U, US, and Yezin (Burma) and from living specimens collected in several Malesian locations (see Map 1). All of the herbarium specimens used in these analyses, as well as representative voucher specimens of species for which field measurements are analyzed (Rogstad numbers deposited at A), are noted in the specimen citations. An attempt has been made to include speci- mens covering the full range of character variability for each species. The known eographic range determined from herbarium specimens for each species of the Polyalthia hypoleuca complex is depicted in maps accompanying the species 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 179 20°N 155°9E 90°E t 1809S Map |. Sites visited. Peninsular Malaysia: Pasoh Forest Reserve (A), Telok Forest Reserve (B), Tasek Bera Reserve (C). Thailand: Khao Chong Forest Reserve (D), Khao Tha Phet Forest Reserve (E). Sarawak: Bako National Park (F), Lambir and Niah national parks (G), Baram River Site (H). Sabah: Sandakan (1). Papua New Guinea: Aluki Village (J), April River (K). descriptions; specimens from the entire known range of each species have been included in the morphometric analyses. Coverage of character variability and geographic distribution is thus most comprehensive for the better-collected species. Measurements of herbarium specimens were recorded from all sheets bearing several intact flowers and/or any fruits and were made with a ruler or with a dissecting microscope fitted with an ocular micrometer when appropriate. The most mature flower or fruit present with all the relevant characters intact was measured. Flowers were denoted as mature or immature based on character- istics determined from field observations. For most of the species, petals of immature flowers are open and radiating, while those of mature ones are erect and closely parallel to the longitudinal axis of the flower (Rogstad, in prep. b). In Polyalthia glauca and probably P. ovalifolia, the outer petals are erect until they turn yellow-orange at maturity, then the distal portion bends outward to 30° or more. Petals were also designated as being mature if label data indicated that they were other than greenish. Fruits were deemed mature if they were at least as large as the smallest fruit of a species for which label data indicated fruit color as deep red to black; fruits of all these species achieve maximum 180 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 size while still green, turning black when mature a number of weeks later (pers. obs.). Specimens with only one or a few flowers were generally not measured for all characters unless the specimen was judged as crucial to the study (e.g., at the edge ofa species range or bearing an unusual character). To avoid unusual phenotypic effects on measurement of leaf characters, an average-size leaf was subjectively chosen from each specimen. Pertinent herbarium label data were recorded for the specimens. The categories of data collected for each herbarium specimen and used variously in the morphometric analyses reported below are listed in APPENDIX 3. A copy of the full data set stored on double-sided, double- density, 5.25-inch diskettes utilizing MS/PC-DOS format has been deposited with Rogstad (1986) in the Botany Libraries of Harvard University. A copy can be obtained free on request, provided the user agrees to share the results of any analyses of these data with the author and sends a self-addressed, stamped envelope containing an appropriate diskette. In the case of living specimens, tree heights were measured with a Suunto PMS/360PC Clinometer following the directions provided. The girth of trees at breast height (GBH) was measured at 1.5 m by tape measure and converted to diameter at breast height (DBH). These data were analyzed by several different methods utilizing the SYSTAT (Wilkinson, 1984; in conjunction with an IBM PC-AT), SAS Version 5 (SAS Institute, Inc., Cary, North Carolina), and BMDP (Health Science Computing Facility, University of California, Los Angeles, California) statistical packages. The last two packages were used as available at the Washington University Computing Facilities, St. Louis. While the univariate and bivariate analyses presented in the results section below are widely used and self-explanatory, my choice of multivariate tech- niques requires justification. As stated above, the species concept I develop with regard to the Polyalthia hypoleuca complex is based on morphological analyses. I begin, therefore, by searching for discontinuities in morphological characters, attempting to apply the widely held tenet that good taxa are best defined not on one character, but on a suite of them. Patterns of variation within and among characters are often very complex. Sokal (1985) argued that many taxa are polythetic, that is, are groups that can be defined by a suite of characters, with some or all of the members lacking one or another character. The literature is replete with disputes over the proper classification of organisms, and these are testimony both to the complex nature of character variation among organisms and to the fact that each investigator carries personal biases into subjective decisions about the importance of con- flicting characters. Sokal and Rohlf (1980) have demonstrated that different taxonomists achieve different subjective classifications based on the same set of taxa, but the differences are greatly reduced when investigators employ standardized morphometric analyses of explicitly coded character states (Sokal & Rohlf, 1970). The task for the taxonomist, then, is to search for nonsubjective signals of morphological discontinuity (at best based on several covarying characters) that rise above the noise of characters exhibiting more or less random variatio A review of cen descriptions, and of the sorting and annotation of 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 181 herbarium specimens by previous researchers of taxa I have included in the Polyalthia hypoleuca complex, indicated that this assemblage was in need of revision due to complex character distributions that did not always result in congruent or stable taxonomic entities. I have used multivariate analyses as an aid in defining the taxa in this group based upon characters and techniques that are repeatable and explicit. Several multivariate statistical techniques have been designed to search for patterns in complex character-state data sets. A review of plant taxonomists’ use of these techniques (Duncan & Baum, 1981) indicates that cluster analysis has been the most widely employed. However, since the first step in my analysis is an attempt to demonstrate whether or not any groups (clusters) exist based upon character variation, and since cluster techniques always give clusters even if a random data set is used, searching for the possible existence of clusters with a method that always yields them seemed questionable. Here I used a two-step approach that does not necessarily sort specimens into groups, but rather summarizes the degree to which character states covary among speci- mens and then, if groups are detected, permits a statistical test of group mem- bership for each specimen. The first step was to examine the data matrix by principal-components analysis, or PCA (SYSTAT; Wilkinson, 1984). Reviews of PCA theory, meth- odological assumptions, and constraints can be found in most texts treating multivariate methods (e.g., Nie et a/., 1975; Neff & Marcus, 1980; and their references), and the following discussion is simplified for brevity. With PCA as utilized here, the raw data matrix—in this case measurements of a set of characters taken from herbarium specimens—is analyzed to derive its corre- sponding correlation matrix (a covariance matrix can also be used, but since this approach 1s strongly influenced by characters of larger size, only correlation matrices have been employed here). From this matrix, principal components (axes) are extracted that are descriptors, in serial fashion from greatest to least, of variance among character states across the specimen set. The first principal component is the axis with the maximum variance. The second component axis 1s derived in the direction of greatest variance orthogonal to the first one, and thus successive axes are computed until all the variance in the data set is accounted for. Each principal component is a mathematical equation (eigen- vector) that includes a term for each character weighted by a coefficient com- mensurate to that character’s contribution to the total variance of that com- ponent. Each specimen can be assigned a numerical value along each component (the character coefficient times the character value for each character of that specimen summed over all of its characters) that summarizes the degree to which its characters covary with the other specimens in characters that heavily influence each component. Specimens more similar to one another in the ways their characters covary are thus placed in closer proximity along the principal components than are those that differ. If strong patterns of character covariance exist, these will be detected in the first three or four components, where the greatest degree of variance is accounted for, the later components accounting for very little of the total variance. If patterns of character distributions are more or less random, 182 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 then the first components will not explain much more variance than later ones. If all specimens are plotted for their scores on three components, specimens with similar patterns of character correlation are placed in groups in the prin- cipal-components space defined by the three components. Use of PCA enables one to search for groups defined by suites of (covarying) characters and, once the principal components are calculated, to identify the characters that con- tribute most to the differentiation of groups along any particular axis. As in- dicated above, groups defined by suites of characters are preferred over groups defined by only one. Two further steps often taken in PCA analyses should be mentioned. First, once all of the data have been analyzed, characters with no high loadings on any component can be removed from the data matrix and the analysis recal- culated. This eliminates characters that vary randomly with respect to strongly covarying characters and therefore contribute only “noise” to the analyses. Groups or loadings on remaining characters may thus be given clearer defi- nition. While I have conducted this type of exploration with the data sets used here, removal of low-weighted characters did not significantly improve the results (and is a somewhat subjective decision in any case), so only results including all of the continuous characters are reported here. Second, once a set of principal components has been determined for a data set, the corresponding component axes can be rotated to search for more meaningful components by simplifying the component structure (Nie et al., 1975). Although rotation may thus improve the interpretability of the component loadings, it did not in the examples discussed below, so only unrotated solutions are presented. Note that PCA does not always give clear groups; decisions of group mem- bership may be difficult and sometimes ultimately rest upon the discretion of the practitioner. As far as | am aware no method has yet been devised for PCA that permits a statistical test of the strength of groups or of group membership by individual cases. When visual inspection of the PCA analyses of the Polyalthia bec complex data set resulted in hypotheses of groups of specimens, I utilized discriminant analysis (again, see standard texts on multivariate ae ee niques: e.g., Nie ef a/. (1975), Neff & Marcus (1980), and their references) to test the statistical strength of the membership of each specimen in the hy- pothesized group. The same data matrix as used above in the PCA can be analyzed by discriminant analysis techniques, although in the latter case each specimen must be designated, before the analysis is undertaken, as belonging to one of the hypothesized groups. The specific technique used here is a jackknifed discriminant analysis, in which the first step is to remove one specimen from the data set. Next, a set of linear discriminant functions (axes) is derived, each maximizing, in decreas- ing order, the separation of the groups. In other words, the first function max- imizes the separation of the groups, the second function is orthogonal to the first and is the next best function at maximizing group separation, and so on, until the maximum number of functions has been derived. If the members of the a priori groups are much more similar to one another than they are to members of other groups, their placement on a discriminant function axis 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 183 should cluster them in relative proximity on that axis away from other groups or specimens. As with PCA above, once the discriminant functions are cal- culated, the characters that contribute most to the differentiation of groups along any particular axis can be identified. If a set of variables is found in the first three or four discriminant functions that provides satisfactory discrimination for the a priori-defined groups, then the excluded specimen can be statistically “‘classified’’ as to the group in which it belongs. This specimen is then returned to the total data set, and the next specimen in the data table is excluded and the above procedure repeated. In this way, each specimen can in turn be removed from the data set and then classified as to group membership in relation to the remainder of the specimens. If groups are well defined, all specimens should be correctly placed with high statistical confidence as belonging to the group to which it was assigned a priori. If no clear groups are present, many individuals will be assigned to groups other than those in which they were first placed as a result of the PCA analyses. It is important to keep in mind that the species assignations indicated in all of the results below have been ultimately decided a posteriori with respect not only to the totality of the morphometric evidence presented here, but also with regard to ecological characteristics to be detailed in related articles (Rogstad, in prep. a, b A valuable feature of both PCA and discriminant analysis is that the per- centage of the total variance or dispersion in the data set explained by each extracted principal component or discriminant function, respectively, is cal- culated. In a totally random data set, little underlying structure os po lccue of variation of variables would be expected, so a particular p p mponent or discriminant function will not explain much more of this variance or dis- persion than the one immediately following. However, if there is structure in the data, then at least the first principal components or discriminant functions derived should explain a disproportionately large portion of the variance or dispersion. In all of the results reported below, the latter possibility will be seen to be the case. Another useful type of information obtainable from both PCA and discrim- inant analysis derives from the calculation of the component loadings and the coefficients for the canonical variables, respectively. These are informative as to which of the characters are most important with regard to the variance or dispersion explained by each principal component or discriminant function. Characters with larger loadings or coefficients on a principal component or discriminant function have patterns of correlated variation with regard to that component or function and can likely be used in suites of correlated characters that best define discontinuous groups (species). Discriminant analysis and PCA are subject to certain constraints and re- strictions. For example, both methods assume multivariate normality for the data. While PCA is thought to be robust against all but extreme deviations from normality (e.g., see Sneath & Sokal, 1973), discriminant analysis also includes the further assumption of equal variance-covariance matrices within groups (Neff & Marcus, 1980). However, use of discriminant analysis is not invalidated even when both assumptions are violated to some degree if errors 184 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 of classification are small (Neff & Marcus, 1980), which is the case for all the results presented below. Only characters that are continuous or meristic over broad ranges are included in the data matrices analyzed by PCA and discrim- inant analysis here. Those that are clearly not normally distributed (e.g., in- formation, such as color, that is designated by arbitrary coding; presence- absence data) are not included in the PCA and discriminant-analysis results described below, although, as will be seen, such data can be used informatively in conjunction with these analyses. Each analysis also includes only specimens with no missing values for any character employed. Additionally, both the principal components and the discriminant functions of the respective methods are based on linear models, and for present purposes this approach has been accepted. Finally, the species treated here have not been equally well collected, and therefore the sample sizes are not equal. While this is not a major problem for PCA, it can affect discriminant analysis. Thus, in the results reported here, all discriminant analyses were performed with the BMDP option PRIOR ad- justed to reflect the sample size of each species. Obviously, the ideal statistical requirements of these methods are only at best approached, rather than met entirely, in the analyses discussed here. This will probably always be the case for systematists’ data sets due to the problems inherent in collecting biological information. Does this indicate that these tests should not be utilized? I maintain that they should be used as exploratory and descriptive tools. The results detailed below are not presented as incontro- vertible systematic truth, but rather as aids that influenced my taxonomic decisions. Further, the characters, methods, and decisions are explicit and repeatable, and other workers can follow the development of concepts, modify the analyses, or contribute additional information in a prescribed manner. At the very least, the presentation of systematic analyses by these methods is heuristic in rapidly conveying to the reader not only the extremes of variation important in defining groups (the usual information provided in systematic treatments), but also the degree to which variation exists within each group. For example, after PCA of the character matrix, data for any desired character are separately available for statistical analysis within selected groups. Graphic depictions of the distribution of individuals according to their scores on various components provide rapid understanding of their dispersion, due to the co- variance of the characters heavily weighted on those components. RESULTS Since over 50 characters from 172 flowering specimens and 16 characters from 230 fruiting specimens were recorded, it would be inefficient to reproduce all of the possible permutations of uni-, bi-, tri-, and multivariate character analyses here. Instead, only a few representative examples will be presented. In the graphic presentation of these results, the species designations have been made after consideration of all of the data and analyses presented here and in subsequent related articles (Rogstad, in prep. a, b). In an evaluation of only two or three characters, a few specimens of a species may exhibit character states of another species for any (or rarely all) of them, but they otherwise 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 185 LEAF W (cm) 124 14 8 10 We B 3& B ‘ 9 2 a B s 84 8 4 G ‘i = G B S Ss 64 G 5 c eae so S s$ a r s® is ass & s mos OD ¥ s = HN fe H H 24 H tf ee pa ee a ae i ete et a a ag ak ago a a a T 5 40 15 20 as 30 LEAF L (cm) Figure 9. Leaf length vs. width for ae of P. hypoleuca complex = P. ovalifolia, D = P. discolor, G = P. glauca, H = P. hypoleuca, M = P. bas aa S = P. sumatrana. conform with the species to which I have assigned them. It will be seen from a comparison of the following bivariate and multivariate analyses that a sys- tematist giving more weight to certain characters than others (as judged from bivariate comparisons) may arrive at different conclusions than those reached y multivariate analyses. Sinclair (1955) noted that although leaves of Polyalthia sumatrana and P. glauca are nearly indistinguishable, those of P. hypoleuca are smaller and have closer veins. However, as is shown in FiGureE 3B, there is some overlap of specimens assigned to both P. sumatrana and P. glauca with P. hypoleuca with respect to leaf size. This is not to say that there are not tendencies in leaf sizes of these species, but this overlap negates leaf size as a simple character by which these species can be delimited absolutely. In fact, if one examines each data point in FiGure 9, disregarding the species designations that have been made based on the sum of the vegetative, floral, fruit, and ecological data, it is difficult to define distinct groups, except perhaps one comprising specimens of P. ovalifolia and another composed of three outlying specimens of P. su- matrana. This overlap has no doubt contributed to problems of classification and identification. Sinclair (1955, p. 322) further indicated that Polyalthia glauca can be dis- tinguished from P. sumatrana by its mature fruit “with thinner [carpel] stalks and pedicels.”’ FiGures 10 and 11 demonstrate that, again, while there are central tendencies in these species, the amount of overlap renders these characters of little use in discriminating clearly between the two species. In Ficures 10 and 11 it can be seen that the length and width of carpel stalks and pedicels from mature fruits are not very useful in discriminating any of the species except perhaps the gro g P. hypoleucaand P. multinervis Diels specimens. Even these are a0 leacly bounded from the P. glauca and P. sumatrana specimens. 186 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 STALK W (mm) 54 Ss s 4 s S Ss : $$ Ss j 3 S 5 $s 38 ae : 3 Ss sa 855 5 3 Ss Ss GC Sae28 5 ‘i D H ao efits = : G s 8 e- ges a 8 Ss Sis G gs § Bs B C D HAG : " Q wpge 8S 2 : HE me S at ee ee a T “> a. OW ae i= 2 =. oe Oe ee T (0) 10 20 30 40 STALK L_ (mm) FiGure 10. Stalk length vs. width for mature carpels of P. hypoleuca complex: B = P. ovalifolia, D = P. discolor, G = P. glauca, H = P. hypoleuca, M = P. multinervis, S = P. sumatrana. An examination of floral characters used by Sinclair (1955) also suggests why herbarium workers have had problems identifying specimens in this group. For example, Sinclair described the petals of Polyalthia glauca as 16-20 mm long, and thus larger than those of P. hypoleuca (8-12 mm long). FiGuRE 12 depicts the relationship of petal length to widest petal width for all the mature PEDICEL L_ (mm) 50 D S D 40 D0 0 ? D G D 5 3 6B “ S s§ D 304 i Sp S88 g , Gegs G B g s Ss SS D 5. Ss G 6G g s@ock e : 6° ss : S P 20 G S S Sis 5s 8s 5 3 ses 5 Sess 8 gs 8 s 258 G ‘ gs s G § S S = 5 10 H G s s LH HAA a HN AAA H 0 ie aa te. eee he ee ee Le ee Ne ee, T pa ae eal a Up Re, Goa he, laa r= 55; carpels > 12 (possibly very rarely = 12 in P. sumatrana). 4. Outer ek narrowly elliptic to elliptic (width/length quotient > 0.2), with dense tomentum on adaxial surface beginning on basal 4; carpel/stamen quo- tient > 0.3; inflorescences borne on branches below the leaves, arising from — . 208 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 tubercles, with peduncle bearing lateral developing flowers or resting buds. .. ee ee ee ree Tey ry ee eee ene et eee men eee ee 3. ence . Outer petals linear to narrowly elliptic (width/length quotient < 0. 2), w moderate to sparse tomentum on adaxial surface absent from at least basal - carpel/stamen quotient < 0.3; aie both from foliate axils and borne on branches below the leaves, well-developed tubercles lacking, the peduncles extremely reduced, bearing | terminal flower and no lateral buds. ......... 4, a 3. Pedicels < 7.5 mm long; stamens < 55; carpels generally < 1 5. Pedicels moderately to densely tomentose; collected only west of New Guinea. mest asse apihogr hentsente uate deprroneds Gat tess g euesou ew eyed eases earn als . hypoleuca. 5. aaa ae (or very sparsely pubescent); collected only in Papua New SA as ca ao ng a es ak otal ee aa eae leeY 6. P. multinervis. Key for Fruiting Material 1. Distal portion of mature carpels globose; placentation ventral; seed(s) biconvex- lenticular, often with irregular indentations on sur race, w ith longitudinal circumfer- slightly raised circumferential area), nO aril scar detectable. 2. Pedicels and youngest shoots with moderate to dense, short, rusty tomentum; laminas usually elliptic to widely elliptic, 7-11.3 cm wide, width/length quotient its ara tae dp an oye eee eee ee ea ae 2. P. ovalifolia. 2. Pedicels glabrous, youngest shoots moderately pubescent at most; laminas nar- rowly elliptic, 2.5—7 cm wide, width/length quotient 0.24-0.34. Stamen scars > 50; mature carpels with stalks > 13 mm long. ............ sj asiagnnen ocr toy gaan Saas ls shear Reto anna ehea laiedehers Aa 1. P. glauca. Distal portion of mature carpels ellipsoid; placentation basal; seeds ellipsoid, never with irregular indentations on surface, with strong longitudinal circumferential ridge, aril scar easily detectable. 4. Pedicels > 9 mm long; stamen scars > 50. ................ 4. P. sumatrana. 4. Pedicels < 9 mm long; stamen scars < 50. 5. Pedicels moderately to densely tomentose; collected only west of New Guinea. autres ee Geta tiie a eh ao ey tore dea hd ad Puen ee emake ee ypoleuca. 5. Pedicels glabrous (or very sparsely pubescent); collected only in Papua New MOORS aetinis Gast aicde fon gr dotonn ea ea ad Pe . P. multinervis. _ Species Accounts As with the characters used in the keys, those included in the species de- scriptions are taken largely from the data sets of mature floral or fruit characters used in the morphometric analyses. These data were collected from dried herbarium specimens (exceptions are noted). The minimum and maximum values of most continuous characters will be given for each species, followed in parentheses by the mean, the standard deviation, and the number of spec- imens for which the character could be evaluated. For noncontinuous char- acters, all specimens have the noted trait unless a percentage breakdown is given. In the latter case, the total number of specimens evaluated for the character will be given. Mean values, standard deviations, and number of herbarium specimens used are not reported when the last figure is less than en. Rather than listing all of the specimens I have examined, I give only selected 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 209 ones that illustrate the range of locations and of morphology for each species; all specimens used in the morphometric analyses are included. A ‘“‘finders list” of all the specimens I have examined is on deposit with the Botany Libraries of Harvard University, and a copy of it will be sent upon request. Most in- formation given for specimens should be self explanatory, with two exceptions. First, where it 1s not clear whether a label identification number refers to a collector or to the distributing institution, I have given the collector’s name or the institutional abbreviation in brackets. Second, a designation as to whether I classified the specimen on the basis of flowers (fl), fruits (fr), or both (fl+ fr) is given immediately following the identification number for each specimen; ster- ile specimens are also indicated (st). If these designations are in capital letters, the specimen has been included in the data sets for the morphometric analyses. Note, however, that different specimens are missing values for different char- acters or are immature, so all of the specimens included in the data sets are not used in all of the data analyses presented above. Local or common names applied to the species as recorded by collectors are listed alphabetically for each species. Different spellings that obviously refer to the same name are listed separately. A general location (or locations, if a name is widespread) is given with each name, with more precise information available in the list of selected specimens for each species. Short descriptions of the habitat ecology are given, but these only very briefly summarize the results of more detailed investigations into comparative habitat ecology, seedling ecology, and floral biology of several species of this group, a full description of which will be detailed subsequently (Rogstad, in prep. a, b). 1. Polyalthia glauca (Hassk.) Mueller, Descr. Notes Papuan Pl. App. 95. 1877, non Boerlage (1899); Uvaria glauca Hassk. Flora 25(Beibl. 2): 31. 1842; Guatteria glauca (Hassk.) Miq. Fl. Ned. Ind. 1: 49. 1855; Monoon glau- cum (Hassk.) Mig. Ann. Mus. Bot. Lugduno-Batavum 2: 19. 1865. Type: Hasskar! s.n. (possibly marked ‘nom. sund. Kaju tinjang” and/ or “Arbor in m. Roeboet’’) (n.v.). Ficures 2A, B; 3B, C; 4C; 6C, D; 8D; 18C; 20A, B; 21A. Saree hypoleuca Mig. Fl. Ned. Ind., Eerste Bijv. 381. 1861. Type: [Sumatra] Pri n, Madang Poelo, Diepenhorst 2095 (u “Unona ee Miq.” ex Hooker f. & Thomson, Fl. Brit. India 1: 63. 1872, nomen nudum. Unona merrittii Merr. Philipp. J. Sci. C. 1: 190. 1906; Polyalthia merrittii (Merr.) Merr. Philipp. J. Sci. C. 10: 250. 1915. Type: Philippines, Mindoro, Bongabong River, February 23, 1906, Whitford 1447 (n.v.). Polyalthia parkinsonii Hutch. Bull. Misc. Inform. 1917: 25. 1917. Synrypes: India Andaman Islands, Long Island, Feb., Parkinson 943 (fr) (kK); without precise locality, Dec., Parkinson 765 (fl) (k), 794 (fl) (K). Tree to 45 m tall, DBH to 56 cm; trunk often fluted at base; pneumatophores (except at drier sites) arising from below soil 0.5-1.5 m from base, ascending up to 0.7 m, then recurving to reenter soil, sometimes emerging again farther away, extending up to 8 m from tree base; bark white, smooth on younger trees and branches but papery and flaking, revealing red, papery and flaking bark 210 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 above height of 3 m or more on boles of many older trees; hoop marks prom- inent even on lower portions of bole; slash on larger trees revealing phloem rays separated from surface by 2 mm or more of woody matrix, phloem-ray apices 0.5-1.5 mm apart. Twigs with bark white, often having small red spots throughout, youngest growth glabrous (39%) to slightly tomentose (61%; n = 31). Leaves with young petiole 3-8 (5.8, 1.4, 31) mm long, glabrous to moderately tomentose abaxially; lamina narrowly elliptic to elliptic, 9-22 (14.3, 3.3, 31) x 2.5-7 (4.4, 1.1, 31) cm, the apex acuminate to apiculate, 0.3-2.1 (0.9, 0.4, 31) cm long, the base rounded to acute, the secondary veins 13 to 39 (21.5, 6.2, 31) per side, occasionally looping more or less regularly to form weak intramarginal vein; leaves when fresh weakly conduplicate with abaxial surface bright white. Inflorescences borne on branches below leaves, arising predominantly from tubercles, 2 to 15 per tubercle; peduncle extremely re- duced, 0.2-1.1 (0.5, 0.3, 23) mm long, usually with only | terminal flower developing, lateral floral buds and developing flowers absent; bracts | to 3, irregularly oriented with respect to branch axis. Flowers with pedicel 7-29 (18.1, 6, 23) mm long, 0.2-0.8 (0.6, 0.16, 23) mm wide at base and 0.6-1.4 (0.9, 0.19, 23) mm at apex, glabrous (90%) to sparsely pubescent (10%; n = 25); bract sometimes (48%; m = 23) present, 4-16 (8.3, 3.2, 11) mm above articulation. Sepals triangular, 0.6-1.9 (1, 0.28, 23) mm long, abaxially mod- erately (22%) to densely (78%; n = 23) tomentose, infrequently caducous at floral maturity. Petals linear (91%) to basally clawed with distal portion linear (9%; n = 23); outer petals 7-18 (10.4, 3.2, 23) x 1-4.5 (2.2, 1, 23) mm, with the pubescence beginning adaxially on basal (83%) or middle (17%; n = 23) ,, sparse throughout (9%), moderate tl ghout (21%), dense throughout (52%), or sparse throughout except abaxially at base where dense (17%; n = 23), the midvein generally not visible abaxially; inner petals 6-18 (10.9, 3.5, 23) x 0.9-3.5 (1.8, 0.7, 23) mm; all petals yellow when stigmas become receptive, very fragrant (pleasant-sickly sweet; similar to Cananga), turning purplish sepia when anthers dehisce. Torus 0.2—0.8 (0.47, 0.18, 23) x 1.1-2 (1.5, 0.26, 23) mm, gynoecial portion concave (4.4%), flat (56.5%), or convex (39.1%; n = 23). Androecium 2-3 (2.5, 0.3, 23) x 0.5-1.4 (1.1, 0.2, 23) mm; stamens 15 to 31 (23.8, 3.6, 23), 0.5-1 (0.78, 0.12, 23) x 0.7-0.9 (0.77, 0.08, 23) mm. Gynoecium 1-2.8 (1.9, 0.4, 23) mm across; carpels 10 to 25 (17.7, 3.1, 23), 0.6-1 (0.78, 0.14, 23) mm long, glabrous (100%; n = 23); stigmas elongate, clavate, 0.4-1.1 (0.6, 0.18, 23) mm long, those of outermost ring reflexed at least 30° out from floral axis. Immature carpels green, turning red, finally blackish purple, the pericarp then soft and fleshy (juice laden), with sweet to bitter-sweet taste. Mature, dry fruits with pedicel 8-35 (22.8, 6.4, 33) mm long, 1.1-3.2 (1.8, 0.56, 33) mm in diameter at apex, glabrous (97%) or sparsely pubescent (3%; n = 33); stamen scars < 40 (100%; n = 21); carpels with the stalk 3-16 (7.9, 2.7, 31) x 1.2-2.7 (1.9, 0.39, 31) mm, the seed-bearing portion globose, 9-25 (15, 3.4, 32) x 11-25 (15.4, 3.3, 32) mm, not ridged, rounded at apex; fruit wall < 1 mm thick, deep red to black. Placentation ventral at maturity. Seed(s) | (or 2) per mature carpel, biconvex-lenticular, with circum- ferential groove (this occasionally within slightly raised area), the aril scar lacking, the testa deep brown to black (often green when fresh), smooth to 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 211 20°N i : pin ‘ We 1k 900E 155°E 18°S Map 2. Distribution of Polyalthia glauca. finely pitted, often irregularly indented in both fresh and dried seeds; rumi- nation acicular. DisTRiBUuTION. Andaman Islands, southern Thailand, peninsular Malaysia, Su- matra, Borneo, New Guinea, Admiralty Islands (Map 2) SELECTED SPECIMENS SEEN. India. ANDAMAN ISLANDs: sine loco, Parkinson 765 (FL) (x), 794 (FL) (k); Bom-ling-la, Parkinson 943 (FR) (kK). Thailand: Khao Tha Phet F.R., Surat Thani, 40 m alt., Rogstad 841 (fr) (A, BKF). Malaysia. TRENGGANU: Kiat, Sungei Nerus, Kuala Trengganu, SFN 40915 (FR) (KEP, L, SING). PERAK: Changkat Jong F.R., 100 m . me 14327 (FR) (A, K, KEP, L), FRI 14438 (FR) (A, KEP, Ls Kuala Lompat Game Reserve, Kuala Krau, FRI 25108 (FR) (KEP). NEGRI SEMBILAN: Pasoh F.R., n i Rogstad 939, 944, 945, 966 (all FL) (all a); Perhutian Tinggi(?), Ridley(?) 10030 (fr) (x): Pasoh F.R., FRI 27582 (FL+ fr) (Fri). JoHORE: Kuala Sedili, Kadim & Noor 195 (fr) (A, SING); Sungai Sedili, Mawai, SFN 29274 (FR) (A, SAR, SING); Ayer Hitam F.R., 50 m alt., KEP 98248 (A, KEP). SARAWAK: Semengoh Arboretum, Kuching, 15 m alt., Brain anak Tada s.n. (FR) (A, KEP, L, SAR); Ulu Mayeng, Kakus, 60 m alt., S 2/710 (FR) (SAR, of Ranau, 472 m alt., SAN 28987 (FL) (L, SAN): Biah trail, Sapong Estate, Tenom Distr., 300 m alt., SAN 50232 [Sadau] (FR) (sANn); Crocker Range F.R., Rayoh, Tenom Distr., SAN 78396 (FR) (KEP, SAN, SAR); Gum-Gum F.R., ier mi N of Sandakan, 140 m alt. SAN 90945 (FL) (SAN); Sogo-Sogo, 11 mi from K. Tongod, SAN 91/2] (FR) (SAN). Singapore: Mandai Road, SFN 34448 (FR) (KEP, SING); Mondai Road, SFN 37120 (FR) 212 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 (KEP, SING). Philippines. Luzon: Baler, Principe Prov., Merrill 1031 (FR) (PNH, US). Camarines Prov., FB 10475 [Curran] (FR) (PNH, Us); Nueva Vizcaya, Luzon Prov., FB 28532 (FR) (PNH, Us); Casiguran Tayabas Prov., BS 45271 (FR) (A, K, PNH, US). SIBUYAN: Magellanes (Mt. Giting-Giting), Capiz Prov., Elmer 12449 (FR) (A, Us). MINDANAO: Mt. Hamiguitan, Davao Prov., 600 m alt., Edafio 1688 (FL) (A, PNH), Butuan Subprov., FB 20757 (FR) (pNH, US). Indonesia. SUMATRA AND ADJACENT ISLANDS: Palembang, 20 m alt., Grashoff 820 (FR) (Bo, L); Tg. Djawidjawi, P. Mendol, Selatpandjang, 5 m alt., bb 21461 (FR) (A, Bo, L); Belimbing, 6 m alt., bb 28490 (FR) (Bo, L); Djepara, NE Lampung, 20 m alt., bb 35755 (FR) (Bo, L). KALIMANTAN: Bangarmassing, Motley 932 (FL) (K); Tdg. Bangko region, near mouth of Mahakam R., 20 m alt., Kostermans 7193 (FR) (A, Bo, L); Nunukan Is. (northern part), Kostermans 8954 (FL+fr) (A, Bo, K); Sg. Mentawir region, near Mentawir Village, Balikpapan Distr., ra ale 10022 (FL) (Bo, KEP). JAVA AND ADJACENT ISLANDS: Udjung Kulon Reserve, Mt. Pajung, 300 m alt., Kostermans c. S. (UNESCO) 165 (FR) (Bo, L). CELEBES: En Ond Malili, ‘Col/2. -446 (FL+FR) (A, Bo); Malili, Cel./2.-491 (FR) (Bo, K). MoLuccas: G. Permatang, Sangowo R., Morotai, Kos- termans 927 (f1+FR) (A, BO, LAE); Tiloppe, Weda, 25 m alt., bb 24853 (FR) (A, BO, L); Morotai, Subdistr. Tobelo Totodokoe, 30 m alt., bb 33737 (FR) (A, Bo, K), bb 33750 (FR) (Bo, lee IRIAN Jaya: Sorong, Fakal, bank of Kasim R., Misool, Pleyte 859 (FL+ fr) (A, NG); Subdistr. Ransiki, Meos Waar Is., 5 m alt., BW 1214 (FR) (4, SING); Hollandia aa 40 malt., BW 4817 (ER) (A, LAE); Radjah Ampat, Lupintol Village, Waigeo Is., 40 m alt., Van Royen 5481 (FR) (L, stnc); Adi Is., Fak-Fak Div., 25 m alt., BW 10173 (FL+fr) (A, Boswezen), sine loco, Zippiel [Zippelius] s.n. (FL) (A, L). Papua New Guinea and adjacent islands. WESTERN Province: Agu R. branch of middle Fly R., near Mipan Village, 30 m alt., Pullen 7394 (FL+ fr) (A, CANB, L); Lake Daviumbu, middle Fly R., Brass 7622 (FL) (A). MANus Province: above Derimbat Village, Admiralty Islands, 100 m alt., LAE 59246 (FR) (A, LAE, US). REGIONAL NAMES. ‘“‘Lanotan,” Ma language, Davao Prov., Philippines, G. E. Edafio 1688. ““Lulusan sowong,” language?, Ranau Distr., Sabah, SAN 28987. ‘“Manitan,” language?, Selangor, peninsular Malaysia, KEP 28833. ““Mempi- sang” (a general name for several Annonaceae), Malay language, SE Kelantan, peninsular Malaysia, FRI 7038. ““Minakoe,” Manikiong language, Meos Waar Is., New Guinea, BW 1214. “‘Paroe,” Selogof language, Waigeo Is., New Guinea, Van Royen 548]. “Pisang-pisang” (a general name for several Annonaceae), Malay language, Tawau, Sabah, SAN 18520. “‘Saselo,” language?, Morotai, Moluccas, bb 33737. “Songyu” (a general name for several Annonaceae), (local?) Thai language, Surat Thani, Thailand, Rogstad 841 Locat use. Polyalthia glauca is thought to be the best source of firewood in the northern region of Sarawak (Motley 932). The wood of larger P. glauca trees may also be used for crate and toy construction in peninsular Malaysia (Timber Utilization Chart, Forest Research Institute, Kepong, Selangor, Ma- laysia). EcoLocy. Specimens have been collected from sea level to 915 m alt. Of species in the Polyalthia hypoleuca complex, P. glauca has the widest geographic range. Although occasional specimens of this species are marked as collected from hillsides, the bulk of the herbarium notations, as well as my own observations, indicate that this species most frequently grows in saturated soils along streams, at seasonally or permanently inundated sites, or in peat swamps. It can be quite common (eight to 15 mature individuals/ha) in such habitats. Exceptions to this pattern can be found in drier regions at the boundaries of the species range (e.g., Thailand; Rogstad 841). Mature trees of this species occupy the middle 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 213 to upper canopy. At Pasoh F.R., peninsular Malaysia, flowering can occur, with varying intensity, at least twice per year. Each period is immediately preceded by the initiation of growth of new branches and leaves and may include four (or more?) separate ten- to 14-day waves of flowering. COMMENTS. I have been unable to locate the type specimen for Polyalthia glauca at A, B, BO, F, GH, K, KEP, L, MO, P, SAN, SAR, SING, U, US, OF WRSL, and neotyp- ification may be = A type specimen for P. merrittii has also not been found at A, B, BO, F, GH, K, KEP, L, LY, MO, P, PNH, SING, U, UC, US, OF WRSL. Miquel’s description of Cun hypoleuca, and the type specimen noted for this name in the synonymy, are both beyond doubt Polyalthia glauca. Hooker and Thomson (1872), in describing Polyalthia hypoleuca, did not men- tion Guatteria hypoleuca anywhere in the text, but they did compare their newly described species to ‘“‘Unona hypoglauca Miq.” This latter name has not been published elsewhere. A syntype of P. glauca (at kK) is labeled Unona hypoglauca in an unidentified hand, and Guatteria hypoglauca by Miquel. This specimen must be responsible for the reference of Hooker and Thomson above and is thus the probable source for the nomen nudum Individuals of Polyalthia glauca are often easy to find i in poorly drained areas due to the bright red bark that develops on the upper bole of many, but not all, trees (the flaking bark of this species makes it a difficult subject for research because it 1s often impossible to climb the slippery trunks) and the pneuma- tophores that may emerge from the soil within a meter or so of the base of the bole, then curve back to and reenter the soil, often again emerging and reentering the soil at a greater distance. A pneumatophore system of one tree may extend up to 8 m from the base of the tree, and because the older pneumatophores have the white bark characteristic of the P. hypoleuca complex, they are easily spotted in the field. The extension growth of emergent pneumatophores has an interesting pattern of periodicity marked by annular scars (see FIGURE 6D). Within the P. hypoleuca complex, such pneumatophores have been found only in P. glauca individuals growing at poorly drained sites. Sinclair (1955) implicitly recognized Polyalthia parkinsonii, although he did not treat it in detail since he thought it occurred outside of the geographic range under consideration. He did suggest, however, that the entity of the P. Aypoleuca complex found in Burma and the Andaman Islands by Kurz (1874) was actually P. parkinsonii, rather than P. sumatrana, as the latter author believed. Since Kurz did not list the specimens on which he based his report (and I have been unable to find any specimens of P. sumatrana collected in Burma or the An- damans), it is likely that he was dealing with specimens more properly assigned to P. glauca. This conclusion finds support in the fact that to date, only P. glauca is known from Thailand (Rogstad 841), thus being the only species of the complex collected to the north and west of Sumatra and peninsular Ma- laysia. Within the Polyalthia hypoleuca complex, specimens of P. glauca are most likely to be confused with P. ovalifolia (restricted to Borneo) or P. discolor (restricted to Papua New Guinea); see the discussion under the latter two species. The curious irregular indentations of the seed surface often seen in herbarium 214 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 specimens of this species can also be observed in fresh material (see FIGURE 8D). All of the more than 230 seeds germinated at Pasoh Forest Reserve had them. Such indentations were well developed in all herbarium specimens with mature fruit collected from peninsular Malaysia, Sumatra, and the Andamans, but they are less prominent in collections from Borneo, present in only two of the four collections from the Philippines, and completely lacking in specimens from New Guinea and the Admiralty Islands. In seeds where the indentations are lacking, the rumination needles characteristic of the Polyalthia hypoleuca complex occasionally become slightly broadened near the longitudinal circum- ferential groove. 2. Polyalthia ovalifolia S. H. Rogstad, sp. nov. FiGureE 3A. P. glauca affinis, sed in ramulis novellis et in pedicellis indumento moderate vel dense tomentoso praeditis (haud moderate vel dense tomentoso in P. glauca); foliis grandioribus (19-26 x 7-11.3 cm versus 9-22 x 2.6-7 cm in P. glauca), latioribis (foliis latitudo/longitudo ratio plerumque 0.34—0.44 versus 0.24—0.34 in P. glauca), et magis coriaceis, differt. Tree to 25 m tall, DBH to 55 cm; trunk often fluted at base; pneumatophores absent; bark white, smooth on younger trees and branches, becoming white mixed with yellow when older; hoop marks prominent even on lower portions of bole; slash on larger trees revealing phloem rays separated from surface by 2 mm or more of woody matrix, phloem-ray apices 0.5—-1.5 mm apart. Twigs with bark white to pale yellow-white, youngest growth always with dense, short, rusty tomentum (n = 20). Leaves with young petiole 6-10 (8, 1.2, 12) mm long, moderately to densely tomentose abaxially; lamina elliptic to widely elliptic, 19-26 (22.5, 1.5, 12) x 7-11.3 (9.6, 1.1, 12) cm, the apex acuminate to apiculate, 0.8-1.7 (1.1, 0.3, 12) cm long, the base rounded, the secondary veins 21 to 36 (29.6, 4.7, 12) per side, wavering and not forming strong intra- marginal vein; leaves when fresh flat, with abaxial surface bright white. Inflo- rescences borne on branches below leaves, arising predominantly from tuber- cles, up to 20 per tubercle; peduncle extremely reduced, 0.7—2.1 mm long, usually with only | terminal flower developing, lateral floral buds and devel- oping flowers absent; bracts | to 3, irregularly oriented with respect to branch axis. Flowers with pedicel 10-18 mm long, 0.7-1.2 mm wide at base and |- 1.5 mm at apex, with dense, short, rusty tomentum; bract 1.1-8 mm above articulation. Sepals triangular, 0.9-1.9 mm long, abaxially densely tomentose, infrequently caducous at floral maturity. veal linear; outer petals 7.5-13 x y on basal 4, moderate throughout except abaxially at base where dense, or dense throughout, the midvein gen- erally not visible abaxially; inner ones 8.5-13 x 2-2.2 mm, all petals (and sepals) pale = strongly fragrant (S 12401), deep burgundy-red when dry. Torus 0.5-1.1 1.8-2.5 mm, gynoecial portion convex. Androecium 0.9-2 x 2.1-2.8 mm; stamens 26 to 36, 0.8-1 x 0.7-0.9 mm. Gynoecium 2-2.1 mm across; carpels 20 to 27, 0.7-1.1 mm long, glabrous or with slight tuft of tomentum only on abaxial apex; stigmas elongate, clavate, 0.5-1.1 mm long, those of outermost ring reflexed at least 30° out from floral axis. Observations * _ 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 215 13°N 108°E 155°E Map 3. Di oleae of Polyalthia discolor (dots), P. ovalifolia (solid stars), and P. multinervis (open sta on immature carpels lacking. Mature, dry fruits (only 2 available) with pedicel 16-27 mm long, 2.2-2.4 mm in diameter at apex, with dense, short, rusty tomentum; stamen scars < 40; carpels with the stalk 3-4 x 1.7-1.9 mm, the seed-bearing portion globose, 15-19 x 14-18 mm, smooth, rounded at apex; fruit wall < | mm thick, deep red to black. Placentation ventral at maturity. Seed(s) 1 (or 2?) per mature carpel, biconvex-lenticular, with circumferential groove, the aril scar lacking, the testa deep brown to black, smooth to finely pitted, not irregularly indented; rumination needles occasionally slightly broad- ened. Type. S. and E. Borneo, Sampit, 5 m alt., 09-20-1925, Buwalda [sheets may be marked F. H. Endert] 7797 (fl) (holotype, A; isotype, Bo). DisTRIBUTION. To date collected only from northwestern and southern Borneo (see MAP SELECTED SPECIMENS SEEN. Malaysia. SAaRAWAK: Maludan Peninsula, Tj. Keranji, Betong Distr., 3 m alt., S 1240] [Anderson] (FL) (Bo, K, SAR); Setapok F.R., Kuching, 15 m alt., S 12967 ae (SAR); Bako Natl. Park, near Kuching, Rogstad 842-845 (st) (A). Brunei: Badas 15 m alt., S 2226 (FL) (FRI, SAR, SING); Badas State Land, Kuala Belait, 13 m ac es Gn (A, KEP, SAN). REGIONAL NAMES. “Dillah,” language?, above Sibu, Sarawak, S 2342. “Dilleh,” language?, Brunei, S 2226. ““Riboei,” language?, Sampit, Kalimantan, Buwalda 216 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 (or Endert?) 7797. ‘Selaut telor,” Malay language(?), Betong Distr., Sarawak, S 12401 [Anderson] Ecotocy. The highest recorded altitude for Polyalthia ovalifolia is 15 m above sea level, but the species is poorly collected. Rather dense populations (up to ten mature individuals/ha) were observed at Bako F.R., Sarawak, where they appeared to be restricted to areas of stunted forest (the closed kerangas (heath) forests of Brunig (1974); canopy at 20-25 m) that grow on sandy soils covered with a spongy root-humus mat (to | m thick). These stunted, closed forests yield black water drainage and are the only habitat where regeneration of this species was found to be occurring at Bako. While this soil formation is remi- niscent of peat, it differs in that it appears on flat or gently sloping hilltops and is well drained. However, P. ovalifolia has also been collected in peat-swamp forests in Sarawak (S 12401 [Anderson], S 12967 [Zen]), although I could not find it in the extensive peat swamp formations adjacent to the Baram River, and foresters working in these swamps (e.g., employees of Forescom Sarawak), when presented with leaves of this species, stated that they had not observed it there although they immediately recognized the leaves of P. hypoleuca, which grows abundantly in that region. ComMENTs. There has been some confusion concerning the collector of the type specimen. Sheets at A and Bo that agree in collection number, date, and specimen material indicate different collectors (E. H. Endert at a, Buwalda at Bo). A comparison of the collecting itineraries (Van Steenis—Kruseman, 1950) indicates that Buwalda was the collector. This species is most likely to be confused with Polyalthia glauca; specimens listed above have usually been identified as P. glauca by various workers, including Sinclair. Polyalthia ovalifolia differs consistently from P. glauca in having moderately to densely tomentose (vs. glabrous) pedicels in flower and fruit, more coriaceous, larger leaves (see FIGURES 3A and 9) that are elliptic to broadly elliptic (vs. narrowly elliptic to elliptic) and usually have a higher width/ length quotient (0.34—0.44, m = 12, vs. 0.24-0.34, n = 31; P < 0.05, unbalanced t-test), and young branches always (vs. rarely) with dense, short, rusty tomen- tum. Although the differences between these two species are not pronounced, they are constant, and no intermediates have yet been found. I have therefore recognized P. ovalifolia as a distinct species, although more material and field observations of both P. ovalifolia and P. glauca in Sarawak and southern Kalimantan are needed. 3. Polyalthia discolor Diels, Bot. Jahrb. Syst. 49: 130, 1913. Type: Papua New Guinea, middle April River, Niksek Village area, 4°42'S, 142°32'E, 100- 150 malt., Sept. 16, 1984, Rogstad 814 (fl+ fr) (neotype, A; 1isoneotypes, K, LAE). Figures 2C, D; 4B; 6B; 7B; 8C; 20C, D. Tree to 35 m tall, DBH to 91 cm; trunk often fluted at base; pneumatophores absent; bark white, smooth on younger trees and branches, with very fine dots of red giving main bole reddish tinge, flaking papery patches scattered on older bark; hoop marks prominent even on lower portions of bole; slash on larger trees revealing phloem rays separated from surface by 2 mm or more of woody matrix, phloem-ray apices 0.5—-1.5 mm apart. Twigs with bark white to pale 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 217 yellow-white, youngest stem growth glavrous (50%) or with moderately dense, short, rusty tomentum (50%; 7 = 20). Leaves with young petiole 3—9 (6.4, 1.7, 14) mm long, glabrous (75%) to moderately tomentose (25%; n = 20) abaxially; lamina narrowly elliptic to elliptic, 10.2-15.8 (13.1, 1.3, 14) « 3.3-5.1 (3.9, 0.55, 14) cm, the apex acuminate to apiculate, 0.4-1.1 (0.75, 0.2, 14) cm long; the base rounded to acute, the secondary veins 17 to 28 (23.3, 3.5, 14) per side, occasionally looping more or less regularly to form weak intramarginal vein; leaves when fresh slightly to strongly conduplicate, with abaxial surface bright white. Inflorescences borne on branches below leaves, arising predom- inantly from tubercles, up to 10 per tubercle; peduncle ca. 1.1-11 mm long, with up to 3 lateral flowers developing or with resting buds; bracts | to 7 (often subtending resting or developing bud), distichous, but irregularly oriented with respect to branch axis on different inflorescences of single tubercle. Flowers with pedicel 18-49 (29.2, 8.3, 11) mm long, 0.8-1.2 (0.94, 0.1, 11) mm wide at base and 1.1—2 (1.6, 0.3, 11) mm at apex, glabrous (75%; n = 12), occasionally with sparse, short, rusty tomentum; bract always present (7 = 11), 2-18 (8.8, 4.9, 11) mm above articulation. Sepals triangular, 1.3-2.2 (1.7, 0.36, 11) mm long, moderately tomentose abaxially, infrequently caducous at floral maturity. Petals narrowly elliptic to elliptic to obovate, occasionally with constriction or narrowing in basal '4 or with distal portion slightly flared (or both); outer petals 16-32 (20, 5.7, 11) x 4.5-7.3 (5.8, 0.83, 11) mm, the pubescence always beginning adaxially on basal '4 (n = 20), moderate throughout except abaxially at base where dense, or dense throughout, the midvein often visible; inner petals 14-38 (22.1, 8.8, 11) x 4-9 (5.4, 1.5, 11) mm; all petals red-purple basally, yellow distally, and noticeably sweet smelling when stigmas become receptive, turning purplish sepia, musky, and sickly sweet when anthers dehisce, rusty to deep burgundy-red when dry. Torus 0.8-1.4 (1, 0.16, 11) x 2.3-3.3 (2.8, 0.32, 11) mm, gynoecial portion flat (100%; » = 3) or convex (100%; 7 = 8). Androecium 1|-2.5 (1.9, 0.4, 11) x 3.6—-5.1 (4.5, 0.4, 11) mm; stamens 77 to 119 (95.3, 14, 11), 0.9-1.3 x 0.6-0.9 mm (” = 11). Gynoecium 2.3-4 (3.1, 0.5, 11) mm across; carpels 31 to 72 (48.2, 13.6, 11), 1-1.4 (1.14, 0.12, 11) mm long, glabrous (90%) or with slight tuft of tomentum only on abaxial apex (10%; m = 20), the outermost ones slightly incurved, distal portion of abaxial surface exposed above androecium, with exposed region slightly swollen and often slightly darker; stigmas globose, 0.2-0.7 mm long (0.45, 0.13, 11) mm in diameter, those of outermost ring not reflexed away from floral axis. Immature carpels green, turning red, finally blackish purple, the pericarp then soft and fleshy (juice laden), with sweet to bitter-sweet taste and often chlo- rinelike odor. Mature, dry fruits with pedicel 24-50 (34.6, 7.2, 14) mm long, 2-4.5 (3, 0.8, 14) mm in diameter at apex, glabrous (100%; n = 14); stamen scars > 40 (100%; mn = 14); carpels with the stalk 10-24 (17.7, 4.5, 14) x 1.5- 2.7 (1.9, 0.36, 14) mm, the seed-bearing portion globose, 15-18 (16.5, 1.3, 14) x 14-19 (16.6, 1.4, 14) mm, smooth, rounded at apex, fruit wall < 1 mm thick, deep red to black. Placentation ventral at maturity. Seed(s) 1 (or 2) per mature carpel, biconvex-lenticular, with circumferential groove, the aril scar lacking, the testa deep brown to black (often greenish when fresh), regular, smooth to finely pitted; rumination acicular, with needles occasionally slightly broadened. 218 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 DisTRIBUTION. New Guinea, and probably also the immediately adjacent 1s- lands (see MApP 3). SELECTED SPECIMENS SEEN. Indonesia. [RIAN JAYA: Warsamsen R.., ae E of Sorong, 60 m alt., BW 2980 (FR) (A, SING); Kebar Valley, Div. W. New Guinea, 450 m alt., BW 7193 (FR) (Bo); Woda, Japen Is., Div. Geelvinkbaai, 50 malt., BW "11207 (FR) (x), BW airstrip, middle April R., 4°42’'S, 142°32’E, 50-150 m alt., Rogstad 814 (FL+fr), 832 (FL+ fr), 833 (FL+ FR) (all A, LAE, K); Sepik R. near Yellow R., NGF 3914 (fr) (A, NGF); [Malu region,] Sepik R. region, Ledermann 6814 (FR) (b, K); Sepik region, Ledermann 10409 (FL) (B, K, L); Koitaki, 460 m alt., Carr 1215] (FL) (SING). WESTERN HIGHLANDS Province: near Rut E.L.G., Jimi Valley, Mt. Hagen, 450 m alt., NGF 38825 (FR) (A, BO, L, LAE). Morose Province: Aluki Village oes km E of Lae), 60 m alt., Rogstad 782 (FL) (A, LAE, K); 7 mi N of Lae, near Butibum R., 65 m alt., Hartley 11863 (FR) (a, CANB); Gnalungumbum, 50 m alt., NGF pl (FL) (L, LAE). NORTHERN PROVINCE: Dobodura area, 125 m alt., NGF 2085 (FR) (L, LAE); between Divinikoari and Horata villages, 50 m alt., Hoogland 3529 (FR) (A, CANB, US); along Girua R., | km NW of Anonda airstrip, 60 m alt., Hoogland 3776 (FR) (A, csiro); Mayu Camp I, jet. of Ugat and Mayu rivers, Mt. Suckling, 305 m alt., LAE 56049 (FR) (A, K, L, LAE, US). CENTRAL Province: Mori R., Cape Rodney, Abau Subdistr., 65 m alt., NGF 38599 (FR) (A, L, LAE); Mori R., Abau Subdistr., 250 m alt., NGF 47840 (FL) (A, BO, L, LAE). WESTERN Province: Orovill Camp, Fly R., Brass 7411] (FR) (A, L). NEW BRITAIN: 7 mi SE of Benim Village, Wariai Subdistr., 300 m alt., NGF 27395 (FL) (L, LAE). REGIONAL NAMES. “Adidionga,” Orokaiva language (Mumuni), Northern Div., Papua New Guinea, Hoogland 3529. “Asiam,” Biak language, Japen Is., Inan Jaya, BW 11210. “Atim,” Bukawa language of the Suling clan, just NE of Lae, Papua New Guinea, Rogstad 782. ‘‘Fawss” or “‘pfahss,”” Niksek language, April R., Papua New Guinea, Rogstad 814. ““Nimotiet,” Kebar language, Div. W New Guinea, Irian Jaya, BW 7193. “Oeloem,” Mooi language, Warsamsen R.., Irian Jaya, BW 2980. “Samoeben,” Biak language, Japen Is., Iman Jaya, BW 11207 Loca use. The timber of Polyalthia discolor is used for house construction in the April River area (Rogstad 833), and for building houses and furniture in the Lae region (Rogstad 782). EcoLoGy. The known altitudinal range is from 35 to 450 m above sea level. Of the 11 collections with details concerning habitat, eight were found on slopes and three at flat sites, but not on constantly inundated soils or those on which a thick root-humus mat had developed, or in peat swamps. This was confirmed by observations on members of a Polyalthia discolor population at April River (tributary to the Sepik), Papua New Guinea (Rogstad, 1n prep. a). Regeneration by seedlings, which was often pronounced, also exhibited this pattern. Flowering begins when individuals reach the middle canopy but is generally prolific only when they achieve a place in the sun in the upper canopy. This species has the slowest-developing flowers of any species in the group, and individuals can have flowers reaching anthesis more or less continuously over periods of up to 1.5 months. Although the other species of the P. hypoleuca complex (except P. ovalifolia, for which no observations of pollinators have 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 219 yet been made) usually attract only one or two pollinators in any numbers, P. discolor flowers consistently drew several species of insects as potential polli- nators. Pigeons and cassowaries have been observed eating the fruit (Pullen 8123). The Annonaceae are generally known for their weak and somewhat flexible wood, a factor that may have contributed to the observation by Hoogland (3529, Northern Div., Papua New Guinea) that Polyalthia discolor was the “most frequent standing tree in a patch of rainforest damaged by a cyclone.” ComMENTS. The type as designated by Diels, Schlechter 18306, was destroyed at Berlin, and I have been unable to locate any duplicates at A, B, BO, F, GH, K, KEP, L, MO, P, SAN, SAR, SING, U, or US. I have also been unable to find any collections of this species from the Ramu region of Papua New Guinea. Diels’s description very clearly refers to the species recognized here. I have chosen as a neotype a collection that both matches the description by Diels and is rep- resentative of this group of specimens. Identification of herbarium specimens of this species has caused much con- fusion due to their similarity in both flowers and fruits to Polyalthia glauca, also from the New Guinea region. This superficial resemblance caused Sinclair to comment (Carr 12151), “This is what Diels calls discolor but it is probably only a variety of glauca.” However, as demonstrated in the morphometric analyses, P. discolor consistently has several unique features that distinguish it with a clarity commensurate with species status in this group, and thus I recognize it as such here. Although Polyalthia discolor and P. glauca are very similar, the flowers of the former have more than 50 stamens (vs. less than 45 in P. glauca), more (vs. always fewer) than 27 carpels, and globose (vs. elongate) stigmas. These two species are much more difficult to distinguish in fruit, as is attested to by the fact that every herbarium specimen of P. discolor with fruit that I have seen has been misidentified as P. glauca. However, although the seed-bearing portions of the mature carpels are globose in both species, the fruit of P. discolor is generally larger than that of P. glauca. While the means for pedicel length and (distal) width and mature-carpel stalk length differ significantly (P < 0.05, unbalanced t-test, n = 14 for P. discolor, n = 31 for P. glauca) for the two species, there is enough overlap to cause confusion. However, specimens (other than immature ones of P. discolor) on the upper or lower ends of these scales can be identified. At any stage of development, the best means of separating fruits of these two New Guinean species is to count the scars left by the abscised stamens. Unfortunately, these are sometimes obscured by the expansion of the torus that accompanies fruit growth. Differences in stamen and carpel number and scars between these two species can be seen in FIGURE 20. Polyalthia discolor and P. glauca are to date known to overlap geographically only in the middle Fly River region and possibly in north-central New Guinea (Rogstad, 1986), although with more comprehensive exploration, the area of overlap is likely to expand, especially since P. glauca has been found on Manus. Further study of the distribution and possible overlap of these species and comparative investigations of their autecologies are of critical importance to understanding the systematics and evolution of this group. 220 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 4. Polyalthia sumatrana (Miq.), Kurz, J. Asiat. Soc. Bengal, Pt. 2, Nat. Hist. 43: 53. 1874, pro parte basionymica tantum; Guatteria sumatrana Miq. Fl. Ned. Ind., Eerste Biv. 380. 1861; Monoon sumatranum (Miq.) Mia. Ann. Mus. Bot. Lugduno-Batavum 2: 19. 1865. Type: [Sumatra,] Pria- man, Diepenhorst H.B. 2342 (lectotype, here selected, U: mature fruit and mature flower fragment). Figures 3B; 4A, D; 8B, D; 18A, B; 21C. Tree to 24 m tall, DBH 4—25 cm; trunk rarely fluted at base; pneumatophores absent; bark white on younger trees and branches, remaining white to pale yellowish white on older trees; hoop marks not prominent on lower portion of bole; slash on larger trees revealing phloem rays reaching surface, phloem- ray apices 0.2-0.7 mm apart. Twigs with bark white to pale yellow-white, youngest growth glabrous (88%) or moderately tomentose (12%; n = 71). Leaves with young petiole 3-12 (6, 1.5, 105) mm long, glabrous (80%) to moderately tomentose (20%; = 71) abaxially; lamina narrowly elliptic to elliptic, 9.2-28 (15.5, 3.2, 106) x 2.5-9.5 (4.5, 1.1, 106) cm, the apex acuminate to apiculate, 0.1-1.6 (0.9, 0.3, 106) cm long, the base rounded to acute, the secondary veins 19 to 80 (34.8, 10.2, 105) per side, not forming regular intramarginal vein; leaves when fresh flat with abaxial surface dull white (white mixed with gray or brown). Inflorescences in foliate axils, up to 4 per axil, and also borne on branches below leaves, tubercles lacking or poorly developed; peduncle ex- tremely reduced, 0.6-2.5 (1.4, 0.4, 65) mm long, usually with only | terminal flower developing, lateral floral buds and developing flowers absent; bracts 1 to 3, oriented perpendicular to branch axis when only | flower per axil. Flowers with pedicel 7-25.4 (15.8, 5.4, 65) mm long, 0.6—1.8 (1.2, 0.3, 65) mm wide at base and 0.8-3 (1.9, 0.4, 65) mm at apex, glabrous (78%) to moderately pubescent (22%; n = 65); pedicel bract sometimes present (43%; n = 65), 4-8 mm above articulation. Sepals triangular, 0.7—3 (1.5, 0.48, 40) mm long, mod- erately (43%) to very densely (57%; n = 40) tomentose abaxially, often caducous (38%; n = 65) at floral maturity. Petals linear (97%) to basally clawed with obovate or flared distal portion (3%; n = 65); outer petals 7.5-52 (28.4, 10.8, 65) x 1.5-6 (3, 0.8, 65) mm, the pubescence beginning adaxially on basal (11%), middle (49%), or distal (40%; n = 65) 14, sparse throughout (62%), moderate throughout (11%), dense throughout (9%), or sparse throughout ex- cept abaxially at base where dense (18%; n = 65), the midvein generally not visible abaxially; inner petals 9-56 (30.8, 11.4, 65) x 1.6-5.7 (2.9, 0.8, 65) mm; all petals turning yellow, or occasionally yellow-white (S 3020), yellow- brown (S 3057), maroon ([SAN] 3626), cream with pale pink tinge (Sinclair et al. 9313) or basal pink flush (SAN 2/222), red (SAN 21459), or gray (SAN 56203) at maturity, ““‘astringently fragrant” (S 38009; only this collection notes odor), tan to deep burgundy-red when dry. Torus 0.9-2.2 (1.4, 0.38, 62) x 1.3-4.4 (2.6, 0.67, 62) mm, gynoecial portion concave (25%), flat (63%), or convex (12%; n = 65). Androecium 1|.1—3.1 (2.7, 0.4, 65) x 3-6.2 (5, 0.75, 65) mm, stamens 80 to 232 (160.4, 33.6, 65), 0.8-1.7 (1.1, 0.15, 65) x 0.4-1.1 (0.78, 0.12, 65) mm. Gynoecium 0.9-4.3 (2.2, 0.65, 65) mm across; carpels 5 to 49 (24.8, 8.8, 65), 0.8-2 (1.2, 0.22, 65) mm long, with small distal, abaxial patch of tomentum (65%) or moderately to densely tomentose (35%; n = 65); stigmas globose (rarely slightly elongate), 0.2—0.8 (0.3, 0.1, 65) mm in diameter, 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 221 150ON 93°E 122°E Map 4. Distribution of Polyalthia sumatrana. those of outermost ring not reflexed away from floral axis. Immature carpels firm, green with irregular red spots or streaks appearing 2-4 weeks before full maturity (such streaking diagnostic for this species within the P. Aypoleuca complex), finally turning deep purple to black, the pericarp then soft and fleshy (juice laden) with sweet to bitter-sweet (pers. obs., Pasoh) or cinnamonlike (W Kalimantan, Schut K. 23) taste or fragrant odor. Mature, dry fruits with pedicel 8-43 (21.7, 6.3, 107) mm long, 1.8—-4.8 (3.2, 0.64, 106) mm wide at apex, glabrous (100%; m = 108); stamen scars > 40 (100%; = 19); carpels with the stalk 5-38 (15.1, 5.4, 105) x 1.3-4.4 (2.7, 0.64, 105) mm, the seed-bearing portion ellipsoid, 13-30 (22, 3.8, 106) x 9-19 (12.6, 2.3, 105) mm, sometimes longitudinally ridged (39%; n = 108), tapered (89%) or blunt or rounded (11%; n = 108) at apex; fruit wall > | mm thick, burgundy to deep red-black. Placentation basal at maturity. Seed(s) | (or 2) per mature carpel, ellipsoid, with circumferential ridge, the aril scar present, the testa deep brown, finely to moderately pitted, not irregularly indented; rumination needles very fine. DISTRIBUTION. From peninsular Malaysia to Sumatra, Borneo, and immedi- ately adjacent islands (see Map 4). SELECTED SPECIMENS SEEN. Malaysia. PERAK: Chior F.R., Sungai Plus-Sungai Siput, 230 malt., FRI 5769 (FR) (KEP); sine loco, Dr. King’s collector (Lanit?) 655 1 (FL) (k); Gunong Bubu, Ria Trong, 400 m alt., FRI 11907 (FR) (KEP, L); Maxwell’s Hill, 180 m alt., SFN 38828 (FL) (kK, SING); Piah F. R., K. Kangsar, Rahim Ismail KEPFN 99836 (FR) (KEP). TRENGGANU: Bukit Bauk F.R., 250 m alt., FR/J 2362 (FR) (KEP); near Geram Galong, Ulu Sungai, 65 m alt., FRJ 8357 (FR) (KEP); Ulu S. Trengan, near K. Petang, Ulu 222 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Trengganu, 275 m alt., FRI 8438 (FR) (A, KEP); Sungai Loh, near K. Datok, 130 m alt., FRI 10690 (FR) (A, KEP); Sekayu F.R., FRI 11854 (FR) (KEP, L); along Sungai Pelong, FRI 14885 (fl+FR) (A, KEP, SAR); Bukit Bauk VJR, Dungun, FR/ 25044 (FR) (KEP); Ulu Brang, 180 m alt., SFN 33740 (FR) (A, KEP, SING); Jarangau S.L., Dungun Distr., KEP 76674 (FR) (KEP). KELANTAN: Kuala Lebir Kechil, Sungai Lebir, FRI 7075 (FR) (A, KEP, L); Sungai Lebir, 2 mi E of Kuala Aring, FR/ 7/05 (FR) (KEP). SELANGOR AND KUALA LumpuR FEDERAL District: Gombak Reserve, Murdock & Hashim 114 (FR) (KEP); Ulu Langat, Bukit Tangkol, K. Pansom, KL 1360 (FR) (KEP), KL 1375 (fl+ FR) (A, Kep); Bukit nga [Peningkot?], KL 1494 (FR) (A, KEP); Sungai Buloh Reserve, [4bu] CF Field ues 3335 (FL) (k, KEP); Ginting Simpoh, Strugnell(?) [KE P?] 11225 (FR) (KEP); Gading ., Ulu Selangor, 370 m, FR/J 1/229(FR)( KeEp, L); Ulu Gombak F.R., Symington [KEP] rary (FR) (KEP); Bukit Lagong, FRJ 26595 (FR) (Kep); Sungai Buloh, Walton [KEP] 29737 (FR) (KEP); Bukit Lajong F.R., 450 m alt., Sow [KEP] 65/47 (FR) (ker); Ulu Gombak F.R., K.L. Distr., KEP 94038 (FR) (4, KEP), KEP 98269 (FR) (KEP), KEP 99610 (FR) (A, KEP). PAHANG: Raub, Sungai Sempam, 920 m alt., Soepadmo 725 (FL+ fr) (KEP, KLU); Ulu Sungai Sepia, near Kuala Aur, Shah & Noor 1904 (FR) (A, KEP, SING); Taman Negara, Bukit Terom, near K. Keniyum, 150 m alt., FR/J 8538 (FL) (keEp); Panching Forest, ca. 15 km N of Panching, Kuantan, 130 m alt., Ogata [KEP?| 10432 (FR) (KEP); Raub, 365 m alt., FRI 14689 (FL) (KEP); Bukit Saup, Jaamat [KEP?] 16522 (FL) (KEP); Chini F.R., S of Tasek Chini, FRI 1727] (FR) (KEP, L); Lesong F.R., 115 m alt., FRI ee (FR) (A, KEP); Simpang Pertang, Ulu Tembeling, SFN 22/12 (FR) (A, KEP, SING); ong F.R., FRI 23910, 28028 (both FR) (both aie Aur F.R., K. Rompin, 50 m alt., KE P 75883 (FR) (KEP); Krau Game Reserve, G. Benom, 300 m alt., KEP 104427 (FR) (A, KEP). NEGRI SEMBILAN: Pasoh F.R., near Simpang Pee Rogstad 508, 509, 527 (all FL) (all at A); Pasoh F.R., 50 m alt., FRI 23496 (FR) (KEP). JoHORE: road to Sungai Kahang, N to Labis, Shah & Sanusi 2107 (FR) (A, KEP, SING); path to Gunong Blumut, Kluang, 150 m alt., Shah & Sanusi 2139 (FR) (A, KEP, SING); Tg. Sedili Kechil, E Johore coast, 30 m alt., FRI 7688 (FL) (A, KEP); Ulu Endau, Labis F.R., FRI 7869 (FL) (KEP, a FRI 7920 (FL) (kep), KEP 110402 (FL) (k, KEP); Kluang F. R., FRI 8701 (FR) (A, EP); NW Gunong Blumut, 548 m alt., FRJ 8820 (FR) (KEP, L); Labis F.R., FRI 14783 (FL+ FR) (A, KEP), FRI 16073 (FR) (A, KEP); Gunong Pulai F.R., 335 m alt., FRI 17639 (FR) (A, KEP); Tanjong Sedili Kechil, SFN 28070 (FR) (KEP, SING); Sungai Kayu Ara, Mawai-Jemulang Road, SFN 29330 (FR) (KEP, SING); Sungai Sedili, below Mawai, SF'N 36979 (FR) (A, KEP, SING); Bukit Mambai F.R., 50 malt., KEP 76295 (FL) (KEP, L); near Mersing F.R., 30 m alt., KEP 77852 (FR) (KEP). SARAWAK: Gunong Mulu, Baram Distr., 550 m, [Anderson & Keng] K 15 (FR) (A, SAR); Sabal Tapang, Ist Div., 130 m alt., Stevens et al. 174 (FR) (A, SAR?); Gunong Api, Baram Distr., 500 m alt., Chew Wee- lek 1106 (FR) (A), 7/67 (FR) (A, L); Ulu Koyan, Mt. Dulit, [Vth Div., 800-1000 m alt., Richards (native collector) 1918 (FL) (a, kK); near Kuching, Haviland 2218 (FL) (k), 2694 (FL) (sAR); Lambir Hills F.R., Miri Distr., 300 m alt., Dan bin Haji Bakar 3020 (FL) (A, SAR); upper Plieran, R. Kenaban, 560 m alt., S 3501 ( (FR) (SAR, Us); 10 km below Belaga, Kapit Distr., HWIrd Div., 500 m alt., Jacobs 5201, 5357 (both FL+FR) (both L, us); Ulu Pelagus, S. Iran, Kapit(?) Distr., 120 m alt., Smythies 9469 (FL+ fr) (k, SAR); Tatau, Ulu Anap, Sungai Takan, [Vth Div., 215 m alt., S 13763 (FL) (A, kK, KEP, SAR); NE Lambir Hills, near Miri, 30 m alt., S 16439 (FL) (A, SAR); Sabal F.R., Serian (A, L, SAR); Pengiran/Chenaning divide, Ulu Mujong, Balleh, 600 m ‘alt., S 2116] (FR) (L, SAR); S Ukong, Miri, 10 m alt., S 2/403 (FL) (A, K, KEP, L, SAR); Kakus, Ulu Mayeng, S 21705 (FL) (A, SAR); Bukit Dam, Ulu Dapoi, Marudi, 1Vth Div., S 22955 (FR) (4, KEP, SAR); Long Dapoi, Tinjar, Marudi, [Vth Div., 100 m alt., S 22971 (FR) (A, KEP, sar); Bok-Tisam, Bukit Mentagai, Marudi, IVth Div., 65 m alt., S 23099 (FR) (A, SAR); Long Dam, Ulu Tinjar, Marudi, [Vth Div., 115 m alt., S 23373 (FR) (A, KEP, SAR); Ulu Dapoi, Long Kelaby, Tinjar, [Vth Div., S 23464 (FR) (A, KEP, SAR); Bukit ju, Ulu Amp, 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 223 Balingian, IIIrd Div., 65 m alt., S 23623 (FR) (A, SAR); Ulu Sungei Bakong, Miri, 50 m alt., S 24409 (FR) (A, SAR); Lambir Natl. Park, Miri, S 25077 (FL) (sar); Kapit, Balleh, Sg. Mengiong, IIIrd Div., S 29699 (FR) (SAR); Ulu Kakus, Anap, IVth Div., S 29974 S 29979 (both FR) (both A, KEP, SAR); Mata Kuching, Ulu Tinjar, S of Dulit Range, 430 m alt., S 38009 (FL+FR) (k, KEP, SAR), 425 m alt., S 38075 (FR) (KEP, SAR); Ulu Sungai Labau, Lambir Hills Natl. Park, 1Vth Div., 1210 m alt., S 38429 (FR) (KEP, SAR). SABAH: sine loco, Agama(?) 550 (FL) (A, PNH); Kinabatangan, Puasa 1069 (FR) (A, NY); Segaliud, Elopura, Cuadra A 1090 (FR) (A, K, KEP, SAN, US); Marutai, Tawau, 3 m alt., Tandom BNBFD 1786 (FL) (A, Kk); Sepilok F.R. SANA 1969 (FR) (A, KEP, SAN); Table Estate, St. Lucia, Tawau, Kadir A 2103 (FL) (A, KEP, SAN); Gomantong, Elopura, San- dakan, 12 m alt., Cuadra A 2118 (FL) (A, K, KEP, SAN); Marutai, 8 m alt., Maidin 2428 (FR) (kK); 7 m alt., Maidin 3057 (FL) (kK); Betotan [Bettotan?], 25 m alt., Orolfo BNBFD 3197 (FR) (A?, SAN), 30 mi WNW of Tawau, 30 m alt., SAN A 3698 (FR) (A, KEP, fo Bettotan, Sandakan, Puasa BNBFD 4544 (FL) (A, k, us); Apas Road, mi 7, 2 Tandom BNBFD 8806 (FR) (a, K); Sepilok F.R., Sandakan, 50 m alt., Sinclair, Kadim, 15263 (FR) (A, KEP, SAN); Sepilok F.R., Sandakan, 15 m alt., SAN 15488 (FL) (A, K, KEP, L, SAN); Kelumpang Balong, Tawau Distr., 80 m alt., SAN 17337 (FR) (SAN); Lalangot, ca. 1.5 mi from Pangkalan, Tawau, 40 m alt., SAN 18718 (FR) (A, KEP, SAN); Sepilo F.R., Sandakan, SAN 21222 (FL+FR) (kK, KEP, SAN), SAN 21303 (FL) (A, KEP, SAN); Tawao, Elphinstone Prov., B.N.B. Elmer 20494 (FR) (A, us), 20877 (FL) (A, K, US), 21878 (FR) (a); Apas Road, mi 13, Tawau, SAN 21459 (FL) (kK, KEP, SAN); Membalua , Tawau Distr., SAN 22752 (FR) (A, KEP, SAN), SAN 22774 (FR) (KEP, SAN); Bukit Pasil, Lungmanis, Sandakan Distr., SAN 24238 (FR) (SAN); Lahad Datu Distr., Takun Kennedy Bay, 125 m alt., SAN 26065 (FL+FR) (SAN); W Sungai Strun, Tawau Distr. 150 m alt., S4N 26875 (FR) (SAN); Beaufort, Beaufort Distr., 30 m alt., SAN 28115 (FR) (SAN); Lohan, Ranau Distr., 1200 m alt., SAN 28758 (FR) (SAN); Mt. Andrassy, Tawau Distr., 300 m alt., SAN 29430 (FR) (SAN); Quoin Hill, Tawau Distr., 125 m alt., SAN 29464 (FR) (KEP, " SAN); Ulu Sungai, Kalumpang, SAN 30480 (FL) (KEP, SAN); mi 15 Quoin Hill Road, Tawau, SAN 30616 (FL) (SAN), 35930 (FL) (kK, SAN); Sepilok F. R.. Sandakan, 15 m alt., Brand [SAN] 30983 (FL) (KEP, SAN); Bay Silabukan F.R., Lahad Datu Distr., 160 m alt., SAN 33432 (FR) (SAN); Halogilot, Beaufort Distr., 300 m alt., SAN 33754 (FR) (SAN); Sepilok F.R., Sandakan, 20 m alt., SAN 36709 (FL) (san); Lokan F.R., Segaliud, Sungai Munyed, Sandakan, 100 m alt., Banang [SAN] 3691] (FL) (KEP, ey Sepilok F.R., Sandakan, 30 malt., P. P. Sam [SAN] 37540 (FL) (SAN); Labuk Road, 0, Sandakan Distr., Meijer [SAN] 37897 (FR) (SAN); Labuk Road trace, mi 81-82, Sandan Distr., 150 m alt., Meijer [SAN] 38773 (FR) (kK, SAN); Bettotan, Sandakan, uasa [SAN] 38920 (KEP); Kabuk [Labuk?] Road, mi 60, Sandakan Distr., 40 m alt., 48900 (FL+FR) (sAN); Montinier Road, mi 12, Kg. Bambangan, 300 m alt., Madius [SAN] 49259 (FL) (L, SAN); Ulu Sungei Kimanis, Papar Distr., 60 m alt., Eging Banang [SAN] 49353 (FL) (A, SAN); Elopura Distr., 20 m alt., Puasa & Enggoh [SAN] 55164 (FL) (KEP); Keningau, mi 8.5, Sook Plain, 350 m alt., "Ahmad Talip [SAN] 55525 ee (A, SAN); Beaufort Hill, Beaufort Distr., S4N 66876 (FR) (L, SAN); Lumaku F.R., dalong, Sipitang Distr., S4N 72361 (fl+ FR) (SAN); Telupid, Sandakan Distr., 100m sit, SAN 75356 (FR) (L, SAN); Gunong Rara F.R., Kalabakan, 500 m alt., SAN 75659: Bengkoka F.R., Kudat Distr., 75 m alt., SAN 76076 (FL) (A, K, L, SAN); Ulu Sungei Dusun (NE of Sepilok), Sandakan Distr., 60 m alt., SAN 77357 (FR) (KEP, SAN); Crocker Range F.R., SAN 78305 (FR) (A, KEP, SAN); Ulu Segama, Tawau Distr., SAN 79179 (FR) 224 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 (KEP, SAN); Sepilok F.R., Sandakan Distr., SAN 83715 (FL+FR) (A, KEP, SAN); Bukit Tiulon, Nabawan, Keningau Distr., 350 m alt., SAN 841741 (FR) (KEP, SAN); Kuala Beatrice, Lahad Datu Distr., 270 m alt., S4N 84916 (FR) (K, KEP, SAN); Denum (Banum?), Ulu Sungai, Lahad Datu Distr., 425 m alt., SAN 8500] (FR) (Kep, SAN); Tawau Hill , Tawau Distr., SAN 88228 (FR) (KEP, SAN); Mt. Tawai [Tawau?] F.R., Karamuak (Kuamut?), Sandakan Distr., 500 m alt., S4N 88722 (FR) (KEP, SAN); Sri Usukan, Tawau Distr., SAN 88974 (FR) (KEP, L, SAN); Lahad Datu Road, Sandakan Distr., 27 m alt., SAN 89169, SAN 89177 (both FR) (both KEP, SAN); Kg. Miruru, Mohd. Gan logging area, Ranau Distr., 600 m alt., SAN 90075 (FL) (A, san); ae an, near Tawau, SAN 91500 (FL) (SAN). Brunei: R. Belait at K. Ingei, BRUN 187 (FR) (BR N, KEP); Kuala Belalong, Temburong, 300 m alt., BRUN 465 (FL) a KEP, L). Singap ore: Sebtor Sarat: Ridley 6348 (FR) (A); Bukit Timah Reserve, SFN 35583 (FL) (A, KEP, SING). Indonesia. SUMATRA AND ADJACENT ISLANDS. Selatan: Rawas, Palembang, 200 m alt., Dumas 1630 (fl+ FR) (Bo, L); Muelok(?), Velve(?), 550 m alt., Grashoff 320 (FR) (Bo, a Marbau, Bilah, Bilah Pertama (Parbasir), Toroes 323 (FR) a Tengah: road from Kua Belilas to Berapit R., Indragiri Uplands, 20 m alt., Buwalda 6669 (FL) (Bo, L). ey Lobok besar, 20 m alt., Kostermans & Anta 264 (FL) (A, BO, KEP); G. Manghol, 50 m alt., Kostermans & Anta 783 (FL) (A, Bo, K); Lobok-besar, G. Pading, 20 m alt., Koster- mans & Anta 942 (FL) (A, BO, K, L). KALIMANTAN: Kutei, R. Tiram, Schut K 23 (FR) (BO, K); Sambodja Village, Balikpapan, 60 m alt., Schut K 33 (FR) (A, Bo); Poeloe Lauet, Soengei Paring, 100 m alt., Verhoef?) 80 (FR) (Bo, L); Djaro Dam, ca. 10 km NE of uara Uja, 100 m alt., Vogel 724, 784 (both FR) (both L); gi muara Ancalong, Long Lees, 100 m alt., Wiriadinata 1128 (FR) (A, Bo); near Mt. Kemoel, W Koetai, 330 m alt., /ndert 3684 (FL) (A, Bo, K); Nunukan Is., 20 m alt., rate 8629 (FR) (A, BO). REGIONAL NAMES. “‘Antoi sembago” (‘‘anto1”’ 1s applied to Cyathocalyx species in peninsular Malaysia), Indonesian?, Palembang, Sumatra, Dumas 1630. ‘‘Ba- netan puteh,” language?, reported by Burkill (1966). ‘“Banitan,” language?, Lobok Besar, Bangka, Kostermans & Anta 264 and 942. “Binhud,” Bandjar- Malay language, Balikpapan, Mentawir R., Kalimantan, Sauveur 99. “Binhut,” Bandjar language, Balikpapan, Kalimantan, Schur K 33. ‘“‘Buah sasak,”’ Malay language(?), E. Kalimantan, Wiriadinata 1128. ‘““Dada guan,” Malay lan- guage(?), Pahang, peninsular Malaysia, Jaamat [KEP] 16522. ‘“‘Delah,” Iban language, Brunei, BRUN 897. “Delasai,”’ Malay language, Brunei, BRUN 897. ‘“Dilah,” Iban language, Brunei BRUN 465. “‘Dilasai,”’ Iban language, Kakus, Sarawak, S 2/705. “‘Dilleh,” Iban language, [Vth Div., Sarawak, S 29979. “Kalamanjat,” Tidong language, Marotai, Sabah, Maidin 3057. ‘‘Karai,” Malay language, Bettotan, Sabah, BNBFD 3626; Kedayan and Sungei languages, Bet- totan, Sabah, BNBFD 4544; dialect?, Kuching, Sarawak, Zen [S] 13022. ‘‘Ka- ray,’ Kedayan language, Bettotan, Sabah, BNBFD 3197. ‘Kayu (or “‘pokok’’) dada bakok”’ (= wood or tree of the drug inducing stupor), Temuan language, Selangor, peninsular Malaysia, KL 690, KL 1360, and KL 1375. “‘Kerai’’ (also applied to other species of Polyalthia outside of the P. hypoleuca complex— e.g., P. lateriflora), Malay language, Sandakan region, Sabah, SAN 4 1969. “Lirap (or sirap?),” Kutei language, R. Tiram, Kutei, Kalimantan, Schut K 23. ‘““Medang bentawar,” language?, reported by Burkill (1966). ““Melada,”’ lan- guage?, Pahang, peninsular Malaysia, F. Gd.(?) & M. Soh(?) [KEP] 15105. “Melian,” language?, Selangor, peninsular Malaysia, F. gd. Abu CF Field No. 3335.**Melilin,” Malay language(?), Selangor, peninsular Malaysia, KEP 996 10: Johore, peninsular Malaysia, Yeob(?) 5873. ““Mempisang,” Malay language (a 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 225 general name referring to several species of Annonaceae), SE. Kelantan, FR/ 7075; Johore, peninsular Malaysia, KEP 77852; Perak, peninsular Malaysia, KEPFN 29911. “Nyebulok,’” Kenyah language, Upper Baram R., Sarawak, Chin See Chung 2737. *‘Pahiding,’’ Kayan language, Upper Plieran, Sarawak, SAR 3501. “Pisang-pisang”’ (pisang = banana), Malay language (a general name referring to several species of Annonaceae), Brunei, Maidin 3057; Pahang, peninsular Malaysia, Hamid [KEP] 10690; Tawau, Sabah, SAN 22752. “‘Pi- sang-pisang bukit,” Malay language, Lahad Datu, Sabah, SAN 26065. ‘“‘Purda hutan,” Malay language, Kuala Belilas, Sumatra, Buwalda 6669. ““Samukau,” Iban language, IVth Div., Sarawak, S 73763. ““Sarbah,” Malay language, Upper Plieran, Sarawak, SAR 3501. ‘“Selangan babi,” Malay language, Tawau, Sabah, Kadir A 2106; Sandakan, Sabah, A. Cuadra A 2118. ‘“‘Selaut,’’ Melanau lan- guage, IVth Div., Sarawak, S 29974. “Selaut gunong,” Malay language(?), Mt. Dulit, Sarawak, Richards 1918. “Selmu,” language?, Lambir Hills, Sarawak, Dan bin Haji Bakar 3020. ‘‘Talooto,”’ Sungei language, Elopura, Sabah, 4. Cuadra A 1090. “Tingo,” tidong language, Marutai, Sabah, Maidin 2428. LOcAL USE. Trees of most species of Annonaceae have bark that can be removed in long strips, yielding excellent fiber. This is probably due to the anastomosing bast fibers of the phloem that continue into the bark. Several specimens of Polyalthia sumatrana (e.g., Kostermans & Anta 264 (Bangka), Tandom BNB 1786 (Sabah)) have notes indicating that the bark 1s often used as a source of fiber. The wood is used for roof timber in the Upper Plieran region of Sarawak and may deter insects (Pickles SAR 3501). Collections by Gadoh anak Umbai (KL 690, 1360, and 1375; peninsular Malaysia) are marked “poisonous” or “stupefying,” and the local names, ““kayu dada bakok” and ‘“‘pokok dada ba- kok” can be loosely translated as “‘the tree of the drug that induces stupor.” KL 1494 notes that the fruits are intoxicating, while Chin See Chung 2737 indicates that in the upper Baram River region of Sarawak, P. sumatrana is recognized as a good source for blowpipe wood. EcoLocy. This species has the highest and widest recorded altitudinal range of any in the Polyalthia hypoleuca complex, growing from near sea level to 1200 m alt. (SAN 28758 from Lohan, Ranau Distr., Sabah). It is found both on drier slopes and in flatter, more poorly drained sites, but apparently not in constantly inundated or peat-swamp habitats. Although occasional individuals are reported to achieve a height of 25 m, P. suwmatrana rarely reaches the mid- canopy, being primarily a denizen of the lower understory. The population of Polyalthia sumatrana at Pasoh F.R., peninsular Malaysia, puts out new vegetative growth synchronously twice per year but flowers ir- regularly, and sometimes periods greater than one year intervene between pop- ulation flowerings. Only one report of floral fragrance exists for this species. However, since the flowers attract beetles thought to feed on rotting fruit, it is probable that they have some fragrance, perhaps one that resembles fermenting fruit, as has been found for other Annonaceae (Rogstad, in prep. b). Obser- vations by M. and D. Leighton (pers. comm.) at Kutei National Park, Kali- mantan, Indonesia, indicate that the flesh (not seeds) of mature fruits of P. sumatrana is eaten by numerous animals, including primates (Pongo pyg- 226 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 moides, Hylobates muelleri, Macaca fascicularis), squirrels (Ratufa affinis, Cal- losciurves prevostii, C. notatus), sun bears, hornbills (Buceros rhinoceros, Ry- ticeros undulatus, R. corrugatus, Arthrococeros malayanus, Anorrhinus galer- itus), and other birds (Calyptomena_ viridis, Calochomphus fuliginosis, Megalaima sp., Ducula aenea, Ptilinopus jambu, and an unidentified bulbul). Hornbills have also been observed feeding on the fruits of P. swmatrana in Sabah, Malaysia (SAN 1/5263). The mature fruits of the members of the P. hypoleuca complex all have a similar taste and color to humans (pers. obs.), and thus each species may attract the wide array of frugivores (and seed dis- persers?) observed for P. sumatrana. Comments. Miquel designated three syntypes: Tevsmann 405 (sterile) and 4007 and Diepenhorst 2342. I have chosen the lectotype indicated above because it has diagnostic portions from mature fruit and a flower; a sheet at Kew pre- viously placed with type materials for Polyalthia sumatrana (but for which no published reference has been found) bears portions of two Teysmann collec- tions. This species varies greatly, often on a local basis. For example, two petal forms have been collected from Bangka: the normal linear form (Kostermans & Anta 264), and one that is obovate or strongly flared distally (Kostermans & Anta 942). The latter shape has also been observed elsewhere (e.g., Agama 550, from Sabah). Research is needed to determine the degree to which flower- color variation is correlated with populations and geography and/or with dif- ferences in floral biology. Leaf size provides another interesting example of local variability. In FIGURE 9 the three specimens of Polyalthia sumatrana with the largest leaves, as well as several of the smaller-leaved ones (e.g., S 16645), all come from Sarawak. The specimen with the largest leaves (S 29979) was collected quite near one exhibiting below-average leaf size (S 29974) at Ulu Kakus, Anap, IVth Divi- sion, Sarawak. Such variation is often observed at a single site and may be due to phenotypic variation. Kochummen (1972) recognized Polyalthia sumatrana var. macrocarpa, based upon one specimen (KEP 104976; KEP) bearing relatively large fruits. When measurements from this specimen are included in the data set that generated Ficure 15 and a PCA of this new data set is calculated, this specimen is indeed a very clear outlier (results not shown). However, it may be a mixed collection. The leaves have clearly been taken from some member of the P. Aypoleuca complex. In contrast, the fruits, which are separate from the leaf-bearing branches, have very wrinkled, unridged mature carpels more like those of P. clavigera than the ellipsoid, prominently longitudinally ridged ones of P. sumatrana. Further, the seeds from KEP 104976 have a circumferential groove rather than a ridge and lack an aril scar; although the interior of the seeds has deteriorated somewhat, there are no rumination needles but rather a series of four rumination plates (in cross section) penetrating the endosperm. In addition, the gynoecial portion of the torus bears dense, minute tomentum, in contrast to the glabrous torus characterizing all the members of the P. hypoleuca complex. These seed and torus characters, like the external appear- 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 227 ance of the fruits, are more consistent with the fruits of the P. clavigera-P. lateriflora alliance of species (Rogstad, unpubl. data). Finally, there is only a single collection of this variety (even duplicates are apparently lacking), despite the fact that a few ‘typical’ specimens of P. sumatrana have been collected from this region. This all suggests that the fruits from another species of An- nonaceae have been mounted with vegetative material of one of the members of the P. hypoleuca complex. Polyalthia sumatrana appears to grow in sympatry with P. ovalifolia, P. glauca, and P. hypoleuca. The leaves of P. ovalifolia are usually much larger and more broadly elliptic than those of P. suwmatrana, and the two species differ in numerous floral and fruit characters as well. Polyalthia sumatrana differs from P. g/auca in several easily detectable characters, including globose (vs. elongate) stigmas, usually many more than (vs. less than) 50 stamens, the ellipsoid (vs. globose) mature carpels, and circumferentially ridged (vs. grooved) seeds with an easily detectable (vs. undetectable) aril scar. It is less readily distinguished from P. hypoleuca, especially if one 1s examining a specimen of the former with only immature flowers. However, the pedicels of P. hypoleuca are usually much shorter (less than 9 mm in fruit) and have dense, short, rusty tomentum in both flower and fruit. If pedicels are not available, P. hypoleuca generally has smaller flowers with fewer stamens (< 45) and carpels (usually < 10 5. Polyalthia hypoleuca Hooker f. & Thomson, Fl. Brit. India 1: 63. 1872, pro parte florif., non Guatteria hypoleuca Mig. Fl. Ned. Ind., Eerste Bijv. Suppl. 381. 1861. Type: Singapore, [18]67, Maingay 1516 (Kew Distr. 50) (lectotype, here selected, k, pro parte). Ficures 3A, B; 6A; 7A; 8D; 18D; 19C, D; 21B. Tree to 50 m tall, DBH to 40 cm; trunk often fluted at base; pneumatophores absent; bark white to pale yellow, smooth on branches and trunk, with finely fissured or cracked patches appearing irregularly on bole; hoop marks not prominent on lower portions of bole; slash on larger trees revealing phloem rays reaching epidermis, phloem-ray apices 0.1-0.5 mm apart. Twigs with bark pale yellow-white, youngest growth glabrous (6%), or with moderate (39%) to dense (55%; n = 31), short, rusty tomentum. Leaves with young petiole 3-7 (4.7, 1.2, 25) mm long, moderately (45%) to densely (55%; n = 31) tomentose abaxially; lamina narrowly elliptic to elliptic, 6.1-12.5 (9, 1.6, 25) x 2-3.5 (2.6, 0.4, 25) cm, the apex acuminate to caudate, 0.6-1.4 (1.1, 0.2, 25) cm long, the base rounded to acute, the secondary veins 25 to 45 (32, 4.9, 25) per side, not forming regular weak intramarginal vein; leaves when fresh flat with abaxial surface dull white (tinged with brown, tan, or gray). Inflorescences in foliate axils, up to 4 per axil, and also borne on branches below leaves, tubercles absent or poorly developed; peduncle extremely reduced, 0.4-1.8 (0.9, 0.3, 26) mm long, usually with only | terminal flower developing, lateral floral buds and developing flowers absent; bracts | to 4, oriented perpendicular to branch axis when only | flower per axil. Flowers with pedicel 1—7.5 (2.7, 1.2, 26) mm long, 0.5-0.9 (0.7, 0.1, 26) mm wide at base and 0.5-1.2 (0.9, 0.4, 26) mm at 228 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 apex, with dense, rust-colored tomentum (100%; n = 26); bract occasionally present (8%; n = 26), 1-2 mm above articulation. Sepals triangular, 0.7-1.2 (1, 0.14, 16) mm long, very densely tomentose abaxially (100%, n = 16), often (38%: n = 26) caducous at floral maturity. Petals linear; outer petals 4-11 (7.1, 1.8, 26) x 1-2.4 (1.4, 0.4, 26) mm, the pubescence beginning adaxially on basal (42%), middle (54%), or distal (4%; 1 = 65) 1s, sparse to moderate through- out (58%), to sparse throughout except abaxially at base where dense (42%; n = 26), the midvein not visible abaxially; inner petals 4-12 (7.5, 2.2, 26) x 0.8-1.9 (1.2, 0.3, 26) mm; all petals turning yellow at maturity, with 5 scent, deep burgundy-red when dry. Torus 0.3-0.8 (0.58, 0.12, 26) x 0.6-1.5 (1, 0.2, 26) mm, gynoecial portion concave (7.7%), flat (57.7%), or convex (34.6%; n = 26). Androecium 0.7-1.8 (1.3, 0.3, 26) x 1.8-2.8 (2.3, 0.3, 26) mm: stamens 25 to 39 (32.3, 4, 26), 0.7-1.1 (0.8, 0.13, 26) x 0.5-0.8 (0.7, 0.07, 26) mm. Gynoecium 0.5-1.5 (1, 0.23, 26) mm across; carpels 3 to 9 (5.4, 1.4, 26), 0.5-1.2 (0.9, 0.17, 26) mm long, glabrous (4%) or moderately to densely tomentose (96%; n = 26); stigmas globose to moderately elongate, 0.2-0.7 (0.4, 0.16, 26) mm long, those of outermost ring not reflexed away from floral axis. Immature carpels green, slowly turning red 2-4 weeks before full maturity, finally deep purple to black, the pericarp then soft and fleshy (juice laden), with sweet or bitter-sweet taste. Mature, dry fruits with pedicel 4-9 (6.5, 1.4, 25) mm long, |-2.3 (1.5, 0.4, 25) mm in diameter at apex, with dense, short, rusty tomentum (96%) or only moderately tomentose (4%; m = 25); stamen scars < 40 (100%; n = 21); carpels with the stalk 2-5 (3.8, 0.9, 25) x 1.2-2.4 (1.5, 0.3, 25) mm, seed-bearing portion ellipsoid, 10-20 (15, 2, 25) x 9-14 (11, 1.1, 25) mm, not ridged, apex blunt or rounded (100%; n = 25); fruit wall < 1 mm thick, deep burgundy to black. Placentation basal at maturity. Seed 1 per mature carpel, ellipsoid, with circumferential ridge, the aril scar present, the testa deep brown, finely to moderately pitted, not irregularly indented; rumination needles very fine. DISTRIBUTION. Peninsular Malaysia, Sumatra, Borneo, and immediately ad- jacent islands (see Map 5). SELECTED SPECIMENS SEEN. Malaysia. KELANTAN: Ulu Lebir Kechil, SE Kelantan, near Trengganu border, 370 m alt., FR/ 44/3, FRI 4414 (both FR) (both kep); Ulu Lebir F.R., KEP 11695 (f1+FR) (KEP); G. Rabong, 610 m alt., FRJ 20623 (FL+FR) (A, K, KEP). Lace Ulu S Trengan, Kg. Melaka, 150 m alt., FR/J 10538 (f+ FR) (A, KEP); Bukit auk F.R., Dungun Distr., 25 m alt., G. H. S. Wood 76094 (FR) (KEP). PAHANG: near ot Sungai Sepia, Shah & Noor MS 2008 (FR) (A, KEP, SING); Lesong F.R., FRI 25181 (FL) (KEP); Bukit Kajang F.R., Raub, Jinal [KEP] 20351 (FR) (KEP); Pelican Road, T. Lapah Tempwiang, Kuantan-Pahang, /aamat & Sow [KEP] 43190 (FR) (KEP). SELANGOR AND KUALA LUMPUR FEDERAL District: Telok F.R., Rogstad 934, 947 (both FL) (both Bangi, Balgooy 2199 (fl+ fr) (KEP, L); Telok F.R., near Klang, Hamid & Yeop (Yeob?) IKEPI 3267 (fl) (KEP); Bukit Changgang, Klang, SEN 33955 (FR) (A, KEP, SING); near Kuala Lumpur, Sudin [KEP] 41882 orca (KEP); Telok F.R., Klang, KEP 45808 (FR) (KEP). NEGRI SEMBILAN: Pasoh F.R ar Simpang Pertang, Rogstad 912, 933, 942 (all FL) (all a); Tampin F.R., 540 m alt.. "ERI 14216 (FR) (KEP); Gunong Angsi F.R., 400 m alt., FRI 14631 (FL) (A, KEP). Jouore: Segamat Wildlife Reserve, FRI 17148 (fi+ FR) A, KEP); Sungai Kayu, SFN 32/89 (FR) (A, KEP, SING); Pengkalan Raja, Pontian, SFN 36644 (fl+ FR), SFN 36697 (FL+FR) (both A, KEP, SING). SARAWAK: Buglgan(?), T. 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 229 15°N 9390E 1220E Map 5. Distribution of Polyalthia hypoleuca. Kelepis(?), S' 690 (FR) (KEP, SAR); S Entulang, Pr S 13120 (FR) ee SABAH: manis F.R., Papar Distr., 3 m alt., S4N A 4600 (FR) (KEP, L, SAN); K. Klias Saratok, Beaufort Distr., SAN 27307 (FR) (KEP, SAN, SAR); Melaliah N.F. Reserve, Sones Distr., 15 malt., SAN 27953 (FR) (KEP, L, SAN); Seratok Camp, Beaufort Distr., 10 m alt., SAN 35236 (FL) (K, SAN); ope rail line Lumat, Beaufort Distr., Binideh [SAN] 58445 (FR) (SAN). Brunei: Seria peat swamp, 8 m alt., S 5892 (fl) (BRUN, oe Badas swamps, Sinclair & Kadim 10460 (fr) (A, eas Kuala Belait, Badas State Land, 7 m alt., SAN 17457 (FL) » KEP, L, SAN). Singapore: Jurong Road, oe 26163 (FL+fr) (A, FRI, SING). Indonesia. MATRA AND ADJACENT ISLANDS: Riouw Ond., Koeantan Distr., Sei Rambei [Su- ae 70 malt., bb 23472 (FR) (A, BO, L). Utara: Masihi F.R., Acahar. E coast, Krukof 4095 (FR) (A, L, US). Tengah: Bengkalis, Sekoedi [E coast], 5 m alt., bb 2/285 (fl) (a, Bo, L); Pulau Gelang, 4 m alt., bb 29766 (f1+FR) (A, Bo, L). Selatan: Palembang, 15 m alt., Zwan WF(?) 1172 (FR +fl) (Bo, L); Belimbingan, 6 m alt., bb 28536 (FL) (A, BO, K). Bangka: Lobok-besar, 20 m alt., Kostermans & Anta 148 (FL+FR) (A, Bo, KEP?), 199 (fl+ FR) (A, Bo), 1072 (FL+FR) (A, Bo, LAE), //7/ (FR) (A, BO, K, KEP). KALIMANTAN: near Sampit, 5 m alt., Kostermans 8134 (FL) (A, Bo, L). REGIONAL NAMES. “Ban aan aandoel,” Indonesian?, Palembang, Sumatra, Zwan 1172. ““Banittan,” Indonesian?, Sumatra, L. B. 72. “Kayu selaut,” Malay lan- guage? (kayu = wood, laut = sea), Sarawak, S 690. ““Mahawai,” language?, ampit, S. Kalimantan, Kostermans 8134. ‘““Manitan,” language?, Pahang, pen- insular Malaysia, KEP 78666. ““Melian,” language? (in Burkill, 1966). ‘““Meli- lin,” language? (in Burkill, 1966). “Mempisang,” Malay language a le name for several species of Annonaceae), Kelantan, peninsular Malaysia, KEP 11695. “Pisang-pisang,” Malay language (pisang = banana) (a ear name for several species of Annonaceae), Pahang, peninsular Malaysia, Jinal [KEP] 230 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 20351; Papar Distr., Sabah SAN A 4600. ‘‘Ridis,” language?, Lobok Besar, Bangka, Kostermans & Anta 1072. “Selaut,” Iban language?, Simanggang, S. Sarawak, S 13/20. “Telinga basing,” language?, Beaufort Distr., Sabah, SAN 27307. “Usai,” language?, Lobok Besar, Bangka, Kostermans & Anta 148. Loca use. As in many Annonaceae, the bast fibers of the phloem and bark anastomose to yield a bark that can be removed in strips and makes excellent rope or fiber. This quality is noted by J. A. R. Anderson (S690), who indicated that the natives of southern Sarawak use the bark of Polyalthia hypoleuca to make string and the wood to make bows. Burkill (1966) noted that this fiber withstands sea water. The wood of larger P. hypoleuca trees may also be used for crate and toy construction in peninsular Malaysia (Timber Utilization Chart, Forest Research Institute, Kepong, Selangor, Malaysia). In addition, Burkill (1966) reported that a decoction of the roots of this species is given after childbirth in peninsular Malaysia. Eco.ocy. Altitudinal range is recorded as from near sea level to 610 m alt. Although Sinclair (1955) indicated that Polyalthia hypoleuca attains a height of 15 m, I have observed several individuals at Pasoh F.R., peninsular Ma- laysia, that have grown to over 30 m; flowering is generally most pronounced in tall individuals with exposed crowns. At Pasoh this species 1s restricted to drier slopes and does not appear to grow at seasonally or constantly inundated sites with little or no peat development. It is somewhat puzzling, therefore, that it (or a cryptic sibling species?) does grow in peat-swamp forests, but only where a thick peat mat has developed; P. Aypoleuca is unable to grow in inundated bare clay soils. Possible reasons for this will be suggested in a sub- sequent article (Rogstad, in prep. a). Ten distinct waves of flowering of varying intensity divided into two main periods were observed at Pasoh over the period of a year. Each main period was immediately preceded by the initiation of a flush of vegetative growth. The waves are very synchronized among the members of the population. Pol- lination appears to be effected, in both the Pasoh and Telok populations, pri- marily by the same species of beetle. ComMENts. Unfortunately, the type designated by Hooker and Thomson (1872) is a mixed specimen consisting of flowering and fruiting branches of Polyalthia hypoleuca and fruiting ones of P. sumatrana. Sinclair (1955) noted this mixed specimen, and here I lectotypify P. Aypoleuca based on it. Polyalthia hypoleuca grows in sympatry with P. glauca, P. ovalifolia, and P. sumatrana. Characters separating P. hypoleuca from P. sumatrana are included in the comments under the latter species. It can be distinguished from P. glauca and P. ovalifolia because the latter two species have elongate, clavate stigmas more than 64 percent of the length of the carpels, stigmas of the outermost ring bent outward at an angle greater than 30°, short (usually less than 10 mm) pedicels in fruit, and seed-bearing portions of the carpels globose at maturity. Due to a very close morphological resemblance, Polyalthia hypoleuca 1s the species most likely to be confused with P. mu/tinervis. As indicated in the PCA analyses of characters (e.g., see FIGURE 17), P. multinervis does differ slightly 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 2a) from P. hypoleuca: it has slightly larger flowers, glabrous or very sparsely pubescent (vs. densely tomentose) pedicels, and mature carpels acute to acu- minate (rather than blunt) at the apex. These two species also differ in floral biology. Populations of P. hypoleuca studied in peninsular Malaysia had very fragrant flowers from the time of stigma receptivity until the petals abscised at anther dehiscence. Flowers at the equivalent stage from numerous individ- uals of P. multinervis studied at April River, Papua New Guinea, lacked de- tectable fragrance. The main pollinator of P. hypoleuca in peninsular Malaysia is of a strikingly different morphology than the main pollinator of P. multinervis in Papua New Guinea (Rogstad, in prep. b). Polyalthia hypoleuca has not been collected east of Borneo, while P. mu/tinervis is thus far known only from the April River region of the Sepik River drainage in Papua New Guinea. The sum of the evidence at hand indicates that slight but constant morpho- logical differences exist between these two entities, that biological differences exist between them such that if they were to grow sympatrically, gene exchange might well not occur, and that they are geographically isolated. The two entities thus appear to conform to the evolutionary species concept defined above, and I therefore recognize them as such here. 6. Polyalthia multinervis Diels, Bot. Jahro. Syst. 52: 178. 1915. Type: [Papua New Guinea,] Sepik River region, April River, Sept. 1912, Ledermann 8835 (fl) (holotype, B; isotypes, K, SING). FiGures 3D; 4E; 8A; 19A, B; 21D. Tree to 30 m tall, DBH to 30 cm; trunk often fluted at base; pneumatophores absent; bark white to pale yellow, smooth on branches and trunk, with finely fissured or cracked patches appearing irregularly on bole; hoop marks not prominent on lower portions of bole; slash on larger trees revealing phloem rays reaching epidermis, phloem-ray apices 0.1—0.5 mm apart. Twigs with bark pale yellow-white, youngest growth glabrous (18%) or with moderate short, rusty tomentum (82%; n = 11). Leaves with young petiole 4-5 (4.6, 0.5, 8) mm long, glabrous (9%) to moderately tomentose (91%; m = 11) abaxially; lamina narrowly elliptic to elliptic, 8.4-10.6 (9.4, 0.7, 8) x 2.3-3.4 (2.8, 0.4, 8) cm, the apex acuminate to caudate, 0.5-1.1 (0.8, 0.2, 8) cm, the base rounded to acute, the secondary veins 16 to 29 (25, 4.1, 8) per side, not forming regular weak intramarginal vein; leaves when fresh flat with abaxial surface dull white (tinged with brown, tan, or gray). Inflorescences in foliate axils, up to 4 per axil, or borne on branches below leaves, tubercles absent or poorly developed; peduncle extremely reduced, 0.4-1.4 (0.8, 0.3, 8) mm long, usually with only | terminal flower developing, lateral floral buds and developing flowers absent; bract(s) 1 to 4, oriented perpendicular to branch axis when only | flower per axil. Flowers with pedicel 2-7 mm long, 0.6—1 mm wide at base and 0.9-1.2 mm at apex, glabrous or rarely with very sparse pubescence; bract always absent. Sepals triangular, |-1.2 mm long, very densely tomentose abaxially, often caducous at floral maturity. Petals linear (100%, m = 18); outer petals 7-— 12 x 1.3-2.1 mm, the pubescence always beginning adaxially on basal 1, sparse to moderate throughout, the midvein not visible abaxially; inner petals 2o2 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 6.6-13 x 1.3-2 mm; all petals turning yellow at maturity, detectable fragrance absent, deep burgundy-red when dry. Torus 0.5-1.1 x 0.9-1.1 mm, gynoecial portion convex. Androecium |.3-2.4 x 2.5-3.2 mm; stamens 26 to 32, 0.9- 1.2 x 0.8-1 mm. Gynoecium 0.7—1 mm across; carpels 2 to 7, 1.1-1.3 mm long, moderately to densely tomentose; stigmas globose to moderately elongate, 0.3-0.6 mm long, those of outermost ring not reflexed away from floral axis. Immature carpels slowly turning red 2-4 weeks before maturity, finally deep purple to black, with soft, fleshy (juice-laden) pericarp tasting sweet or bitter- sweet. Mature, dry fruits with pedicel 6-10 mm long, |.2-1.9 mm in diameter at apex, glabrous or with very sparse, short pubescence; stamen scars < 40; carpels with stalk 2-5 x 1.4-1.9 mm, seed-bearing portion ellipsoid, 18-22 x 7-12 mm, not ridged, acutely tapering at apex; fruit wall < 1 mm thick, burgundy to black. Placentation basal at maturity. Seed | per mature carpel, ellipsoid, with circumferential ridge, the aril scar present, the testa deep brown, finely to moderately pitted, not irregularly indented; rumination needles very fine. DisTRIBUTION. Sepik River basin of Papua New Guinea in middle to lower April River region (see Map 3 SELECTED SPECIMENS SEEN. Papua New Guinea. East SEpIK Province: Niksek Village, April River airstrip, middle April R. (tributary to Sepik R.), 4°42’S, 142°32’E, 50-150 m alt., Rogstad 813 (FL+fr), 817, 818, 819 (all FL), 820 (FL+FR), 828 (FL+fr), 829 (FL+FR) (all a, LAE); [April R. region, Sepik Valley, ca. 15-20 km upriver from Ambunti,] Ledermann 88335 (fl) (B, K, SING). ws REGIONAL NAMES. “‘Fawss”’ or “‘pfahss,”’ Niksek language, April R., Papua New Guinea (Rogstad 829). LocAL USE. Timber of this species is used for house construction in the April River region of Papua New Guinea (Rogstad 829). EcoLocy. Polyalthia multinervis is thus far known only from 80 to 200 m alt. since it has only been collected from the April River region. More information on all aspects of this species is obviously needed. In the forests around Niksek Village (April River airstrip), flowering individuals are generally small to me- dium-size trees. However, their crowns are at or near canopy level ena they grow almost exclusively in somewhat stunted forest on a thick, “peat humus mat (mat depth may exceed | m) that develops mostly in nee areas and yields black water drainage. Areas with this root-humus mat are not true peat swamps; they are well drained and are not even intermittently inundated. Tracts on which the mat develops may be ecologically similar to the closed kerangas (heath) forests of Sarawak described by Brunig (1974), but this can only be determined by comparative investigations. Polyalthia multinervis ap- pears to be a late secondary species, not growing in the full sun of open gaps but also not regenerating in the deepest forest. The most prolific regeneration can be found on well-developed root mats on banks well above the April River. As in P. hypoleuca, flowering in this species appears to occur in several distinct waves each year. Although mature individuals vary in the number of waves in which they flower, flowering ofall plants within a wave 1s highly synchronized (see also P. hypoleuca, and Rogstad, in prep. b). 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 233 COMMENTS. Diels (1915) was apparently not familiar with Polyalthia hypoleuca since he made no mention of its similarity to P. Mes Despite this close resemblance, I have maintained the two as distinct speci under P. hypoleuca). The New Guinea region is very ey collected, and nis this classification scheme may require review as more information accumulates. The type material designated by Diels (1915) that I have seen bears only immature flowers. However, enough characters are present to distinguish this entity clearly from Polyalthia discolor and to confirm its identity with specimens I have collected in the April River region. Diels also mentioned the only other collection of this species previous to mine, Ledermann 9843, stating that it bore fruit. I have been able to locate but one fruiting sheet of this collection (at L), and it bears only an apparently mature fruit pedicel with the developed carpels missing. The stamen scars number less than 50, and the pedicel agrees with P. multinervis as found at April River. Uncertain Taxa The possibility that a putative variety of Polyalthia sumatrana, var. macro- carpa, may be based on a mixed collection has been discussed above under that species. Two specimens from the Andulau Forest Reserve, Kuala Belait District, Brunei (Sinclair & Kadim 10454, SAN 17486), and one from R. Belait at K. Ingei, Brunei (BRUN 187), are most closely allied, on the basis of flower and fruit characters, with Polyalthia sumatrana. However, they differ in their ped- icels that are moderately tomentose and are occasionally fully articulated about 8-11 mm above the articulation here interpreted as separating the peduncle from the pedicel. The latter character is variable, however, even on a single specimen. Since “‘typical’’ specimens of P. sumatrana can also be found in this region (e.g., BRUN 465, BRUN 897), and since no population has been found that exhibits constant and discrete variation in these characters (as has been seen in P. ovalifolia, for example), I have not recognized this variation formally. ACKNOWLEDGMENTS I wish to thank P. Ashton, F. Bazzaz, J. Croft, D. Frodin, R. German, T. Givnish, E. Henty, E. Kellogg, J. Koek-Noorman, Samad bin Latif, P. Maas, B. Mischler, F. Ng, J. and T. Niksek, Salleh Mohd. Nor, L. Roseman, E. Schmidt, C. Sperling, G. Staples, P. Stevens, E. Taylor, Y. Tho, F. and S. Urschitz, and A. Weitzman. I also wish to thank the staffs of the Harvard University Herbaria, the Forest Research Institute Malaysia (particularly the Pasoh Forest Reserve) and the Forest Departments of Sabah and Sarawak, the Missouri Botanical Garden, the Royal Forest Department of Thailand, and the Forest Department of Papua New Guinea. This project was supported by the National Science Foundation (BSR-8218184), the Anderson and Atkins funds of Harvard University, Sigma Xi, the Explorer’s Club, and personal funds. Any errors are the fault of the author alone. 234 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 LITERATURE CITED AARSSEN, L. W. 1983. Ecological combining ability and competitive combining ability in plants: toward a general evolutionary theory of coexistence in systems of com- ae Amer. Naturalist 122: 707-731. 985. Interpretation of the evolutionary consequences of competition in plants: Sacre approach. Oikos 45: 99-109. Mae: R. 1982. Species diversity and phasing of disturbance. Ecology 63: 289-293. Acevepo, M. F. 1980. Tropical rainforest dynamics: a simple mathematical model. Pp. 219-227 in J. 1. FuRTAbo, ed., ae <— and development. International Society of Tropical Ecology, Kuala Lum On Horn’s Markovian model nes dynamics with particular reference to tropical forests. Theor. Populat. Biol. 19: 230-250. Acren, G. I., & T. FAGERSTROM. 1983. Limiting dissimilarity in plants: randomness prevents exclusion of species — similar competitive abilities. Oikos 43: 369-375. ARMBRUSTER, W. S. 1985. Patterns of character divergence and the evolution of re- productive ecotypes of balechorpia ai (Euphorbiaceae). Evolution 39: 733- 752. ARMSTRONG, R. A., & R. MCGEHEE. 1980. Competitive exclusion. Amer. Naturalist 115: 151-170. Asuton, P. 8. 1969. Speciation among tropical forest trees: some deductions in the light of recent evidence. Biol. J. Linn. Soc. 1: 155-196. 977. A ana ie : rain forest research to evolutionary theory. Ann. Missouri Bot. Gard. 64: 6 79. Some geographic ieee in ers variation in the Asian tropics and their possible significance. Pp. 35-48 in K. Larsen & L. B. HOLM-NIELSEN eds., Tropical botany. Academic Press, New York. 1984. Biosystematics of tropical forest plants: a problem of rare species Pp. 497- 518 in W. F. Grant, ed., Plant biosystematics. Academic Press, New AUBREVILLE, A. 1938. La forét coloniale: les foréts de l'Afrique occidentale frangaise. Ann. Acad. Sci. Colon. Paris 9: 1-245. 1971. Regeneration patterns in the closed forest of Ivory Coast. Pp. 41-55 in S. R. Eyre, ed., World vegetation types. Columbia Univ. Press, New Baker, H.G. 1959. Reproductive methods as ea in speciation in flowering plants. Cold Spring Harbor Symp. Quant. Biol. 24: 177-199. 1970. Evolution in the tropics. eae 2: 101-11 Barton, N. H., & B. CHARLESWORTH. 1984. Genetic ice founder effects, and speciation. Ann. Rev. Ecol. Syst. 15: 133-164. BENTHAM, G., & J. D. Hooker. 1862. Anonaceae. Genera plantarum. Vol. 1, pp. 20- 29, 955-958. Reeve and Co., London BERNSTEIN, H., H. C. Bverty, F. A. Hopr, & R. MicHop. 1985. Sex and the emergence of species. J. Theor. Biol. 117: 665-690. BRoKAW, N. V. L. 1985. Gap-phase regeneration in a tropical forest. Ecology 66: 682- 6 Brunic, E. F. 1974. Ecological studies in kerangas forests of Sarawak and Brunei. Borneo ea Bureau for the Sarawak Forest Department, Kuching, Sarawak, East Malay BUCKLEY, R. 1983. A possible mechanism for maintaining diversity in species rich communities: an addendum to Connell’s hypotheses. Oikos 40: 312. BurGer, W. C. 1974. Ecological See eee in some congeneric species of Costa Rican flowering plants. Ann. Missouri Bot. Gard. 61: 297-306. 1980. Why are there so many eee of flowering plants in Costa Rica? Brenesia 8. —. 1981. Why are there so many kinds of flowering plants? BioScience 31: 576- 81. 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 235 BurkiLL, I. H. 1966. A dictionary of the economic products of the Malay Peninsula. 2 vols. Ministry of Agriculture and Co-operatives, Kuala Lumpur. Busu, G. L. 1981. Stasipatric speciation and rapid evolution in animals. Pp. 201-218 in W. R. ATCHLEY & D. S. WooprurF, eds., Evolution and speciation. Cambridge Univ. Press, Cambridge, England. Carson, H. L., L. S. CHANG, & T. W. LytTrLe. 1982. Decay of female sexual behavior under parthenogenesis. Science 218: 68-70. CHAN, H. T. 1980. Reproductive biology of some Malaysian dipterocarps. Pp. 169- 175 in J. I. FurTAbo, ed., Tropical | and development. International Society of Tropical Ecology, Kuala Lumpu & S. APPANAH. 1980. Re productive aes of some Malaysian dipterocarps I: flowering biology. Malaysian Foreste 32-143. CHESSON, P. L., & R. R. WARNER. 1981. ss ieat m variability promotes coexis- tence in lottery competitive systems. Amer. Naturalist 117: 923-943 Comins, H.N., & I. R. Nosie. 1985. Dispersal, variability, and transient niches: species coexistence in a uniformly variable environment. Amer. Naturalist 126: 706-723. ConNELL, J. H. 1978. Diversity in tropical rainforests and coral reefs. Science 199: y, & L. J. Wess. 1984. Compensatory recruitment, growth, and mortality as pace maintaining rain forest tree diversity. Ecol. Monogr. 54: 141- 164 Corner, E. J. H. 1949. The annonaceous seed and its four integuments. New Phytol. 48: 332-364. . 1954. The urs 2 tropical forest. Pp. 34-46 in J. S. Huxxey, A. C. Harpy, & E. B Rb, eds., Evolution as a process. Allen & Unwin, Ltd., London. 1970. Fhe ce of ao deltoidea;, a recent invasion of the Sunda Shelf. Phil. Trans., Ser. B. 256: 281-318. 1976. The seeds of dicotyledons. 2 vols. Cambridge Univ. Press, Cambridge, En gland. CULLEN, J. 1968. Botanical problems of numerical taxonomy. Pp. 175-183 in V. H. HEyYwoop, ed., Modern methods in plant taxonomy. Academic Press, New York. Darwin, C. 1859. On the — of species. Facsimile ed. (1964). Harvard Univ. Press, Bee Massachus Davis, J. I., & A. J. a a. 1985. Morphological variation and speciation. Syst. Bot. 10: 417-425 DiamMonb, J. M. 1973. Distributional ecology of New Guinea birds. Science 179: 759- 769 Diets, L. 1915. Neue Anonaceen von Papuasien. Pp. 177-186 in C. LAUTERBACH, ed., Beitraége zur Flora von Papuasien, IV. Wilhelm Engelmann, Leipzig. DoszHANSsKY, T. 1950. Evolution in the tropics. Amer. Sci. 38: 209-221. Dover, G. A., & R. B. FLAVELL. 1982. Genome evolution. oo Press, London. DRANSFIELD, J. 1981. Palms and Wallace’s line. Pp. 43-56 in T. C. WHITMORE, ed., Wallace’s line and plate tectonics. Clarendon Press, Ox ord. Duncan, T., & B. R. BAUM. 1981. Numerical phenetics: its uses in botanical system- atics. Ann. Rev. Ecol. Syst. 12: 387-404. ELDREDGE, N., & J. CRACRAFr. 1980. Phylogenetic patterns and the evolutionary pro- cess. Columbia Univ. Press, New York. ELton, C.S. 1975. Conservation and the low population density of invertebrates inside Neotropical rainforest. Biol. Conservation 7: 3- Ewe , J., ed. 1980. Tropical succession. Biotropica, Sipe 1. 12: 1-95. Feporov, A. A. 1966. The structure of the tropical rain forest and speciation in the humid tropics. J. Ecol. 54: 1-11. On speciation in the humid tropics: some new data. Gard. Bull. Singapore 24: 127-136. Fiscuer, A. G. 1960. Latitudinal variations in organic diversity. Evolution 14: 64-81. 236 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Fircn, W. M., & W. R. ATCHLEY. 1985. Evolution in inbred strains of mice appears rapid. Science 228: 1169-1175. FLEMING, T. H. 1985. Coexistence of five sympatric Piper (Piperaceae) species in a tropical dry forest. Ecology 66: 688-700. Fries, R. E. 1959. Annonaceae. Pp. 1-170 in H. Metcuior, ed., Nat. Pflanzenfam. ed. 2. 17all. Duncker & Humblot, Berlin. Gan, Y. Y., F. W. RoBertson, & E. SOEPADMO. 1981. Isozyme variation in some rain forest trees. Biotropica 13: 20-28. Gentry, A. H. 1976. Bignoniaceae of southern Central America: distribution and ecological specificity. Biotropica 8: 117-131. ae | . Patterns of Neotropical plant species diversity. Evol. Biol. 15: 1-84. GLAZIER, D. : 1987. Energetics and taxonomic patterns of species diversity. Syst. Zool. 36(1): 62-71. eee R. 1940. The material basis ofevolution. Yale Univ. Press, New Haven, Gort ies, L. D. 1984. Genetics and morphological evolution in plants. Amer. Natu- ralist 123: 681-709. 1985. Reply to Coyne and Lande. /bid. 126: 146-150. GoTTSBERGER, G., & I. SILBERBAUER-GOTTSBERGER. 1984. Pollen units, pollen shape and apertural position in the Annonaceae—a reassessment. Beitr. Biol. Pflanzen 59: 73. Grant, P. R. 1972. Convergent and divergent character displacement. Biol. J. Linn. Soc. 4: 39-68. GRANT, V. 1963. The origin of adaptations. Columbia Univ. Press, New York. 1971. Plant speciation. Columbia Univ. Press, New York. Gruss, P. J. 1977. The maintenance of species-richness in ae communities: the importance of the regeneration niche. Biol. Rev. 52: 107-14 Haber, W. A., & G. W. Frankie. 1982. Pollination of Luehea (Tiliaceae) in Costa Rican deciduous forest. Ecology 63: 1740-1750. HALLE, F., R. . OLDEMAN, & P. B. TOMLINSON. 1978. Tropical trees and forests. Springer- Verlag, New York. Harbin, G. 1960. The competitive exclusion principle. Science 131: 1292-1297. Harper, J. 1977. spi biology of plants. Academic Press, New York. Harti, D. L., D. E. DyKHuizen, & A. M. Dean. 1985. Limits of adaptation: the evolution of selective eat ty. Genetics 111: 655-674 HENNIG, W. 1966. ps at systematics. Univ. Illinois Press, Urban Heywoop, J. S., & T. H. FLem 1986. Patterns of allozyme variation in siues Costa Rican species of Piper. poiapies 18: 208-213. Ho.trum, R. E. 1953. Evolutionary trends in an equatorial climate. Symp. Soc. Exp. Hooker, J. D., & T. THOMSON. 1872. ames Pp. 45-94 in J. D. Hooker, ed., The flora of British India. L. Reeve and Co., London Husse._, S$. P. 1979. Tree dispersion, abundance, a diversity in a tropical dry forest. Science 203: 1299-1309. Huston, M. 1979. A general hypothesis of species diversity. Amer. Naturalist 113: 81-101. HutcHINson, G. E. 1959. Homage to Santa Rosalia, or why are there so many kinds of animals? Amer. Naturalist 93: 145-159. Hutcuinson, J. 1917. Decades Kewenses. Bull. Misc. Inform. 1917: a fa 1964. The genera of flowering plants. Vol. 1. Oxford Univ. Pre ondon. Jacons, M. 1962. Pometia (Sapindaceae), a study in variability. Lee 6: 109- 144 JANZEN, D. H. 1980. avo of seed-attacking beetles in a Costa Rican deciduous forest. J. Ecol. 68: 929-952 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 231 Jones, C. E. 1978. Pollinator constancy as a pre-pollination isolating mechanism be- tween sympatric species of Cercidium. Evolution 32: 189-198. Jone, K. 1980. A cytoembryological approach to the study of variation and evolution in rainforest tree species. Pp. 213-218 in J. I. FurTADo, ed., Tropical ecology and development. International Society of Tropical Ecology, Kuala Lumpur. Kaur, A. 1980. Cytological and embryological studies on some Dipterocarpaceae. Pp. 207-212 in J. I. Furtabo, ed., Tropical ecology and development. International Society of ee Ecology, Kuala Lumpur. C. O. Ha, K. Jona, V. E. SANps, H. T. CHAN, E. SOEPADMO, & P. S. AsH 1977. Moot may be widespread among trees of the climax rain forest. Seine 271: 440-442. Kimura, M., & T. OHTA. 1971. Theoretical aspects of population genetics. Princeton Univ. Press, Princeton, New Jersey Weiss. 1964. The nue stone model of population structure and the decrease of genetic correlation with distance. Genetics 49: 561- 576. Kina, J. L. 1984. Selectively neutral alleles with significant ph pic effects: a steady- state bel Evol. Theory 7: 73-79. H. Juxes. 1969. Non-Darwinian selection. Science 164: 788-791 KLOPFER, : H.,& R.H. MacArtuur. 1961. On the causes of tropical species diversity: niche evetan: Amer. Naturalist 95: 223-226. KOCHUMMEN, K. M. 1972. Annonaceae. Pp. 61-99 in T. C. WuiTMorE, ed., Tree flora of Malaya. Vol. 1. Longman Group, Ltd., London. Kora, K. 1958. On the periodicity of tree growth in the tropics with reference to the mode of branching, the leaf fall and the formation of the resting bud. Gard. Bull. KRIEBER, M., & M. R. Rose. 1986. Molecular aspects of the species barrier. Ann. Rev. Ecol. Syst. 17: 465-485. Kurz, S. 1874. Contributions towards iii of the Burmese flora. J. Asiat. Soc. Bengal, Pt. 2, Nat. Hist. 43: 39- LANDE, R. 1976. Natural selection and in genetic drift in phenotypic evolution. Evolution 30: 314-334. Lana, G. E., & D. H. Knicut. 1983. Tree growth, mortality, recruitment, and canopy gap formation during a 10-year period in a tropical moist forest. Ecology 64: 1075- 108 LEENHOouTS, P. W. 1967. A conspectus of the genus A/lophylus (Sapindaceae). Blumea : 301-358. Leicu, E. G., Jk. 1982. Why are there so many kinds of tropical trees? Pp. 63-66 7 Leicu, Jr., A. S. RAND, & D. M. Winpsor, eds., The ecology of a aes for est. Smithsonian Institution Press, Washington, D. C. 9. Le THomas, A. | ae nonacées. Flore du Gabon. Vol. 16. Muséum National d’His- toire Naturelle, P 1981. Ocal characters of the pollen grains of African Annonaceae and their significance for the phylogeny of primitive angiosperms. Pollen & Spores 22: 267-342; 23: 5-36. 1988. Variation de la région aperturale dans le pollen des Annonacées. Taxon 37: 644-656. Levin, D. A. 1979. The nature of plant species. Science 204: 381-384. Lewis, H. 1966. Sahai in flowering plants. Science 152: 167-172. LIEBERMAN, D., M. BERMAN, R. PERALTA, & G. S. HARTSHORN. 1985. Mortality patterns and ates turnover rates ina wel tropical forest in Costa Rica. J. Ecol. 73: 24 Lovejoy, T. E. 1975. Bird diversity and abundance in Amazon forest communities. Living Bird 13: 127-191. Maas, P. J. M., E. C. H. VAN HEUSDEN, J. KOEK-NOORMAN, A. K. VAN SETTEN, & L. Y. T. 238 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 WestRA. 1986. Studies in Annonaceae. VII. New species from the Neotropics and miscellaneous notes. Proc. Kon. Ned. Akad. Wetenschappen, ser. C. 89: 249-278. MasBeRLEY, D. J. 1979. The species of Chisocheton (Meliaceae). Bull. Brit. Mus. (Nat. 6 MacArTHuR, R. H. 1957. On the relative abundance of bird species. Proc. Natl. Acad. 1 Ma cecoT, G. 1959. Les modéles stochastiques en génétique de population. Publ. Inst. Statist. Univ. Paris 8: 173-210. MarTIN, T. E. 1985. Resource selection by tropical frugivorous birds integrating mul- tiple interactions. Oecologia 66: 563-573. MAYNARD SmiTH, J. 1970. Population size, polymorphism, and the rate of non-Dar- winian evolution. Amer. Naturalist 104: 231-237. Mayr, E. 1963. Animal species and evolution. Harvard Univ. Press, Cambridge, Massachusetts. —. 1970. Populations, species, and evolution. Harvard Univ. Press, Cambridge, Massachusetts. Mou.ton, M. P. 1985. Morphological similarity and coexistence of congeners: an experimental test with introduced Hawaiian birds. Oikos 44: 301-305. Nerr, N. A., & L. F. Marcus. 1980. A survey of multivariate methods for systematics. American Museum of Natural History, New No, F. S. P. 1983. oe principles of tropical lowland rain forest conservation. Pp. 359-375 in S. L. Sutton, T. C. Wuitmore, & A. C. CHADWICK, eds., Tropical rain forest: ecology and Sean Blackwell Scientific Publications, London Nig, N. H., C. H. Hutt, J. G. Jenkins, K. STEINBRENNER, & D. H. BENT. 1975. Statistical OxKapba, H., & K. UEDA. 1984. Cytotaxonomical studies on Asian Annonaceae. PI. Syst. Evol. 144: 165-177 PERIASAMY, K., & B. G. L. Gyan 1961. Studies in the Annonaceae II. The devel- opment of ovule and seed in Cananga odorata and Miliusa wightiana. J. Indian Bot. Soc. 40: 206-216 PickeTT, 8S. T. A. 1980. Non-equilibrium coexistence of plants. Bull. Torrey Bot. Club 107: 238-248. Poors, M. E. O. 1968. Studies in Malaysian rainforest. J. Ecol. 56: 143-189. Prance, G. T. 1972. Chrysobalanaceae. Flora Neotropica, Monograph 9. Hafner Pub- lishing Co., New York PULLIAM, H. R. 1985. Foraging sere resource partitioning, and the coexistence of sparrow species. Ecology 66: 1829-1836. Rao, P. R. M. 1975. Seed anatomy of Artabotrys odoratissimus with a discussion on chalaza, integumentary bundles, and ruminate endosperm. Phytomorphology 25: 215-228. . 1979. Seed and fruit anatomy in Polyalthia longifolia (Annonaceae). Bot. Jahrb. Syst. 101: 83-89. 1982. Seed and fruit seas > 4simina triloba, with a discussion on the affinities of Annonaceae. [hid. 103: 47-57. RAvEN, P. H. 1976. Systematics and a population biology. Syst. Bot. 1: 284-316. Reicu, P. B., & T. H. HINCKLEY. 1980. Water relations, soil fertility, and plant nutrient composition of a pygmy oak ecosystem. Ecology 61: 400-416. RicHarps, P. W. 1969. Speciation in the tropical rain forest and the concept of the niche. Biol. J. Linn. Soc. 1: 149-153. Rick.ers, R. E. 1977. Environmental heterogeneity and plant species diversity: a hy- pothesis. ee Naturalist 111: 376-381. Rinco, J., D. op, R. RocKweLi, & H. Dowse. 1985. An experiment testing two hypotheses -. speciation. Amer. Naturalist 126: 642-661. 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 239 Roastab, S. H. 1986. A biosystematic investigation of the Polyalthia hypoleuca com- plex (Annonaceae) of Malesia. Unpubl. Ph.D. thesis, Harvard Univ., Cambridge, Massachusetts —. Inprep.a. The biosystematics and evolution of the Polyalthia hypoleuca species complex (Annonaceae) of Malesia. II]. Comparative spatial distribution ecology. ag ee to J. Trop. Eco In prep. b. The bio osystematics and evolution of the Polyalthia hypoleuca species eae eee of Malesia. IJ. Comparative flowering ecology. To be sub- mitted to Ecology. n prep. c. New chromosome counts in Polyalthia (Annonaceae). To be sub- mitted to Ann. Missouri Bot. Gar & A. LE THomas. In prep. The biosystematics and evolution of the Polyalthia hypoleuca species complex (Annonaceae) of Malesia. IV. Pollen and its phylogenetic implications. To be submitted to Pollen & Spores Rose, M. R., & W. F. Doouitt_e. 1983. Molecular biological hani of speciation. Science 220: 157-1 Rotn, J. L., Jr. 1981. Epidermal studies in srs Annonaceae and related families. Unpubl. Ph.D. thesis, Indiana Univ., Bloomington. SCHOENER, T. W. 1974. Resource Beer ene in ecological communities. Science 185: gi —. 1983. Field experiments on interspecific competition. Amer. Naturalist 122: 240-285. sae hai Ae K. 1970. Behavior and genetic variations in natural populations. Amer. 00 53-66. SETTEN, iG mi VAN, & J. KOEK-NOORMAN. 1986. Studies in Annonaceae. VI. A leaf- anatomical survey of genera of Annonaceae in the Neotropics. Bot. Jahrb. Syst. 108: Soson, G. G. 1961. Principles of animal taxonomy. Columbia Univ. Press, New ee J. 1955. A revision of the Malayan Annonaceae. Gard. Bull. Singapore 14: ae a 1985. Gene flow in natural populations. Ann. Rev. Ecol. Syst. 16: 393- eee P. H., & R. R. Soka. 1973. Numerical taxonomy. W. H. Freeman, San Snow, D. W. 1965. A possible selective factor in the evolution of fruiting seasons in tropical forest. Oikos 15: 274-281. SOKAL, R. R. 1985. The continuing search for order. Amer. Naturalist 126: 729-749. & F. J. Routr. 1970. The intelligent ignoramus, an experiment in numerical taxonomy. Taxon 19: 305-319. & An experiment in taxonomic judgment. Syst. Bot. 5: 341-365. STEBBINS, G. L. 1950. ‘Variation and evolution in plants. Columbia Univ. Press, New York. 1974. Flowering plants: evolution above - species level. Belknap Press of Harvard Univ. Press, Cambridge, Massachuset STEENIS, C. G. G. J. vAN. 1969. Plant speciation in ies with special reference to the theory of non-adaptive saltatory evolution. Biol. J. Linn. Soc. 1: 97-133. 1976. Autonomous evolution in plants. Gard. Bull. Singapore 24: 103-126. . 1978. Patio ludens and extinction of plants. Notes Roy. Bot. Gard. Edinburgh 36: 317-323. . 1981. Rheophytes of the world. Sijthoff & Noordhoff, Alphen aan den Rijn, Netherlands. STEENIS-KRUSEMAN, M. J. vAN. 1950. Malaysian plant collectors and collections. Fl. Males. I. 1: 1-639 240 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 1958. Malaysian plant collectors and collections. Supplement I. /bid. I. 5: CCXXXVH—CCCXI. STERN, K., & L. RocHe. 1974. Genetics of forest ecosystems. Springer-Verlag, New York ork. STEVENS, P. F. 1980. A revision of the Old World species of Calophyllum (Guttiferae). J. Arnold Arbor. 61: 117-699. 1986. Evolutionary classification in botany, 1960-1985. bid. 67: 313-339. STILES, F. G. 1975. Ecology, flowering phenology, and hummingbird pollination of some Costa Rican ae onia species. Ecology 56: 285-301. StronG, D. R., Jr. 1977. Epiphyte loads, tree falls, and perennial forest disruption: a mechanism for maining higher tree species richness in the tropics without ani- mals. J. Biogeogr. 4: 215-21 TANNER, E. V. J. 1982. Species diversity and reproductive mechanisms in Jamaican trees. Biol. J. Linn. Soc. 18: 263-278. TERBORGH, J. 1985a. Habitat selection in Amazonian birds. Pp. 311-338 in M. L. Copy, ed., Habitat selection in birds. Academic Press, New . The vertical component of plant species diversity in temperate and tropical forests. Amer. Naturalist 126: 760-776. VOHRERMEEE, nh ee J. Stout, & G. MiLter. 1974. Growth rates of Welfia georgil, irissima, and Iriartea Ate under various conditions in a natural rain ea in Costa Rica. Principes 18: 148-154. VERDCOURT, B. 1969. The re aa genus Polyalthia Blume (Annonaceae) in Africa. Adansonia, II. 9: 87- Vink, W. 1970. The ee of the Old World. I. Pseudowintera and Drimys— morphology and taxonomy. Blumea 18: 225-354. Wana, M. 1985. Ultrastruktur und systematische Bedeutung des Pollens bei Boca- geopsis, Ephedranthus, Malmea und Unonopsis (Annonaceae). Pl. Syst. Evol. 150: 165-177. WALKER, J. W. 1971. Pollen saul phytogeography, and phylogeny of the An- nonaceae. Contr. Gray Herb. 202: 1-131. WaL.tace, A. R. 1878. Tropical nature. Macmillan & Co., Lon WALLACE, B. 1985. Reflections on the still “hopeful monster.” aes Rev. Biol. 60: 31-42 WENT, F. W. 1973. Competition among plants. Proc. Natl. Acad. U.S.A. 70: 585-590. WERNER, P. 7 1979. Competition and coexistence of similar species. Pp. 287-310 in _ T. RIG, S. Jatin, G. B. JoHNson, & P. H. RAvEN, eds., Topics in plant sopulation biology. Columbia Univ. Press, New York. WHEELWRIGHT, N. T. 1985. Competition for dispersers, and the timing of flowering and fruiting in a guild of tropical trees. Oikos 44: 465-477 WHIFFIN, T. 1978. Geographical variation in tropical tree species. Pp. 31-54 in P. B. TOMLINSON & M. H. ZIMMERMANN, eds., Tropical trees as living systems. Cambridge Univ. Press, Cambridge, England Wuite, F. 1962. Geographic variation and speciation in Africa with particular reference to Diospyros. Pp. 71-103 in D. NIcHOLs, ed., Taxonomy and geography. Systematics Association, London. WHITMORE, T. C. 1984. Tropical rain forests of the Far East. ed. 2. Clarendon Press, Oxford. WHITTAKER, R. H. 1967. Gradient analysis of vegetation. Biol. Rev. 42: 207-264. Witey, E.O. 1981. Phylogenetics: the theory and practice of phylogenetic systematics. John Wiley and Sons, New York. WILKINSON, L. 1984. SYSTAT manual. Systat, Inc., Evanston, Illinoi WILLIAMS, W. 1964. Genetical principles and plant breeding. Blackwell’ s Scientific Publications, Oxfor Wona, M. 1983. Understory phenology of the virgin and regenerating habitats in Pasoh Forest Reserve, Negeri Sembilan, West Malaysia. Malaysian Forester 46: 197-223 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 241 Won, Y. K., & T. C. WuitmMore. 1970. On the influence of soil properties on species distribution in a Malayan lowland dipterocarp rain forest. Malayan Forester 33: 42- 54 WRIGHT, S. 1931. Evolution in Mendelian populations. Genetics 16: 97, 98. WRIGHT, S. J., & S. P. HusBett. 1983. Stochastic extinction and reserve size: a focal species approach. Oikos 41: 466-476. Yap, S. K. 1982. The phenology of some fruit tree species in a lowland dipterocarp forest. Malaysian Forester 45: 21-35. APPENDIX 1. Specimens examined for leaf papillae* and for seeds with spiniform endosperm ruminations.+ Anaxagorea javanica Blume var. a Corner: Elmer 21131 (A). A. luzonensis Gray: Ramos BS 13623 Annona muricata L. and A. squamosa L.: specutiens from living collections at Fairchild Tropical Garden, Coral Gables, Florida. Artabotrys harmandii Finet & Gagnep.: Pierre 423 (A). A. siamensis Miq.: Rogstad 959; Sargent s.n., 16 Oct. 1903 (mo). A. suaveolens Blume: Maxwell 81-206 (mo); Rogstad 831. A. trichopetalus Merr.: Elmer 20489 (a). A. zeylanicus Hooker f. & Thomson: Saldanha 16365 (A); Waas 934 (A). Asimina triloba (L.) Dunal: specimen in living collection at Arnold Arboretum, Jamaica Plain, Massachusetts. Bocageopsis canescens (Spruce) R. E. Fries: Prance et al. 5929, 8792 (both Mo). B. mattogrossensis (R. E. Fries) R. E. Fries: Maas et al. 6242 (Mo B. multiflora (Martius) R. E. Fries: Prance et al. 22714, 25798 (both Mo); Silva & Rosario 49 (A Cleistochlamys kirkii (Bentham) Oliver: Pereira & Correia 2351 (mo); De Carvatho 1045 Cleistopholis glauca Pierre ex Engler & Diels: Leeuwenberg 6467 (mo); Louis 12459, 0). M C. patens (Bentham) Engler & Diels: Darko 732 (Mo), Louis 7993 (Mo). C. staudtii Engler & Diels: Bates 1568 (mo). Cremastosperma anomalum R. E. Fries: Correa & Dressler 887, 1060 (both Mo); Stern et al. 107 (A). C. cauliflorum R. E. Fries: Bristan 1341 (mo); Diaz & Jaramillo 1229 (mo); Lleras et al. 16879 (Mo). C. macrocarpum Maas: Liesner & Gonzalez se (MO). C. pedunculatum (Diels) R. E. Fries: Klug 372 6 (A). Desmos cochinchinensis Lour. and D. d. lus (BI ) Safford: speci in living collections at Kebun Raya, Bogor, Indonesia. Enantia chlorantha Oliver: Bates 1959 (mo); Leeuwenberg 7355 (Mo); Zenker 44] (Mo). E. polycarpa (DC.) Engler & Diels: Baldwin 10696 (Mo). *See also APPENDIX 2 most species several specimens have been examined, but usually only one representative collection is noted. Rogstad collections are deposited at a 242 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 Ephedranthus amazonicus R. E. Fries: Prance et al. 11555, 18787 (both mo). 2. guianensis R. E. Fries: Gentry et al. 29075 (mo). E. pisocarpus R. E. Fries: Schatz et al. 732 (Mo). Goniothalamus giganteus Hooker f. & Thomson: Rogstad 985, G. tortilipetalus Henderson: Rogstad 566. Lettowianthus stellatus Diels: Seuisei S1027 (mo). Malmea depressa (Baillon) R. E. Croat 74646 (mo); Lundell 4852 (Mo). M. obovata R. E. Fries: Riedel s.n. (A). sh ee Engler a Diels: De Wilde 1942 A, 2735 (both mo); Zenker 02 Teenie fruticosum Lour.: Scortechini 1946 (SING). Monocarpia marginalis (Scheffer) James Sincl.: De Vogel 4492 (mo). Monodora myristica Dunal: specimen growing at the Forest Research Institute, Kepong, Selangor, Malaysia. Oncodostigma monosperma (Hooker f. & Thomson) James Sincl.: Rogstad 920. Oncodostigma sp.: Rogsta Pi aaa a Arukoffii R. E. res: Oe 5326 (MO). O. lanceolatum R. E. Fries: Krukoff 6909 (mo). 0. oa R E. Fries: Krukoff 8214 wo) andra acuminata Diels: Croat 19729 (mo). O. "asbeckit (Pulle) R. E. Fries: Mori & Boom 15170 (mo). O. eneura Diels: Croat 18738 (Mo). O. espintana (Spruce) Baillon: Gentry, Aronson, & Ramirez 26744 (MO). O. laurifolia Rich.: Duss 4180 (mo). O. leucodermis (Spruce) Warm.: Liesner 6983 (Mo); Liesner & Clark 8949 (Mo). O. longipetala R. E. Fries: Holdridge 6313 - O). O. xvlopioides Diels: Huashikat 1257 (mo haeanthus sp.: Rogsta Bes a Li fasciculata (Wilde) Boutique: Germain 2396 (Mo); Toussaint 2151 (Mo). P. glabrescens Oliver: Thomas 510 (mo a) Q = = = = Q a = 3 cas) .3 ima, = 2) = £2 . Oo a TR a Q & Polyalthia angustissima Ridley: specimen in living collections at Singapore Botanical Garden arden P. br unneifolia James Sincl.: James Sinclair 24179 (a). P. bullata King: FRI 11643 (KEP) P. capuronii Cavaco & Keraudren: Capuron SF 11795 (p). P. cf. cauliflora Hooker f. & Thomson: Rogstad 997. P ; P. emarginata Diels: Capuron 20977-SF (p). P. gigantifolia Merr.: Ramos & Edario 36966 (SING P. glauca (Hassk.) Mueller: Rogstad 939, 944, 945, 966. P. gracilis Burck: Versteeg 1408 (Bo). P. heteropetala Diels: Ghesquier 4942 (P). P. hypoleuca Hooker f. & Thomson: Rogstad 912, 916, 942. P. jenkinsii (Hooker f. & Thomson) Hooker f. & Thomson: Rogstad 968. P. lateriflora (Blume) King: Rogstad 931. P. longifolia (Sonn.) Thwaites: Rogstad 960. P. macropoda King: Kostermans 7360 (SING). P. michaelii C. T. White: NGF 77111 (LAE). P. multinervis Diels: Rogstad 813, 817-819, 829. P. obliqua Hooker f. & Thomson: Rogstad 935. 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 243 oblongifolia Burck: Ramos 1595 (A). oligosperma (Danguy) Diels: Thouvenot 62 (P). oliveri Engler & Diels: Baldwin ae (mo); Enti 1701 (mo); Oldeman 856 (Mo). ovalifolia Rogstad: Buwalda 7797 pedicellata A. C. Smith: A. C. Smith 67 (BO). rumphii (Blume) Merr.: Rogstad 922. sclerophylla Hooker f. & Thomson: Rogstad 930. stenopetala (Hooker f. & Thomson) Ride: Dev 165 (KU). panna: (Engler) Verdc.: Harris 3635 suaveolens Engler & Diels: Gilbert 8500 ( ee Leeuwenberg 5082, 7322 (both Mo). suberosa (Roxb.) Thwaites: ae from living collections at Fairchild Tropical Garden, Coral Gables, Flori P. sumatrana (Miq.) Kurz: eal 508, 509, 527. P. thorelii Finet & Gagnep.: Pierre 1506 (SING). Polyalthia sp.: Rogstad 987. Polyceratocarpus gossweileri (Exell) Paiva: Breyne 752 (Mo). P. microtrichus (Engler & Diels) Ghesq. ex Pellegrin: Zenker 478 (Mo). P. parviflorus (Engler & Diels) Ghesq.: Bos 6684 (Mo); Jacques-Georges 16832 (MO). Pseudephedranthus fragrans (R. E. Fries) Aristeg.: Maguire et al. 60189 (mo). Pseudoxandra coriacea R. E. Fries: Prance, Steward, Ramos, & Pinheiro 11468 (mo). P. cuspidata Maas: C. C. Berg et al. 757 (mo); Prance & Sylva 58673 (MO). P. guianensis R. E. Fries: A. C. Smith 2665 (mo P. lucida R. E. Fries: Stevermark & Delascio 129373 (Mo). P. polyphleba (Diels) R. E. Fries: Gentry, Vasquez, Andrade, Horna, & Stern 28807 (Mo); Krukoff 8409 (Mo). Pseuduvaria spp.: NGF 77112 (Lak); Rogstad 924. Ruizodendron ovale (Ruiz & Pavon) R. E. Fries: Begazo 61 ee Klug 3798 (mo). Stelechocarpus cauliflorus (Scheffer) R. E. Fries: Rogstad 998 Unonopsis floribunda Diels: Gentry 7452 (Mo). panamensis R. E. Fries: Johnston 1643 (Mo U. pittieri Saff.: Contreras 10031 (mo); Lent 2289 (MO). U. spectabilis Diels: Maas et al. 6229 (mo). Uvaria boniana Finet & Gagnep.: W. r. Tsang 23823 (MO). U. calamistrata Hance: Lau 468 (mo). U. confertiflora Merr.: Elmer 21081 (A). U. grandiflora Roxb.: Rogstad 629. U. javana Dunal: Elmer 20857 (mo). U. littoralis Blume: De Vogel 3895 (Mo). U. lucida Bentham: Faden 74/1251 (mo), Pawek 12051 (mo). a mendesii J. Paisa: Mendes 639 (Mo). U. muricata Pierre & Engler: Gentry & Pilz 32795 (Mo). U. osmantha Diels: Mendes 673 (mo U. ovata A. DC.: Jacques-Georges 5850 (Mo). U. poggei Engler & Diels: Robyns 4227 (mo). U. sabrida Oliver: Gossweiler 10417 (Mo). U. sofa Elliot: Jacques-Georges 14602 (mo). Xylopia elliptica Maingay ex Hooker f. & Thomson: Yeob 5037 (SING). X. ferruginea ere f. & Thomson) Hooker f. & Thomson: Rogstad 703. Xylopia sp.: Rogstad 816. Unidentified oe Rogstad 928, 961, 969. Lo 244 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 APPENDIX 2. Species of Polyalthia examined for leaf papillae* and for thickened, straight anticlinal walls in adaxial epidermal peels.t P. amygdalina (A. Gray) ose E. H. Bryan, Jr. 509. P. borneensis Merr.: Elmer 21733 P. cauliflora Hooker f. .& Thomson (sensu lato): Kiah 32101; Mohd. Shah 167; S 16620. P. ee Miq.: bb J P. cinnamomea Ho ne ee & Thomson: Henderson 35168. P. clemensorum Ast: Poilane 29657. P. consanguinea Merr.: Chun & Tso 44029. P. corticosa Finet & Gagnep.: Poilane 19623. P. debilis Finet & Gagnep.: Pierre 289. P. evecta (Pierre) Finet & Gagnep.: Poilane 17313. mlerantha Boerl.: Buwalda 7241]. microtus Migq.: SAN A modesta Finet & Gagnep.: Poilane 17782. motleyana (Hooker f.) Airy-Shaw: Chew Wee Lek 1008. nemoralis DC.: Pételot 5955. obliqua Hooker f. & Thomson: Hallier . oligosperma (Danguy) Diels:** ue an . oliveri Engler: Thomson 109. parviflora Ridley: Poilane 704. . plagioneura Diels: Poilane 29564. . rumphii (Blume) Merr.: FRI 14190. simiarum (Hooker f. & Thomson) Hooker f. & Thomson: Khant D.R. 1174. - stenopetala (Hooker f. & Thomson) Ridley: Kiah 36/62. suaveolens Engler & Diels:** Kennedy 1544. subcordata (Blume) Blume: Korthers s.n. . suberosa (Roxb.) ae as 5185. . tenuipes Merr.: Richards 2 P. venosa Merr.: Ramos & ae 36558. Se ae es 8 Se a "Oe *See also APPENDIX |. le similar in structure and distribution to those of the Polyalthia hypoleuca complex were lacking in all species listed; the two species with similar thickened anticlinal walls are marked with double ene All specimens are at A APPENDIX 3. Characters used in the morphometric analyses of the Polyalthia hypoleuca complex.* I. Floral and vegetative characters. A. Continuous (measured; in mm unless noted) and meristic (counted) characters used in multivariate data sets. 1. Androecium width. 2. Torus width *For measurements of mature leaves, an average-size leaf on older growth was subjectively chosen. 1989] ROGSTAD, POLYALTHIA HYPOLEUCA 245 Pedicel distal width. Stamen number. Carpel number. Inner-petal length. ORI AA RY 10. Stamen length. 11. Gynoecium width. 12. Torus height. 13. Inner-petal eae 14. Outer-petal w 15. Stamen inner nee length. 16. Carpel length 17. Pedicel length. 18. ee es gynoecium emergence above androecium. 9. Flo mber per leaf axil. 20. es ee distal peduncle articulation to base of pedicel bract. 21. Stigma length. 22. Outer theca length. 23. Peduncle length. 24. Stamen width. 25. Lamina length (in cm). 26. Lamina maximum width (in 27. Leaf drip-tip length (measured from inflection point of distal lamina border and border of drip-tip; in cm). . Leaf primary vein number (counting only largest class of veins branching directly from mid-vein). 29. Petiole length. B. Characters used in conjunction with results of multivariate analyses. Number ers. — No co le of latest flush of vegetative growth bearing flow 2. Flowers in ae of present leaves = 1; of abscised leaves = 2. 3. Tree height (in 4. Bole diameter at he height (in cm). 5. Inflorescence type: flowers with reduced peduncles (< 2 mm long), lateral buds and flowers lacking, large tubercles built up of old inflorescence scars lacking = 1; similar to 1, but tubercles present = 2; peduncles generally longer than 2 mm, lateral buds (and/or developing lateral flowers) and tubercles present = 6. Peduncle bract number 7. Lowest peduncle ee nosiion soe respect to stem bearing inflorescence: abaxial = 0; adaxial = 1; abaxial and adaxial on same specimen = 2; all angles (as on tuber a )= 3. 8. Medial pedicel bract: present = 0; lacking = 1. 9. Pedicel color: tan-brown = 1; dark red-black = 10. Peduncle and pedicel pubescence: lacking = 0; present only on peduncle = 1; tomentum = 11. Sepal pubescens. (abaxial surface): lacking = 0; present on margins only = 1; present over entire surface = 12. Outer petal malas no ean tion, linear = 1; basal portion clawed, distal penuon linear = 2; basal portion clawed, distal portion obovate or flared = no ee hee obovate or elliptic = 13. Portion of inner petal i aaa begins | on adaxial surface: basal '4 = |; middle 4 = 2; distal 4 = 3. JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 14. Color of adaxial, glabrous (basal) portion of inner petal: light tan-brown = 1; brownish red to wine red = 2; dark ads) black = 3; deep purple = 4. 15. Inner petal pubescence: sparse throughout = |; moderate throughout = 2; dense throughout = 3; sparse throughout except dense on abaxial surface at 16. Mid-vein of outer petal on abaxial surface: not detectable = 0; easily seen 17. Gynoecium emergence: all tig dep d below androecium = 0; stigmas level with androecium = |; stigmas Leeks androecium = 2: stigmas and portion of carpels emergent above androec ium = 3. 18. Stigmas: central ones Gepiessoy = = |; all level = a central ones raised = 3; all separate, radiating 19. Gynoecial area oPiorus concave = 0; level = 1; convex = 20. Apices of stamen connectives: concave = 0; flat = 1; mixed = 2. 21. Inner thecae of stamens: not touching = 0; touching = 1. 22. Color of carpel: all tan = 0; distal, pape surface deep red-brown, becoming 5 tan basally = 1; all red-brown = 2; all dark reddish black = 23. Carpel pubescence: lacking = 0; oresehi on abaxial apex only = 1; present on abaxial apex with some on ridged sides and/or stigmatic crest = 2; present throughout = 3. 24. Shape of stigma: spherical = 0; elongate = 1. 25. Outer ring of carpels with very strongly curved carpels: no = 0; y 26. pubescence on youngest stems: lacking = 0; sparse = 1; fc Gc heavy Dh Pubescen on petiole of smallest leaf: lacking = 0; sparse = 1; moderate to heavy = I]. Fruit and ee characte A. Continuous (in mm ee ota) and meristic characters used in multivariate data sets. 1. Pedicel length. 2. Pedicel distal width. 3. Mature carpel stalk len 4. Mature carpel stalk distal width. 5. Mature carpel ee including only seed-bearing portion above stalk. 6. Mature carpel ia 7. Lamina length (in m). mina maximum nen (in cm), 9. Leaf drip-tip length (measured from inflection point of distal lamina border and border of drip-tip; in cm). 10. Leaf primary vein number (counting only largest class of veins branching directly from mid-vein). B. Characters used in conjunction with multivariate data analyses. —_ : jee pubescence: lacking = 0; sparse or irregular = 1; dense tomentum 2. ae carpel apex: rounded or blunt = 0; acute = 1. 3. Dried mature carpels wa longitudinal ridges: no = 0; yes = 1. 4. Stamen scars: < 40 = 0; > 40 = 1. 5. Tree height (in m). 6. Bole ieee at breast height (in cm). 1989] PRICE, PINACEAE 247 THE GENERA OF PINACEAE IN THE SOUTHEASTERN UNITED STATES! ROBERT A. PRICE? PINACEAE Lindley, Nat. Syst. Bot. ed. 2. 313. 1836, nom. cons. (PINE FAMILY) Evergreen [or deciduous], monoecious trees (occasionally shrubs), usually pyramidal in form. Leaves linear to acicular [rarely lanceolate], spirally ar- ranged (often apparently 2-ranked), sessile or short-petiolate; foliage leaves either borne on long-shoots or clustered tightly on short-shoots. Pollen cones (microsporangiate strobili) bearing spirally arranged, bilaterally symmetrical microsporophylls; microsporangia 2 on the abaxial surface of each sporophyll, nonsaccate in Larix and Pseudotsuga]; prothallial cells 2. Ovulate cones with ic Fl fthe Soutl United States, a long-term ee made a TOT the AJOL fous the support of National Science Foundation Grant BSR-8415769 (C. E. Wood, Jr., cipal investigator), under which this account was prepared, and BSR-8415637 (N. G. Miller, ener inv ee The 127th in ues series, uae paper follows the format established in Br first one (Jour. Arnol b. 39; 296-346. d to the present. The area covere d by the Generic Flora ene er and South Carolina, ee Florida, Tennessee, Alabama, nee Arkan and Louisiana. The descriptions are based primarily on the plants of this area, with info Sr ation about extraregional members of a family or genus in brackets. The references that I have not verified are marked with asterisks. I thank Carroll Wood and Norton Miller for the opportunities afforded by participation in the Generic Flora project and for their guidance and editorial help in the study. I am grateful to Ihsan °y, this study, and I wish to thank the librarians and curatorial staffs of these institutions. Elizabeth Schmidt and Stephen Spongberg added their editorial expertise to the manuscript. The illustrations were drawn by Karen Stoutsenburger in 1977 under the supervision of Carroll Wood, who also made the dissections. Living, liquid-preserved, and some of the dry materials used were variously collected by Walter Judd, i z include an unattributed sheet [A] with mature seeds of Pinus Strobus utilized by C. E. Faxon for the illustration in Sargent’s Si/va of North America, as well as specimens collected by Adams & Wherry (2252 [GH]), R. M. Harper (2023, 2027 [Gu]), and J. D. Smith (s.., Aug. 1881 [GH]). The diagrammatic cross sections of leaves of Pinus are after J. W. Lanyon’s A Card Key to Pinus Based on Needle Anatomy. The pollen grain of P. Strobus is after a photograph ae by Norton Miller. 2University Herbarium, Depart ment of Botany, University of California, Berkeley, California 94720. ase address reprint requests % Dr. Norton Miller, Room 3132 CEC, New York State Education ae Albany, New York 12230. © President and Fellows of Harvard College, 1989. Journal of the Arnold Arboretum 70: 247-305. April, 1989. 248 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 70 spirally arranged, strongly flattened bract-scale complexes; bracts included or exserted, separate from the scales for most of their length; mature ovuliferous scales relatively thin to strongly thickened and woody (in Pinus); ovules 2 on the adaxial surface of each scale, the micropyles directed toward the cone axis; archegonia few per ovule, not clustered. Seeds with an elongate terminal wing partially folded around the seed body [wing short or absent in some species of Pinus]; embryo straight, cotyledons 2—1 2{—18]. Chromosome number 2” = 24 [26 and 44 in one species each]. (Including Abietaceae Walpers; Abietineae Rich.) Type GENUS: Pinus The largest and most economically important family of conifers, with ten genera and approximately 220 species, the Pinaceae are restricted almost en- tirely to the Northern Hemisphere? both at present and as fossils (Florin, 1963). Three small genera are confined to eastern Asia: Kefeleeria Carr., a genus of perhaps nine species (Cheng & Fu) much in need of further study, and the monotypic Cathava Chun & Kuang (described in 1958 from southern China) and Pseudolarix Gordon (golden larch, false larch). Cedrus Trew, comprising the true cedars, consists of four closely related species occurring disjunctly in North Africa, the Near East, Cyprus, and the Himalayas and is widely cultivated in temperate areas, including the southeastern United States. All six remaining genera are represented by species native to the United States. Only two of them have no species indigenous to our area: Larix Miller, larch, with about ten species in northern North America and Eurasia, and Pseudotsuga Carr., Doug- las fir, with seven or eight species in western North America and eastern Asia. The family Pinaceae is a well-defined natural group, distinguished particu- larly by characters of its ovulate cones. These have well-developed scales that are distinct from the subtending bracts, two inverted ovules per scale, and a prominent terminal seed wing developing from the cone scale. It 1s supported as a monophyletic group by shared features unique among the gymnosperms, including the pattern of proembryogeny (Doyle, 1963; Singh), protein-type sieve-cell plastids (Behnke), and the absence of biflavonoid compounds (Geiger & Quinn). On the basis of morphological comparisons (e.g., Eckenwalder; Hart), im- munological comparisons of seed-protein extracts (Price & Lowenstein), and phloem polypeptide profiles (Alosi & Park), there is no evidence that the Pi- naceae are closely related to any of the other extant families of conifers. Rather, the family has a long fossil history as a distinct lineage, probably dating to the Triassic (Miller, 1977), as is also the case for such other major extant coniferous groups as the Araucariaceae, the Podocarpaceae, and the Cupressaceae-Tax- odiaceae lineage. The combination of apparently SaeyeS cae (bisaccate pollen with two prothallial cells, chromosome number = 24, and lack of bract-scale fusion) and uniquely derived features ache of the Pinaceae suggests that its origin should be sought among early conifer groups preceding the extant families (see Florin, 1951). Thus a number of European workers sOne species, Pras Merkusti Jungh & De Vries, is native from Indochina south of the Equator to Sumatra 1989] PRICE, PINACEAE 249 (e.g., Emberger; Lebreton & Sartre; Norin) have placed the Pinaceae in the unifamilial order Pinales. The monotypic Sciadopitys Sieb. & Zucc., Japanese umbrella pine (Scia- dopityaceae), has sometimes been associated with the Pinaceae on the basis of its unusual “double-needles.” These superficially resemble longitudinally fused pairs of Pinus needles but in fact differ from them considerably in pattern of development (see Roth). Other morphological and cytological evidence suggests that Sciadopitys diverged at an early time from the cupressaceous-taxodiaceous lineage, since Sciadopitys—as well as the Cupressaceae sensu Jato (including Taxodiaceae)—differs from the Pinaceae in having nonsaccate pollen without prothallial cells, substantial fusion of bract and scale in the mature ovulate cone, multiple ovules per scale, and lateral seed wings derived from the seed coat (Florin, 1951; Singh; Sporne). In its proembryogeny Sciadopitys appears to be more primitive than either group (Doyle & Brennan; Singh). Immuno- logical comparisons of seed-protein extracts (Price & Lowenstein) and overall patterns of secondary-product chemistry (Hegnauer, 1962, 1986) also indicate that Sciadopitys is quite isolated from the Pinaceae. The ten genera of Pinaceae are clearly delimited, but subfamilial groupings have been controversial. Pinus, in recognition of its unusual short-shoots (needle fascicles) and its distinctive, usually highly thickened cone scales, is often placed by itself in subfam. Pinoideae. Vierhapper and a number of later authors (e.g., Florin, 1931, 1963; Kriissmann; Pilger & Melchior) divided the remaining genera into two groups based on presence or absence of strongly condensed vegetative short-shoots that bear the majority of the foliage leaves. The genera with marked shoot dimorphism (Cedrus, Larix, Pseudolarix) have been placed in subfam. Laricoideae Pilger & Melchior, with the remaining genera (long- shoots only) in subfam. Abietoideae Ascherson & Graebner emend. Pilger & Melchior or in equivalent groups of lower rank. The relatively recently dis- covered genus Cathaya, which has less-marked shoot dimorphism (see photo in Chun & Kuang, 1958), was placed by Florin (1963) in the Abietoideae. However, other morphological characters show little concordance with shoot dimorphism, so division of the family on this basis alone is highly artificial. Another subfamilial division, first proposed by Van Tieghem (1891), placed together those genera with resin canals in the primary root adjacent to each protoxylem pole (Pinus, Picea, Larix, Pseudotsuga (and Cathaya; see Hu & Wang), comprising the Pinées or Epixylocéles of Van Tieghem, here termed the pinoid group) and those with a single central resin canal in the primary root (Cedrus, Tsuga, Abies, Keteleeria, and Pseudolarix, comprising the Cedrées or Myélocéles of Van Tieghem, here termed the abietoid group). This classi- fication appears to be a natural one, based on the concordance of root anatomy with other morphological features. Each of the pinoid genera regularly exhibits both vertical and horizontal resin canals in the wood and lacks resin vesicles in the seed coat, while the abietoid genera have much more localized resin canals in the wood following wounding (“traumatic resin canals,” usually in the vertical system) and have resin vesicles in the seed coat (Jeffrey; Phillips; Price et al.). Immunological comparisons of seed-protein extracts have also yielded precisely the abietoid and pinoid groups of Van Tieghem (Prager et al.; 250 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Price et a/.). Thus recognition of two subfamilies, the Pinoideae and the Abie- toideae, corresponding to the groups of Van Tieghem—or of three subfamilies, with Pinus placed in a monogeneric Pinoideae and the remaining pinoid genera in the Laricoideae—seems to be the most reasonable alternative. Within the pinoid group, Larix and Pseudotsuga appear to be closely related sister-genera on the basis of their shared possession of nonsaccate pollen, a highly modified micropylar apparatus at time of pollination, fiber-sclereids in the bark, and similar, relatively asymmetric karyotypes, all of which seem to be derived characters within the family. They also group together in immu- nological comparisons of seed proteins (Prager ef a/.; Price et al.). Chromosome counts have been obtained for all ten genera and approximately 75 percent of the species of Pinaceae (see particularly Khoshoo, 1961). The number 2n = 24, which is apparently primitive for the conifers,* has been found in all but two species (Pseudotsuga Menziesii (Mirbel) Franco, 2 = 26, and Pseudolarix amabilis (Nelson) Rehder, 2n = 44). Counts are available for six of the remaining seven species of Pseudotsuga; all have 2n = 24 (Doerksen & Ching; El-Kassaby ef a/.). Karyotypes have been obtained for all genera of the Pinaceae (see particularly Chu & Sun; Khoshoo, 1962; Kuo et al.; Sax & Sax; see also Saylor, 1964, 1972, 1983, cited under Pinus); they vary from largely symmetrical (11 of 12 chromosomes metacentric or submetacentric in Pinus, Cathaya, and Cedrus; nine of 12 metacentric in Picea and Tsuga) to extremely asymmetric in Pseudolarix. Pseudotsuga and Larix both have asym- metric karyotypes—six metacentric and six subtelocentric chromosomes, with one metacentric chromosome replaced by a subtelocentric and an additional short telocentric chromosome in P. Menziesii. Pseudolarix amabilis has a par- ticularly unusual karyotype, with two metacentric and 20 shorter subtelocentric chromosomes (Sax & Sax), implying an aneuploid increase involving fission of ten of the original 12 chromosomes (Ehrendorfer; Khoshoo, 1959). Given their long fossil record, the Pinaceae have shown remarkably little change in chromosome number, a situation shared with the other conifer families except the Podocarpaceae (Ehrendorfer; Khoshoo, 1962), but one differing markedly from most angiosperm groups. Individuals or cells with polyploid chromosome numbers have occasionally been reported in the Pinaceae (Khoshoo, 1959), but polyploid plants apparently seldom survive under natural conditions. The Pinaceae are characterized by a specialized variant of the coniferous pattern of proembryogeny (Dogra, 1980; Doyle, 1963; Singh). Two free-nuclear mitotic divisions yield four nuclei that come to lie in a more or less planar arrangement. A third mitotic division yields two tiers of four nuclei that rapidly become separated by transverse walls. The Pinaceae differ from the other conifers in that the lower of the two tiers produced by vertical division of the original upper or ‘‘open” tier does not elongate to form the suspensor. Instead, the tier of cells below this, produced by vertical division of the lowest original tier, forms the suspensor, while only the lowest of the four tiers forms the embryonal mass. In Pseudotsuga (at least in P. Menziesii) apparently only the 4This number is also found in Taxus (Taxaceae), Cephalotaxus (Cephalotaxaceae), some Podocar- paceae, and the outgroup Ginkgo L. (see discussions in Ehrendorfer, and Khoshoo, 1962). 1989] PRICE, PINACEAE 251 lower of the original two tiers divides, yielding a three-tiered proembryo (Allen & Owens; Singh). Although only a single embryo usually matures in each seed, polyembryony is characteristic of many conifers, including all genera of Pinaceae. Simple polyembryony, where embryos are formed in more than one archegonium per ovule, occurs relatively frequently. Cleavage polyembryony, where multiple embryos are formed by the splitting up of the products of a single zygote, is also a regular feature in Pinus, Cathaya, Cedrus, Tsuga, and Keteleeria (Doyle & Brennan; Wang & Chen). It also occurs in a limited percentage of embryos in some species of Abies but has not been found in Pseudolarix, Picea, Larix, or Pseudotsuga. In the last two genera, unequal growth rates often cause two of the units to overtop the others (although the four embryonal units do not separate from one another, a situation termed “incipient cleavage’”’ by Doyle & Brennan). The Pinaceae have the highest average number of cotyledons and the greatest range in this number of any family of seed plants (Butts & Buchholz). The highest numbers are found in Pinus, which has the entire range of numbers for the family (2-18, with intraspecific averages ranging from ca. 4 to 14), and in Cedrus (5-14, averaging ca. 9 or 10). Detailed discussions of the vascular- bundle patterns in the cotyledons and embryos are presented in Boureau (1939) and Ferré (1952). Ovulate-cone anatomy in the Pinaceae has been studied by Aase, Miller (1976), Radais, and Van Tieghem (1869), with emphasis on distribution of the vasculature and resin canals. Pinus appears to be unusual in having partial fusion of the bract and scale traces at their bases. The arrangement of the resin canals associated with these traces seems to be fairly specific to each genus (Miller, 1976; Radais), with those in the cones of Pinus restricted to the area abaxial to the scale trace, perhaps because of the basal fusion of the traces. The axial vascular cylinder of the cone is reduced in thickness and much dissected in the genera with abscising cone scales (Abies, Cedrus, Pseudolarix; Miller, 1976). Pollen grains are bisaccate in the Pinaceae, except in Larix and Pseudotsuga, where they are spheroidal and nonsaccate, and Tsuga sect. TSUGA, where the saccae are reduced to a subequatorial frill (G. Erdtman, 1957, 1965; Ueno). Saccate pollen is evidently the primitive state both in the Pinaceae and among the conifers generally (see Florin, 1951; Millay & Taylor), but the saccae have been lost in all other extant coniferous groups except the Podocarpaceae, where pollen may be bi-, tri-, or nonsaccate. The Pinaceae are wind pollinated, as are the other conifers, and often release huge volumes of pollen. Specific mechanisms of pollination have been reviewed by Dogra (1964), Doyle (1945), and Singh and have recently been elegantly demonstrated through scanning electron, as well as light, microscopy for a number of western North American taxa by Owens and coworkers (Allen & Owens; Owens & Molder; and additional references cited under the individual genera). A pollination-drop mechanism has been substantiated only in Pinus and Picea and is lacking in the other genera examined to date (Cathaya, Ke- teleeria, and Pseudolarix are unstudied in this regard). It is found in most other 252 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 gymnosperms and is presumably the primitive state for the Pinaceae (see Doyle, 1945; Singh). In Pinus and Picea, pollen is caught on a sticky film on the inner surfaces of the two integumentary arms and, when contacted by the pollination drop, is then drawn into the micropylar canal, after which it floats or is pulled up to the nucellus (Owens ef al., 1981, 1987, cited under Pinus and Picea, respec- tively). Pseudotsuga and Larix have a highly specialized mechanism in which the unwinged pollen is caught between hairs on the highly expanded outer- integument lip and then drawn into the micropylar canal by collapse of the surface of the lip (Allen & Owens; Owens & Molder). In Cedrus and Abies, pollen is caught on the inner surface of the funnel-shaped integumentary area around the micropyle, and the nucellus grows out to the pollen (Doyle, 1945; Owens & Molder, 1977, cited under Abies). In Tsuga Mertensiana (Bong.) Carr., of sect. HESPEROPEUCE, the situation is similar, but the integument tip is very asymmetric; in sect. TsuGA the pollen lands primarily on the bract surfaces rather than the ovule tip, and then the pollen tubes grow over to the nucellus (Doyle & O’Leary). This unusual situation is otherwise seen only in the Araucariaceae and in Saxegothaea Lindley (Podocarpaceae), which also have unwinged pollen (Doyle, 1945). Doyle (1945) related the pollination mech- anisms of Pinaceae to the groupings of Van Tieghem (1891) in an evolutionary diagram, showing series leading from Pinus and Picea to Larix and Pseudotsuga on the one hand, and Cedrus, Abies, and Tsuga on the other. Cathaya might be expected to have a mechanism similar to that of Pinus and Picea, and Keteleeria and Pseudolarix to have ones similar to those of the other abietoid genera. A characteristic feature of the Pinaceae is the formation of ectomycorrhizal associations with a variety of fungi, including a number of basidiomycetes, such as Armillaria matsutake and Boletus edulis, that are highly sought for their edible mushrooms. The presence of mycorrhizae appears to help buffer the plants against various types of environmental stress, and thus mycorrhizal associations have been much investigated in regard to commercial tree pro- duction. Because of the economic importance of the Pinaceae, numerous studies of the chemical composition of the leaves, wood, and bark have been made on many species (see reviews by Hegnauer, 1962, 1986; Norin; Squillace; Von Rudloff), with particular emphasis on Pinus, Picea, Abies, and Pseudotsuga. The chemistry of Cedrus has recently been reviewed by Agrawal & Rastogi. Significant gaps remain in our knowledge of the chemosystematics of even the best-known genera, however. The Pinaceae contain a diversity of terpenoid compounds in the bark, wood, leaves, and cones, particularly in the characteristic oleoresins of the resin canals or vesicles. The bulk of the volatile portion of the stem and leaf oleoresins is usually a complex mixture of monoterpenes. These impart much of the char- acteristic fragrance associated with Pinaceae. Terpenoid and hydrocarbon pro- files of the oleoresins or turpentines (the steam-distillable portion thereof) often show significant differences among species and have been widely used in che- mosystematics of the Pinaceae (see particularly the reviews by Squillace and 1989] PRICE, PINACEAE 258 Von Rudloff and the papers by Mirov, Von Rudloff, Zavarin, and their co- workers cited under the individual genera). Sesquiterpenoid compounds have received much less attention than mono- terpenoids, although the former may also be of considerable chemosystematic utility. Cedrus is particularly notable for its array of unusual structural types including the atlantones, himachalenes, and allohimachalenes (Agrawal & Ras- togi; Norin). Diterpene resin acids are major constituents of the “rosin” remaining after removal of the volatile components of pinaceous oleoresin by steam distillation (Hegnauer, 1962, 1986; Norin). Resin acids of the abietane and pimarane types have been reported from woods of Pinus, Picea, Larix, Pseudotsuga, and Abies (H. Erdtman, 1963; Norin) and from the cone oleoresins of Cedrus (Norin). These compounds are to be expected in the other genera if appropriate oleo- resins are examined. Both pimarane and abietane terpenoids occur widely in the conifers (Hegnauer, 1986), although the abietane type was once thought to be specific to the Pinaceae (H. Erdtman, 1963). Several other types of diter- penoids have also been found in the Pinaceae, including labdane terpenoids in various species of Pinus, Picea, Larix, and Abies, and macrocyclic compounds of the thunbergene type in Pinus, Picea, Larix, and Pseudotsuga (Norin). Pseu- dolaric acids, a very unusual structural class of diterpenes, have been isolated from the roots of Pseudolarix amabilis, these constitute the active components of antifungal Tne gaGInes:) in poor = in Cinna (Zhou et al.). Serratene trit tic bark components of Pinus and Picea (Hegnauer, 1962, 1986; Norin) and the morphologically similar Cathaya (He et al.) but have not been found in the other genera. Unusual lanostane triter- penoids have been isolated from the bark or needles of a few species of Abies (Hegnauer, 1962, 1986; Norin), while other tetracyclic triterpenes have been found in Larix, Pseudotsuga, Pinus, and Picea. Alkaloids do not occur frequently in the Pinaceae but have been detected in some species of Pinus, Picea, and Keteleeria (Hegnauer, 1962; Willaman & Schubert). The pyridine alkaloids pinidine and a-pipecoline have been isolated from needles of several species of Pinus but are apparently absent in many others. The Pinaceae are unusual in their almost total lack of biflavonoid com- pounds, otherwise ubiquitous among the conifers and Ginkgo (Geiger & Quinn). A single biflavonoid compound, abiesin, has recently been reported to occur in low concentration in the needles of Abies spectabilis (D. Don) Spach (Heg- nauer, 1986). Anthocyanins (3-glucosides of cyanidin (red) and delphinidin (blue), either singly or in combination) are important in the coloration of the young pollen and seed cones of the Pinaceae (Santamour, 1966). They are also involved in the transient spring coloration of Picea needles (Santamour, 1967). Cyclitols (sugar 5-methyl ethers) such as sequoyitol and pinitol occur widely among the conifers, including the Pinaceae, and are of interest in that pinitol is the sugary substance exuded from the trunk of sugar pine (P. Lambertiana Douglas) and the related P. Ayacahuite Ehrenberg (Hegnauer, 1962). The re- lated compound D-1-O-methylmucoinositol occurs widely in the leaves of the 254 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 other families of conifers but is apparently absent in those of the Pinaceae (Dittrich et al/.). Jamieson & Reid surveyed the fatty acids of conifer leaves, including rep- resentatives of seven genera of the Pinaceae. These uniformly exhibited a series of 18-carbon polyunsaturated acids that were absent in the other conifers ex- amined. Many species of Pinaceae are of great economic importance, with the family supplying over half of the world’s timber (Sporne). Pinus, Pseudotsuga, Picea, Larix, Abies, Tsuga, and Cedrus are all of major significance in this regard. These genera are important sources of ornamental trees, and numerous hor- ticultural varieties have been developed (L. H. Bailey; Dallimore & Jackson; Den Ouden & Boom; Kriissmann). The attractive golden larch (Pseudolarix amabilis), grown as an ornamental in China, is suitable to much of our region. Species of Abies (e.g., A. balsamea and A. Fraseri), Picea, and Pinus are fre- quently used as Christmas trees. The Pinaceae, particularly Pinus, are also economically important as the source of turpentine, pitch, and rosin used in shipbuilding (often termed ‘“‘naval stores’’), as well as in other industries. In addition, the leaf and stem resins of several genera have been used by native peoples of the United States and Canada in treating respiratory ailments and aiding wound healing, as well as for other medicinal purposes (Arnason ef al.; Krochmal & Krochmal). REFERENCES: Aase, H. C. Vascular anatomy of the megasporophylls of conifers. Bot. Gaz. 60: 277- 313. 1915. AGRAWAL, P. K., & R. P. Rastoci. Chemistry of true cedars. Biochem. Syst. Ecol. 12: 133-144. 1984. [Cedrus.] ALLEN, G. S., & J. N. OweEns. e life history of Douglas fir. x + 139 pp. Ottawa, Canada. 1972. [Pseudotsuga ioe numerous illustrations, including seven color plates. ] ALOSI, M. C., & R. B. PARK. A survey of phloem polypeptides in conifers. Curr. Topics Pl. Biochem. Physiol. 2: 250. 1983. ARNASON, T., R. J. Hespa, & T. Jouns. Use of plants for food and medicine by native peoples of eastern Canada. Canad. Jour. Bot. 59: 2189-2325. 1981. [Pinaceae, 2202, 2216, 2315-2 BAILEY, I. W. The ee of wood in the Pineae. Bot. Gaz. 48: 47-55. pl. 5. 1909. BAILEY, L. . The cultivated conifers in North America. ix + 404 pp. /58 pis. New York. ; BAILLON, 7 Coniféres. Hist. Pl. 12: 1-45. 1894. [Pinaceae, 16-22, 27-29, 44.] BANNAN, M. W.. Vertical resin ducts in the secondary wood of the Abietineae. New Phytol. 35: 11-46. p/. 1. 1936. [Results of experimental wounding on resin-canal distribution in Pinaceae; response is greater in pinoid than in abietoid gen ae Barner, H., & H. CHRISTIANSEN. The formation of pollen, the pollination mechan and the determination of the ea favorable time for controlled ser ererenars in Larix. Silvae Genet. 9: I-11. 1960. BEAN, W. J. Trees and shrubs hardy in the British Isles. ed. 8. 4 vols. London. 1970- 1980. [Abies, 1: 143-169; Picea, 3: 176-199; Pinus, 3: 209-248; Tsuga, 4: 622-630.] BEHNKE, H.-D. Sieve element plastids of Gymnospermae: their ultrastructure in relation to ora Pl. Syst. Evol. 123: 1-12. 1974. [Protein-type plastids seen only in Pinaceae. | See: G. & J. D. Hooker. Abietineae. Gen. Pl. 3(1): 438-442. 1880. [Pinaceae.] 1989] PRICE, PINACEAE 290 Bosrov, E.G. On intergeneric hybridization in the family Pinaceae. (In Russian; English summary.) Bot. Zhur. 68: 857-865. 1983. [Forestry Abstr. 48: 564. 1987; suggests origins of pinaceous taxa—e.g., portions of Larix, Picea, and Tsuga—by this mech- anism; needs independent experimental verification. ] Boureau, E. Recherches anatomiques et expérimentales sur l’ontogénie des Epatucs des Pinacées et ses rapports avec la phylogénie. Ann. Sci. Nat. Bot. XI. 1: l BucuHorz, J. T. Polyembryony among Abietineae. Bot. Gaz. 69: 153-167. 1920. [See also Am. Jour. Bot. 7: 125-145. 1920.] —. The pine embryo and the embryos of related genera. Trans. Illinois Acad. Sci. 23: 117-125. 1931. . A comparison of the embryogeny of Picea and Abies. Madrono 6: 156-167. 1942. —. Generic and subgeneric distribution of the Coniferales. Bot. Gaz. 110: 80-91. 1948 Butts, D. ,& J. T. BucuHorz. Cotyledon number in conifers. Trans. Illinois Acad. Sci. 33: 58- ce 1940. [Data from population samples; literature review. Campo, M. vAN, & J. SrvAK. Structure alvéolaire des pollens 4 ballonets des eae Pollen Kee 14: 115-141. 1972. [Alveolar structure of pollen saccae in the Pin ceae CAMPO- Dup LAN, M. van. Recherches sur la phylogénie des Abiétinées d’aprés lenes grains de pollen. Trav. Lab. Forest. Toulouse, Tome II, Sect. I, Vol. IV, Art 1950. CARRIERE, E. A. Traité général des coniféres. ed. 2. xii + 910 pp. Paris. 1867. [Pinaceae, 239-589; first treatment to use the nine known genera with their modern circum- scriptions; 7swga raised to rank of genus in ed. | (1855).] CHAMBERLAIN, C. J. Gymnosperms: structure and evolution. xi + 484 pp. Chicago. 1935 CHANG, Y.-P. Bark structure of North American conifers. U. S. Dep. Agr. Tech. Bull. 1095. ii + 86 pp. 1954. [Pinaceae, 12-48, figs. 3-28; detailed ce aa compar- isons for six genera; numerous photomicrographs; keys to gen ee W.-C., & L.-K. Fu, eds. Gymnospermae. (In Chinese; ee anes portions imarily in English and Latin.) Fl. Reipubl. Pop. Sinicae 7. xiv + 544 pp. 1978. (Pinaceae, 32-281; significant new treatments of the Chinese species, including Ke- teleeria (9 species) and Tsuga (5 species); many useful illustrations of little-known taxa.] CuHowpuury, C. R. The morphology and embryology of Cedrus Deodara (Roxb.) Loud. Phytomorphology 11: 283-304. 1961. The embryogeny of conifers: a review. [bid. 12: 313-338. 1962. CHowDHURY, K. A. Abies and Picea: morphological studies. Council of Sci. and Industr. Res., Bot. Monogr. 9. viii + 46 pp. New Delhi. 1974. [Critical review, Phytomor- phology 25: 124. 1974.] CHRISTIANSEN, H. The formation of pollen, the pollination mechanism, and the deter- mination of the most favorable time for controlled pollination in Pseudotsuga Men- ziesii. Silvae Genet. 11: 89-124. 1962. [See also Jbid. 18: 97-104. 1969, which contains the unusual claim that the male gametes are unique among conifers in being ciliated and rapidly motile; needs verification. Cuu,C. A brief introduction to the Chinese species of the genus Pseudotsuga. Davidsonia 12: 15-17. 1981. [Key, map, and distributional information; five species recognized for China and eight worldwide.] Cuu, C.-C., & C.-S. SuN. Chromosome numbers and morphology in Cathaya. (In Chinese; English summary.) Acta Phytotax. Sinica 19: 444-446. 1981. [2n = 24.] Cuun, W.-Y., & K.-Z. KuAnc. A new genus of Pinaceae—Cathaya Chun et Kuang, gen. noy., from southern and western China. (In Russian and Latin; English sum- mary.) Bot. Zhur. 43: 461-476. 1958. [Extensive illustrations; see also the later 256 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 treatment in Acta Bot. Sinica 10: 245, 246. pls. 1-3. 1962, in which one of the two original species is placed in synonymy.] Coker, W. C., & H. R. Totten. Trees of the southeastern states. ed. 3. vili + 419 pp. Chapel ae North Carolina. 1945. [Pinaceae, 15—42.] Coutter, J. M., & C. J. CHAMBERLAIN. Morphology of gymnosperms. ed. 2. xi + 466 pp. ae 1917. CRITCHFIELD, W. B. Impact of the Pleistocene on the genetic structure of North American conifers. Pp. 70-118 in R. M. LANNER, ed., Proceedings of the Eighth North Amer- ican Forest Biology Workshop. Logan, Utah. 1984. [Reviews variation in electro- phoretic alleles, terpenoids, and morphology for selected species of Picea, Abies, Pinus, and Pseudotsuga.] DA.uimor_E, W., & A. B. Jackson. A handbook of Coniferae and Ginkgoaceae. ed. 4 (revised by S. G. HARRISON). xix + 729 pp. 46 p/s. London. 1966. DEN Oupen, P., & B. K. Boom. Manual of cultivated conifers. x + 2 unnumbered + 526 pp. The Hague. 1965. DickINson, T., P. KNow.es, & W. H. PARKER. Data set congruence in northern Ontario tamarack (Larix laricina, Pinaceae). Syst. Bot. 13: 442-455, 1988. [Compares mor- phometric and isozyme data. Dittricu, P., M. Grett, & O. KANDLER. D-1-O-methyl-mucoinositol in ae plants. Phytochemistry 11: ae 250. 1971. [Absent from the leaves of Pinac DoerksEN, A. H., & K. K. CHtnG. Karyotypes in the genus yee rece Sci. 18: 66-69. (2n = 24 in five Asian species , ae P. D. Pollination mechanisms in gymnosperms. Pp. 142-175 in P. K. K. Narr, , Advances in palynology. Lucknow. 1964. [General review relying heavily on oy of DoYLeE.] are ene Sea and taxonomy —an assessment. Glimpses in Plant Research 5: 114-128. 0. Doy te, J. Developmental eer in pollination mechanisms in the Coniferales. Sci. Proc. oy. Dublin Soc. 24: 43-62. 1945. [Evolutionary interpretation of pollination mech- aay : Pinaceae. ] mbryogeny in Pinus in relation to that in other conifers—a survey. Proc. Irish Acad. B. 62: 181-216. 1963. & M. BRENNAN. Cleavage polyembryony in conifers and taxads—a survey II. Cupressaceae, Pinaceae and conclusions. Sci. Proc. Roy. Dublin Soc. A. 4: 137-158. 1972 & A. Kane. Pollination in 7suga Pattoniana and in species of Abies and Picea. Sci. Proc. Roy. Ras ae II. 23: 57-70. pis. 2, 3. 1943. [T. Pattoniana = T. Mertensiana (Bong.) C - M. O’LEaARY. Pollination in Tsuga, at us, Pseudotsuga and Larix. Sci. Proc. y. Dublin Soc. II. 21: 191-204. pls. 4, 5. 1935. FcxeNwate ,J.E. Re-evaluation of C ee and Taxodiaceae: a proposed merg- r, Madrofio 23: 237-256. 1976. [Comparisons among the families of conifers.] ae ee F. Evolutionary significance of chromosomal differentiation patterns in gymnosperms and primitive angiosperms. Pp. 220-240 in C. B. Beck, ed., Origin and early evolution of angiosperms. ai York. 1976. ie A. W. Coniferae. Jn: A. ENGLER & K. PRANTL, Nat. Pflanzenfam. II. 1: 28- 16. 1889. “(With contributions by anne an PRANTL. EL- Peres Y.A., A.M. CoLaANGELLI, & O. SzikLar. A numerical analysis of karyotypes in the genus Pseudotsuga. Canad. Jour. Bot. 6: 536-544. 1983. [Compares karyo- types of all species except P. brevifolia Cheng & Fu.] & E. E. Wuire. Isozymes and forest trees: an annotated pipopraphy: Canad. Forest Serv. Pacific Forest Res. Ctr. Info. Rep. BC-X-267. 1985.* EMBERGER, L. Gymnospermes. Traité de oo systématique 2(1): 383-459. 1960. [Pinaceae placed in a monofamilial Pina ErpTMAN, G. Pollen and spore a taxonomy. Vol. 2. Gymnospermae, 1989] PRICE, PINACEAE 251 Pteridophyta and les Cr uaer as +iv+ 151 a 5 pls. Stock- holm. 1957. [Pinaceae, 6, 11, 13, 27, 28, 32-36, 39, 40, 43, 44.] Vol. 3. Gymnospermae, ae ae aa 191 pp. - pls. Stockholm. ges [Pinaceae, 11, 13, 14, 22-25, 45-48, 54-60, 69-71, 78-80; pls. J, 8-19, 21. ERDTMAN, H. Conifer chemistry and taxonomy of conifers. Pp. 1-27 in K. Kratzi & G. Bitiek, eds., Proceedings of the Fourth International Congress of Biochemistry, Vienna, 1958. Vol. 2. New York. 1959. [Pinaceae, 2-12.] . Some aspects of chemotaxonomy. Pp. 88-125 in T. Swain, ed., Chemical plant taxonomy. London. 1963. [See particularly pp. 106-112. Ferre, Y. DE. Les formes de jeunesse des Abiétacées. Trav. Lab. Forest. Toulouse, Tome II, Sect. I, Vol. III, Art. I: 1-284. 1952. [Morphology of cotyledons and primary and secondary leaves compared. ] . GAUSSEN. Le rameau phylétique: Pinus, Pseudolarix, Keteleeria. Bull. Soc. Hist. Nat. Toulouse 80: 108-118. 1945. [Reprinted as Trav. Lab. Forest. Toulouse, Tome I, Vol. IV, Art. VIII: 1-11. 1945.] FITSCHEN, J. Handbuch der Nadelholzkunde. xv + 765 pp. Berlin. 1930. [Revised 3rd of BEISSNER’s Handbuch der Nadelholzkunde.] FLorRIN, R. Untersuchungen zur Stammesgeschichte der Coniferales und Cordaitales. Sv. Vet.-akad. Handl. II. 10: 1-588. p/s. 1-58. 1931. [Classic treatment of vegetative anatomy of the conifers; Pinaceae, 93-111, 306-347, pls. 34-39; generic relation- ships and classification of the Pinaceae, 476-483. Evolution in cordaites and conifers. Acta Horti Berg. 15: 285-388. 1951. [Classic study of the evolution of the coniferous ovulate cone; see also Biol. Rev. 19: 367— 389. 1954.] The systematics of the gymnosperms. Pp. 323-403 in E. L. Kessez, ed., century of progress in the natural sciences, 1853-1953. San Francisco. 1955. [His- torical review, mostly of morphology —. The ee of conifer and ee genera in time and space. Acta Horti Berg. 20(4): 121-312. Fious, F. Bees et évolution d’une groupe d’Abiétinées. Trav. Lab. Forest. Toulouse, Tome J, Vol. IJ], Art. XVI: 1-286. 1936. [Classification, with useful historical discussion; putative evolutionary relationships in the Pinaceae; see also Ibid., To ee nee III, Art. I: 1-16. 1938.] FowE is, H. A., compiler & reviser. Silvics of forest trees of the United States. U. S. Dep. Agr. ae Handb. 271. vi + 762 pp. 1965. [Maps by E. L. Littte, J GaAussEN, H. Abiétacées. /n: Les gymnospermes: actuelles et fossiles. Trav. Lab. Forest. Toulouse, Tome II, Sect. I, Vol. I, Chap. XI: 1-716. 1960-1967. [Pinus (fasc. VI), 1-272. 1960; Pinus (suite), Cedrus, Abies (fasc. VII), 273-480. 1964; Pseudolarix, Keteleeria, Larix, Pseudotsuga, Pityites, Picea, Cathaya, Tsuga (fasc. VIII), 481- 672. 1966; Additions et corrections (fasc. IX), 673-716. 1967; Bibliographie, Tome II, Vol. I, Partie III: 1-180. 1979 (bibliography for the entire WS GeIGcer, H., & C. Quinn. Biflavonoids. Pp. 692-742 in J. B. HARBORNE, T. J. MABRY, & H. Marry, eds., The flavonoids, pt. 2. London. 1975. [Pinaceae unique among conifers in lacking Seenatas ] GIANNASI, D. E., & D. J. WFORD. Biochemical systematics II. A reprise. Evol. Biol. 26: 25-248. 1986. Terpenoids : in Pinaceae, pp. 66-71.] Grecuss, P. Identification of living gymnosperms on the basis of xylotomy. 263 pp. 350 pls., 8 tables. Budapest. 1955. [Pinaceae, 183-249, pls. 196-345. . Xylotomy of the living conifers. 172 + 329 pp. Budapest. 1972. aie 109- 159, pls. 10, 115-165, figs. 95-145; an extension of GreGuss, 1955 Hamrick, J. L., J. B. Mitron, & Y. B. LINHART. Levels of genetic variation in trees: influence of life history characters. Pp. 35-41 in M. T. CoNnkKLE, tech. coord., Pro- ceedings of the Symposium on Isozymes of North American Forest Trees and Forest Insects. U. S. Dep. Agr. Gen. Tech. Rep. PSW-48. 1981. [Compares isozyme vari- ability in Abies, Picea, Pinus, and Pseudotsuga. | 258 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Harpin, J. W. Studies of the southeastern United States flora. I]. The gymnosperms. Jour. Elisha Mitchell Sci. Soc. 87: 43-50. 1971. [Keys; valuable commentary on Pinaceae. ] Hart, J. A. A cladistic analysis of conifers: preliminary results. Jour. Arnold Arb. 68: 269-307. 1987. [Both intergeneric and interfamilial cladograms; character states tabulated for all genera of Pinaceae.] E, G.-F., Z.-W. MA, W.-F. Yin, & M.-L. CHENG. On s in relation to the systematic position of Cathaya (Pinaceae). ‘Acta Phot Sinica 19: 440- 443, 1981. [Found in bark of Cathaya, Pinus, and Pic HEGNAUER, R. Chemotaxonomie der Pflanzen. Band 1. Thallophyten, Bryophyten, Pte- ridophyten und Gymnospermen. 517 pp. Basel and Stuttgart. 1962. [Pinaceae, 373- 410, 480, 481.] Band 7. Nachtrage zu Band | und Band 2. xi1 + 804 pp. 1986. [Pinaceae, 504-523, 801; extensive literature review. ] HickeL, R. Graines et plantules des coniféres. Bull. Soc. Dendrol. France 19: 13-115; 20: 134-204. 1911. [Pinaceae, 52-115 (seeds), 163-204 (seedlings).] Hitt, T. G., & E. De Fraine. On the seedling structure of gymnosperms. II. Ann. Bot. 23: as 227. pl. XV. 1909. Ho, R.-H., & O. SzikLar. On the pollen morphology of Picea and Tsuga species. sia Palynol. 12: 31-40. 1972. [Emphasizes morphology of 7. Mertensiana in compariso to that of other taxa.] Hu, Y. S., & F. H. Wanc. Anatomical studies of Cathaya (Pinaceae). Am. Jour. Bot. 71: 727-735. 1984. [Thorough treatment of vegetative features and comparison with other genera. ] _J. Yao. Transfusion tissue of gymnosperm leaves. Bot. Jour. Linn. Soc. 83: 263-272. 1981. Hunt ey, B., & H. J. B. Birks. An atlas of past and present pollen maps for Europe: 0-13000 years ago. xiv + 667 pp. Cambridge, England. 1983. [Abies, 22, 73-89, Larix, 35, 253-263; Picea, 39, 285-305; Pinus, 41, 42, 306-336.] Jamieson, G. R., & E. H. Rem. The leaf lipids of some conifer species. Phytochemistry 11: —— 1972. [Includes seven genera of Pinaceae.] JEFFREY, E. C. Sg eae anatomy and phylogeny of the Coniferales. Part 2. The Abietineae. oe oston Soc. Nat. Hist. 6: 1-37. 1905. [Wood anatomy of Pinaceae ae the seats of VAN TIEGHEM; groups treated as tribes Abieteae and Pine ee T. N. Polyploidy in gymnosperms. Evolution 13: 24-39. 1959. [A very rare intraspecific phenomenon in Pinaceae.] Chromosome numbers in gymnosperms. Silvae Genet. 10: 1-9, 1961. [Com- prehensive summary of the literature. Cytogenetical evolution in the gymnosperms—karyotype. Pp. 119-135 in P. MaAuesHwakri, B. M. Jour, & I. K. VasiL, eds., Proceedings of the Summer School of Botany, Darjeeling. gue me 1962. Konar, R. N A and Abies Pindrow ice. Bot. Jahrb. 101: 267-297. 1980. _ Operor. Recent work on reproductive structures of living conifers and ae review. Bot. Rev. 35: 89-116. 1969. [Pinaceae, 90-95.] KROCHMAL, A., & C. KROCHMAL. A field guide to medicinal plants. vii + 274 pp. New York. 1984. KrijssMANN, G. Handbuch der Nadelgehdlze. ed. 2 (revised by H.-D. Warpa). 6 un- numbered + 396 pp. /60 pis. Berlin and Hamburg. 1983. [Also published in English as Manual of cultivated conifers (M. E. Epp, translator). 361 pp. 160 pis. Portland, Oregon. 1985. Kuo, S. R., T. T. Wana, & T. C. HUANG. Karyotype analysis of some Formosan gymnosperms. Taiwania 17: 66-80. 1972. [Includes Keteleeria, Tsuga, Pseudotsuga, Picea, and Pinus in Picea Smithiana 1989] PRICE, PINACEAE 299 LEBRETON, P., & J. SARTRE. Les pinales, considerées d’un point de vue chimiotaxino- mique. Canad. Jour. Forest Res. 13: 145-154. 1983. [Flavonoids in representatives of nine genera of Pinaceae. Li, H. L. Present distribution and habitats of the conifers and taxads. Evolution 7: 245- 261. 1953. LittLeE, E. L., Jk. Atlas of United States trees. Vol. 1. Conifers and important hardwoods. ep. Agr. Forest Serv. Misc. Publ. 1146. v + 9 pp. 2 base maps, 200 species maps, 9 overlay maps. 1971. [Detailed range maps for all American species of Pinaceae. . Checklist of United States trees (native and naturalized). U. S. Dep. Agr. Forest Serv. Agr. Handb. 531. iv + 375 pp. 1979 cis Society field guide to North American trees. Eastern region. 714 pp. New York. 1980. Lotova, L. I. On ne correlation of the anatomical features of the wood and phloem in the Pinaceae. (In Russian.) Vestn. Moskov. Univ. Biol. 6. 1: 41-51. 1975. MAHESHWARI, P., & H. SinGH. The female gametophyte of gymnosperms. Biol. Rev. 42: 88-130. 1967. MenuRa, P. N., & T. N. KHosHoo. Cytology of conifers. I. Jour. Genet. 54: 165-180. pl. 5. 1956. [Chromosome counts and karyotypes in Pinus, Cedrus, Picea, and Abies; all n = 12. MERGEN, F. Microsporogenesis and macrosporogenesis in Pseudolarix amabilis. Silvae Genet. 25: 183-188. 1976. Miiiay, M. A., & T. N. Taytor. Evolutionary trends in fossil gymnosperm pollen. Rev. Palacobot. Palynol. 21: 65-91. 1976. MILLER, C. N., Jr. Early evolution in the Pinaceae. Rev. Palaeobot. Palynol. 21: 101- 117. 1976. [Compares ovulate-cone anatomy e om and fossil genera. . Mesozoic conifers. Bot. Rev. 43: 217-280. —. Current status of Paleozoic and Mesozoic oes Rev. Palaeobot. Palynol. 37: 99-114. 1982. [Preliminary phylog Pp f fossil and extant genera . Pityostrobus pubescens, a new species of pinaceous cones from the Late Cre- taceous of New Jersey. Am. Jour. Bot. 72: 520-529. 1985. [Includes anatomical comparisons of the extant genera of Pinaceae. ] MIRANDA, V., & M. CHAPEKHAR. SEM study of the inner periclinal surface of leaf cuticles in the family Pages? me Jour. Linn. Soc. 81: 61-78. 1980. NIEMANN, G. J., & H. H. vAN GENDEREN. Chemical relationships between Pinaceae. Biochem. Syst. Ecol. 8: 339-240. 1980. [Compares flavonoid “fingerprints” of species in nine genera. ] Norn, T. Some aspects of the chemistry of the order Pinales. Phytochemistry 11: 1231- 1242. 1972. [Valuable review emphasizing wood-extractive chemistry of the Pina- ceae. Owens, J. N., & M. MoLper. Sexual reproduction of Larix occidentalis. Canad. Jour. Bot. 57: 2673-2690. 1979. . Smmpson. Pollen from conifers native to British Columbia. Canad. Jour. Forest Res. 16: 955-967. 1986. [SEM and LM eae and descriptions; in- cludes Abies, Tsuga, Pseudotsuga, Larix, Picea, and Pin Pant, D. D., & N. Basu. A comparative study of the ee of Cathaya argyrophylla Chun & Kuang and three species of Keteleeria Carriére. Bot. Jour. Linn. Soc. 75 271-282. 1977. PARLATORE, P. Coniferae. DC. Prodr. 16(2): 361-521. 1868. [Pinaceae, 364, 365, 377- 431. PuiLuips, E. W. J. Identification of softwoods by their microscopic structure. Dep. Sci Industr. Res. London Forest Prod. Res. Bull. 22. ui + 56 pp. 3 pls. 1948. [Excellent comparative treatment of wood anatomy for all genera except Cathaya; a similar account in Jour. Linn. Soc. Bot. 52: 259-320. 1941.] 260 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Pircer, R. Pinaceae. Jn; A. ENGLER, Nat. Pflanzenfam. ed. 2. 13: 271-342. 1926. {Definitive separation of ee aon Taxodiaceae and Cupressaceae; introduction of the modern system of families. ] . MELcHIoR. Gymnospermae. /”; H. MELCHIOR, Engler’s Syllabus Pflan- zenfam. ed. 12. 1: 312-344. 1954. eens 329-332, figs. 132, 133. Pracer, E. M., D. P. FowLER, & A. CW. ILSON. Rates of evolution in conifers (Pinaceae). Saar 637-649. 1976. [ logical ison and el phoretic ch acterization of seed proteins. i; Price, R. A., & J. M. Lowenstein. An immunological comparison of the Sciadopity- aceae, Taxodiaceae, and Cupressaceae. Syst. Bot. (in press). [Seed-protein extracts compared by radioimmunoassay; Sciadopitys relatively isolated and not close to Pinaceae. } _ J. OLSEN-STOJKOVICH, & J. M. LoweNSTEIN. Relationships among the genera of Pinace eae: an immunological comparison. Syst. Bot. 12: 91-97. 1987. [Immunolog- ical comparison of seed proteins supports the classification of VAN TIEGHEM (1891).] Rapais, M. L. Anatomie comparée du fruit des coniféres. Ann. Sci. Nat. Bot. VII. 19: 165-368. pls. I-15. 1894. uae vascular anatomy of the ovulate cone.] RADFORD, A. E., H. E. AHLES, & C. R. BELL. Manual of the vascular flora of the Carolinas. Ixi + 1183 pp. Chapel Hill, North eee 1968. [Pinaceae, 34-40; dot maps, illustrations. ] ReHDeER, A. Manual of cultivated trees and shrubs hardy in North America, exclusive of the subtropical oe warmer temperate regions. ed. 2. xxx + 996 pp. New York. 1940. [Pinaceae, 8-4 . Bibliography es trees and shrubs hardy in the cooler temperate regions of the Northern Hemisphere. xl + 825 pp. Jamaica Plain, Massachusetts. 1949. [Detailed nomenclatural treatment; Pinaceae, 6-41. Rotu, I. Histogenese und morphologische Deutung der Doppelnadeln von Sciadopitys. Flora 152: 1-23. 1962. [‘‘Double-needles” of Sciadopitys are not comparable to a fused pair of needles. ] SANTAMOUR, F. S., JR. New chromosome counts in Pinus and Picea. Silvae Genet. 9: 87, 88. 1960. [Picea (15 species) and Pinus (12 species), all n = 12.] —. Anthocyanins of conelets in the Pinaceae. Forest Sci. 12: 429-431. 1966. —. Anthocyanins in spruce foliage. Bull. Morris Arb. 18: 41, 42. 1967. SARGENT, C. §. Silva of North America. Vol. 11. Pinus. vii + 163 pp. pls. 538-592. Boston and New York. 1897 [1898]. Vol. 12. Abietineae after Pinus. vii + 144 pp. pls. 593-620. Boston and New York. 1898 [1899]. [Excellent illustrations of Abies, Picea, Pinus, se Pseudotsuga, and Larix; much useful information about uti- lization of the g Sax, K., & H. J. Sax. Chromosome number and morphology in the conifers. Jour. Arnold Arb. 14: 356-374. p/. 75. 1933. [Chromosome numbers and karyotypes for most ae - Pinaceae. ] Scuopr, J. M. The embryology of Larix. Illinois Biol. Monogr. 19(4): 1-97. 1943. Seitz, F. W. ae ay rent eons bei Holzpflanzen. Silvae Genet. 1: 22- 32. 1951. [Review of chromosome counts, mainly from the literature; includes a few otherwise-unpublished counts by ae author for Abies and Picea (all n = 12).] Smmak, M. Karyotype analysis of Larix Griffithiana Carr. Hereditas 56: 137-141. 1966. {See also Stud. Forest. Suecica 17: 1-15. 1964 StnGH, H. Embryology of gymnosperms. Band 10, Teil 2 in K. LinspAuER, Handbuch der Pflanzenanatomie. ed. 2. xii + 302 pp. Berlin. 1978. [Male and female game- tophytes: pollination oe. embryogeny; extensive bibliogra Sivak, J. Les caractéres de diagnose des graines de pollen a ballonets. Pollen Spores 17: 349-421. 1975. [Extensive scanning electron microscopic comparisons of pinaceous o r 1989] PRICE, PINACEAE 261 SMALL, J. K. Manual of the southeastern flora. xxii + 1554 pp. New York. 1933. SNAJBERK, K., & E. ZAVARIN. Mono- and sesqui-terpenoid differentiation of Pseudotsuga of the United States and Canada. Biochem. Syst. Ecol. 4: ‘159- 163. 1976. [Cortical SPORNE, K. R. The morphology of gymnosperms. ed. 2. GD pp. London. 1974. SQUILLACE, A. E. Analysis of monoterpenes of conifers by gas-liquid chromatography. Pp. 126-157 in J. P. Mixscue, ed., Modern methods in forest genetics. Berlin. 1976. [Extensive bibliography cross referenced by taxa.] SRIVASTAVA, L. M. Secondary piloe au in the Pinaceae. Univ. Calif. Publ. Bot. 36: 1- 142. 1963. [35 plates of micr of species in six genera; emphasis on western North America. 1 Stairs, G. R. Monoterpene composition in Larix. Silvae Genet. 17: 182-184. 1967. STERLING, C. Structure of the male gametophyte in gymnosperms. Biol. Rev. 38: 167- 203. 1963 TAYLoR, R. J. The relationship and origin of Tsuga heterophylla and Tsuga Mertensiana based on phytochemical and morphological interpretations. Am. Jour. Bot. 59: 149- 157. 1972. [Leaf phenolics of Tsuga and Picea species a Teou, S. B., & H. Rees. Nuclear DNA amounts in populations of Picea and Pinus species. Heredity 36: 123-137. 1976. TIEGHAM, P. vAN. Anatomie comparée de la fleur femelle et du fruit des Cycadées, des Coniféres et des Gnetacées. Ann. Sci. Nat. Bot. V. 10: 269-304. 1869. [Important early contribution on ovulate-cone anatomy of the Pinaceae, with emphasis on vasculature Structure et affinités des Abies et des genres les plus voisins. Bull. Soc. Bot. France 38: 406-415. 1891. [Defines subfamilial groups based on resin-canal position in the primary root; provides synopsis of Abies species based on leaf anatomy.] Ueno, J. Some palynological observations of Pinaceae. Jour. Inst. Polytech. Osaka Univ. er. D. 9: 163-177. pis. 1-3. 1958. [All genera except Cathaya compared.] VIERHAPPER, F. Entwurf eines neuen Systemes der Coniferen. Abh. Zool.-Bot. Ges. Wien 5(4): 1-56. 1910. [Classification of Pinaceae emphasizing shoot dimorphism.] Von RuptorrF, E. Volatile leaf oil analysis in chemosystematic studies of North Amer- Wana, F.-H. Life history of Keteleeria. I. Strobili, development of the gametophytes and fertilization in Keteleeria Evelyniana. Am. Jour. Bot. 35: 21-27. 8. & T.-S. CHEN. The embryogeny of Cathaya (Pinaceae). (In ene English summary.) Acta Bot. Sinica 16: 64-69. p/s. 1-3. 1974. [Similar to that of Pinus.] Watts, W. A. ees history of the eastern United States 25,000 to 10,000 years ago. Pp. saa E. WRIGHT, ed., Late- saat environments of the United States. Vol. 1. The i Pleistocene (S. C. Porter, ed.). Minneapolis. 1983. [Dis- cusses Current ae a and Quaternary distribution of Abies, Larix, Picea, Pinus, and Tsuga. WILLAMAN, J. J.. & B. G. ScuuBert. Alkaloid-bearing plants and their contained al- kaloids. U. S. Dep. Agr. Agr. Res. Serv. Tech. Bull. 1234. 287 pp. 1961. [See also WILLAMAN, J. J., & H. L. Li, Lloydia 33(Suppl.). 286 pp. 1970. WODEHOUSE, R. P. Pollen grains. xv + 574 pp. New York. 1935. [Pinaceae, 256-268, fig. 78, pls. 2, 3.] ZHOU, B.-N., B.-P. YING, G.-Q. SONG, Z.-X. CHEN, J. HAN, & Y.-F. YAN. Pseudolaric acids from Pseudolarix ee Pl. Med. 47: 35-38. 1983. [Structures of anti- fungal agents from bark of P. amabilis (P. Kaempferi) characterized.] 262 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 KEY TO THE GENERA OF PINACEAE IN THE SOUTHEASTERN UNITED STATES General characters: monoecious evergreen [or winter-deciduous] trees (rarely shrubs); leaves alternate, linear to acicular, containing resin canals; pollen cones with spirally arranged, flattened microsporophylls; sporophylls with 2 abaxial microsporangia; pollen bisaccate (sometimes nonsaccate); ovulate cones erect or pendent at maturity, bract-scale complexes spirally arranged, flattened; bracts largely free from cone scales, included or exserted: ovules 2 per cone scale, adaxial, with micropyles facing the cone axis; seeds with 1 terminal wing (rarely wingless). A. Foliage leaves in fascicles of 2-5, each ee subtended by a scale leaf; cone ae thickened at apex, often armed with a DICK Gs censredus ace betwee ose 1. Pin A. Foliage leaves borne singly, without g scale | : ] thickened at apex, unarme B. Twigs roughened by a leaf bases (sterigmata); ovulate cones not erect at ee falling as au ce uarees in cross section, acute tipped, not abruptly igeaiee at base; co SHOOlEIECl> sesnc ba yee eee a ee aoe ee Picea. C. I eaves See blunt tipped, abruptly petiolate at base; leading ae droop- saat cheth ascites east iseeeance a anata halateeh ase eee iets etacitan ae Aen easing aa Reema . Tsuga. B. Twigs. ve roughened by leaf bases; leaf scars circular, flush with ae ovulate cones erect at maturity, cone scales shed individually. ............. 4, Abie 1. Pinus Linnaeus, Sp. Pl. 2: 1000. 1753; Gen. PI. ed. 5. 434. 1754. Evergreen trees (or less commonly shrubs). Bark deeply furrowed to thin and scaly. Wood with axial and radial resin canals and ray tracheids regularly present; axial parenchyma absent. Primary leaves spirally arranged; secondary (foliage) leaves borne in fascicles of 2, 3, or 5 [rarely to 8 or solitary] on highly condensed short-shoots subtended by a scale leaf and surrounded at base by a sheath of bud scales or their remnants; foliage leaves acicular [to linear or rarely lanceolate], entire or minutely serrulate, usually much elongated ([2—]3-50 cm); fibrovascular paca single (subg. STROBUS) ¢ or Pouble ASUvE: Pinus); resin ducts 2 to several. Pollen indrical, clustered around the bases of the current year’ S leafy shoots, each bore in the axil of a bract; microsporophylls many per strobilus; lly dehiscent; sporophyll apex prolonged into a crest; pollen ‘bisaccate. Ovulat te cones terminal or subterminal (sometimes appearing to be lateral) on young lateral shoots, borne singly or more commonly clustered, maturing in the second year [less commonly in the third year, or after ca. 1 year in the tropics]; mature cones usually ovoid to cylindrical, radially symmetrical to oblique, sessile or short stalked, pendulous; bracts much smaller than scales at maturity; ovuliferous scales persistent, obovate to oblong, very woody (or only slightly lignified in some members of subg. StROBUS), with exposed portion (apophysis) thickened and bearing in a terminal or abaxial position the remnant of the exposed portion (umbo) from the first year’s development. Seeds with a terminal wing [this sometimes short or absent]; body rounded on the sides and usually acute at base: seed coat hard and somewhat thickened, without resin vesicles. Cotyle- dons ca. 4-11 [2-18], denticulate or entire. Chromosome number 2” = 24. (Including Strobus Opiz, Ducampopinus A. Chev.) LECTOTYPE SPECIES: P. syl- 1989] PRICE, PINACEAE 263 vestris L.; see Britton, N. Am. Trees, 5. 1908. (Classical Latin name for pine.) — PINE Pinus, with approximately 100 species, is the largest genus of conifers and the most widespread and economically important genus of trees in the Northern Hemisphere (Critchfield & Little; Mirov, 1967). It occurs from arctic or sub- arctic areas of North America and Eurasia to subtropical and even tropical areas of Central America and Asia, with a few species reaching approximately 12°N latitude in the highlands of Nicaragua and Vietnam. The range of one species, P. sylvestris L., extends north of the Arctic Circle in Scandinavia, while that of P. Merkusii Jungh & De Vries crosses the equator in Sumatra. The greatest number of species (ca. 65) are native to North America, with particular concentrations in Mexico, California, and the southeastern United States. Thir- ty-four species are native to Eurasia, with three of these ranging into North Africa, and one (P. canariensis C. Smith) endemic to the Canary Islands. Several species are now widely grown in the Southern Hemisphere. The complex history of classification within Pinus has been reviewed by Mirov (1967) and Little & Critchfield. Many of the currently recognized sub- divisions date to the important 1838 classification of Loudon, which divided the genus into three sections based on number of needles per fascicle, and then into 15 groups corresponding to subsections. The division of the genus into two main groups based on the number of vascular bundles in the needle (one vs. two), which has been central to most recent classifications, dates to 1893, when Koehne treated these groups as sections Haploxylon and Diploxylon (= subg. Srrosus (D. Don) Lemmon and subg. Pinus, respectively). Shaw’s (1914) worldwide monograph has been the basis from which most modern classifications of the genus have been developed. He utilized characters from a diversity of plant parts—including morphology of the shoots, leaves, ovulate cones, ovuliferous scales, and seeds, and also data on wood anatomy derived from the research of I. W. Bailey (1910). Shaw’s treatment is notable in emphasizing proposed evolutionary transformation series for seed and cone characters in the formulation of the classification. Some of the assignments of character polarity are open to question, and not enough attention was given to character conflict and the possibility of convergence, but many of his groups appear to be natural ones. Shaw divided the genus into sections Haploxylon and Diploxylon, under which he recognized two subsections each and thirteen “groups” designated by names and roman numerals. He later (1924) merged the related white-pine groups Strobi and Flexiles under the former name. The yellow pines (subg. Prnus) from our region, along with a diversity of species from elsewhere in North and Central America and a few from Europe, were placed in his groups Australes and Insignes, depending on the degree of serotiny and cone asymmetry More recent treatments, beginning with Pilger’s in 1926 and including that of Little & Critchfield (1969), which is used here, have raised Shaw’s infra- generic groups in rank so that his sections now correspond to subgenera, his subsections to sections, and his groups to subsections. Later classifications have diverged most from Shaw in their treatment of groups in subg. Pinus. Duffield’s 264 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 studies of crossability among species in Shaw’s subsect. Pinaster (comprising the bulk of subg. Pinus) have resulted in important rearrangements reflected in Little & Critchfield’s classification. Further study of the diverse Mexican pine flora has also resulted in the publication of a number of new species in both subg. Pinus (see particularly Martinez, 1948) and subg. Strosus, where classification of the pion pines has recently been in great flux (see Eguiluz Piedra, 1988; Zavarin, in press Little & Critchfield’s classification recognizes three subgenera, PINUS, STROBUS, and DucAmpopiNus (A. Chev.) Ferré. The last includes only the unusual Viet- namese Pinus Krempfii Lecomte, which has two strongly flattened needles per fascicle but otherwise fits well into subg. Srrosus (see Erdtman et a/.; Van der urgh). Little & Critchfield rigorously applied the rules of botanical nomen- clature in their choice of names and ranks and utilized information on cross- ability and, to some degree, chemistry and karyotype in recognizing five sections (essentially the subsections of Shaw plus a new section for P. Krempfii) and fifteen subsections, three of which are represented in our area. Another recent classification that strongly emphasizes characters of wood anatomy, along with more traditional morphological characters, is presented in Van der Burgh. It has been modified somewhat by Farjon. Van der Burgh did not split the genus into two main groups representing haploxylon and diploxylon pines but instead implied (see his fig. /) that various groups of diploxylon pines have been derived separately from primitive haploxylon pines (see also Farjon, fig. /, p. 202). He thus divided the genus into eight sections, two among the haploxylon pines and six among the diploxylon pines, and 21 subsections. Many of the sections and subsections are similar to those recognized by Little & Critchfield, but Ficure |. Pinus subg. Strosus sect. SrRoBus subsect. Strosus. a-s, P. Strobus: a, up of shoot “a last season’s leaves, new growth, and 2 ovulate cones at time of , diagr single fibrovascular bundle, endoderis enclosing transfusion tissue, and 2 resin ducts, 4: f, g, lateral and abaxial views of microsporophyll, showing the 2 abaxial sporangia and their dehiscence, < 10; h, pollen grain, showing the 2 saccae, x 250; 1, ovulate cone at time of pollination, x : j, detail of ovulate cone, showing scales with ovules and subtending bracts, x 10; k, detail in lateral view of single cone scale and subtending bract at time of pollination, 1 ovule with 2 micropylar appendages visible, x 12; 1, m, adaxial and abaxial views of cone scale and bract at time of pollination, showing mi- cropylar appendages on the 2 ovules, x 12; n, “whorl” of 4 cones at end of first growing season, all surrounding fascicles of leaves removed, lateral and terminal buds visible, x %: 0, cone scale with 2 developing seeds at time of pollination of the next year’s cones, showing remnants of the micropylar appendages and development of wings, x 3; p, mature cone, pendent on tree, seeds already shed, < ‘2; q, mature seed, after wing has separated from cone scale, x 1; r, longitudinal section of seed, wing removed, showing embryo surrounded by tissue of megagametophyte (stippled), micropyle facing base, x 5; s, embryo, showing suspensor and numerous cotyledons, x 1989] PRICE, PINACEAE 203 266 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 some—particularly in the heterogeneous sections PARRYA Mayr and PINEA Endl.—have been divided more finely. Another quite different classification that emphasizes resin-canal distribution in the needles, cotyledonary formulae (see Ferré, 1953, 1965), and pollen size is given by Gaussen. Pinus is readily distinguished from the other genera of Pinaceae by its elon- gate foliage leaves borne in tight clusters of two to five [one to eight] needles on extremely reduced short-shoots, and its distally thickened, woody, often spine-tipped cone scales. As discussed under the family, Pinus is linked to the other pinoid genera (Picea, Cathaya, Larix, Pseudotsuga) by features of wood and primary root anatomy. Pinus shows the greatest overall similarity to Picea and Cathaya, which are the only other Pinaceae containing serratene triter- penoids. Pinus has the longest well-established fossil record of any of the extant genera of Pinaceae, dating back to the Lower Cretaceous. Miller (1976, 1977) has observed that early fossil cones of the Pinaceae are often similar to those of Pinus in form and anatomy. This led him to suggest that early evolution of the family was distinctly Pinus-centered. Immunological comparisons of seed proteins (Prager ef al.; Price et al.), however, do not provide evidence that Pinus diverged before the other nine genera and instead place it within the pinoid group of genera. The 11 species native to our region fall into two of the subgenera recognized by Little & Critchfield. Subgenus Strosus (subg. Haploxylon (Koehne) Rehder) is characterized by leaves with a single fibrovascular bundle, five needles per fascicle [one to five in sect. PARRYA Mayr] with a deciduous sheath at the base [rarely persistent in sect. PARRYA], and a nondecurrent subtending bract. Mem- bers of sect. Srrosus, including the white pines (subsect. SrroBus) and the stone pines (subsect. CEMBRAE Loudon), have five needles per fascicle and cone scales with the umbo terminal and lacking an apical point or bristle. Subsection Strosus, which differs from subsect. CEMBRAE in having seed cones that are dehiscent at maturity and winged seeds, comprises 14 species in North and Central America and Eurasia (Little & Critchfield). Our only species, Pinus Strobus L. (Strobus Strobus (L.) Small), eastern white pine, white pine, 2n = 24, occurs broadly in the northeastern United States and the Great Lakes states, in southern Canada from extreme southeastern Manitoba to New- foundland, and south in the Appalachian Mountain region to northern Georgia and South Carolina (Little, 1971, maps 73N, E). This species, characteristic of areas with a cool, moist climate, occurs most commonly on well-drained sandy soils, although it also grows on a number of other soil types (Fowells). Huge numbers of these trees were cut for timber in the last two centuries, so few large first-growth individuals, which may reach 40-50 m in height, remain (Peattie). In our area the species occurs up to 1500 m altitude in the Appalachian Mountains of Tennessee, the Carolinas, and Georgia. Disjunct populations are also found in central North Carolina (Chatham and Lee counties; Coker & Totten) and central Tennessee (Cheatham County). A related taxon from highland areas of southern Mexico and Guatemala has often been treated as P. Strobus var. chiapensis Martinez (see, for example, Sharp). Andresen (1964, 1966) has treated it as a distinct species, P. chiapensis (Martinez) Andresen, on the basis of quantitative studies indicating that the 1989] PRICE, PINACEAE 267 taxon is less similar to P. Strobus than is the western white pine (P. monticola Douglas ex D. Don). According to Andresen (1966), P. chiapensis differs quan- titatively from P. Strobus in having longer leaves with a greater number of serrations per unit length, ovulate cones with longer stalks and more scales per cone, and qualitatively in having undulate, truncate cone-scale apices and in lacking recurved cone scales adjacent to the stalk. Chemical comparisons of these taxa will be of great interest in providing independent assessments of relationship. Subgenus Pinus (subg. Diploxylon (Koehne) Rehder, subg. Eupitys (Spach) Rehder) (leaves two or three [rarely four to eight] per fascicle, fibrovascular bundles two per needle, basal sheath of fascicle persistent [rarely deciduous], the subtending bract decurrent; seed cones symmetrical or oblique; cone scales with umbo abaxial, often armed with a sharp prickle) comprises at least 62 species (Little & Critchfield) distributed throughout virtually the entire range of the genus. Two subsections of sect. Pinus (seeds with a long [or rarely short] detachable wing, leaf fascicles with a persistent basal sheath) have species native to our area Subsection AUSTRALES Loudon (leaves two or three per fascicle, resin ducts internal or medial; seed cones symmetrical, generally opening at maturity; cone scales usually with persistent prickles) comprises 11 species in the eastern United States, the Caribbean islands, and adjacent Central America; eight of them are native to our region. Pinus palustris Miller (P. australis Michx. f.), longleaf pine, longstraw pine, Georgia pine, yellow pine, 2” = 24, is found on well-drained sandy soils of the Atlantic and Gulf coastal plains and sometimes up to 600 m altitude on foothills of the adjacent Piedmont, from southeastern Virginia to south-central Florida and west to eastern Texas (Little, 1971, map 65E). The species has been heavily exploited for timber and turpentine produc- tion, and it has been estimated that by 1930 only ten percent of its original volume of timber remained (Snyder ef a/.). Young trees of P. palustris are highly fire tolerant, so the species is often found in pure stands on burned areas. Pinus palustris is a large tree (to ca. 30 m) with the longest leaves (ca. 25— 45 cm) and largest cones (ca. 15-25 cm long) of any pine in the eastern United States. Leaves are three per fascicle, and cones are nonserotinous and have small prickles on the cone scales. The winter buds are distinctively large and whitened. This is one of very few pine species in which the seedlings go through an unusual “‘grasslike”’ stage. The stem thickens but elongates very little during the first few years, and the densely overlapping needles protect the growing point (see discussion in Mirov, 1967). This form of seedling development is Se to a few subtropical or tropical pines and conveys substantial fire resistan ne hybrids between Pinus palustris and P. Taeda (P. x Sondereggeri Chapman) occur with some regularity when the species grow in mixed popu- lations (Snyder ef al/.), but the two species nevertheless remain quite distinct overall. Hybrids of P. palustris and P. Elliottii var. Elliottii occur less frequently in southern Georgia and northern Florida (Mergen, 1958a; Snyder et a/.). Pinus Taeda L., loblolly pine, oldfield pine, North Carolina pine, 2n = 24, is found on the Coastal Plain and Piedmont from southern New Jersey south JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 268 NG INOAI/Z LAR. \ if SF Ly \) / CA ef TELA gD AY Pl AS E a a a RY pas Ne ae —— tans i = = —<—_— = > = ss Et 1989] PRICE, PINACEAE 269 to central Florida and west to southern Arkansas, eastern Texas, and extreme southeastern Oklahoma (Little, 1971, map 75E). It is a common tree in South Carolina, Georgia, Alabama, and eastern Mississippi, and north to southern Tennessee. It grows in both poorly drained clay (the name “loblolly” means mud puddle; Little, 1980) and well-drained sandy soil. It rapidly invades aban- doned fields on the Coastal Plain, but to a much lesser degree in the Piedmont, where P. echinata is the characteristic oldfield pine (Fowells; Little, 1980). Loblolly pine increases in size very rapidly and is the most widely utilized commercial pine in the southeastern United States, being very widely grown in plantations Loblolly pine is a large tree (up to 45 m) with stout but relatively flexible needles ca. 15-25 cm long, usually three per fascicle, and cones ca. (5—)7-13 cm long, with short, stout prickles. It differs from the related Pinus rigida Miller and P. serotina Michx. in having larger, more readily deciduous cones with larger prickles (ca. 2-3 mm long) and wider cone scales (ca. 12-15 vs. 6-10 mm), and nonresinous winter buds (Little et a/.). In addition to the hybrids with P. palustris noted above, P. Taeda forms natural hybrids with P. rigida, P. serotina, and P. echinata (Dorman & Zobel; Little et a/.; Smouse & Saylor, 1973b), but the species remains relatively distinct even in areas of narrow sympatry. Pinus echinata Miller, shortleaf pine, yellow pine, 2” = 24, is the most widely distributed of the southern yellow pines, occurring from New Jersey and ex- treme southeastern New York south to northern Florida and west to eastern —E 2. Pinus subg. PINUS sect. PINUS subsect. iS are a—n, P. Tigida: a, tip of ae ak last season’s leaves of pollen, and new vegetative lateral and terminal shoots, the pats just beginning to expand, the subtending bracts evident, x '2; b, diagrammatic cross section of needle let shove the 2 BUOvasetay bundles, nearby sclerenchyma (dotted, ae 6 resin the end x 15;c, single ubtending bract before sporangial dehiscence, x 2: d,e, %4 view of abaxial side of Bee anal showing 2 microsporangia, before and after dehiscence: x 10; f, shoot tip after shedding of pollen, showing persistent fascicles, cone of preceding season, new lateral and terminal growth (fascicles only beginning to expand), and microsporangiate strobili just after shedding of pollen, x 2; g, expanding fascicle of 3 er leaves with scale leaves and subtending bract (decurrent on stem) below, = 2; h, tip of shoot with terminal bud, lateral soa with terminal eons ous betore pollination) ane) new esos of leaves just be O21. t) at time of pollination, x a ; ae “adaxial and abaxial views of cone scale with 2 ovules (each with 2 micropylar appendages) and subtending bract at time of pollination, x 12; 1, cone scale with developing seeds (from second year’s cone in “f?’) at time of pollination of new cones, showing shriveled micropylar appendages and evident wing of seed, x 3; m, mature cone with seeds shed, x 12; n, seed with wing (abaxial view), x 1. o-r, P. pungens: 0, mature cone with seeds shed, x '4; p, — with wing (adaxial view), x 1; q, detail of seed and base of wing (abaxial view), < 3; r, diagrammatic longitudinal section of seed, the seed coat hatched, ee eer embryo (showing 2 cotyledons) un- shaded, x 6. 270 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Texas, eastern Oklahoma, Arkansas, and southern Missouri (Little, 1971, map 52E). It grows at moderate elevations (to ca. 1000 m) in the Appalachian Mountains and commonly invades abandoned fields on the Piedmont. It is less common on the Coastal Plain (Coker & Totten). The species can grow on many kinds of soils but is best developed on fine sandy loam with good drainage (Fowells). Young plants are capable of surviving repeated fires by sprouting from the root collar (Fowells; Mattoon). Pinus echinata is a large tree (ca. 20-30 m) with two or occasionally three slender, flexible needles ca.7—13 cm long per fascicle, small (ca. 4-6 cm), conical or narrowly ovoid cones that open at maturity but are long persistent, and cone scales with a small, slender prickle. Pinus glabra Walter, spruce pine, cedar pine, Walter’s pine, 2 = 24, one of the least common of the southern yellow pines, grows in moist lowland areas of the Coastal Plain from eastern South Carolina south to northern Florida and west to southern Alabama and eastern Louisiana, particularly along river courses (Little, 1971, map 58E). It is a large tree (to ca. 30 or more m high) with relatively slender and short needles (ca. 4-10 cm long) in fascicles of two, and small, conical to narrowly ovoid cones (ca. 3-6 cm long) with small, easily shed, slender scale prickles. The characteristic bark is smooth and grayish, in contrast to that of the related P. echinata, which 1s irregularly flaking and reddish brown. Pinus serotina Michx. f. (P. rigida Miller var. serotina (Michx. f.) Loudon, P. rigida subsp. serotina (Michx. f.) R. T. Clausen), pond pine, marsh pine, pocosin pine, 2n = 24, occurs in somewhat surprising habitats for a pine— sandy swamps and shallow ponds—along the Coastal Plain from southern New Jersey and Delaware south to central and northwestern Florida, often forming nearly pure stands (Little, 1971, map 74£). It is a medium-sized tree (generally 10-20 m high) with flexible needles ca. 15-20 cm long and usually three per fascicle. It is characterized by short (ca. 4-6 cm), rounded to ovoid cones that remain unopened on the tree for several years and cone scales that have only a weak prickle. It intergrades morphologically with the closely related P. rigida, from which it differs modally in ecology. Pinus rigida Miller, pitch pine, 2” = 24, is a small to medium-sized tree that occurs from southern Maine and extreme southern Quebec and Ontario through New England and the mid-Atlantic states and south, mostly in the Appalachian Mountains, to northern Georgia and South Carolina. It is tolerant of dry, rocky soils but also occurs on sands and sometimes in swampy soils. The needles are usually three per fascicle, stout, and ca. 7-13 cm long; the cones are ovoid, 3- 7 cm long, long persistent, occasionally serotinous, and generally with sharp, slender prickles. Pitch pine is highly resistant to fire and other forms of injury and sprouts from roots and stumps after damage, becoming very irregular in form (Ledig & Fryer, 1974; Little, 1980). Dwarfed plants of the species are characteristic of parts of the Pine Barrens of New Jersey, where the high in- cidence of fire may have selected for cone serotiny, a feature otherwise not typical of the species (Givnish; Ledig & Fryer, 1972). A number of authors have discussed the geographic variation of Pinus rigida in relation to that of P. serotina, with which it intergrades in New Jersey, 1989] PRICE, PINACEAE 271 Delaware, and Maryland (Clausen; Ledig & Fryer, 1974; Smouse & Saylor, 1973a, b). The two are largely allopatric and differ modally in ecology and morphology but show clinal variation in a number of characters. Smouse & Saylor (1973a, b) found pond and pitch pines to group much more closely in multivariate analyses than loblolly and shortleaf pines did (all of these taxa being subject to natural hybridization) and suggested that P. serotina should be treated as a subspecies of P. rigida. Other recent authors (Ledig & Fryer, 1974; Little, 1980) have continued to treat them at the level of species. Pinus pungens Lamb., Table Mountain pine, prickly pine, hickory pine, 27 = 24, isa small to medium-sized tree of the Appalachian area from southeastern Pennsylvania south to extreme northeastern Georgia and eastern Tennessee, and also locally in New Jersey, Delaware, and the District of Columbia (Little, 1971, map 66E). In our area the species occurs on dry, rocky slopes and ridges of the Appalachians up to ca. 1200 m altitude. The species has two stiff, usually twisted needles per fascicle and ovoid, long-persistent cones that differ from those of the other eastern pines in their very stout, incurved prickles. The cones are at least sometimes serotinous (C. E. Wood, pers. comm.). Pinus pungens is named “hickory pine’’ because of its tough, hickorylike branches. Pinus Elliottii Engelm. (P. caribaea sensu Small, non Morelet; P. palustris sensu Small, non Miller; P. heterophylla (Ell.) Sudw., 1893, non K. Koch, 1849), slash pine, swamp pine, 2” = 24, occurs along the Coastal Plain from central South Carolina to southern Florida and west to southern Mississippi and eastern Louisiana. It is also naturalized in eastern Texas. Little & Dorman (1952b, 1954) distinguished two geographic varieties: var. densa Little & Dor- man, of central and southern Florida, differs from var. Elliottii in having a grasslike seedling stage and a lower average number of needles per fascicle. Rangewide geographic variation in morphological and physiological traits was thoroughly examined by Squillace (1966), who used both field and common orchard studies. There are significant differences between the varieties in the means of several characters, but the patterns of variation are largely continuous within and between them, and there is no increase in variability in the area of transition in central Florida. Variety E//iottii was originally largely confined to ponds, pond margins, and other poorly drained habitats (Fowells), but it has become more widespread in its range because natural fires have been sup- pressed. South Florida slash pine (var. densa) occurs in dryish flatwoods and on thin soil over limestone (Hardin; Squillace, 1966). Because of its thicker stems and better protected growing points in young plants, it shows a greater degree of fire tolerance than does var. Elliottii (Squillace, 1966). Pinus Elliottii is a medium-sized to large tree (to 30 m or more) with stout, stiff needles ca. 15-30 cm long, both two and three per fascicle, and narrowly ovoid cones ca. 6-15(-18) cm long with relatively short prickles. It is closely related to P. caribaea Morelet, of the Caribbean Islands and Central America, under which it was formerly treated, but differs in having longer (1-2 vs. < | mm) cone-scale prickles, broader seeds with more readily detachable wings, and needles in fascicles of two and three, rather than consistently three or more (Little & Dorman, 1954). Pinus Elliottii can be crossed artificially with several species including P. caribaea in subsect. AUSTRALES, and also with P. clausa 212 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 in subsect. CONTORTAE (Critchfield, 1963; Saylor & Koenig), but under natural conditions it hybridizes only occasionally with P. palustris (Mergen, 1958a). Subsection ConTorTAe Little & Critchfield (leaves two per fascicle, short (2- 8 cm), often twisted, with mostly medial resin ducts; seed cones small (3-8 cm long), symmetrical [or oblique], often serotinous, long persistent; cone scales with [or without] a persistent prickle) is quite similar in morphology to subsect. AUSTRALES, and the two subsections may be more appropriately combined once their evolutionary relationships are better understood. Subsection CONTORTAE comprises four species native to the United States and Canada, of which two, P. clausa (Engelm.) Vasey ex Sarg. and P. virginiana Miller, are native in our region. A third species, P. Banksiana Lamb., jack pine, which now has a native range north of our region in Canada and the Great Lakes states, occurred in parts of the southeastern United States during the Late Pleistocene (Watts, 1970, 1980) Pinus clausa, sand pine, scrub pine, spruce pine, 2” = 24, occurs on well- drained sandy soils from extreme southwestern Alabama east through the panhandle of northwestern Florida, as well as in scattered localities through much of peninsular Florida, except in the extreme south (Little, 1971; Ward, 1963). It often forms dense pure stands and is the dominant tree in some 200,000 acres in Marion Co., Florida (Ward, 1963). Two hi es have been defined: the Ocala race (var. clausa), with serch nous Sones. aides almost all of the peninsular populations, and the Choctawhatchee race (var. immuginata Ward), with cones opening at maturity, encompassing most of the northwestern populations (Little & Dorman, 1952a; Ward, 1963; Wheeler ef al.). The two cone morphs occur intermixed in several areas of northwestern Florida (Ward, 1963), and no other obvious morphological differences are present, so formal designation as varieties seems unwarranted. The closely related Pinus virginiana Miller, Virginia pine, scrub pine, poverty pine, Jersey pine, 2” = 24, geographically separated by over 100 km from P. clausa, occupies a broad area from Long Island and New Jersey west to Ohio and southern Indiana and south to northern South Carolina and Georgia, central and northern Alabama, and northeastern Mississippi (Little, 1971, map 77E:). In our region it is found primarily in the Piedmont and at moderate elevations (usually below 1000 m) in the Appalachians. It occurs very occa- sionally on the Coastal Plain in the Carolinas and Georgia (Coker & Totten). Tending to grow on dry, shallow soils, including clay, loam, or (less commonly) sandy soils near the coast, it often invades old fields on the Piedmont as a pioneer species (Kellison & Zobel). It is a medium-sized tree with relatively short (ca. 4-8 mm), stout needles and narrowly ovate, persistent but nonserot- inous cones. It differs from P. c/ausa primarily in having longer (ca. 5 mm) more slender cone-scale prickles, and stouter, more often twisted needles (Coker & Totten; Little, 1980), as well as in its edaphic tolerances. Pinus virginiana 1s very similar in isozyme profiles to P. clausa (Wheeler et al.), with which it is readily crossable (Critchfield, 1963; Kellison & Zobel). Both show strong cross-incompatibility with the two remaining species of sub- sect. CONTORTAE (P. contorta Douglas ex Loudon and P. Banksiana). The electrophoretic evidence is consistent with the hypothesis that P. virginiana is 1989] PRICE, PINACEAE 273 a relatively recent derivative of an ancestor similar to P. clausa, and Wheeler et al. have suggested that the two taxa be treated as ecogeographic subspecies. Preliminary data on turpentine composition (Mirov, 1961) indicating differ- ences in terpenoid profiles between these species need further confirmation. Chromosome numbers have been reported for approximately 90 species of Pinus; with the exception of occasional aberrant individuals, all have 2n = 24 (see particularly Khoshoo, 1961; Santamour, 1960; Sax & Sax; Saylor, 1964, 1972, 1983; Styles & Khosla). Karyotypes have been reported for 87 species in the systematic surveys of Saylor (1964, 1972, 1983), and chromosome con- striction patterns have been examined in detail by Pederick for a much smaller sample of species. The smallest chromosome is definitely heterobrachial (with a submedian centromere) 1n Pinus, while the other 11 are approximately iso- brachial, except in subsect. Pinus (subsect. Sy/vestres of Saylor, 1972), in which the second smallest chromosome is also somewhat heterobrachial. Patterns of arm lengths and arm-length ratios are reasonably concordant with the Little & Critchfield classification of subg. Prnus and lend some support to the sep- aration of the southern yellow pines into subsects. AUSTRALES and CONTORTAE. Patterns are more variable within some subsections in subg. STROBUS, partic- ularly among the white and stone pines of sect. STROBUS As noted in the species discussions above, natural hybridization occurs be- tween a number of species of pines within the subsections recognized by Little & Critchfield. In many cases, however, interspecific hybridization is greatly restricted or prevented by a combination of phenological and physiological barriers (see discussions in Critchfield, 1963, and Mirov, 1967). Natural hybrids between species in closely related subsections occur only infrequently (Mirov, 1967)—for example, between Pinus Coulteri D. Don and P. Jeffreyi Grev. & Balf. (Zobel). Zones of hybridization between certain species—for example, P. Banksiana and P. contorta in western Canada (Dancik & Yeh; Mirov, 1956b; Moss; Wagner et a/.; Zavarin et al., 1969)—have been studied intensively through a number of biochemical techniques, as well as morphological com- parisons. Thorough investigations (see particularly Critchfield, 1963, 1966, 1967, 1977, 1986; Duffield) of artificial crossability among pine species have had a major impact on the systematics of the genus. These results have been particularly important in evaluating relationships among the diploxylon pines (see Critch- field, 1963; Duffield; Little & Critchfield) and have led to the treatment of the southern yellow pines under subsections AUSTRALES and CONTORTAE, rather than intermixed with the western North American closed-cone pines of subsect. OOcARPAE Little & Critchfield as in the classifications of Shaw (1914) and Pilger (1926). The species in subsect. AUSTRALES are all linked directly or indirectly by successful crosses (Critchfield, 1963), although some individual combinations (e.g., P. palustris x P. echinata) seem to be highly cross-incom- patible. Crossability within a number of other subsections (e.g., among the pines in sect. Strosus; Critchfield, 1986) is much more restricted and may sometimes show unpredictable geographic patterns. No successful crosses be- tween species representing different subgenera have ever been m Pinus is notable for the great diversity in the size of its cones and: the form 274 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 of its cone scales (Sargent, 1897; Shaw, 1914). The morphological complexity of the cone-scale apex relates directly to the multi-year development of the ovulate cone, with the areas left exposed in the first-year conelet (the umbo) and in the mature cone (the apophysis) both being evident. The extremely stout prickles seen in several species of subg. Pinus (e.g., P. pungens and P. Sabiniana Douglas ex D. Don) presumably act as a deterrent to seed predation (McCune). Klaus (1980), who has surveyed cone-scale form and ornamentation in re- lation to the systematics of Pinus, emphasized the central position of sect. ParRYA in the evolution of the genus. This section of subg. Srrosus (which includes the lacebark, bristlecone, and pifion pines) has an abaxial (‘dorsal’) umbo, often with an apical point or bristle, and thus provides a morphological link with subg. Prnus, which invariably has an abaxial umbo. The same pattern of relationship is seen in wood anatomy (Hudson; Van der Burgh). The amount of toothing on the walls of the ray tracheids is greater in subg. Pinus than in sect. StRoBuS, while the species of sect. PARRYA span most of the range in denticulation of the genus. Other types of characters separate the haploxylon (subg. SrRosus) and diploxylon (subg. Prnus) pines more sharply. Pollen grains of species in subg. Strosus have a strongly ornamented germinal zone, while those of species in subg. Pinus have a relatively smooth one (Sivak). The profiles of heartwood phenolic compounds (H. Erdtman, 1959; Norin), discussed be- low, also separate the haploxylon pines (including P. Krempfti) from the dip- loxylon ones. Pinus shows great diversity in seed size and in the degree of seed winging. Loss of a seed wing in the stone pines (subsect. CEMBRAE) and the pifion pines (subsect. CEMBROIDES) is part of a morphological syndrome related to seed dispersal by corvid birds, especially jays and nutcrackers, or to a lesser degree by rodents (Lanner, 1982; McCune). Cone serotiny occurs in a number of North American species in subg. Pinus (particularly in subsects. CONTORTAE and OGcARPAE, but also in a few species of subsects. AUSTRALES and SABINIANAE Loudon (McCune; Mirov, 1967)). Many such species are polymorphic for serotiny, so some seed dispersal can occur in the absence of major fires (McCune). A simple genetic basis for poly- morphism in this character has been demonstrated in some species (Sittmann & Tyson). Many of the serotinous species have asymmetric cones, a character that may have been overemphasized in Shaw’s (1914) classification, given its potential for convergence. Mirov (1958, 1961) did a preliminary survey of turpentine composition for the majority of species of Pinus. Most species have a steam-distillable oleoresin fraction composed predominantly of various monoterpenoids, but in some it is made up almost entirely of hydrocarbons such as n-heptane. Turpentine composition does not separate major groups such as subgenera but is often valuable in the determination of interspecific hybrids, since closely related species (e.g., P. Banksiana and P. contorta in subsect. CONTORTAE) may differ considerably in their profiles. Leaf-oil composition is reviewed for a number of North American species by Von Rudloff, and wood or leaf oleoresins have been studied for a number of groups in subg. Strosus by Zavarin and coworkers (see, for example, Zavarin et a/. (1982) on subsect. BALFOURIANAE, and Zavarin (in press) and Zavarin & Snajberk (1980, 1987) on subsect. CEMBROIDES). 1989] PRICE, PINACEAE 24D The composition of heartwood phenolics has been surveyed for over 80 species of Pinus and has proved to be very useful in distinguishing the hap- loxylon and diploxylon pines (H. Erdtman, 1959, 1963; Norin). Members of subg. Pinus have a characteristic, relatively simple pattern that includes the stilbene compound pinosylvin and its methyl ethers and the flavanone deriv- atives pinocembrin and pinobanksin. The members of subg. StRoBus also have these compounds, as well as dihydropinosylvin and various flavone derivatives. Some groups within subg. Strosus also contain C-methylated flavonoids. It is postulated that the enzymatic capabilities needed to produce these additional groups of compounds have been lost in the common ancestor(s) of subg. Pinus (H. Erdtman, 1959; Norin). Pinus Krempfii, although treated as a separate subgenus by Little & Critchfield, has a pattern of compounds characteristic of subg. Strosus (Erdtman et al.). Results of allozyme or isozyme electrophoresis have been used 1n systematic or evolutionary studies in several species groups of Pinus (e.g., Conkle et al. on the P. brutia—P. halepensis complex; Millar et al. on the California members of subsect. OGCARPAE; Steinhoff et a/. on P. monticola; Wheeler et a/. on subsect. CONTORTAE), and a number of similar studies are now in progress. Levels of within-population variation for electrophoretic alleles are generally high in pines and other long-lived, wind-pollinated trees (Hamrick et a/.), but they are extremely low in some species—for example, P. resinosa Aiton (Fowler & Morris) and P. Jorreyana Parry (Ledig & Conkle). Pines are of exceptional economic importance for their timber, which is used for construction, pulpwood, and numerous other purposes. Among the eastern North American species, Pinus Elliottii, P. palustris, and P. Strobus yield particularly high quality lumber for construction, although P. Taeda is now more widely available. The rapid-growing Pinus radiata D. Don (a narrow endemic in coastal California and nearby islands) is now one of the most widely cultivated timber trees in the Southern Hemisphere. Several species (e.g., P. echinata, P. Elliottii, and P. palustris in our area) are tapped for stem oleoresin, from which commercial turpentine, tar oils, rosin, and pitch are obtained by distillation (Dallimore & Jackson; Mirov, 1961, 1967). Pine-leaf oil, utilized medicinally, is also obtained from various species by distillation. Highly es- teemed edible seeds (pine nuts) are obtained from large- seeded species (e.g., are of horticultural importance, and several are prized for bonsai. REFERENCES: Under family references see I. W. BAILEY; L. H. BAILEY; BEAN; BOUREAU; BUCHHOLZ, 1920, 1931; Butts & BUCHHOLZ; CAMPO & SIVAK; CAMPO-DUPLAN; CARRIERE; CHANG; C. R. CHowpuury, 1962; Coker & TOTTEN; CRITCHFIELD; DALLIMORE & JACKSON; Dov _e, 1945, 1963; DovLeE & BRENNAN; G. ERDTMAN, 1957, 1965; H. ERptMan, 1959, 1963; FERRE; Florin, 1931, 1951, 1963; FLlous; FOwELLs; GAuSsSEN, 1960; GREGUSS, 1955, 1972; Hamrick et al.; HARDIN; HEGNAUER. 1962, 1986; HickEL; HUNTLEY & Birks; KHOSHOO, 1961; KRUSSMANN; LITTLE, 1971, 1980; MILLER, 1976, 1977, 1985; NorIn; Owens & SIMPSON; PHILLIPS; PILGER; PRAGER et al.; PRICE et al.: ; RADAIS; REHDER, 1940, 1949: Santamour, 1960, 1966; SARGENT, 1897; SAx & SAX; SINGH: SIVAK; SQUIL- LACE; TEOH & REES; UENO; VON RUDLOFF; and WATTS 276 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 The literature on Pinus is very extensive, particularly with regard to chemistry, ecology, economic uses, genetics, and silviculture. The selected references below are weighted toward broad-based systematic studies, review papers, and studies on the species of our region. Mrrov (1967) provides a useful compilation of references on the genus up to the , mid 1960’s. Abpe, L. B., & A. S. CRAFrs. Phloem of white pine and other coniferous species. Bot. Gaz. 100: 695-722. 1939. ALLEN, R., & A. B. WARDROpP. The opening and shedding mechanism of the female cones of Pinus radiata. Austral. Jour. Bot. 12: 125-134. pls. 1-8. 1964. [See also ANDRESEN, J. W. The taxonomic status of Pinus chiapensis. Phytologia 10: 417-421. 1964. [Treated as distinct from P. Strobus A multivariate analysis of the Pinus chiapensis- monticola—Strobus phylad. Rho- dora 68: 1-24. 1966. Anprews, E. F. Agency of fire in propagation of longleaf pines. Bot. Gaz. 64: 497-508. 1917. [Dense covering of needles protects the growing point of young plants from fire.] AXELROD, D. I. History of the maritime closed-cone pines, Alta and Baja California. Univ. Calif. Publ. Geol. Sci. 120. x + 144 pp. 1980. [History of subsect. O6CARPAE as inferred from fossil material.] enozoic history of some western American pines. Ann. Missouri Bot. Gard. 73: 565-641. 1986. [Distribution maps; photographs of ovulate cones of extant and extinct (fossil) species. ] BaiLey, D. K. Phytogeography and taxonomy of Pinus subsection Balfourianae. Ann. Missouri Bot. Gard. 57: 210-249. 1970. [P. /ongaeva separated from P. aristata.] A study of Pinus subsection Cembroides |: The single-needle pinyons of the Californias and the Gre at Basin. Notes Roy. Bot. Gard. Edinburgh 44: 275-310. 1987. [Proposes a new species, P. californiarum. | BAILEY, I. W. Anatomical characters in the evolution of Pinus. Am. Nat. 44: 284-293. 1910. BARNETT, J. P., & B. F. McLemore. Cone and seed characteristics of sand pine. U. S. Dep. Agr. Forest Serv. Res. Pap. SD-19. 13 pp. 1965. [Reports responses of cones of P. clausa to heating.] BucHHOLz, J. T. Suspensor and early embryo of Pinus. Bot. Gaz. 66: 185-228. pls. 6- 1918. —. Embryological aspects of hybrid vigor in pines. Science 102: 135-142. 1945. CuHapman, H. H. The place of fire in the ecology of pines. Bartonia 26: 39-44. 1952. [Reviews response to fire of eastern North American species, including six from our region. CLAUSEN, R. T. Contributions to the flora of New Jersey. Torreya 39: 125-133. 1939. [Pinus serotina treated as a subspecies of P. rigida.] ConkLe, M. T., G. SCHILLER, & C. GRUNWALD. Electrophoretic analysis of ene and phylogeny of Pinus brutia and closely related taxa. Syst. Bot. 13: 411-424. 1988. Cote, W. A., & A.C. Day. Wood ultrastructure of the southern yellow pines. ent Univ. Coll. Forestry Tech. Publ. 95. 70 pp. 1969. [Extensive SEM comparisons of ten species; few consistent differences observed.] CRITCHFIELD, W. B. Geographic variation in Pinus contorta. Maria Moors Cabot Found. Hybridization of the southern pines in California. Southern Forest Tree Im- provement Comm. Publ. 22: 1-9. 1963. [Crossability among the species of subg. Pinus from the cea United States.] ——. Crossability and relationships of the California bigcone pines. U. S. Dep. Agr. 1989] PRICE, PINACEAE Zid Forest Serv. Res. Pap. NC-6: 36-44. 1966. [P. Coulteri, P. Jeffrevi, P. Sabiniana.] . Crossability and relationships of the ee -cone pines. Silvae Genet. 16: 89- 97. 1967. [Relationships within subsect. OGCARPAE. ] Hybridization of foxtail and bristecone pines. Madrofio 24: 193-212. 1977. P. aristata, P. Balfouriana, P. longae Crossability and relationships of Washoe pine. Ibid. 31: 144-170. 1984. [P. washoensis. | The late Quaternary history of lodgepole - jack pines. Canad. Jour. Forest Res. 15: 749-772. 1985. [P. contorta, P. Banksi . Hybridization and classification of the se pines (Pinus section Strobus). Taxon 35: 647-656. 1986. . KIntock. Sugar pine and its hybrids. Silvae Genet. 35: 138-145. 1986. [P. Dambenana .L. Lirte, Jr. Geographic distribution of the pines of the world. U.S. Dep. Agr. Forest Serv. Misc. Publ. 999: 1-97. 1966. [Excellent maps of all species then known; treatment using the new classification detailed by Lirrte & CRITCHFIELD.] DALLIMORE, W. The yellow pines of North America. Bull. Misc. Inf. Kew 1921; 330- 335. 1921. Dancik, B. P., & F.C. YEH. Allozyme variability and evolution of lodgepole pine (Pinus contorta var. latifolia) and jack pine (P. Banksiana) in Alberta. Canad. Jour. Genet. Cytol. 25: 57-64. 1983. DE See . J. Coniferales. Fl. Malesiana. ser. 1. 10(3): 337-453. 1988. [Pinaceae, 447-453; P. Merkusii, P. kesiya Royle ex Gordon.] DEVALL, W. B. The taxonomic status of Pinus caribaea Mor. Proc. Florida Acad. Sci. 5: 121-132. 1941. [Comparisons of P. palustris and P. Elliottii (under P. caribaea).] Doak, C. C. Evolution of foliar types, dwarf shoots, and cone scales of Pinus, with remarks concerning similar structures in related forms. Illinois Biol. Monogr. 13: 1-106. 1935. Dorman, K. W., & A. E. SQuILLAcE. Genetics of slash pine. U. S. Dep. Agr. Forest Serv. Res. Pap. WO-20. 20 pp. 1974. [P. Elfiottii.] & B. J. ZoBeL. Genetics of loblolly pine. U. S. Dep. Agr. Forest Serv. Res. Pap. WO-19. 21 pp. 1973. [P. Taeda. Dov Le, J., & M. O'Leary. Pollination in Pinus. Sci. Proc. Roy. Dublin Soc. 21: 181- 190. pl. 3. 1935. DuFFIELD, J. W. Relationships and species hybridization in the genus Pinus. Silvae Genet. 1: 93-97. 1952. [Results of an extensive program of artificial hybridization ith emphasis on subg. Pinus.] EGuILuz PreprA, T. Geographic variation in needles, cones and seeds of Pinus Tecu- numanii in Guatemala. Silvae Genet. 33: 72-79. 1984. . Distribucién natural de los pinos en México. Notas Técn. Centro Genét. Forest. Chapingo 1: 1-6. 1988. [Includes 52 species.] ENGELMANN, G. Revision of the genus Pinus, and description of Pinus Elliottii. Trans. Acad. St. Louis 4: 161-189. 1880. [Treats the species under two sections and eight subsections. ERDTMAN, H., B. KIMLAND, & T. NorInN. Wood constituents of Ducampopinus Krempfii (Lecomte) Cae (Pinus Krempfi Lecomte). Phytochemistry 5: 927-931. 1966 [P. Krempfii shows pattern of heartwood phenolics typical of subg. StRosus.] Ewers, F. W., & R. SCHMID. “Longevity of needle fascicles of Pinus eee (bristlecone pine) and other North American pines. Oecologia 51: 107-115. 1981. FarJON, A. Pines: drawings and descriptions of the genus Pinus. i pp. Leiden. 1984. [Illustrations; coarse-scale maps; classification and tables of morphological modified from VAN DER BuRGH, with aes of phylogeny by VAN DER BURG and schematic evolutionary diagrams by FARJON.] 278 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Fercuson, M.C. Contributions to the knowledge of the life history of Pinus with special reference to sporogenesis, the development of the gametophytes and fertilization. Proc. Wash. Acad. Sci. 6: 1-202. pls. 1-24. 1904. [Data primarily from P. Strobus.] FERNALD, M. L. The confused bases of the name Pinus palustris. Rhodora 50: 241-249. 1948. ara that P. australis Michx. be applied; see WARD, 1974, for counter- arguments. ] Ferré, Y. DE. Quelques particularités anatomiques d’un pin indochinois: P. Krempfii. Trav. me Forest. Toulouse, Tome I, Vol. IV, Art. XXV: 1-6. figs. 1-4. 1948. sion du genre Pinus en quatre sous-genres. Compt. Rend. Acad. Sci. Paris 236: 196- 228. 1953. —. Structure des plantules et systématique du genre Pinus. Bull. Soc. Hist. Nat. Toulouse 100: Pak 280. 1965. [Types of cotyledons, based on arrangement of vas- cular | t of those in Butts & BUCHHOLZ; reprinted in Trav. Lab. Forest. Toulouse, Tome II, Sect. I, Vol. II, Art. TH: 1-50. 965. l FLORENCE, L. Z., & eS R. Hicks, JR. Further evidence for introgression of Pinus Taeda wit iability and variation in resistance to Cronar- tium fusiforme Silvae Genet. 29: 41-43, 1980. [Disc electrophoresis with general protein Fow er, D. P. "Effects of inbreeding in red pine, Pinus resinosa Ait. Silvae Genet. 13: 170-177. 1964. [Effects are limited due to high homozygosity; see also Ibid. 14: 12- 23, 37-46. 1965.] R. W. Morris. Genetic diversity in red pine: evidence for low genic hetero- zygosity. Canad. Jour. Forest Res. 7: 343-347. 1977. [No detectable eee variation; variation for null alleles has been reported by F. W. ALLENDoRF, K. L. KNupsEN, & G. M. BLAKE (Genetics 100: 497-504. 1982).] Furnier, G. R., & W. T. ApAMs. Geographic patterns of allozyme variation in Jeffrey pine. Am. Jour. Bot. 73: 1009-1115. 1986. [P. Jeffreyi. Grvnisu, T. H. Serotiny, geography, and fire in the Pine Barrens of New Jersey. Evolution 35: 101-112. 1981. [Factors affecting frequency of serotiny in Pinus rigida, see also Lepic & Frvrer, 1972.] GREENWOOD, M. S. Gene exchange in loblolly pine: the relation between pollination mechanisms, female receptivity and pollen viability. Am. Jour. Bot. 73: 1443-1451. 1986 Grierson, A. J. C., D. G. Lona, & C. N. PAGE. Notes relating to the flora of Bhutan: Ill. Pinus bhutanica: a new 5-needle pine from Bhutan and India. Notes Roy. Bot. Gard. Edinburgh 38: 297-310. 1980. GrirFINn, J. R. Cone morphology in Pinus Sabiniana. Jour. Arnold Arb. 45: 260-273. 1964. Guries, R. P., & F. T. Lepic. Genetic diversity and population structure in pitch pine (Pinus rigida Mill. ). Evolution 36: 387-4 HA ter, J. R. Variation and hybridization in ponderosa and Jeffrey pines. Univ. Calif. Publ. Bot. 34: 123-166. 1962. Hariow, W. M. The identification of the pines of the United ge native and intro- duced, by needle structure. New York State Coll. Forestry Tech. Publ. 32: 1-21. pls. J-19. 1931. [Extensive illustrations of cross-sectional anatomy. ] , W. A. Core, JR., . Day. The opening mechanism of pine cone scales. jour Forestry 62: 538- 540. 1964, Haupt, A. W. Odgenesis and mc in Pinus Lambertiana and P. monophylla. Bot. Gaz. 102: 482-498. 194 HErRBIN, G. A., & K. SHARMA. ae on plant cuticular waxes. V. The wax coatings of pine needles: a taxonomic survey. Phytochemistry 8: 151-160. 1969. Hupson, R. H. The anatomy of the genus Pinus in relation to its classification. Jour. Inst. Wood Sci. 6: 26-46. 1960. [Comparative wood anatomy.] 1989] PRICE, PINACEAE 219 JAHRIG, M. Beitrage zur Nadelanatomie und Taxonomie der Gattung Pinus L. Will- denowia 3: 329-366. 1962. [New classification based on needle anatomy; agrees poorly with those based on other characters. JEFFREY, E. C. On the structure of the leaf in Cretaceous pines. Ann. Bot. 22: 207-220. pls. 13, 14. 1908 are: R.C., & B. J. ZoBEL. Genetics of Virginia pine. U. S. Dep. Agr. Forest Serv. Pap. WO-21. 10 pp. 1973. [P. virginiana. ] KENG, HL. & E. L. Litt.e, Jk. Needle characteristics of hybrid pines. Silvae Genet. 10: 131-146. 1961. [Detailed descriptions and anatomical illustrations for 42 interspe- cific hybrids in Pinus. ] Kraus, W. Saccusdifferenzierungen an Pollenkérnern ostalpiner Pinus-Arten. Osterr. Bot. Zeitschr. 120: 93-116. 1972. n the taxonomic significance of tectum sculpture characters in alpine Pinus species. ue 17: 161-166. 1978. Neue Beobachtungen zur Morphologie des Zapfens von Pinus und ihre Bedeu- tung fiir die Systematik, Fossilbestimmung, Areaigestaltung und Evolution der Gat- tung. Pl. Syst. Evol. 134: 137-171. 1980. [Relationships based on umbo and mucro characters of the cone scales. KoEHNE, E. Deutsche ios 601 pp. Stuttgart. 1893. [Proposes sections Hap- loxylon and Diploxylor Konar, R. N. The ee clowy and embryology of Pinus Roxburghii Sar. with a com- parison with Pinus Wallichiana Jack. Phytomorphology 10: 305-319. 1960. & S. RAMCHANDANI. The morphology and embryology of Pinus Wallichiana Jack. Phytomorphology 8: 328-346. 1958. KONDRATYUK, E. M. An attempt to utilize the characteristics of pollen for classification of pine tree species. (In Ukrainian.) Visn. Akad. Nauk Ukrajins’k. RSR Bot. Sad. 3: 25-27. 1961.* KruGMan, S. L. The leuco- anthocyanin distribution in the genus Pinus. Forest Sci. 5: 169-173. 1959. [Ubiquitous in leaf, wood, and bark samples of the genus.] LANGLET, O. A cline or not a cline—a question of Scots pine. Silvae Genet. 8: 13-22. 1959. [P. sylvestris] Lanner, R. M. Natural hybridization between Pinus edulis and Pinus monophylla in the American Southwest. Silvae Genet. 23: 108-116. 1974. [See also D. K. BAILEy, 1987. Adaptations of whitebark pine for seed dispersal by Clark’s nutcracker. Canad. Touy Forest Res. 12: 391-402. 1982. [Discusses loss of seed wing and cone opening mechanism in . albicaulis.] Lecomte, H. Un pin remarquable de |’Annam. Bull. Mus. Hist. Nat. Paris 2: 193-197. 1921. [P. Kempf) Lepic, F. T., & M. T. ConKLE. Gene diversity and genetic structure in a narrow endemic Torrey pine (Pinus Torreyana Parry ex Carr.). Evolution 37: 79-85. 1983. [Exhibits very low diversity in electrophoretic alleles. ] & J. Fryer. A pocket of variability in Pinus rigida. Evolution 26: 259-266. 1972. elgg variation in frequency of cone serotin netics of pitch pine. U. S. Dep. Agr. Forest Serv. Res. Pap. WO- 27. 14 pp. 1974. 1p rigida. Littte, E. . , JR. Key to Mexican species of pines. Carib. Forester 23: 72-77. 1962. . CRITCHFIELD. ana ce the genus Pinus (pines). U. S. Dep. Agr. mae Serv, Misc. Publ. 1144: 1-51. 1969. [Reviews previous eee and presents a new synthetic classification. DorMAN. Geographic differences in cone-opening in sand pine. Jour. Forestry 50: 204, 205. 1952a. [Two races of P. clausa distinguished by presence or absence of serotiny.] ——.. Slash pine (Pinus Elliottii), its nomenclature and varieties. Ibid. 918- 280 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 923. 1952b. [Variety densa from southern Florida distinguished by presence of grasslike seedling stage & Slash pine (Pinus Elliotti’) including South Florida slash = No- menclature and description. U. S. Dep. Agr. Southeast Forest Exper. Sta. Pap. 36. li + 82 pp. 1954. [Thorough morphological comparisons of P. Elliottii ve its relatives. | , 5. LIvTLe, neo T. Doo.ittLe. Natural hybrids among pond, loblolly, and pitch pines. U. S. Dep. Agr. Forest Serv. Res. Pap. NE-67. 22 pp. 1967. [Detailed dis- cussion a and distribution of P. rigida, P. serotina, P. Taeda, and their hybrids. & F. 1. RiGHTer. Botanical description of forty artificial pine hybrids. U. S. De Agr. Forest Serv. Tech. Bull. 1345. iii + 47 pp. 1965. [Excellent illustrations of cones of F, hybrids and parental taxa.] Loock, E. E. M. The pines of Mexico a British Honduras. S. Afr. Dep. Forestry Bull. 35: 1-244. 1951.* ae J.C. Arboretum et fructicetum Rermuaane oe 1838. [Pinus, 4: 2152- 2292: an important early classification of the gen MARTINEZ, M. Los pinos mexicanos. ed. 2. 368 pp. Meats City. 1948. Masters, M. T. A general view of the genus Pinus. Bot. Jour. Linn. Soc. 35: 560-659. pls. 20-23, 1904. MASTROGIESEPPE, R. J., & J. D. Bee A study of Pinus Balfouriana Grev. Balf. (Pinaceae). Syst. Bot. 5: 86-104. 1980. Mattoon, W. R. Life history of the Fait pine. U. S. Dep. Agr. Bull. 244: 1-46. pls. 1-10. 1915. [P. echinata McCung, B. Ecological diversity in North American pines. Am. Jour. Bot. 75: 353- 368. 1988. are Se from multivariate comparisons among 34 species. ] McWILuIAM, J. The role of the micropyle in the pollination mechanism of Pinus. Bot. Gaz. oe 109- 7 1958. . Bioelectric phenomena in relation to pollination in Pinus. Silvae Genet. 8: 59- 959a. 61. . Interspecific incompatibility in Pinus. Am. Jour. Bot. 46: 425-433. 1959b. ——— & F. MerGEN. Cytology of fertilization in Pinus. Bot. Gaz. 119: 246-249. 1958. MERGEN, F. Genetic variation in needle characteristics of slash pine and in some of its hybrids. Silvae Genet. 7: 1-9. 1958a. [P. Elliottii.] Natural polyploidy in slash pine. Forest Sci. 4: 283-295. 1958b. [Partially polyploid chimeric seedlings found at very low frequency in nursery planting; illus- trates their aberrant leaf anatomy and reviews literature on pea) in Pinaceae. ] . Ecotypic variation in Pinus Strobus L. Ecology 44: 716-727. . Stairs, & E. B. SNYDER. Natural and es ere x shortleaf pine hybrids i in Mississipp1. es Sci. 11: 306-314 J& sporogenesis in pes ec ae and Pinus Taeda. Silvae Genet. 12: 128-130. 1966. [Micrographs of meiotic stages; commentary on phe- nology under conditions of sympatry. i Mixvcar, C. I. The Californian closed co ines (subsection Odcarpae Little and Critch- field): a taxonomic history and review. _ Taxon 35: 657-670. 1986. ,5S.H uss, M. T. CoNKLE, & R. D. WESTFALL. Allozyme differentiation sare biosystematics of the Californian closed-cone pines (subsect. Odcarpae). Syst. Bot. 13: 351-370. 1988 Mirov, N. T. The terpenes (in relation to the biology of the genus Pinus). Ann. Rev. Biochem. 17: 521-540. 8. . Taxonomy and chemistry of the white pines. Madrono 12: 81-89. 1953a. . Chemical fears of Diploxylon pines. Silvae Genet. 2: 93-96. 1953b. [Prelim- inary comparisons of oleoresin terpenoid and hydrocarbon profiles for the majority of species in the genus. 1989] PRICE, PINACEAE 281 . ees and flowering of pines. Forest Sci. 2: 328-332. 1956a. mposition of turpentine of lodgepole x ie pine hybrids. Canad. Jour. Bot. 34: 443-457. 1956b. [P. contorta x P. Banksian . Distribution of turpentine components a oan species of the genus Pinus. Pp. 251-268 in K. V. THIMANN, ed., The physiology of forest trees. New York. 1958. Composition of gum turpentines of pines. U. S. Dep. Agr. Forest Serv. Tech. Bull. oe v + 158 pp. 1961. [Ninety-four species, including all those in our area.] nology of tropical pines. Jour. Arnold Arb. 43: 218, 219. 1962. [Length of productive cycle decreased by lack of winter dormancy.] The genus Pinus. viii + 602 pp. New York. 1967. [Chapters on paleobotany, ocaeee morphology, sd i and ecology, chemistry, taxonomy; 105 species recognized. ] RANK, & E,. ZAVARIN. Chemical eee of P. Elliottii var. Elliottii turpentine and its possible relation to taxonomy of several pine spec hyto- chemistry 4: 563-568. 1965. [Preliminary investigation of terpenoid here in- dicates greatest similarity to the Caribbean species of subsect. AUSTRALES. ] Monr, C. Timber pines of the southern United States, together with a discussion of ie ra of their wood, by F. Roth. U. S. Dep. Agr. Bull. 13. 176 pp. 27 pis. 7. [P. echinata, P. Elliottii, P. glabra, - palustris, P. Taeda; maps, excellent Tee wood anatomy; notes on P. serotina. Moss, E. H. Natural pine hybrids in Alberta. Cun Jour. Res. C. 27: 218-229. 1949. [Pinus a x P. Banksiana.] Owens, J. N., & M. Morper. Seed-cone differentiation and sexual reproduction in western an pine (Pinus monticola). Canad. Jour. Bot. 55: 2574-2590, 1977. , 8. J. Simpson, & M. Motper. Sexual reproduction of Pinus contorta. I. Pollen development, the pollination mechanism and early ovule development. Canad. Jour. Bot. 59: 1828-1843. mee ,& ual reproduction of Pinus pee Post-dormancy OV wile embryo, and seed Gea oe scat Ibid. 60: 2071-2083. Panetsos, C. P. Natural hybridization Fi Pinus ieee ae Pinus brutia in Greece. Silvae Genet. 24: 163-168. PEATTIE, D. C. White pine; once king re primeval forests from Maine to Minnesota, the noble eae has exerted a powerful influence in American history. Sci. Am. 178(6): 48-53. 1948. [P. Strobus.] Peperickx, L. A. Chromosome ee between Pinus species. Silvae Genet. 19: 171-180. 1970. [Detailed karyotyp1 mpari of secondary constriction patterns for nine species. ] Prus-GLOWACKI, W., J. SZWEYKOWSKI, & R. NowAk. Serotaxonomical investigation of the European pine species. Silvae Genet. 34: 162-170. 1985. [Twelve species of subg. PINUS were well separated from two of subg. STRoBus.] RzeEDowSsKI, J., & L. VELA. Pinus Strobus var. chiapensis en la Sierra Madre del Sur de México. Ciencia 24: 211-216. 1966. Sax, K. Meiosis in interspecific pine hybrids. Forest Sci. 6: 135-138. 1960. SayLor, L. C. A karyotypic analysis of selected species of Pinus. Silvae Genet. 10: 77- 4. 1961. [Compares karyotypes of P. Strobus, P. Taeda, P. palustris, P. eee and P. oe aryot analysis of Pinus—group Lariciones. Ibid. 13: 165-170. 1964. [Karyotypes ioc for the 19 species of subsect. LARICIONES. ] Karyot ype analysis of the genus Pinus—subgenus oe Ibid. 21: 155-163. 1972. [Forty-six species, including all southern yellow pine Karyotype analysis of the genus Pinus—subgenus oo Ibid. 32: 119-124. 1983. [Twenty-two species, including P. Strobus.] & R. L. Koenic. The slash x sand pine hybrid. Silvae Genet. 16: 134-138. 282 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 1967. [P. Elliottii x P. clausa, an artificial hybrid between subsects. AUSTRALES and CONTORTAE. ra B. W. SmitH. Meiotic irregularity in species and interspecific hybrids of Pinus. m. Jour. Bot. 53: carey 1966. oo G., M. T. Conkie, & C. GRUNWALD. Local differentiation among Mediter- ranean populations oe Aleppo ie in their isoenzymes. Silvae Genet. 35: 11-19. 1986. [P. halepensis. | SEIKEL, M. K., S. S. HALL, L. C. epee: & R. - eee Chemotaxonomy as an aid in differentiating wood of eastern and western white pine. Am. Jour. Bot. 52: 1046-1049. 1965. [P. aie ne P eae are separable by relative concen- trations of certain flavanones.] SHarp, A. J. Pinus Strobus a of the United States. Jour. Elisha Mitchell Sci. Soc. 62: 629, 630. 1946. [Ecology of P. Strobus var. chiapensis.] SHaw, G. R. The pines of Mexico. Publ. Arnold Arb. 1. 29 pp. 22 pls. 1909. enus Pinus. Ibid. 5. 96 pp. 1914. [Important seee ose treatment of the ee ears illustrations of comparative morphology; 39 plates n the genus Pinus. Jour. Arnold Arb. 5: 225-227. 1924. — a a H. Tyson. Estimates of inbreeding in Pinus Banksiana. Canad. Jour. t. 49: 1241-1245. 1971. [Simple genetic basis for cone serotiny. ae P.E.,& L.C. Saytor. Studies of the Pinus rigida-serotina complex. I. A stud of geographic variation. Ann. Missouri Bot. Gard. 60: 174-191. 1973a. see gas comparisons of morphology; P. serotina treated as a subspecies of P. ri & ——— udies of the Pinus rigida-serotina complex. II. Natural cael among the Pinus rigida-serotina complex, P. Taeda and P. echinata. Ibid. 60: 192- 203. 1973b. Snyper, E. B., R. J. Dinus, & H. J. | Genetics of longleaf pine. U. S. Dep. Agr. Forest Serv. Res. Pap. WO-33. a pp. 1977. [P. palustris. ] & J. M. HAMAKER. Needle characters of hybrids of some species of southern pine. Silvae Genet. 27: 184-188. 1978. [Multivariate comparison of characters dif- ferentiating parents and hybrids pence P. echinata, P. Elliottii, P. palustris, P. SquiLLace, A. E. Geographic variation in slash pine. Forest Monogr. 10: 1-56. 1966. [Morphological and physiological variation in P. Elliottii.] & O. O. WELLS. Geographic variation of monoterpenes in tel oleoresin of loblolly pine. Silvae Genet. 30: 127-135. 1981. [Maps of the geographic variation for major terpene components in P. Taeda.] STEAD, J. W. Studies of variation in Central American pines V: a numerical study of variation in the Pseudostrobus group. Silvae Genet. 32: 101-115. 1983. STEINHOFF, R. J., & J. W. ANDRESEN. Geographic variation in Pinus flexilis and Pinus Si mis and its bearing on their taxonomic status. Silvae Genet. 20: 159-167. 197 D. G. Joyce, & L. Fins. Isozyme variation in Pinus monticola. Canad. Jour. Pores Res. 13: 1122-1131. 1983. StyLes, B. T., & P. K. Kuosta. Jn: IOPB chromosome number reports XLVII. Taxon 24: 143-146. 1975. [Chromosome numbers, all 2” = 24, for 16 species of Pinus.] THomson, R. B. The spur shoot of the pines. Bot. Gaz. 56: 362-385. pls. 20-23. 1914. Tomuinson, P. B. The biology of trees native to tropical Florida. ix + 480 pp. Published by the author, Harvard Univ. Printing Office, Allston, Massachusetts. 1980. [Pi- naceae, 61-67; P. Elliottii var. densa, including two full-page illustrations. ] VAN DER BurGu, J. H6lzer der niederrheinischen Braunkohlenformation, 2. Hélzer der Braunkohlengruben “Maria Theresia” zu Herzogenrath, “Zukunft West” zu Esch- weiler und “Victor” (Ziilpich Mitte) zu Ziilpich. Nebst einer systematisch-anato- mischen Bearbeitung der Gattung Pinus L. Rev. Palaeobot. Palynol. 15: 73-275. 1973. [An important account of the systematics of Pinus, with a revised classification 1989] PRICE, PINACEAE 283 emphasizing features of wood anatomy; micrographs of many extant and fossil species WAGNER, D. B., G. R. FURNiER, M. A. SAGHAI-MAROOF, S. M. WILLIAMS, B. P. pases & W. ALLARD. Chloroplast DNA polvinorphisiis in lodgepole and jack p and then hybrids. Proc. U. S. Natl. Acad. Sci. 84: 2097-2100. 1987. [P. a P. Banksiana.] Warp, D. B. Contributions to the flora of Florida—2, Pinus (Pinaceae). Castanea 28: ‘a 10. 1963. [Key, maps, ecological notes; P. clausa var. immuginata published for the open-coned race of sand pine. On the scientific name of the longleaf pine. Rhodora 76: 20-24. 1974. [P. palustris Miller, rather than the later names P. australis Michx. f. or P. longifolia Salisb. See FERNALD for opposite argument.] Watts, W. A. Late-Quaternary vegetation history at White Pond on the inner Coastal Plain of South Carolina. Quaternary Res. 13:187-199. 1980. [Late Pleistocene fossil needles of P. Banksiana from the southeastern United States; see also Ecology 51: 17-33. 1970, which records this species from northwestern Georgia contempora- neous with the main Wisconsin glaciation. ] WELLS, O. O. Geographic variation in ponderosa pine. I. The ecotypes and their dis- tribution. Silvae Genet. 13: 89-103. 1964. [Five ecogeographic races recognized based on common orchard — es.] , W. I. NANce, & B. A. THIELGES. Variation in needle traits in provenance tests of Pinus Taeda and P. ie Silvae Genet. 26: 125-130. 1977. [The two species are well separated in discriminant analyses of needle characters despite broadly overlapping geographic ranges. WHEELER, N. C., R. P. Guries, & D. M. O’MALLEY. Biosystematics of the genus Pinus, subsection Contortae. Biochem. Syst. Ecol. 11: 333-340. 1983. [Electrophoretic ave aac ed comparisons of P. virginiana, P. clausa, P. contorta, and P. Ban aaa J. W. Species hybridization in the white pines. Forest Sci. 5: 210-222. 1959. Wu, C. L. The taxonomic revision and ai eea cca study of Chinese pines. (In Chinese; English summary.) Acta Phytotax. Sinica 5: 131-163. pls. 24-27. YOsHIE, F., & A. SAKAI. Types of Florin rings, RR patterns of epicuticular wax, and their relationships in the genus Pinus. Canad. Jour. Bot. 63: 2150-2158. 1985. [The subgenera Stropus and Pinus are largely distinct in the form of their stomatal complexes and their epicuticular wax patterns. ZAVARIN, E. Taxonomy of pinyons based on terpenoids from wood oleoresin and on morphological and other biological characters. Proc. 2nd Pifion Symp., Chapingo, Mexico. (In press.) [Reviews systematics of subsect. CEMBROIDES. , W. B. CrITCHFIELD, & K. SNAJBERK. Turpentine composition of Pinus contorta x Pinus Banksiana hybrids and hybrid derivatives. Canad. Jour. Bot. 47: 1443- 1453. ee JBERK. Oleoresins of pinyons. Agr. Food Chem. 28: 829-834. 1980. Pe ae the terpenoid composition of wood and gum turpentines from eleven taxa.] ——— & ———. Monoterpene differentiation in relation to the morphology of Pinus culminicola, ‘Pinus Nelsonti, Pinus Pinceana and Pinus Maximartinezii. Biochem. Syst. Ecol. 15: 307-312. 1987. [Compares wood monoterpenes and morphological features for these unusual Mexican species of subsect. CEMBROIDES. . BAILEY, & E.C. ROCKWELL. Variability in essential oils and needle resin canals of Pin us longaeva from eastern California and western Nevada in relation [Compares wood and leaf terpenoids of the bristlecone and foxtail pines and the Mexican P. Rzedowskii, which is chemically distinct from the species of subsect. BALFOURIANAE. | 284 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 ZoseL, B. The natural hybrid between Coulter and Jeffrey pines. Evolution 34: 405- 413. 1951. [P. Coulteri x P. Jeffreyi.] 2. Picea A. Dietrich, Fl. Gegend Berlin, 794. 1824. Pyramidal evergreen trees (or shrubs under harsh conditions) with slender, more or less whorled branches. Branchlets pubescent with simple trichomes or glabrous, roughened by persistent leaf bases. Bark thin and scaly or some- times furrowed. Wood with axial and radial resin canals and ray tracheids regularly present; axial parenchyma absent. Leaves spirally arranged, linear, 4-sided and then stomatiferous on all sides [or dorsiventrally flattened and stomatiferous on the lower—but morphologically adaxial—surface], acute to acuminate, often sharply pointed [or sometimes obtuse or emarginate]; fibro- vascular bundle apparently single [or often double]; resin canals 2 [rarely 1], laterally placed; leaf bases persistent, ultimately woody. Pollen cones (micro- sporangiate strobili) subsessile [or stalked], borne singly in axils of leaves of previous year; microsporophylls many per strobilus, each with apex prolonged into a broad, nearly circular crest; microsporangia longitudinally dehiscent; pollen bisaccate. Ovulate cones borne at tips of previous year’s growth, ma- turing in | year, before pollination erect on short stalks; mature cones ovoid to cylindrical, 2-7[-15] cm long, sessile or short stalked, pendulous; bracts much smaller than scales at maturity; ovuliferous scales persistent, obovate to oblong, with apex rounded [to rhombic], entire to denticulate [erose]. Seeds with thin terminal wing; body rounded on the sides and usually acute at base; seed coat without resin vesicles. Cotyledons ca. 6 [4-15], denticulate or entire. Chromosome number 2” = 24. Type species: Picea rubra A. Dietr., nom. illeg., = P. Abies (L.) Karst.; see Britton, N. Am. Trees, 54. 1908. (Classical Latin name for a conifer, from pix, pitch.) —SPRUCE. A well-defined genus of approximately 35 species, Picea occurs widely in the boreal and temperate areas of the Northern Hemisphere, from the arctic of North America and Eurasia south to the higher mountains of Mexico and southern China. The greatest number of species (ca. 24) occurs in eastern and central Asia, where the taxonomic situation is particularly complex (see Cheng & Fu; Rehder, 1940; Schmidt-Vogt, 1977; Wright). Seven species are native to the United States and Canada; three of these occur in the eastern part, with only one, P. rubens Sarg., ranging south into our region. Sections within Picea are not easily distinguished on the basis of overall morphology, chemistry, or crossability. Willkomm divided the genus into two sections, Eupicea Willk. (= sect. PicEA) and OMoRIKA Willk., based on the cross-sectional form and stomatal arrangement of the needles. Mayr divided the genus into three sections, Morinda (= sect. PICEA), CAsicTA, and OMORIKA, on the basis of cone-scale shape and texture, in addition to the needle characters noted above. A number of later authors (Bobrov, 1970; Dallimore & Jackson; Gaussen, 1966; Kriissmann; Rehder, 1940) have used this three-section clas- sification, but there has been considerable disagreement over the placement of several species (see comparisons in Schmidt-Vogt, 1977). Wright, in a world- wide treatment utilizing both crossability data and comparisons of overall morphological similarity, found no clear-cut groups among the species. Thus 1989] PRICE, PINACEAE 285 neither his treatment nor the more recent monograph by Schmidt-Vogt (1977) divided the genus into sections. Preliminary comparisons of monoterpene pro- files (Schantz & Juvonen; Von Rudloff, 1975), leaf phenolics (Wellendorf & Kaufmann), and electrophoretic alleles (Wellendorf & Simonsen), as well as further crossability data (Fowler; Mikkola, 1969, 1972), also did not yield groups concordant with the earlier morphological sections. Smaller groups of morphologically similar species recognized as series by Bobrov (1970) seem to be of greater utility. Our only species, Picea rubens (P. australis Small; P. rubra (Du Roi) Link, non P. rubra A. Dietr.), red spruce, yellow spruce, he-balsam, 2n = 24, a prominent forest tree of the cooler parts of northeastern United States and adjacent Canada, occurs from eastern Ontario east to Nova Scotia, south to Pennsylvania, northern New Jersey, and Delaware, and locally south in the Appalachians to the higher mountains of eastern Tennessee and western North Carolina (Little, 1971, maps 41N, E). The species occurs most commonly on well-drained, rocky slopes, as compared to the closely related black spruce (P. mariana (Miller) BSP.), which is generally found on wet, boggy soils where it overlaps the range of the red spruce in the northeastern United States and southeastern Canada (Gordon, 1976; Manley & Ledig). Red spruce differs from black spruce in having lustrous dark or bright green (rather than glaucous) leaves and ovoid-cylindrical, reddish brown, ovulate cones that are usually deciduous at maturity (vs. ovoid, dull gray ones that persist for several years). Both red and black spruce have sharply pointed four-sided leaves and pubescent branchlets, and both are unusual in the genus in having terminal buds with awl-shaped points on their outer scales. Monoterpene pro- files are also very similar for the two species (Von Rudloff, 1967a, b; Wilkinson & Hanover). Reports of P. mariana in the southern Appalachians (e.g., by Small) are erroneous and based on material of P. rubens. Hybridization between red and black spruces is well documented (Gordon, 1976; Manley; Morgenstern & Farrar) in areas where the two come in contact in southern Canada, but the prevalence of intermediate plants is apparently limited by the habitat differences between the species, relatively low crossability between them (Gordon, 1976), and natural selection against hybrids and hybrid derivatives. Manley & Ledig reported a lack of heterosis and found that pho- tosynthetic rates of hybrids and backcrosses were significantly lower than those of the parental species under all environmental regimes examined. Khalil, however, did find heterosis in hybrids from the lowlands of New Brunswick. The significant geographic variation in morphological and physiological char- acters within Picea rubens found in Khalil’s long-term provenance studies appears to reflect selection along latitudinal and altitudinal gradients, as well as effects of hybridization and backcrossing. During late Pleistocene times additional species of Picea occurred in the southeastern United States (Watts). Large cones (up to 10 cm long) that appear to represent an undescribed species similar to P. glauca (Moench) Voss, white spruce, have been found as fossils at sites in Louisiana, Tennessee, and Georgia (Critchfield; Delcourt & Delcourt). Picea comprises a natural group, as is indicated by the overall similarity and crossability among its species (Wright), and is appropriately treated as a distinct 286 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 genus. It differs from the other pinoid genera primarily in lacking their unusual derived features, as detailed in the family treatment. It also differs from each of these genera in having leaves that are squarish in cross section and with prominent, persistent bases. As noted in the family treatment, Picea, Cathaya, and Pinus are apparently the only genera in the Pinaceae that produce the unusual serratene triterpenoids (He ef al.,; Hegnauer, 1986). Pinus and Picea are quite similar in the details of their pollination mechanisms. Picea differs from Pinus and Cathaya in lacking cleavage polyembryony (Doyle & Brennan) and in having a more asymmetric karyotype. Chromosome numbers have been reported for 22 of the 36 species of Picea recognized by Schmidt-Vogt (see particularly Kuo er a/.; Santamour, 1960; Sax & Sax: Seitz), with all species having 2” = 24. Aneuploid and polyploid seed- lings have been found at very low frequency in nursery plantings of P. Abies (Kiellander), but these presumably would not survive in nature. Supernumerary (B) chromosomes have been reported from some populations of P. sitchensis (Bong.) Carr., P. glauca, and P. obovata Ledeb. (P. Abies var. obovata (Ledeb.) Fellm.) but are lacking in others (Herzog; Moir & Fox; Pravdin er al.). Increase in nuclear-DNA content with latitude has been reported for P. glauca and P. sitchensis by Miksche (1968, 1971), but this has not been confirmed in wider sampling by Teoh & Rees. Intergradation or more limited hybridization among species is an important factor in the taxonomic complexity seen in Picea. Wright proposed that a number of Asian taxa of uncertain status were the variable products of inter- specific hybridization. Among the North American species, P. glauca and P. Engelmannii Parry show an extensive zone of intergradation in western Canada and Montana (Daubenmire; La Roi & Dugle; Ogilvie & Von Rudloff; Roche), leading some authors (e.g., T. M. C. Taylor, 1959) to treat these taxa as eco- geographic subspecies. Relatively frequent hybridization in regions of sympatry is also seen between the more obviously distinct species P. sitchensis and P. glauca (Daubenmire; Nienstadt & Teich; Roche), as well as P. rubens and P. mariana, discussed above. In contrast, the widely sympatric P. glauca and P. mariana only rarely hybridize (Little & Pauley) and are very difficult to cross artificially (Fowler). The closely related P. Engelmannii and P. pungens Engelm. may hybridize to a limited degree in areas of sympatry in the southern Rocky Mountains, but no evidence for this was seen in the electrophoretic study of Mitton & Andalora. Extensive programs of artificial interspecific crossing have been undertaken in Picea (Bongarten & Hanover; Fowler; Gordon, 1976; Mikkola, 1969, 1972; Wright), resulting in a substantial number of verified hybrid combinations, including several involving species from different continents. The infrageneric groupings proposed by Mikkola (1969) and Fowler on the basis of crossability agree poorly with those based on leaf anatomy. For example, P. mariana and P. rubens are more readily crossable with the morphologically dissimilar P. omorika (Panéié) Purkyne than they are with one another (Gordon, 1976), perhaps because of selection against hybridization in the two partially sympatric species (Fowler). Prezygotic barriers to hybridization (e.g., failure of the pollen tube to penetrate the nucellus) and early postzygotic barriers have been detailed for a number of interspecific combinations by Mikkola (1969). 1989] PRICE, PINACEAE 287 Detailed expositions of the comparative leaf anatomy of Picea are given by Marco and Colleau. Species assigned to sect. OMORIKA have needles dorsiven- trally flattened to varying degrees and are unusual in generally having stomata only on the adaxial leaf surface, which faces downward due to twisting of the leaf base. Picea sitchensis, however, often has stomata on all four sides of the leaf (Marco). Colleau attempted to limit sect. OMoRIKA to those species with a foliar stomatal density greater than 80 per mm?, but it is unclear that this yields a more natural group. Resin canals in species of sect. OMORIKA are near the abaxial leaf surface, while those in the other species are usually adjacent to the lateral angles of the leaf. The leaf mesophyll is differentiated into palisade and mesophyll cells in some species with flattened needles and in some with quadrangular ones (Marco). Another interesting feature of the leaves is the variation in color from bluish to greenish in some species—for example, P. pungens (Colorado blue spruce)— due to differences in the amount and structure of the surface wax (Hanover & Reicosky). Leaf-terpenoid patterns in Picea have been critically reviewed by Von Rud- loff (1975), following the earlier treatment of Schantz & Juvonen. Von Rudloff found relatively minor differences between some of the species assignable to sects. OmoriKA (P. omorika and P. Br Wenigne Si Watson) ane meee (P. rubens and P. mariana), but greater ones to the latter section or to sect. CasicTa. Schantz & Juvonen also found differences in leaf- and stem-terpenoid composition to be greater within than among the sections of Picea. The degree of geographic or intrapopulational variation in leaf-terpene composition is quite limited in some species (e.g., P. mariana and P. rubens; Von Rudloff, 1975) but quite variable in others (such as P. Engelmannii). Leaf-stilbene spot patterns and their intensities were surveyed for 23 species of Picea by Wellendorf & Kaufmann, who found no clear-cut subgroups within the genus. It has been alleged that sect. OMorIKA is distinct in its leaf-stilbene profile, but this has not been substantiated (Hegnauer, 1962, 1986). Overall leaf-phenolic profiles have proved useful in differentiating Picea species in the western United States and Mexico (La Roi & Dugle; R. J. Taylor & Patterson). A number of species of Picea are heavily utilized for pulpwood, lumber for construction, and specialized wood products, with P. Abies in Europe, and P. sitchensis and P. glauca in North America, being of particular economic im- portance. Acid rain has recently had serious effects on spruce forests in central Europe (e.g., the Black Forest in Germany) and is also adversely affecting P. rubens populations from northern Vermont southward in the Appalachian Mountains. Wood of P. Abies and P. glauca has been valued for violins and for the sounding boards of keyboard instruments, while the high strength-to- weight ratio of the wood of P. sitchensis made it particularly valuable for aircraft construction (Dallimore & Jackson). Spruces are frequently grown as orna- mentals in the cooler parts of the North Temperate Zone, with Norway spruce (P. Abies) and Colorado blue spruce (the glaucous form of P. pungens) the most widely cultivated in the United States. Spruce beer, prepared from P. rubens and P. mariana in the United States and P. Abies in Europe by boiling the leafy shoots with flavorings and sugar, was used as an antiscorbutic on sea voyages, while spruce chewing gum was obtained from the resin of the former two species before chicle became widely utilized (Dallimore & Jackson; Little, 288 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 1980). The purified resin of P. Abies, known as Burgundy pitch, has been used in medicinal plasters (Dallimore & Jackson), and the pitch of P. rubens was similarly utilized in the United States (Krochmal & Krochmal). REFERENCES: Under family references see particularly L. H. BAILEY; BEAN; BOUREAU; BUCHHOLZ, 1920, 1931, 1942: Butts & BUCHHOLZ; CAMPO & SIVAK; CAMPO-DUPLAN; CHANG; CHENG & Fu: C. R. CHowpHury, 1962; K. A. CHOWDHURY; CRITCHFIELD; DALLIMORE & JACKSON, Dov_e, 1945, 1963; DoyLE & BRENNAN; DoyLe & KANE; G. ERDTMAN, 1957, 1965; Ferré; FLorIn, 1931, 1963; FLous; FoweLis; GAussEN, 1966; GreGuss, 1955, 1972; He et al.; HEGNAUER, 1962, 1986; HickEeL; Ho & SzIKLAI; HUNTLEY & BirRKS; KHOSHOO, 1961, 1962; Konar & NAGMANI; KROCHMAL & KROCHMAL; KRUSSMANN; Kuo et al.; LittLe, 1971, 1980; MAHESHWARI & SINGH; MEHRA & KHOSHOO; MILLER, 1976, 1977, 1985; Norin; Owens & Simpson; eee PILGER; REHDER, 1940, 1949; SANTAMOUR, 1960, 1966, 1967; x: SEITZ; SINGH; SIVAK; SMALL; STERLING; TAYLOR; TEOH & REES; UENo: VON a 1975a: and WATTS. Bosrov, E. G. Generis Picea historia et systematica. (In Russian; nomenclature pri- marily in Latin and English.) Nov. Syst. Pl. Vasc. 7: 5-40. rei [New classification of Picea, with nine series recognized under the sections of MAyr.] . Die introgressive By aeicouen in der Gattung Picea ‘ Dietr. Symp. Biol. Hung. 12: 141-148. BONGARTEN, B. C., & J. W. HANoveER. Hybridization among white, red, blue, and white x blue spruces. Forest Sci. 28: 129-134. 1982. [Artificial crosses and backcrosses, al verification based on morphology and terpenoid composition Brooks, J. E., J. H. Borpen, H. D. Pierce, Jr., & G. R. Lister. Seasonal variation in foliar and bud se laa in Sitka spruce. Canad. Jour. Bot. 65: 1249-1252. 1987. BULARD, C. Quatre épicéas de la Chine centrale: P. ascendens, P. brachytyla, P. com- planata, P. Sargentiana. Bull. Soc. Hist. Nat. Toulouse 82: 177-186. 1947. [Re- printed in Trav. Lab. Forest. Toulouse, Tome I, Vol. IV, Art. XX: 1-10. 1947.] Burey, J. Karyotype analysis of Sitka spruce, Picea sitchensis (Bong.) Carr. Silvae Genet. 14: 127-132. 1965. Coteau, C. Anatomie comparée des feuilles de Picea. Cellule 67: 185-253. pls. 1-3. l DAUBENMIRE, R. Taxonomic and ecologic relationships between Picea aug uca and Picea Engelmannii. Canad. Jour. Bot. 52: 1545-1560. 1974. [Extensive intergradation occurs where the two species meet in the northern Rocky Mountains. DeLcourt, P. A. & H.R. Deccourt. The Tunica Hills, Louisiana- epee late glacial locality for spruce and deciduous forest species. Quaternary Res. 7: 218-237. 7. [Discusses large-coned Picea similar to P. glauca.] Fow.er, D. P. The hybrid black x Sitka spruce, implications to phylogeny of the genu Picea. Canad. Jour. Forest Res. 13: 108- - 1983. [Groupings based on ee bit, agree poorly with those from leaf anatom L. Rocue. Genetics of Engelmann nee e. U. S. Dep. Agr. Forest Serv. Res. Pap. WO-30. ii + 13 pp. 1975. [P. Enge/mannii, map, oe data.] Gaussen, H., & M. LACASSAGNE. Les épicéas du Kamchatka. Bull. Soc. Hist. Nat. Toulouse 59: 190-202. 1930. Gorpon, A. G. Ecology of Picea chihuahuana Martinez. Ecology 49: 880-896. 1968. . The taxonomy and genetics of Picea rubens and its relationship to Picea mariana. Canad. Jour. Bot. 54: 781-813. 1976. [Detailed comparison of Canadian populations indicates limited hybridization in areas of sympa 1989] PRICE, PINACEAE 289 Guepes, M., & P. Dupuy. Morphology of the seed-scale complex in Picea Abies (L.) Karst. (Pinaceae). Bot. Jour. Linn. Soc. 68: 127-141. 1974. [Teratological devel- opment of the cone scale.] HANOvER, J. W. Genetics of blue spruce. U.S. Dep. Agr. Forest Serv. Res. Pap. WO 28. 11 + 12 pp. 1975. [Picea pungens, map, scanning electron a of leaf surface, photos of cone variation.] . Retcosky. Surface wax deposits on Picea pungens and other conifers. Am. jour Bot. 58: 681-687. 1971. [Comparisons of green and glaucous needles by scanning electron microscopy. ] HERzoG, S. Investigations on chromosomes of Siberian spruce (Picea obovata Ledeb.). Silvae Genet. 36: 42-44. 1987. [No B chromosomes found in a sample of 38 trees. ] KuauiL, M. A. K. Genetic variation in red spruce (Picea rubens Sarg.). Silvae Genet. ra 164-171. 1987. [Summary of results of 20 years of common garden experiments nvolving numerous provenances; review of literature on the P. rubens—P. mariana He ee KIELLANDER, C. L. Polyploidy in Picea Abies. Hereditas 36: 513-516. 1950. [Found in a very low percentage of seedlings. ] KossuTH, S. V., & G. H. FEcHNER. Incompatibility between Picea pungens Engelm. and Picea Engelmannii Parry. Forest Sci. 19: 50-60. 1973. [Strong but not fully complete barriers to crossing seen at stages prior to fertilization; based on results of a single reciprocal cross. ] mane. M. Etude morphologique, anatomique et systématique du genre Picea. v. Lab. Forest. Toulouse, Tome II, Sect. VI, Vol. II, Art. I: 1-292. 1934. [Forty- ne species described and illustrated; ae now placed in synonym La Rot, G. H., & J. R. DuGLe. A systematic and genecological study of Picea glauca and P. Enge/mannii, using paper sea tograms of needle extracts. Canad. Jour. Bot. 46: 649-687. 1968. [Comparison of phenolic profiles indicates substantial in- tergradation in western Canada.] LinpqulisT, B. 1948. The main varieties of Picea Abies (L.) Karst. in Europe, with a contribution to the theory of a forest vegetation in Scandinavia during the last Pleistocene glaciation. Acti Horti Berg. 14: 249-342. 1948. [See also Moe.] LitT.e, E. L., Jr., & S. S. PauLey. A natural hybrid between black and white spruce in Minnesota. Am. Midl. Nat. 60: 202-211. 1958. [Partially fertile hybrid between the widely sympatric P. glauca and P. mariana reported for the first time; detailed morphological comparisons of the species. Man _ey, S. A. M. The occurrence of hybrid swarms of red ae black spruces in central New Brunswick. Canad. Jour. Forest Res. 2: 381-391. 1972. F. T. Lepic. Photosynthesis in black and red ene and their hybrid deriv- atives: oe isolation and hybrid adaptive inferiority. Canad. Jour. Bot. 57: 305-314. 1979. MANNERS, G. E., & E. P. Swan. aes in the bark of five Canadian Picea species. Phytochemistry 10: 607-610. 197 Marco, H. F. The anatomy of pale needles. Jour. Agr. Res. 58: 357-368. pls. I-7. 1939. Mayr, H. Monographie der Abietineen des Japanischen Reiches. viii + 104 pp. 7 pls. Munich. 1890. [Divides Picea into three sections: Morinda Mayr (= sect. Picea), CasicTaA Mayr, and OmorIKA Willk.; also treats Abies, Tsuga, Larix, and Pinus for Japan.] MERGEN, F., J. BURLEY, & G. M. FurNnIvAL. Embryo and seed development in Picea glauca (Moench) Voss after self-, cross-, and wind-pollination. Silvae Genet. 14: 188-194. 1965. MIKKOLA, ey Observations on interspecific sterility in Picea. Ann. Bot. Fenn. 6: 285- 969. [Causes of failed seed set in artificial crosses; compares relationships oa on morphology and crossability.] 290 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 —. Crossability between Picea Omorika (Panic) Purkyne and P. glauca (Moench) Voss. Ibid. 9: 33-36. 1972. MIKSCHE, J. P. Quantitative study of intraspecific variation of DNA per cell in Picea | Mitton, J. B., & R. ANDALORA. Genetic and morphological relationships between blue spruce, Picea pungens, and Engelmann spruce, Picea Engelmannii, in the Colorado Front Range. Canad. Jour. Bot. 59: 2088-2094. 1981. [Electrophoretic comparisons of a polymorphic enzyme system provided no evidence of hybridization; multi- variate ia inaea of morphological characters fail to give complete separation, how Mog, D. The post-glacial immigration of Picea Abies into Fennoscandia. Bot. Not. 123: 61-66. 1970. Morr, R. B., & D. P. Fox. Supernumerary chromosome distribution in provenances of Picea sitchensis (Bong.) Carr. Silvae Genet. 26: 26-33. 1977. [B chromosomes occur in varying frequencies in the populations below 53°N lat.; similar B chromosomes also seen in P. glauca. ] MorGENSTERN, E. K. Note on chromosome morphology in Picea rubens Sarg. and Picea mariana Saco B.S.P. Silvae Genet. 11: 163, 164. 1962. & J. L. FARRAR. Introgressive hybridization in red oe and black spruce. Univ. eae Fac. Forestry Tech. Rep. 4. 46 pp. 1964. NieENSTADT, H., & A. TeIcH. Genetics of white spruce. U.S. Dep. Agr. Forest Serv. Res. Pap. WO-15. iv + 24 pp. 1971. [P. glauca; map, crossability data. Oaitvig, R. T., & E. VoN RUDLOFF. Chemosystematic studies in the genus Picea (Pi- naceae). IV. The introgression of white and Engelmann spruce as found along the Bow River. Canad. Jour. Bot. 46: 901-908. 1968. Owens, J. N., & M. D. Biaxe. The pollination mechanism of Sitka spruce (Picea sitchensis). Canad. Jour. Bot. 62: 1136-1148. 1984. & M. MoLper. ine reproduction of white spruce (Picea glauca). Canad. Jour. Bot. 57: 152-169. ; aa a ants of Sitka spruce (Picea sitchensis). Ibid. 58: 886- & mOL, fae PSON, & G. E. CARON. The pollination mechanism of Engelmann spruce (Picea oe Canad. Jour. Bot. 65: 1439-1450. 1987. Pace, C. N., & K. D. RUSHFORTH. Picea Farreri, a new Poe conifer from upper Burma. Notes Roy. Bot. Gard. Edinburgh 38: 129-136. 1980. Pravpin, L. F., G. A. ABATUROVA, & O. P. SHERSHUKOVA. ae analysis of European and Siberian spruce and their hybrids in the U.S.S.R. Silvae Genet. 25: 89-95. 1976. [Similar karyotypes in P. Abies, P. obovata, and their hybrids; B chromosomes only in P. obovata.] Rocu_e, L. A genecological study of - genus Picea in British Columbia. New Phytol. 68: 505-554. pl. 1, foldout fig. 19 D. P. Fow.Ler. Genetics Be ae spruce. U. S. Dep. Agr. Forest Serv. Res. Pap. WO-26. iii + 15 pp. 1975. [P. sitchensis; map, poetnttra data.] ee L.O. Picea. Pp. 587-597 in Seeds of woody plants in the United States. U. S. Dep. Agr. Agr. Handb. 450. 1974 sae M. von, & S. JUVONEN. Chemotaxonomische Untersuchungen in der Gattung a. Acta Bot Fenn. 73: 1-51. 1966. [Preliminary comparisons of leaf, bark, and oo terpenoids; see discussion in VON RUDLOFF, 5.] ScHMIDT-VocT, H. Die Fichte. Bd. I. Taxonomie. Verbreitung. Morphologie. Okologie. Waldgesellschaften. xvill + 647 pp. Hamburg. 1977. [Taxonomy and distribution of the species of Picea of the world, maps, cone and habit photographs, extensive 1989] PRICE, PINACEAE 291 bibliography, followed by a ee treatment of the Norway spruce, P. Abies, continued in Bd. II/l. xvi + 563 p J SincH, H., & J. N. Owens. Sexual ia of Engelmann spruce (Picea Engel- mone: Canad. Jour. Bot. 59: 793-810. 1981. Supo, $8. Anatomical studies on the wood of Picea with some considerations on their geographical distribution and taxonomy. Bull. Gov. Forest. Exper. Sta. Meguro, Tokyo 215: 39-130. 1968.* Tay_or, R. J., & T. F. PAtrerson. Biosystematics of Mexican spruce species and populations. Taxon 29: 421-469. 1980. [Univariate distributions and discriminant analyses of morphological data; comparisons of phenolic profiles; P. mexicana Mar- tinez a variety of P. Engelmannii, P. chihuahuana Martinez is distinct.] , 5. WILLIAMS, & R. DAUBENMIRE. Interspecific relationships and the question of introgressive hybridization between Picea Engelmannii and Picea pungens. Canad. Jour. Bot. 53: 2547-2555. 1975. [Multivariate comparisons of morphology show intermediate individuals in both allopatric and sympatric populations; hybridity not clearly demonstrated TayLor, T. M. C. The taxonomic relationship between Picea glauca (Moench) Voss and P. Pram Parry. Madrono 15: 111-115. 1959. [Treats both as subspecies Von RuDLoFF, E. Chemosystematic studies in the genus Picea (Pinaceae). I. Introduc- tion. Canad. Jour. Bot. 45: 891-901. 1967a. [Includes preliminary comparisons of leaf terpenoids in P. glauca and P. mariana.) hemosystematic studies in the genus Picea (Pinaceae). II. The leaf oil of Picea glauca and P. mariana. [bid. 45: 1703-1714. 1967b. WELLENDORF, H., & U. KAUFMANN. Thin layer chromatography of fluorescent phenolic com eoeate comparisons among 23 species give no clear-cut groups; concentrations within spots tend to increase with latitude. ] & V. SIMONSEN. A chemotaxonomic study in Picea with isozymes in the seed endosperm. A preliminary report. Pp. 182-193 in D. Rupin, ed., Proceedings of the conference on biochemical genetics of forest trees. Umea, ‘Sweden. 1978. [Pre- liminary comparisons of 20 species based on presence/absence of alleles in seven enzyme systems; cluster diagram shows only limited agreement with morphological groups. WILKINSON, R. C., & J. W. HANovER. Geographical variation in the monoterpene composition of red spruce. Phytochemistry 11: 2007-2010. 1972. [P. rubens WILLKOMM, H. M. Forstliche Flora von Deutschland, Osterreich und der Schweiz. ed. 2. Leipzig. 1887.* [Divides Picea into sections based on leaf form. WRIGHT, J. W. Species crossability in spruce in relation to distribution and taxonomy. Forest Sci. 1: 319-349. 1955. [Important paper summarizing morphology and geo- graphic distribution of the species of Picea, as well as results of artificial hybridiza- YEH, F. C., M. A. KHALIL, Y. A. EL-Kassasy, & D. C. Trust. Allozyme variation in Picea mariana from Newfoundland: genetic diversity, population structure, and analysis of differentiation. Canad. Jour. Forest Res. 16: 713-720. 1986. 3. Tsuga (Endlicher) Carriére, Traité Conif. ed. 1. 185. 1855. Evergreen trees with pendulous leading shoots and branches. Bark usually furrowed and scaly. Wood pale, without normal resin canals; ray tracheids and axial parenchyma regularly present. Leaves spirally arranged (often appearing 2-ranked due to twisting of leaf bases), linear, flattened and bearing 2 whitened 292 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 abaxial stomatiferous bands [or quadrangular and bearing stomata above and beneath], narrowing abruptly to a short petiole, abscising at maturity from the ultimately ligneous leaf bases; apex obtuse and often notched, or acute; mar- gin entire or minutely serrulate; fibrovascular bundle double; resin canal single, abaxial to the vascular cylinder. Pollen cones (microsporangiate strobili) small (ca. 1 cm long), subglobose, stalked, borne singly in leaf axils of previous year; microsporophylls prolonged at tip into a short crest or knob; microsporangia subglobose, transversely dehiscent; pollen grains with saccae reduced to a cir- cular “‘frill’’ [or bisaccate in sect. HESPEROPEUCE]. Ovulate cones terminal on short branches, maturing the first year; mature ovulate cones relatively small (1-4[-7] cm long), usually pendulous [rarely erect], with cone scales persistent; bracts shorter than scales [to somewhat protrusive in T. /ongibracteata]; ovu- liferous scales suborbicular to oblong, entire or minutely lacerate. Seeds with delicate, obliquely oblong terminal wings; seed coat thin, bearing resin vesicles; cotyledons 2-7. Chromosome number 2” = 24. LecroTyPE SPECIES: 7suga Sieboldii Carr.;> see Britton & Brown, Illus. Fl. No. U.S. & Canada, ed. 2. 1: 62. 1913. (Name in Japanese for one of the species.) — HEMLOCK. A genus of ten or more species of forest trees in the Temperate Zone of eastern and western North America and eastern Asia west to the Himalayas, both at low elevations and in mountain areas up to near timberline. Four well- marked species occur in the United States, including two in our region, and at least six grow in Asia. Over ten species have been described from China (Down- ie; Flous, 1936b, 1937), of which several are doubtfully distinct and have been placed in synonymy by Cheng & Fu. Thorough study of their patterns of variability is badly needed. Fossils indicate that the genus was more widespread and diverse in the past, having occurred in Europe well into the Tertiary (Florin, 1963: Kirchheimer; Sivak, 1973). The species of Tsuga have generally been divided into two sections: TSUGA (sect. Eutsuga Engelm., sect. Micropeuce (Spach) Schneider), with pollen saccae reduced to a frill and stomata only on the lower (abaxial) leaf surface, and ‘heng in southern China. The latter species, which is unusual in the genus in having the ovulate cones strongly ascending at maturity and the cone-scale bracts exserted, has otherwise been treated in sect. Heopeuce Keng & Keng, and in subg. Paleotsuga Miki, which was based on fossil material. French workers of the Toulouse school, beginning with Campo-Duplan & Gaussen, have proposed that the two species of sect. HESPERO! are actually stabilized intergeneric hybrids, as discussed below. Section TsSuGA is represented in our area by two species. Tsuga canadensis (L.) Carr. (J. americana (Miller) Farw.), Canada hemlock, eastern hemlock, ‘Abies Araragi Sieb., Verh. Batav. Genootsch. 12: 12. 1830 (Tsuga Araragi (Sieb.) Koehne), is an earlier name for 7. Sieboldii. It was published without a usable description, illustration, or type and was only secondarily validated by Siebold & Zuccarini, Fl. Jap. 2: 15. 1842, under Abies Tsuga Sieb. & Zucc. I thus accept the widely used name 7. Sieboldii Carr. as correct. 1989] PRICE, PINACEAE 299 hemlock-spruce, 2” = 24, is a widely distributed species in the northeastern United States and adjacent Canada (Little, 1971, maps 91N, E) that occurs west to Wisconsin, eastern Minnesota, eastern Ohio, and southern Indiana. It is widely distributed in the Appalachian region of our area, ranging south to northern Georgia and northwestern Alabama and west to central Tennessee. It occurs in moist valleys and ravines and sometimes on steep, usually north- facing slopes at lower elevations in the northern part of its range, and on moist slopes and streamsides up to 1500 m altitude in our region. Ecogeographic variation in physiological and morphological features is pronounced between northern and southern provenances of the overall range and between north and south of the Tension Zone in Wisconsin (Eickmeier et a/.; Ruth), although discrete infraspecific taxa have not been proposed on this basis. Tsuga cana- densis is characterized by distichous, minutely serrulate leaves with rounded apices, pubescent young branchlets, and short-stalked, brownish to grayish ovulate cones with suborbicular scales. It differs from the rather similar western North American 7. heterophylla (Raf.) Sarg. in having a broader tree crown and narrower, more sharply defined white stomatal bands on the abaxial leaf surface. Tsuga caroliniana Engelm., Carolina hemlock, crag hemlock, southern hem- lock, 2n = 24, is a much more narrowly distributed species, occurring locally on dry slopes and ridges and along streamsides in the Appalachian Mountains from western Virginia and northeastern Tennessee south to extreme north- eastern Georgia and northwestern South Carolina (Little, 1971, map 94E), generally at higher elevations than 7. canadensis. It differs from the latter in having clearly spirally arranged, entire leaves (see FiGure 3b) and larger (2- 3.5 cm long), yellowish, early-deciduous cones with oblong (vs. broadly round- ed) scales (see FiGure 3k, 1, r). Although the two species are reported occa- sionally to occur sympatrically (e.g., in Stokes County, North Carolina, and Roanoke County, Virginia; Coker & Totten; C. E. Wood, pers. comm.), despite differences in habitat preference, they are very distinct morphologically, and hybridization between them has not been reported. Tsuga is distinguished from all other genera of the Pinaceae by its abruptly petiolate leaves with only a single resin canal abaxial to the vascular tissue. The genus has often been treated as similar to Picea because the twigs of both are roughened by the persistent leaf bases, the cones are usually pendulous at maturity, and prominent short-shoots are lacking. The two genera differ, how- ever, In embryogeny and in their pollination mechanisms (Owens & Blake; Owens & Molder), as well as in the characters of root, stem, and seed anatomy that separate T’suwga and the other abietoid genera from Picea and the other pinoid genera. Campo-Duplan & Gaussen and subsequent workers of the Toulouse school in France have proposed that 7suga Mertensiana is an intergeneric hybrid of the western North American 7. heterophylla and Picea sitchensis (Bong.) Carr., and that 7. /ongibracteata is a hybrid of Tsuga and Keteleeria. These hypotheses were based primarily on purported intermediacy for a series of morphological characters and irregularities in the bisaccate pollen of the two hemlock species. No case of natural or artificial hybridization has been confirmed for either of 294 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 1989] PRICE, PINACEAE ya the pairs of genera, however, and it is far more likely that the two taxa involved are merely surviving species that have retained primitive states (e.g., bisaccate pollen) instead of the derived ones seen in sect. TsuGA. Comparison of leaf- terpenoid profiles (Von Rudloff, 1975a, b), immunological distances from seed proteins (Prager et al.; Price et al.), and patterns of embryological development (Owens & Molder) all show 7. Mertensiana to be quite similar to other species of Tsuga but not to Picea. Chromosome numbers, all 2 = 24, are known for six species, including all four North American ones (Kuo e¢ al.; Sax & Sax; Vabré-Durrieu, 1954b). The karyotype of nine approximately isobrachial and three heterobrachial chro- mosomes is similar to that of Picea (see Khoshoo, 1962; Sax & Sax) Interspecific hybridization 1s extremely infrequent among the North Amer- ican species. Tsuga x Jeffreyi (Henry) Henry is a cultivated plant of unknown origin that is morphologically intermediate between 7. Mertensiana and T. heterophylla. Attempts at artificial hybridization between these taxa have been unsuccessful (Meagher; Taylor), but a portion of the infrequent morphological intermediates from areas of near sympatry have proved to be intermediate in leaf-phenolic profile (Taylor). Comparative pollen morphology of 7suga has been treated in detail by Campo, G. Erdtman (1957, 1965), Sivak, and Ueno (1957). The species of sect. TSUGA are unique in the Pinaceae in having the pollen saccae reduced to a circular frill and in often having spinules atop the verrucate surface orna- eae (G. Erdtman, 1957, 1965; Sivak, 1973). The spinules are apparently absent in 7. canadensis and are relatively small in T. eee and T. het- erophy ie (Sivak, 1973). The size of the two saccae is quite variable within the species of sect. HESPEROPEUCE (Campo; Ho & Sziklai), and fossil Tsuga pollen FiGure 3. Tsuga. a—o, 7. caroliniana: a, branchlet at time of shedding of pollen, microsporangiate strobili at upper right, 2 ovulate strobili terminating branches below, x 1; b, leaf, showing characteristic entire margin, x 5: c, part of twig after leaf abscission, showing persistent leaf bases, x 6; d, microsporangiate seo with bud scales at base, x 6; e-g, microsporophyll with dehisced sporangia, viewed from below, from above, and from side, x 10; h, twig (leaves removed) with a ovulate cone (subtended by bud scales), very young leaves beginning to show at tip of short twiglet at mght, x 3; i, cone-sale bract (abaxial view) at time of pollination, the smaller ovuliferous scale above completely hidden, x 10; j, adaxial view of ovuliferous scale and bract at time of pollination, micropyles of the 2 ovules below, x 10; k, mature cone in moist condition, cone scales appressed, x 1; 1, mature dry cone, seeds already shed, impressions of seed abaxial views, bract showing at base inn, = 1; 0, seed with wing, adaxial view, x 3. p— u, 7. canadensis: p, branchlet with mature unopened cone (compare leaf arrangement with “a,” noting especially the smaller appressed leaves on upper side), x 1; q, leaf, ae serrulate margin and abrupt narrowing to petiole, x 5; r, mature dry cone with seeds shed (compare to “!””), x 1; s, mature cone scale, abaxial view, with bract at base, x 1; t, seed with wing, adaxial view, x 3; u, diagrammatic longitudinal section of seed, with seed coat hatched, megagametophyte stippled, and embryo (with 2 cotyledons) unshaded, x 10. 296 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 has been reported to show a wide range of variation between saccate and frilled forms (Kirchheimer; Wodehouse). In accord with the differences in pollen form, the pollination mechanisms of Tsuga Mertensiana and members of sect. TsuGA are quite dissimilar, al- though both lack the pollination-drop mechanism of Picea and Pinus. The relatively smooth pollen of 7. Mertensiana is caught on microdroplets on the integumentary flaps (Owens & Blake). In contrast, the spinulose pollen surface of species such as 7. heterophylla apparently aids in its adherence to the waxy surface of the bracts, which are much more prominent than the cone scales at this stage of development (Colangelli & Owens; Owens & Blake). Although this unusual pollen-capture mechanism has also been reported in 7. dumosa (D. Don) Eichler, it needs to be documented in other species of sect. TSUGA (Doyle & O’Leary). Leaf anatomy of 7suga canadensis has been studied in detail by Gambles & Dengler. Crystals, presumably of calcium oxalate, on the outer surfaces of the leaf-parenchyma cells are an unusual feature to be sought in other species of the genus Relatively few comparative chemical studies of Tsuga have been made (see Hegnauer, 1962, 1986). Von Rudloff(1975a, b), comparing leaf-terpene profiles of each of the North American species except 7. caroliniana, found 7. Mer- tensiana to be more or less intermediate between 7. heterophylla and T. can- adensis and not at all similar to Picea sitchensis. Leaf-phenolic pigments of seven species of 7suga were compared by Taylor, yielding a cluster diagram agreeing rather poorly with morphological data. Immunological comparisons of seed-protein extracts (Prager et a/.; Price et a/.) indicate that 7. heterophylla and 7. Mertensiana are more similar to one another than to representatives of any other pinaceous genus. Both of our species of Tsuga are utilized as ornamentals in the eastern United States, and numerous cultivars of 7. canadensis have been selected (Den Ouden & Boom; Kriissmann). Bark of Tsuga is rich in tannins, and that of 7. can- adensis has been heavily exploited for use in the tanning industry. Wood of most 7Tsuga species is relatively soft and is used most often for crates and paper pulp, while that of 7. heterophylla is often employed for construction purposes. REFERENCES: Under zones! references see L. H. BAILey; BEAN; Butrs & BUCHHOLZ; CARRIERE; CHANG; CHENG & Fu; C. R. CHowpHurRy (1962); CoKkER & ToTTEN; DALLIMORE & JACKSON; Den Opes Boom; Doy te, 1945, 1963; DoyLe & KANE; DovLe & O’ LEARY; G. ERDTMAN, 1957, 1965; FERRE; FLorIN, 1931, 1963; FLous, 1936a; FOwELLS; GAUSSEN, 1966; Greauss, 1955, 1972; HEGNAUER, 1962, 1986; HICKEL; Ho & SzIKLAI; KHOSHOO, 1961, 1962; KRUSSMANN; Kuo et al.; LitTLe, 1971, 1980; MAHESHWARI & SINGH; MILLER, 1976, 1977, 1985: PHILLIPs;: PILGER: PRAGER ef al.; Price et a/l.; REHDER, 1940, 1949; SARGENT; SAX & SAX; SINGH; SIVAK, 1975; TAYLOR; UENO; VON RupDLorF, 1975a; and WODEHOUSE BoRMANN, F. H., & R. B. Platt. A disjunct stand of hemlock in the Georgia Piedmont. Ecology 39: 16— 93, 1958. [T. canadensis. | 1989] PRICE, PINACEAE Phe Campo, M. VAN. Quelques pollens d’hybrides d’Abiétacées. Silvae Genet. 4: 123-126. 1955. [Illustrates pollen of 7. longibracteata and T. Mertensiana in comparison with that of artificial interspecific hybrids in Abies CaMPO-DuPLAN, M. VAN, & H. GAuSSEN. Sur quatre hybrides de genres chez les Abié- tinées. Bull. Soc. Hist. Nat. Toulouse 83: 95-109. 1948. [Proposes that 7. longi- bracteata and T. Mertensiana are actually stabilized intergeneric hybrids; see cri- tiques in DUFFIELD and Owens & MOLDER. CHENG, W.-C. A new 7suga from southwestern China. Contr. Biol. Lab. Sci. Soc. China Bot. 7: 1-3. 1932. [7. longibracteata.] COLANGELLI, A. M., & J. N. Owens. Pollination mechanism in western hemlock. (Ab- stract.) West. Forest Genetics Assoc. 1984 Ann. Meeting, 20. 1984. [Pollen adheres preferentially to bracts of the ovulate cone in 7. heterophylla. Downig, D. G. Chinese species of Tsuga. Notes Roy. Bot. Gard. Edinburgh 14: 13-19. pl. 194. 1923. [Seven species recognized, including five new ones, most of which are now placed in synonymy under 7. yunnanensis (Franchet) Masters or 7. dumosa (D. Don) Eichler.] DUFFIELD, J. W. Review of “Sur quatre hybrides de genres chez les Abiétinées” [VAN CaAmMpo-DupLAN & GAUSSEN, 1948]. Jour. Forestry 48: 440. 1950. EICKMEIER, W., M. ADAMS, & D. LESTER. Two physiological races of Tsuga pea in Wisconsin. Canad. Jour. Bot. 53: 940-951. 1975. [Physiological and mo log- ical differences between seedlings from northern and southern provenances. a FITSCHEN, J. Die Gattung Tsuga. NEU ree Denere’ oes roe 1-12. 22 Pateen species recognized; key; descriptions; habit ph i ies i tin FLous, F. Espéces nouvelles du genre foe Bull. Soc. Hist. Nat. Toulouse 69: 409- 415. 1936b. [7. Blaringhemi (presumably = T. diversifolia (Maxim.) Masters), 7. crassifolia (= T. Mertensiana (Bong.) Carr.; isotype, uc!), 7. tchekiangensis (T. cue aes Pritzel var. ieheontenss (Flous) Cheng & L. K. Fu). ion du genre Tsuga. Ibid. 71: 315-450. 1937. [Eighteen species recognized, one a sees hybrid; several are doubtfully distinct.] GAMBLES, R. L., & N. G. DENGLER. The leaf anatomy of hemlock, Tsuga canadensis. Canad. Jour. Bot. 52: 1049-1056. 1974. [Excellent LM and SEM photographs. ] Ho, R.-H., & J. N. Owens. Microstrobilate morphology, microsporogenesis, and pollen fo nmation in western hemlock. Canad. Jour. Forest Res. 4: 509-517. 1974. [T. heterophylla.] JAMES, R. L. Carolina hemlock—wild and cultivated. Castanea 24: 112-134. 1959. [T. caroliniana. KIRCHHEIMER, F. Uber Tsuga-Pollen aus dem Tertiar. Planta 22: 171-179. 1934. MEAGHER, M. D. Studies of variation in hemlock (7suwga) populations and individuals from southern British onan. Unpubl. Ph.D. dissertation, Univ. British Colum- 4 Murri__, W. A. The es of the archegonia and Sa in the hemlock spruce (Tsuga canadensis Carr.). Ann. Bot. 14: 583-607. pls. 31, 1900. Owens, J. N., & M. D. BLAKE. Pollen morphology and development 7 the pollination recha mist in Tsuga heterophylla and T. Mertensiana. Canad. Jour. Bot. 61: 3041- 3048. 1983. M. Mo.per. Sexual Ages of mountain hemlock (Tsuga Mertensiana). Canad. Jour. Bot. 53: 1811-1826. 19 Rut, R. H. 7Jsuga (Endl.) Carr. Pp. a 827 in Seeds of woody plants in the United States. U. S. Dep. Agr. Agr. Handb. 450. 1974 SHaw, A.C. The essential oil of Tsuga canadensis (L.) Carr. Jour. Am. Chem. Soc. 73: 2859-2861. 1951. SivaAk, J. Observations nouvelles sur les grains de pollen de Tsuga. Pollen Spores 15: 397-457. 1973. [Detailed comparison of pollen from extant and fossil species of sect. TSUGA.] 298 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 STANLAKE, E. A., & J. N. Owens. Female gametophyte and embryo ry oe in western hemlock (7'suga heterophylla). Canad. Jour. Bot. 52: 885-893. UENO, J. Relationships of genus 7suga from pollen morphology. Jour. an ase Osaka Univ. D. 8: 191-196. pls. 7, 2. 1957. Vasré-Durriezu, A. L’hybride Tsugo-Picea Hookeriana et ses parents. Etude des plan- tules. Bull. Soc. Nat. Hist. Toulouse 89: 47-54. 1954a. [The purported hybrid 7. Mertensiana (Bong.) Carr. is more or less intermediate between 7. Seer ats and Picea site eee in seedling morphology. e Tsugo-Picea Hookeriana et ses parents. Etude chromosomique et Bo ae “bid 401-405. 1954b. [2n = 24 for both Tsuga heterophylla and T. Mertensiana.] Von Rup torr, E. Chemosystematic studies in the genus Tsuga. Leafand twig oil analysis of western hemlock. Canad. Jour. Bot. 53: 933-939. 1975b. [Terpenoid profiles of T. heterophylla differ from those of 7. canadensis and T. Mertensiana; see also VON Rup LorF (1975a), cited under family references. ] Abies Miller, Gard. Dict. abr. ed. 4. Vol. 1 (alph. ord.). 1754. Pyramidal evergreen trees (rarely shrubs at high elevation or latitude) with well-separated ‘‘whorls” of branches. Bark with frequent resin blisters, smooth on younger trunks, becoming fissured with age. Wood pale, without normal resin canals (occasionally with traumatic ones); ray tracheids only occasionally present; axial parenchyma present. Leaves spirally arranged (often appearing 2-ranked due to twisting of the leaf bases), linear to linear-lanceolate, flattened and bearing 2 whitened abaxial stomatal bands [less often triangular or quad- rangular in cross section and with stomata frequent above, as well as below], obtuse and often notched at apex or acute [to sharply pointed in Abies brac- teata], fibrovascular bundle double; resin canals 2 [rarely to 12], in the medial parenchyma [or near the leaf margin]; leaf scars circular, nearly flush with twig [to somewhat raised]. Pollen cones (microsporangiate strobili) short stalked, pendent, borne singly 1n axils of leaves of previous year; apex of microspo- rophyll prolonged into a short knob; microsporangia opening transversely; pollen bisaccate. Ovulate cones borne on upper side of topmost branches, maturing the first year; mature cones erect, ovoid [to cylindrical], short stalked or sessile; bracts fimbriate, with apiculate to long-attenuate tips, exserted [or more commonly included]; ovuliferous scales fan shaped, with margin deeply indented near base; bracts and scales abscising from the persistent cone axis. Seeds compressed; body triangular with acute base, the thin terminal wing broadest and often oblique at apex, partially folded over open surface of seed; seed coat thin, with large resin vesicles. Cotyledons [2—]4 or 5[-8 or rarely 14]. Chromosome number 2n = 24. Lectotype species: Abies alba Miller (Pinus Picea L., not A. Picea Miller); see Britton, N. Am. Trees, 74. 1908. (Classical Latin name for fir or related conifers.) — Fir, SILVER FIR. a A genus of approximately 40 species, Abies is widely distributed in temperate and boreal portions of the Northern Hemisphere, with the largest number of species 1n mountainous areas. The centers of distribution are in eastern and central Asia (ca. 17 species), southern Europe and North Africa (ca. eight), the western United States and Canada (eight), and Mexico and Guatemala (ca. six). Abies sibirica Ledeb. ranges north of the Arctic Circle in the U.S.S.R., 1989] PRICE, PINACEAE 299 while A. /asiocarpa (Hooker) Nutt. and A. balsamea (L.) Miller extend well into the subarctic of North America. Abies balsamea and A. Fraseri (Pursh) Poiret occur in eastern North America, but only the latter is native to our region. Infrageneric groups in Abies are based on a limited number of morphological characters, which are probably subject to considerable homoplasy, and thus a number of conflicting classifications have been proposed (see comparisons in Critchfield, 1988; Liu). Liu’s recent monographic treatment, which is followed here, recognizes two subgenera, PsEUDOTORREYA (Hickel) Franco (including only the unusual California endemic 4. bracteata D. Don) and Apies, continuing the precedent of Franco (1942, 1950). Within subg. Apres, Liu recognized 14 sections based on ovulate-cone, leaf, and branchlet characters. Each of these sections is quite restricted in geographic range, usually to part of a continent. Franco (1950) divided subg. Apres into seven sections and seven series, with the 11 terminal taxa very similar in composition to those of Liu. The North and Central American species of subg. Asres fall into three sections (one in- cluding two series) in Franco’s classification and five in Liu’s. The less-formal classification of Gaussen, in which the genus is divided into five sections (one for A. bracteata only) and 14 species groups, differs in having sections that include both American and Eurasian species. A significantly different classi- fication by Matzenko (1964, 1968) split the genus into four sections and 18 series based largely on relative size and form of the bracts and scales in the ovulate cone. This treatment often agrees poorly with the preponderance of morphological data and utilizes names without regard to nomenclatural prior- ity. The terminal groups recognized by Liu, Franco, and Gaussen for the North American taxa have proved useful in predicting additional character states such as crossability and distribution of wood crystals. The sectional classifications of the entire genus seem to be fairly arbitrary, however, and should be evaluated using further data on crossability and biochemistry. Subgenus Abies is characterized by emarginate to acute (vs. acuminate and callose-tipped) leaf apices, ovate to globose (vs. fusiform) winter buds, and included or exserted bracts without a long-aristate middle lobe. Section BAL- SAMEAE Engelm. emend. Liu (leaves with median resin canals; fibrovascular bundles distinct; ovulate cones oblong to ovoid) comprises three species of the United States and Canada: 4. balsamea, A. lasiocarpa, and the sole species in our region, A. Fraseri (A. balsamea var. Fraseri (Pursh) Spach), Fraser fir, southern balsam fir, she-balsam. Fraser fir is restricted to higher elevations (ca. 1200-2040 m) in the Appalachian Mountains of southwestern Virginia, eastern Tennessee, and western North Carolina. It occurs in pure stands on cool, moist slopes above 1500 m altitude on Mount Rogers, Roan Mountain, Grandfather Mountain, Mount Mitchell, and Clingman’s Dome, and in mixed stands with red spruce (Picea rubens) and various hardwoods at lower elevations (Coker & Totten; Liu). In recent years several populations of A. Fraseri (e.g., on Mount Mitchell) have declined significantly in size, possibly as a result of the effects of acid rain. Abies Fraseri is closely related to A. balsamea, differing most prominently in having the bracts of the ovulate cone strongly exserted and reflexed (vs. 300 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 included or somewhat exserted but not strongly reflexed). Abies Fraseri tends to have more lines of stomata per band on the leaves (Fulling, 1934; Rehder, 1940), and its young branchlets pubescent with generally reddish (vs. grayish) hairs. Plants of 4. balsamea from New England and eastern Canada with somewhat exserted bracts were treated by Fernald as var. phanerolepis. Al- though bract/scale ratio 1s apparently positively correlated with elevation, it exhibits considerable variation within the northeastern populations (Jacobs ef al.; Lester, 1968; Myers & Bormann), making this variety not very distinct. Similar plants with a tendency toward exserted bracts are also found scattered in the mountains of West Virginia and northern Virginia in the area between the allopatric ranges of A. balsamea and A. Fraseri. Several studies (Jacobs et al.; Robinson & Thor; Thor & Barnett) have compared these populations to A. Fraseri and eastern A. balsamea and have found intermediacy but no sig- nificant increase in variability in regard to morphological features, terpenoid profiles, or electrophoretic alleles. This is consistent with the hypothesis that these populations are remnants of a previous continuum from which A. Fraseri has diverged, rather than products of hybridization. Abies Fraseri shows sig- nificant differences from A. balsamea in the percentage composition of several monoterpenes in the wood and balsam oleoresins (Thor & Barnett; Zavarin & Snajberk, 1972), but only minimal divergence in electrophoretic alleles (Jacobs et al.), The two taxa are highly crossable (Hawley & DeHayes, 1985a, b) but are kept distinct by geographic isolation. Given the number of other species in the genus that are subject to partial intergradation, A. Fraseri is treated here as a separate species, albeit presumably a recent derivative of A. balsamea. Abies is a distinct genus characterized by a unique combination of characters that separate it from the other abietoid genera of Pinaceae. Cedrus and Pseu- dolarix differ from Abies in their pronounced shoot dimorphism, Keteleeria and 7’suga differ in their ovulate cones with persistent scales; and Tsuga differs in its cones that are terminal and usually pendulous at maturity. Abies, Kete- leeria, and Pseudolarix all have cone scales that are strongly indented at the base, with the seed apex extending beyond the cone scale and the seed wing consequently partially folded over the abaxial surface of the seed, a condition most pronounced in Abies. Abies and Keteleeria are also similar in having circular, only slightly raised leaf scars, but Kereleeria differs in its clustered pollen cones, along with its persistent cone scales. Immunological comparison of seed proteins indicates that Abies is most similar to Cedrus and Keteleeria (Price et al.). While the genus is a distinct one, species delimitation within Abies is often problematic. Intergradation in morphology and chemistry occurs in a number of areas where related species are geographically adjacent. This could be due to incomplete evolutionary differentiation of taxa or to hybridization between distinct ones, and both of these situations probably occur in Abies (Critchfield, 1988). In addition to the A. balsamea—-A. Fraseri complex discussed above, intermediate populations have been documented in other North American species pairs: A. concolor—A. grandis in sect. GRANDES Engelm. (Hamrick & Libby; Von Rudloff, 1976; Zavarin et al., 1977a), A. balsamea-A. lasiocarpa in sect. BALSAMEAE Engelm. (Hunt & Von Rudloff: Parker et a/.; Von Rudloff, 1975a), and A. magnifica-A. procera in sect. Nosices Engelm. (Liu). Inter- 1989] PRICE, PINACEAE 301 mediate populations between related species are also known in Europe—for example, involving A. alba and A. cephalonica (Mitsopoulos & Panetsos)—and Asia (Jain; Liu). Disagreements persist as to the distinctness of several Chinese taxa (cf. Cheng & Fu; Gaussen, 1964; Liu; Rushforth, 1983, 1984). Studies of morphological and chemical variation in the Chinese firs are badly needed. Several artificial interspecific hybrids have been reported in Abies (Critch- field, 1988; Hawley & DeHayes, 1985a, b; Klaehn & Winieski; Mergen et al.), but many of the early reports were not verified using F, morphology or chem- istry. Critchfield’s (1988) recent review indicates that at least for the North American taxa, interspecific hybrids can frequently be produced within the sections of Liu, but seldom between them. None of the crosses of American and exotic species attempted by Critchfield succeeded, except for that of J. concolor with Mexican A. religiosa, although earlier reports of unverified crosses of American and Eurasian species need further investigation. In some cases failure of crosses due to embryo abortion or pollen-tube inhibition have been documented (Kormutak; Kormutak & Dubovsky). Chromosome counts (all 2 = 24) have been obtained for 22 of the 39 species recognized by Liu and for one putative interspecific hybrid (see especially Khoshoo; Mergen & Burley); Abies Fraseri apparently has never been counted. Natural polyploidy within species seems to be very rare. Karyotypes have been investigated by several authors (e.g., Mehra & Khoshoo; Mergen & Burley; Sax & Sax). According to Mergen & Burley, who compared six species of diverse taxonomic affinities and analyzed the effects of different chemical treatments, the three chromosomes with the smallest short arm are clearly heterobrachial and the two with the next longest ones are sometimes heterobrachial. This compares to earlier reports of five heterobrachial chromosomes by Mehra & Khoshoo and Sax & Sax. Secondary constrictions were found by Mergen & Burley to be variable in position and number, even within individual trees. The pollination mechanism of Abies entails adherence of the pollen to mi- crodroplets on the funnel-shaped integument tip, followed by slight infolding of the tip (Doyle & Kane; Owens & Molder; Singh & Owens, 1981, 1982). The pollen is then dormant for up to two months before germinating and growing into the nucellus (megasporangium). Syngamy occurs within a week of pollen germination in the North American species studied by Owens and coworkers. Inefficiencies in the pollination mechanism may be a factor in the high fre- quency of unfilled seeds seen in several species of firs (Franklin; Owens & Molder). Seed crops also tend to vary substantially in size over two- to several- year cycles (Franklin). Variation in the form of the bracts and their length relative to the scales in the mature ovulate cones is particularly marked among—and sometimes with- in—species of Abies (Liu). The color of young cones at the time of pollination also varies strikingly, ranging from green or yellowish to various shades of purplish, bluish, or reddish among and sometimes within species (Liu; Sturgeon & Mitton). Features of leaf anatomy, including number and position of resin canals, presence and distribution of hypodermal cells, degree of separation of the fibrovascular bundles, and distribution of stomata, are often useful in separating species of Abies (Fulling, 1934; Liu), although care must be taken to assure 302 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 sufficient and comparable samples, since there can be considerable intraspecific variability. Distribution of calcium-oxalate crystals in the ray parenchyma of the wood has been compared among the nine species of Abies native to the United States (Kennedy et a/.). Rhomboid and elongate crystals are both regularly present in wood samples from five of the western species but are sparse or absent tin the western A. amabilis (Douglas) Forbes and in the three species of sect. BALSAMEAE (except for elongate crystals in some samples of A. /asiocarpa). Rhomboid or elongate crystals have also been reported from the wood of several Asian species (Greguss, 1955, 1972). Terpenoid profiles from leaf, wood, or bark oleoresins have been used ina number of systematic comparisons among or within species of Abies (e.g., Von Rudloff, 1976; Von Rudloff & Hunt; Zavarin & Snajberk, 1965, 1972; Zavarin et al., 1973, 1977a, b, 1978), with emphasis on the North American taxa (see reviews by Critchfield, 1984; Giannasi & Crawford; Von Rudloff, 1975a). These studies have been particularly useful in assessing intergradation among species and in highlighting geographic races within such widespread species as 4. con- color, A. balsamea, and A. lasiocarpa. A survey of cortical terpenoids in the Japanese and Taiwanese firs (Zavarin et al., 1978) yielded a diversity of mono- terpenoid and sesquiterpenoid compounds and showed partial agreement with previous morphological classifications. A particularly interesting sesquiterpenoid derivative isolated from the wood of Abies balsamea is juvabione, the “paper factor,” which exhibits juvenile hormone activity in certain hemipteran insects (Williams). Juvabione and re- lated todomatuic-acid derivatives with juvenile hormone activity have been found in several other Abies species and appear to be produced in response to woolly-aphid attack in some (Puritch & Niholt). This group of compounds has also recently been found in the wood of Cedrus Deodara (see Hegnauer, 1986), which is of interest in light of the morphological similarities between these genera. Himachalene sesquiterpenoids characteristic of Cedrus have also been found in Abies (Zavarin et al., 1978) Several species of Abies are important as ornamental trees, particularly in cool-temperate areas. Abies balsamea and others, including A. Fraseri, are highly sought after as Christmas trees. Several species of Abies are important sources of pulpwood, and fir wood is used for a variety of purposes not requiring great durability. “Canada balsam,”’ employed as a mounting medium in mi- croscopy and as an optical cement, is an oleoresin obtained from the bark blisters of 4. balsamea, while “Strasbourg turpentine,”’ used in varnishes and artists’ paints, is a refined oleoresin from Abies alba. Bark and leaf oleoresins from Abies have also been utilized medicinally by native peoples in North America to treat colds and to aid wound healing, among other purposes (Ar- nason et al.; Krochmal & Krochmal). REFERENCES: Under family references see ARNASON et al.; L. H. BAILEY; BEAN; BOUREAU; BUCHHOLZ, 1920, 1931, 1942; Burrs & BUCHHOLZ; CAMpo & SIVAK; CAMPO-DUPLAN; CHA NG; CHENG & Fu: C. R. CHowpuHury, 1962; K. A. CHowpDHURY; COKER & TOTTEN; ae 1989] PRICE, PINACEAE 303 1984; DALLIMoRE & JACKSON; Doy Le, 1945, 1963; DoyLe & KANE; G. ERDTMAN, 1957, 1965; FERRE; Firorin, 1931, 1963; FLous; GAussEN, 1964; GIANNASI & CRAWFORD; 1961; KonaR & NAGMANT; KROCHMAL & KROCHMAL; KRUSSMANN; LITTLE, 1971, 1980: MAHESHWARI & SINGH; MEHRA & KHOSHOO; MILLER, 1976, 1977, 1985; Norin; OnENs & SIMPSON; PHILLIPS; PILGER; PILGER & MELCHIOR; PRICE et al., 1987; RERDER. 1940, 1949; SARGENT; SAX & Sax; SIVAK; TIEGHEM; UENO; Von RUDLOFF, 1975a: and WATTS. Boivin, B. Abies balsamea (Linné) Miller et ses variations. Nat. Canad. 86: 219-223. 1959. [Suggests that 4. /asiocarpa be included as a ea ie s.] CLARKSON, R. B., & D. E. FAIRBROTHERS. A serological a of eastern North American Abies (Pinaceae). Taxon 19: P0217: 1970. ae of seed proteins of A. Fraseri and A. balsamea by immunology and disc electro- phoresis.] CRITCHFIELD, W. B. Hybridization of the Catone firs. Forest Sci. 34: 139-151. 1988. [Reviews crossability among f the genus. ] DanliELs, J. D. Variation and intergradation i in the grand fir-white fir complex. 235 pp. Unpubl. Ph.D. dissertation, Univ. Idaho, Moscow. 1969.* [4. grandis, A. concolor.] DeVore, J. E. Fraser fir in the Unicoi Mountains. Castanea 37: 148, 149. 1972. [Range extension to Haw Knob on the North Carolina/Tennessee border. ] ENGELMANN, G. A synopsis of the American firs (Abies Link). Trans. Acad. Sci. St. Louis 3: 593-602. 1878. FERNALD, M. L. A new variety of Abies balsamea. Rhodora 11: 201-203. 1909. [Abies balsamea var. phanerolepis. FRANCO, J. Do A. Sub-géneros e seccaes do género Abies Mill. Bol. Soc. Port. Ci. Nat. 13(Suppl. 2): 163-170. figs. I-35. 1942. Abetos. (In Portuguese.) Anais Inst. Sup. Agron. Lisboa 17: 1-260. figs. 1-109. 1950, [Monographic treatment of Abies; keys, illustrations for the 20 species culti- vated in Portugal.] FRANKLIN, J. F. Abies. Pp. 168-183 in Seeds of woody plants in the United States. U. S. Dep. Agr. Agr. Handb. 450. 1974. FULLING, E. H. Identification, by leaf structure, of the species of Abies cultivated in the United States. Bull. Torrey Bot. Club 61: 497-524. pls. 26-32. 1934. [Keys and extensive illustrations. ] . Abies intermedia, the Blue Ridge fir, a new species. Castanea 1: 91-94. 1936. [For plants from Virginia and West Virginia that are intermediate between J. bal- samea and . oo ] HAMRICK, J. L., & W. J. Lippy. Variation and selection in western U.S. mont p I. White fir. ‘Siac Genet. 21: 29-36. 1972. [4. concolor.] HAwLey, G. J., H. DeHayes. Hybridization among several North American firs. I. Giant Canad: Jour. Forest Res. 15: 42-49. 1985a. [4. balsamea, A. Fraseri, A. lasiocarpa.] Il. Hybrid verification. Jbid. 50-55. 1985b. Hunt, R. S., & E. Von RupDLorF. Chemosystematic studies in the genus Abies. I. Leaf and twig oil analysis of alpine and balsam firs. Canad. Jour. Bot. 52: 477-487. 1974. [Significant differences between the terpenoid profiles of 4. /asiocarpa and of A. balsamea, and between eastern and western populations of the latter.] Hutcuinson, A. H. The male gametophyte of Abies. Bot. Gaz. 57: 148-153. 1914. ——. Fertilization in Abies balsamea. Ibid. 60: 457-472. 1915. —. Embryogeny of Abies. Ibid. 77: 280-288. 1924. Jacoss, B. F., C. R. Wertu, & S. I. GUTTMAN. Genetic relationships in Abies (fir) of eastern Aue States: an electrophoretic study. Canad. Jour. Bot. 62: 609-616. 1984. [Com s A. Fraseri to eastern A. balsamea.] JAIN, K. K emer hybridization in the West Himalayan silver firs. Silvae Genet. 25: 107-109. 1976. [A. spectabilis, A. Pindrow.] KENNEDY, R. W., C. S. R. Sastry, G. M. BARTon, & E. L. Etuis. Crystals in the wood 304 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 of the genus oe pare eae to Canada and the United States. Canad. Jour. Bot. 46: 1221-1228. pl. 7. 1968. Kim, Y. D. Sern ae pare of the genus Abies in Korea. (In Korean.) Jour. Korean Forestry Soc. 83.* . Km. ae on the morphological and anatomical characteristics of the genus . Abies in Korea. Jour. Korean Forestry Soc. 62: 68-75.* KLAEMN, F. U., & J. A. Winteski. Interspecific Aenea in the genus Abies. Silvae Genet. 11: 130-142. 1962. [Results of artificial crosse KormutTak, A. Cytological study on crossability of some Abies species. Biol. Bratislava 36: 245-251. 1981.* . Dusovsky. Embryo inviability in aes concolor var. Lowiana x Abies alba crossing. Biol. Bratislava 39: 3-14. 1984. Laina, E. V. The genus Abies and recognition species Scot. Forestry 10: 20-25. 1956. [Keys to vegetative and reproductive mater Lams, W. H. A conspectus of North hee ae firs (exclusive of Mexico). Proc. Soc. Am. Foresters 9: 528-538. 1914. Lester, D. T. Variation in cone morphology of balsam fir, Abies balsamea. Rhodora 70: 83-94. 1968. . Geographic variation in leaf and twig monoterpenes of balsam fir. Canad. Jour. Forest Res. 4: 55-60. 1974. Liu, T.-S. A monograph of the genus Abies. 608 pp. Taipei. 1971. [Includes detailed illustrations ven distribution maps. Martinez, M. Tres especies nuevas mexicanos del género Abies. Anal. Th Biol. Méx. 13: -634. 1942. [4. durangensis, A. mexicana, A. Vejari, figs. 1-10.] os Abies mexicanos. Ibid. 19: 11-104. 1948. opens Ko, A. E. The firs of the Eastern Hemisphere. (In Russian; English summary.) Acad. Sci. USSR, Inst. Bot. V. L. Komarov. Flora et Syst. Pl. Vasc. 13: 1-103. 1964. [New generic classification emphasizing bract and scale characters, extensive bib- liography. . Series novae generis Abies Mill. Acad. Sci. USSR, Inst. Bot. V. L. Komar Novitates Syst. Pl. Vasc. 1968: 9-12. 1968. wan and species for the ae and series in the classification of MATZENKO (1964 Meura, P. N. Conifers of the Himalayas with particular reference to the Abies and Juniperus complexes. Nucleus 19: 123-139. 1976. [4. Forestii, A. Pindrow, n= 12.] MERGE ae & J. BuRLey. Abies karyotype ree eee Silvae Genet. 13: 63-68. 1964. [A. cephalonica, A. x Borisii-regis, A. lasiocarpa, A. firma, A. procera, A. gua- pe aes all n = 12. . _ & B. A. Simpson. Artificial hybridization in Abies. Zichter 34: 242- 251. 1964. * D. T. Lester. Microsporogenesis in Abies. Silvae Genet. 10: 146-156. 1961. Mitsopouos, D. J., & C. P. PANETsOs. Origin of variation in fir forests of Greece. Silvae Genet. 36: 1-15. 1987. [Chemical and ony gues comparisons of A. alba, A. cephalonica, and putative hybrid derivative Myers, O., JR., & F. H. BoRMANN. Phenotypic eee in Abies balsamea in response to altitudinal and geographic gradients. Ecology 44: 429-436. 1963. Owens, J. N., & M. Motper. Sexual reproduction of Abies amabilis. Canad. Jour. Bot. PARKER, W. H., J. MAzeE, & G. E. BRADFIELD. Implications of morphological and an- atomical variation in Abies balsamea and A. lasiocarpa from western Canada. Am. Jour. Bot. 68: 843-854. 1981. [Multivariate and univariate comparisons indicate significant peace ] Puritcu, G. S., & W. W. NiHOLT. Occurrence of juvabione-related compounds in grand fir and Pacific silver fir eid by balsam woolly aphid. Canad. Jour. Bot. 52: 585- 1989] PRICE, PINACEAE 305 587. 1974. [Compounds related to ses appear to be produced in response to aphid attack on A. grandis and A. amabilis. RAMSEUR, G. S. saa for the ai Aone an Abies (fir or balsam). (Abstract.) ASB Bull. 8: 31. REHDeER, A. The firs Bice and Guatemala. Jour. Arnold Arb. 20: 281-287. 1939. Rosinson, J. F., & E. Toor. Natural variation in Abies of the southern Appalachians. Forest Sci. 15: 238-244. 1969. Rotter, K. J. Resin canal via in the needles of balsam, alpine and Fraser firs. Forest Sci. 12: 348- ae RUSHFORTH, K. D. Notes on oon silver firs 3. Notes Roy. Bot. Gard. Edinburgh 43: 269-275. 1986. [See also Ibid. 41: 333-338. 1983; Ibid. 535-540. 1984. SINGH, H., & J. N. Owens. Sexual reproduction in subalpine fir (Abies lasiocarpa). Canad. Jour. Bot. 59: 2650-2666. 1981. & Sexual reproduction in grand fir (Abies grandis). Ibid. 60: 2197-2214. 1982. pace ee K. B., & J. B. Mitton. Cone color polymorphism associated with elevation white fir, Abies concolor, in southern Colorado. Am. Jour. Bot. 67: 1040-1045. 0. THOR, E., & P.E. BARNETT. Taxonomy of Abies in the southern Appalachians: variation Viculk, M.-T., & H. GAussen. Révision du genre Abies. Bull. Soc. Hist. Nat. Toulouse 57: 369-434. 1928; 58: 245-564. 1929. [Reprinted in Trav. Lab. Forest. Toulouse, Tome IJ, Art. I: 1-386. 1929.] Von Rup.orr, E. Chemosystematic studies in the genus Abies. I. Leaf oil analysis of grand fir. Canad. Jour. Bot. 54: 1926-1931. 1976. [Terpenoid variation in south- central Oregon consistent with introgression involving 4. concolor. . S. Hunt. Chemosystematic studies in the genus Abies. II. Leaf and twig oil analysis of amabilis fir. Canad. Jour. Bot. 55: 3087-3092. 1977. Warp, D. S. The first record of the Fraser fir. Castanea 27: 78, 79. 1962. WILLIAMS, C. M. Hormonal interactions between plants and insects. Pp. 103-132 in E. SONDHEIMER & J. B. SIMEONE, eds., Chemical ecology. New York and London. 1970. [Juvenile hormone activity of “paper factor” from Abies balsamea. ZAVARIN, E., & K. SNAJBERK. Chemotaxonomy of the genus Abies. I. Survey of the terpenes present in Abies balsams. Phytochemistry 4: vie ee 1965. [Preliminary comparison of 17 species, including 4. ba/samea and A. & eographical a of monoterpenes ie Abies balsamea and A. Fraseri. Ibid + 1407-14 1972. ,& WwW ne Monoterpene variability of Abies amabilis cortical oleoresin. Biochem. Syst. Ecol. 1: 87-93. 1973. ,& . Terpenoid chemotaxonomic studies of Abies grandis. Ibid. 5: 89-93. 1977a. Relation of cortical monoterpenoid composition of Abies to tree age and size. erence: 16: 770-772. 1977b. ,& . Lee. Chemical relationships between firs of Japan and Taiwan. Biocheni: Syst. Eccl. 6: 177-184. 1978. [Cortical terpenoids of six species.] JOURNAL OF THE ARNOLD ARBORETUM INSTRUCTIONS FOR AUTHORS General policy The Journal of the Arnold Arboretum is a refereed botanical quarterly that presents the results of original research in systematic and evolutionary botany and related disciplines. Critical, in-depth reviews of major books published in the above fields are also included. In addition to Arboretum staffand associates, botanists worldwide are encouraged to submit manuscripts to the Journal. We particularly welcome those that record the results of studies utilizing Harvard University Herbaria (A and GH) collections or focusing on geographic areas of traditional interest to Arboretum staff. Submission of manuscripts Manuscripts should be submitted in triplicate to Ms. E. B. Schmidt, Man- aging Editor, Journal of the Arnold Arboretum, 22 Divinity Avenue, Cambridge, Massachusetts 02138. A copy of the manuscript should be retained so that when reviews and/or editorial suggestions are received, any necessary correc- tions can be made and the appropriate portions of the paper resubmitted. For ease of editing, an outline of the paper (not to be published) showing the basic structure of the manuscript should be included. Preparation of manuscripts Papers should be triple spaced throughout (including title, abstract, text, citation of specimens, footnotes, acknowledgments, bibliography, and figure legends), on bond (not erasable) paper, with wide margins on all four sides. Nothing should be underlined except generic and infrageneric scientific names, italics when present in a quotation, and the collector and collection number of all specimens cited. FORM AND STYLE. This can be determined from a recent issue of the Journal. The title should be as short as possible; it should usually contain the name of the family concerned but not authorities of scientific names. Each author’s current address should be given as a footnote to his name. An abstract should be included. Abbreviations should be employed only when two or more letters will be saved, and with the exception of units of measure, compass directions, and herbarium designations, they should always be followed by a period. Metric measurements should be used when possible. Authority names should be given for all generic and infrageneric taxa the first time they are creas in the text unless they are included in the formal taxonomic treatm Acknowledgments should be placed at the end of the paper Sabie the bib- liography. Footnotes should be kept to a minimum and should be numbered consecutively throughout the paper, with the exception of those appearing in tables. Here either standard symbols or lower-case letters should be used. In case of question, the latest edition of Words into Type, by Marjorie E. Skillin ef a/. (Prentice-Hall), should be consulted. CITATION OF SPECIMENS. Currently accepted geographic names should be used, with spelling according to a standard source. Names of countries should be in English and should be typed in regular capital and lower-case letters. Below the country level, names may be in the language of the country involved. If this option is taken, careful attention should be paid to consistency, spelling, and accent marks. Arrangement of areas should be consistent (geographic or erie racine Soa the former) within a paper. ata obtained from the label should not be changed unless it is obviously wrong or meee critical information; in this case, additions or corrections should be bracketed. When a collector has a common surname, his initials should be given. With specimens from the Indo-Malesian area, care should be taken to de- termine whether collections are institutional or not. In the case of an institu- tional collection, the institutional series and number (e.g., LAE 20257) should be given. This, in addition to the location, is all the information needed for brief, unambiguous citation. If it is desirable to include the ite he information should be placed after the institutional series and num e.g., LAE 20257, Foreman or LAE 20257 (Foreman)). The list areas eae be arranged by institutional numbers, where applicable, for ease of use. As an alternative to lengthy specimen citations, a list of specimens examined, prepared in the normal Journal style, may be provided to the Managing Editor. This list will be deposited in the Library of the Arnold Arboretum and Gray Herbarium, and copies will be prepared and distributed to the libraries of the New York Botanical Garden and the Missouri Botanical Garden. It will be available by interlibrary loan from the cooperating institutions; a brief state- ment in the published paper will notify readers of its availability. Specimen citations will still be necessary for nomenclatural purposes (e.g., citation and designation of types). ILLUSTRATIONS AND LEGENDS. Reference must be made in the text to all maps, figures, and plates. Insofar as possible, their sequence should be determined by the order in which they are mentioned. Plates (illustrations grouped together at the end of an article) and figures (illustrations scattered through the text of an article) should be prepared with Journal page proportions in mind. The maximum size after reduction is 4.25 by 6.5 inches (10.8 by 16.5 cm) for figures, and 4.25 by 6.75 inches (10.8 by 17.1 cm) for plates. A figure may occupy any portion of the length of a page; plates should be more or less full- page size. To a mailing and handling, mounted illustrations must be of a manageable si Line eee fens habit and plant parts should be lettered in the same order (e.g., habit, leaf, inflorescence, whole flower, calyx, corolla) on each illustration aes a paper. Pho s should be trimmed, grouped appropriately, and mounted with no space eee them on stiff white cardboard with a margin of at least 1 inch left on all four sides. The author’s name and the figure number(s) should be noted on the back of each illustration. To prevent bending or other damage, the art work should be wrapped carefully and shipped flat. Three clear copies of each illustration should be included for review purposes. Legends should be written in telegraphic style (see back issues for examples). They should be grouped in numerical sequence on a separate page, rather than placed below each figure or plate. Illustrations of each type (i.e., figures, plates, or maps) should be numbered consecutively and separately, figures and maps with Arabic numerals and plates with Roman numerals. For example, a paper could include figures 1-3, maps 1-3, and plates I-VII. If the illustrations are to be figures, each figure should be numbered separately, with the numbers running in order through the text. Subdivisions of the figures should be indi- cated with letters. Thus, figures 1-5 may appear on one page, figures 6-8 on the following page, and Figure 9A—C on the next. In the case of plates, each page of photographs is a separate plate; the individual photographs comprising the plates are numbered or lettered consecutively. Illustrations with dark back- grounds should have white letters or numbers and vice versa. Whenever pos- sible, scales should be included in the illustrations; any magnifications necessary in the legends should be calculated to include reduction of illustrations to our page size. Illustrations are not returned to the author after publication unless this is requested. TABLES. Titles for tables should be short, with all explanations placed in footnotes. Tables should be as simple as possible and must be neatly typed. Long and/or complicated tables can be photographed directly if they are in good order and the copy is clear enough (in this case the copy should not be triple spaced, but should have spacing appropriate to contents and headings); this eliminates the chance for error and the need for proofreading. Tables should be numbered consecutively using Arabic numerals. Each table should be cited in the text. BIBLIOGRAPHY. The Guide to Citation of Botanical Literature in the Inter- national Code (all editions through 1972) should be followed. When possible, reference should be made to past issues of the Journal for form. Titles should be abbreviated according to Botanico-Periodicum-Huntianum style. Runovers should be indented. Except in cases where confusion would result, only authors’ initials are used in place of first and middle names. All typing should be done in regular capital and lower-case letters, and nothing should be underlined except generic and infrageneric scientific names. Titles of articles and books should not be capitalized except for the first word, scientific names, and proper nouns and adjectives. Author’s alterations Charges for extensive alterations in proofs will be billed to the author. Page charges Authors are requested to help defray printing costs. Although actual printing costs are much higher, $20.00 per page is the customary charge, and authors are expected to make every effort to pay. Under special circumstances the fee may be reduced or waived altogether, if this is agreed upon in advance. Ability or inability to pay will in no way affect acceptance or handling of a manuscript. The Journal of the Arnold Arboretum is a refereed botanical journal that presents the results of original research in systematic and evolu- tionary botany and related disciplines. Beginning in 1988, critical, in- depth reviews of major books published in the above fields will also be included on a periodic but timely basis. In addition to Arboretum staffand associates, we encourage botanists worldwide to submit manu- scripts to the Journal. We particularly welcome those that record the results of studies utilizing Harvard University Herbaria (A and Gu) collections or focusing on geographic areas of traditional interest to Arboretum staff. A guide for potential contributors, periodically pub- lished in the Journal, is available on request. Journal of the Arnold Arboretum April, 1989 CONTENTS OF VOLUME 70, NUMBER 2 The Biosystematics and Evolution of the Polyalthia hypoleuca Com- plex (Annonaceae) of Malesia, I. Systematic Treatment. SEVEN Fl POGSTALS cues vipiedeos saaaseierriadsaneanev as 153-246 The Genera of Pinaceae in the Southeastern United States. MiGseeT As PRICE ac 6 odik sh casas Se eea le ns oe eo beReewuve eas 247-305 Volume 70, Number 1, including pages 1-152, was issued 10 January 1989. JOURNAL OF THE ARNOLD ARBORETUM HARVARD UNIVERSITY VOLUME 70 NUMBER 3 ISSN 0004-2625 Journal of the Arnold Arboretum The Journal of the Arnold Arboretum (ISSN 0004-2625) is published quarterly in January, April, July, and October for $70.00 per year, plus $10.00 or $15.00 postage for addresses outside of the United States, by the Arnold Arboretum of Harvard University. It is printed and distributed by Allen Press, Inc., 1041 New Hampshire Street, Lawrence, Kansas 66044. Second-class postage paid at Lawrence, Kansas. POSTMASTER: send address changes to Journal of the Arnold Arboretum, % Allen Press, Inc., P.O. Box 368, Lawrence. Kansas 66044 Subscriptions and remittances should be sent to Journal of the Arnold Arboretum, 1041 New Hampshire Street, Lawrence, Kansas 66044, U.S.A. Claims will not accepted after six months from the date of issue. Volumes |—51, reprinted, and some back numbers of volumes 52-56 are available from the Kraus Reprint Corporation, Route 100, Millwood, New York 10546 U.S.A EDITORIAL COMMITTEE S. A. Spongberg, Editor E. B. Schmidt, Managing Editor P. S. Ashton K. S. Bawa P. F. Stevens C. E. Wood, Jr. Printed at Allen Press, Inc., Lawrence, Kansas COVER: The stylized design appearing on the Journal and the offprints was drawn by Karen Stoutsenberger. THIS PUBLICATION IS PRINTED ON ACID-FREE PAPER. JOURNAL OF THE ARNOLD ARBORETUM VOLUME 70 JuLy 1989 NuMBER 3 THE GENERA OF SUBORDER APOCYNINEAE (APOCYNACEAE AND ASCLEPIADACEAE) IN THE SOUTHEASTERN UNITED STATES! THOMAS J. ROSATT? Suborder Apocynineae L. Benson ex Rosatti,*? comprising the Apocynaceae Juss. and Asclepiadaceae R. Br.. is recognized here as a compromise between conflicting traditional and more recent views about the systematics of these plants. Although most botanists have maintained the “‘apocynads” and the 'Prepared for the Generic Flora of the Southeastern United States, a long-term project made possible by grants from the National Science Foundation and at this writing supported by BSR-8303100 and BSR-8415637 (Norton G. ae Ce investigator), under which this account was prepared, and BSR-8415769 (Carroll E. Wood, Jr., principal investigator). This treatment, the 128th in the series, follows the format established in the i one (Jour. Arnold Arb. 39: 296-346. 1958) and continued to the present. The area covered by the Generic Flora includes North gu South Carolina, Georgia, Florida, Tennessee, Alabama, Mississippi. Arkansas, and Louisiana. e descripuons are based primarily on the a of this area. with fafecmartn about eat members of a family or genus in brackets. References | have not verified are marked with an asteris Norton Miller tie anol moos Coseixe recognition for maintaining the Genet Flora project. and for their man in particular. Gordon C. Tucker should be canonized for having calmly endured my almost daily practice of dribbling ideas around him and then slam- dunking them into the computer without even a passing acknowledgment of the integral role he played in their development. Ihsan A. Al-Shehbaz, at Harvard, and George K. Rogers, now at the Missouri Botanical Garden, have continued to be valuable colleagues in matters regarding the Generic Flora and otherwise. Walter S. Judd provided a good deal of critical information about species not native to but occurring without cultivation in Florida. Donald J. Drapalik freely shared his ious about species of Matelea and Gracies in the Southeast and provided liquid-preserved material of special importance to these investigations. Warren D. Stevens conveyed helpful insights regarding taxonomic concepts at and see the level of genus. | thank Li Ping-T’ao for sending a copy of the treatment of the Apocynaceae and Asclepiadaceae he coauthored for the Flora of the Mee 's Republic of C the many fine illustrations considerably enhanced my otherwise meager knowledge about these plants in that part of the world. Elizabeth A. Shaw helped with the Latin diagnosis. This manuscript, as well as the others | have authored in the Generic Flora series, has benefited from the editorial expertise of Elizabeth B. Schmidt and Stephen A. Spongberg. In Berkeley James M. Affolter, Robert A. Price, and Toni M. Rosatti helped with the considerable task of reading proof, © President and Fellows of Harvard College, 1989. Journal of the Arnold Arboretum 70: 307-401. July, 1989. 308 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 ‘‘asclepiads” as distinct families since Robert Brown first separated them early in the nineteenth century, an increasing number (e.g., Hallier; Judd, pers. comm.; Stebbins; Stevens; Thorne, 1976, 1981) believe that the two groups grade into each other so gradually that it would be more appropriate to treat them as a single family; in addition, the peculiar gynoecium (carpels free in the ovule-bearing region but connate above) present nearly throughout is nearly unknown in other flowering plants. Despite these facts, even botanists (includ- ing this author) supporting unification of the two families usually continue to recognize them as distinct for reasons of practicality and tradition; each is a and William J. Stone provided technical computing assistance. These investigations were based on collections of the libraries and/or herbaria associated with the New York te > Mu seum, Harvard Univ ersily. the University of California at Berkeley, and the University of F ustrations were prepared by Karen Stoutsenberger (KS) and ioe Z "eas (RAW), saris under the direction of Kenneth R. A apg. Carroll Wood, and George K. Brizicky, from ae Robertson. Wood. Brizicky, A. H. Curtiss, R. A. Howard, N. G. Miller, H. B. — + ere Barks: G. V. Nash. F. Prichard, and R. J. Eatc ahs esanen is published as Contribution ere 600 of the New York State Science Service. Biol New York State Museum, The State Education Department, Albany, New York 12230. ie nt eee Jepson . am Department of Botany, University of € auferdia. Berke- ley, California 947 2¢ Article 16.1 of the International Code of Botanical Nomenclature (ICBN, 1988) states that “names of taxa above the rank of family are automatically typified if they are based on generic names,” which this name is. and that “for such automatically tas names, ... the name of a suborder which includes the type of the adopted name of an order” 1s to be * ‘based on the e generic name equivalent to that type. but without the citation of an at ae s name 1 Gentianales. not Apocynales. so in this case author citation is necessary. Benson pr vnineae in 1979, but his name was not validly published because it was not accompanied either by a Latin description or kee or by a reference toa Ce iously and effectively published one (i.e it was not published in accordance with Art. 36.1 [ICBN, 1988], the effective date of which is January, 1935). The following ts ae to correct this situation. Apocynineae L. Benson ex Rosatti (Apocynineae L. Benson, Pl. Classification, ed. 2. 263. 1979, nomen invalidum). Herbae vel frutices lactifert: folia oe opposita: flores sleet hermaphroditi, penta- meri: calyx partitus persistens: corolla sympetala; stamina 5, epipetala: carpella 2, supera, ad basim plerumque libera. ad apicem connata: eta marginales, rare Se vel axiales; fructus ple- rumque folliculus, rare drupa vel capsula: semina apice plerumque comosa. Trees to herbs with latex. Leaves simple. usually opposite, entire {rarely toothed or ope often with aolleiere cae Inflorescences terminal but often appearing interpetiolar, cymose, or flowers sometimes solitary. Flowers perfect [rarely carpellate or staminate pncnone yt: ‘actinomorphie 0 or nearly so, 5- se 4-Jmerous except carpels 2 ‘ squamellae basally; corollas variously sympetalous: tubes often appendaged adaxially, overlapping: stamens epipetalous, alternating with corolla lobes, filaments free and short or connate around gynoecium, unappendaged or bearing abaxially 1, 2, [or more] coronas; coronas discoid or comprising 5 free or connate, variously shaped segments. aes sometimes appendaged adaxially; anthers free. variously agglutinated to cach other and to style-sugma head, or connate into anther head around er stigma head: pollen released as individual grains or tetrads, or in pollinia; nectaries “xternal to ovaries, absent. or within stigmatic chambers; carpels free or sometimes connate in ovule- ane region, free or connate in stylar region, and connate apically into la! modified style- e, parietal [or axile] stigma head: ovaries Saad to subinferior, placentae marginal when ovaries {re when connate; ovules | to many per carpel. Fruits follicles, sometimes dru oe {or berries]; seeds in follicles ee each with a tuft of trichomes at apex [or in at both ends] [or arillate}], those in drupes [or berries] naked. those in capsule winge 1989] ROSATTI, APOCY NINEAE 309 very large and complex group (i.e., one in which many infrafamilial taxa have been recognized) that has long been accepted at the level of family Recognition of the Apocynineae as proposed here is ee a, consistent with Stevens’s (p. 67) assertion that “almost everyone ... would agree that what is represented by Apocynaceae and Asclepiadaceae (including Periplo- caceac) is a single taxon at some classificatory level; the cumulative similarities far outweigh the differences.” However, acceptance here within a single taxon, the Apocynineae, of only two families, the Apocynaceae and the Asclepiada- ceae, is at odds with his belief that there are five groups involved (which he recognized as subfamilies) and that the greatest discontinuity lies not between the Apocynaceae and the Asclepiadaceae, as is implied by the arrangement adopted here, but between the Plumerioideae K. Schum. and the Apocynoideae, the two subfamilies of Apocynaceae recognized here. Two other alternatives to unification of the Apocynaceae and the Asclepia- daceae into a single family would also reflect the close relationship between the two, but in my view they are not as acceptable as the one adopted here. The better of these would involve establishment of a superfamily but is un- desirable because there is no recommended name termination for such a taxon in the International Code of Botanical Nomenclature, and its implementation therefore could cause confusion. The other alternative would consist of in- cluding the two families in their own order, the Apocynales, as was adopted by Hutchinson (1926), among others. However, such an arrangement would impart nothing about the relationships between these two and the other families usually included in the Gentianales, an order of relatively consistent circum- scription by most modern authors; therefore, it, too, is undesirable. In summary, recognition of the Apocynineae as a suborder within the Gen- tianales is adopted here not only because it allows a rather commonly accepted order to be maintained but also because it conveys the closeness of relationship between the Apocynaceae and the Asclepiadaceae in a way that has been specifically addressed by the International Code of Botanical Nomenclature. As indicated above, there appears to be nearly uniform agreement among modern systematists regarding the circumscription and, to a lesser extent, the infraordinal structuring of the Gentianales. Dahlgren, Takhtajan, Thorne (1976, 1981), and Wagenitz cach included at one level or another and in various sequences groups recognizable as the Apocynaceae, Asclepiadaceae, Gentia- naceae, Menyanthaceae, Loganiaceae, Buddlejaceae, Rubiaceae, and various other taxa. A most notable exception to this uncharacteristically high degree of conformity is the view of Cronquist, who considered the Rubiales to be a link between his Gentianales (excluding the Buddlejaceae and Menyanthaceae but otherwise delimited as above) and Dipsacales, but too large to be included as a peripheral group in either. Benson placed the Rubiales even farther away, in the fourth of four “groups” of dicotyledons, and included the Adoxaceae, Caprifoliaceae, Dipsacaceae, Rubiaceae, and Nesnrmanig in ane oreeh ene Gentianales (including the Oleaceae but othe of the suborders Gentianineae, Oleineae, oad Apocynineae, ane ie as defined here. The Gentianales, in the usual sense and as accepted in this treatment, may 310 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 be defined as Asteridae (according to Cronquist, a subclass in which more than 99 percent of the species are distinguished from other dicotyledons by sym- petalous flowers and stamens equal in number to and alternate with, or some- times fewer than, the corolla lobes) that are characterized by a combination of traits including diversity of habit, although the plants are commonly woody and usually nonglandular; simple, entire, opposite or whorled leaves: colleters associated with petiole bases, leaf axils, and/or adjacent areas of the stem; five or sometimes four sepals, petals, and stamens but only two carpels; corolla lobes that overlap and are twisted either to the right or left, at least in aesti- vation; epipetalous stamens; and superior or inferior ovaries. According to Wagenitz, the order occupies an isolated position among other sympetalous groups and 1s of unknown ancestry. As accepted here, the Apocynineae can be distinguished most readily from the rest of the Gentianales by their laticifers, which usually produce a white liquid and are lacking in the rest of the order, including the Rubiaceae. Oth- erwise, the suborder is generally set apart by a combination of character states including funnelform, salverform, or urceolate corollas, sometimes with strong- ly reflexed lobes; anthers that are adherent to or fused with the apical portion of the gynoecium (except in the Plumerioideae): carpels that are free in the ovule-bearing region (except in some Plumerioideae) but variously united above and often enlarged and elaborated terminally; superior ovaries; and follicular fruits (except in some Plumerioideae) containing comose seeds (except in most Plumerioideae). The Apocynaceae and Asclepiadaceae also share a number of other features, including those having to do with morphology, embryology, anatomy, and chemistry; many of the frequently cited differences may not hold after the plants are more thoroughly known. These similarities and differences are sum- marized below, and in some cases they are discussed more fully in the treat- ments of the individual families and genera. Several traits commonly used to separate the Apocynaceae and Asclepia- daceae, including the widely presumed absence or presence, respectively, of special structures for pollen transport (see below), in fact represent continua or involve a number of exceptions. As pointed out by Cronquist, in various aspects of floral morphology there are evolutionary progressions from the Plu- merioideae to the Apocynoideae of the Apocynaceae, and from there into the Asclepiadaceae, in which the sequence proceeds from the Periplocoideae R. Br. ex Endl. to the Secamonoideae Endl. and the Asclepiadoideae The most characteristic features of the Asclepiadaceae are related to polli- nation, the specialized mode of which is usually cited in attempts to explain the large number of species in the family. Within the Apocynaceae, however, a comparable number of species exists in which pollination is less specialized. According to Cronquist, the equal numbers of species in the two families are not due to equal levels of pollination-mode effectiveness, but instead to the fact that the number of species in the Periplocoideae has remained small be- cause the mechanism of mass pollen transfer (particularly the pollinia) is im- perfectly developed in the group. The extensive elaborations and the various fusions characterizing the flower 1989] ROSATTI, APOCYNINEAE S41 in the Asclepiadaceae appear to represent, at least in some cases, culminations of evolutionary trends apparent in the Apocynaceae. The coronas in the Apoc- ynaceae generally arise as relatively simple appendages in the throat of the corolla, while those present in most Asclepiadaceae are more complex struc- tures derived from the abaxial side of the staminal filaments or from the area of fusion between the filaments and the corolla. In the Plumerioideae anthers are completely fertile, unappendaged, and free from each other as well as from the style-stigma head,* whereas in the Apocynoideae they are characteristically fertile only in the apical half or so, variously appendaged (at the base and sometimes also apically), and adherent to one another as well as agglutinated to the style-stigma head. Anthers in the Periplocoideae are mostly unappend- aged and usually free but are sometimes adherent to one another and appressed against the style-stigma head, a condition comparable to that described for the Apocynoideae; in the Secamonoideae and the Asclepiadoideae they, as well as other floral parts including the filaments, are elaborate (and hence only partially fertile) and fused to each other as well as to the style-stigma head, thereby contributing to formation of the so-called gynostegium of the Asclepiadaceae. Anther dehiscence is longitudinally introrse (and sometimes slightly latrorse) throughout the Apocynaceae and in some Asclepiadaceae (Periplocoideae, Sec- amonoideae) and apical in the majority of the Asclepiadaceae (i.e., most As- clepiadoideae). According to Woodson (1930), anthers are essentially four- locular (tetrasporangiate) in the Apocynaceae. However, whereas they remain so throughout development in most of the family, in Apocynum L., Poacynum Baillon, and Trachomitum Woodson (recognized here as the tribe Apocyneae, see below) the partition between adaxial and abaxial pairs ruptures by maturity, yielding a seemingly bilocular anther. This condition was considered by Wood- son (1930) to be indicative of an affinity with the Asclepiadaceae, in which anthers are tetrasporangiate but bilocular in the Periplocoideae and Secamo- noideae, and bisporangiate and bilocular in the Asclepiadoideae (see primarily Cronquist; Davis; and Maheswari Devi, 1964). The account of Safwat differs from the foregoing in stating that anthers are four-locular in the Periplocoideae and Secamonoideae (see discussion of the Asclepiadaceae). he apparent lack in the Apocynaceae and universal presence in the Ascle- piadaceae of special structures for pollen transport out of the flower has been widely believed to be the single most reliable difference between the families, and it may indeed be. However, Schick (1980) has suggested that in the Apoc- ynaceae secretions of the style-stigma head that were at one point in evolution (i.e., before development of the concealed stigma) stigmatic presently hold pollen in place and perhaps even contribute to its attachment to insect visitors. This possibility, together with the fact that it is now known that pollen-trans- porting structures of the Asclepiadaceae are actually hardened, acellular secre- 4Several terms including “‘stigma.” “stigma head,” “stigmatic head,” and “‘style head” have been used for this structure, primarily beca use it is usually elaborated to various degrees and is only pantially suemane: oe stigma head” is used here to reflect the idea that it appears to be precisely style nor the i as these terms are usually used, but to both the stigma and the a portion ae style. ale JOURNAL OF THE ARNOLD ARBORETUM [vVoL. 70 tions of the style-stigma head, suggests that even in this regard the families are indistinct. (For more detailed accounts, see family discussions.) The two meiotic divisions of pollen mother cells are simultaneous in most Apocynaceae, Periplocoideae, and Secamonoideae, whereas only the successive type has been reported in the Asclepiadoideae (see Safwat in the discussion of the Asclepiadaceae, to follow). Pollen is released as individual grains in most of the Apocynaceae, but as tetrahedral, isobilateral, or decussate tetrads in Apocynum, Poacynum, and Trachomitum, although grains or tetrads in the family are sometimes loosely coherent (Cronquist), they do not form pollinia. In the Asclepiadaceae, pollen is released as linear (or sometimes tetrahedral, isobilateral, rhomboid, or T-shaped) tetrads that are loosely coherent in the Periplocoideae but firmly enclosed in pollinia in the Secamonoideae and As- clepiadoideae. According to Brewbaker, the pollen is binucleate when shed in some genera of both families (e.g., d//amanda L., Catharanthus G. Don, Nerium L., Rau- volfia L., Urechites Mueller-Arg. [= Pentalinon Voigt], Vinca L. in the Apoc- ynaceae; several extraregional genera of Asclepiadaceae) and trinucleate in others (¢.g., Amisonia Walter, Apocynum, Carissa L., Thevetia L.; Asclepias L.): such reports are not consistent with the accounts of Cronquist, Davis, and others, who have indicated that the grains in both families are in all cases trinucleate. According to Cronquist, pollen in the Apocynaceae is tricolporate or bi- or triporate, sometimes with granular instead of the usual columellar infratectal structure. It is notable that such aspects of pollen morphology and structure evidently have not been reported for the Asclepiadaceae Inflorescences in both families are essentially cymose (for discussion of those of the Apocynaceae, see Woodson, 1935b: for those of the Asclepiadaceae, see Nolan), but while they are usually raceme- or panicle-like in the Apocynaceae, they are often umbel-like in the Asclepiadaceae. Flowers in both families display most of the features generally associated with insect pollination. They are perfect throughout the suborder, except that those in the Asclepiadaceae are rarely imperfect and are five-merous with respect to the number of sepals, petals, and stamens, although some Apocy- naceae are in this sense four-merous. The gynoecia are bicarpellate in all but the tribe Pleiocarpeae K. Schum. (Apocynaceae), in which they are usually three- to five- (but sometimes two-, six-, seven-, or eight-)carpellate. Anthers are held within the fused portion of the corolla in parts of both families, but they are exserted in tribe Parsonsieae K. Schum. of the Apocynaceae and in most of the Asclepiadaceae. In some Apocynaceae, including Apocynum, and many Asclepiadaceae the corolla lobes are strongly reflexed. Filaments are distinct in the Apocynaceae and Periplocoideae but are usually connate into a short sheath around the style (to which they are coherent) in the Asclepiadoi- deae; filaments are unappendaged in the Apocynaceae but are variously elab- orated in the Asclepiadaceae. Nectaries are present in most Apocynaceae and, according to Good, espe- cially frequent in the Apocynoideae. Although they may number either two or five and vary from completely free to totally fused into an unlobed, annular 1989] ROSATTI, APOCYNINEAE 313 structure, they are always located around the base of the gynoecium (see V. S. Rao & Ganguli). Nothing comparable exists in the Asclepiadaceae, where nectar is produced in the stigmatic chambers and sometimes held in the corona (see discussion of Asclepiadaceae). While some apocynaceous flowers (i.e., those of most members of the tribe Pleiocarpeae) are unique within the suborder in having three or more carpels, others qualify as such by virtue of their compound (yet bicarpellate) ovaries: those in the Melodininae K. Schum. are bilocular and have axile placentae, while those in the Landolphiinae K. Schum. and the Allamandeae Pichon are unilocular with parietal placentae. The ovule-bearing region in the asclepia- daceous gynoecium is uniformly composed of two free carpels. Ovaries are wholly superior throughout both families, with the exception that they are partially sunken into the receptacle in some members of each (e.g., Plumeria L., Himatanthus Willd. ex Schultes, Apocynaceae). The non-ovule- bearing region of the gynoecium is more or less completely united throughout the Apocynaceae (except, for example, in double flowers of Tabernaemontana divaricata (L.) R. Br.; see Raghuvanshi & Chauhan), although the styles of many genera are bilobed in transverse section. In the Apocynineae there is a progression from a generally typical style and a small, relatively simple, stig- malike style-stigma head in more primitive Apocynaceae to an enlargement and elaboration of the style-stigma head in more advanced members of the Apocynaceae and all Asclepiadaceae. In Apocynum the region may represent an evolutionary diversion in the sense that it is not differentiated into a style and a style-stigma head but into a single massive, sessile, rhomboid structure, the clavuncle (see FiGureE Se, i). The non-ovule-bearing region of the gynoecium in the Asclepiadaceae appears to be generally free in the stylar region (1.e., there are two distinct styles), a condition in sharp contrast to that described for the Apocynaceae. This distinction between the two families is not often made and needs to be investigated further. Ovules are usually numerous in each ovary or locule and are commonly pendulous throughout the Apocynineae but may be reduced to as few as two or even one in the Apocynaceae and Asclepiadaceae, respectively. Whereas some authors (e.g., Corner, Davis) have indicated that the ovules are anatropous and massively unitegmic in both families, Cronquist stated that they are am- phitropous, anatropous, or hemianatropous in the Apocynaceae but anatropous in the Asclepiadaceae. They are tenuinucellar in the Apocynaceae in all accounts I have seen, but “pseudocrassinucellar” in the Asclepiadaceae, accordin Davis. That is, although the archesporial cell in investigated Asclepiadaceae functions directly as the megaspore mother cell, as in tenuinucellate ovules, the megaspore mother cell is deeply embedded in a thick nucellar cap; except that no primary parietal cell is formed, this condition is comparable to that seen in the crassinucellate ovule. In members of the Apocynineae, cytokinesis accompanies meiosis in the megaspore mother cell, and the chalazal megaspore of the resulting linear (or, in some Asclepiadaceae, T-shaped) tetrad develops into a Polygonum-type megagametophyte. The synergids are pyriform throughout the suborder, and an associated filiform apparatus has been reported in at least one (extraregional) 314 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 species of Asclepiadaceae. In the megagametophytes of both families, the polar nuclei fuse before fertilization, the antipodals are ephemeral (except in Cero- pegia L. and Cynanchum L., of the Asclepiadaceae, in which they simulate the egg apparatus), and starch grains are commonly present. Accessory megaga- metophytes have bi reported in ovules of several extraregional genera of Apocynaceae (Ra Endosperm formations is nuclear throughout the suborder (but see Mahes- wari Devi, 1964), although cell walls form centripetally early in the develop- ment of the embryo. In the Apocynaceae a central cavity free of endosperm develops, but in the Asclepiadaceae the tissue remains cellular throughout. According to Cronquist, the endosperm is oily in both families, and copious to scanty in the Apocynaceae but uniformly scanty in the Asclepiadaceae. Endosperm haustoria in the Apocynaceae have been reported in species of Carissa and Catharanthus (Maheswari Devi, 1971, 1974). Evidently through- out the suborder, embryogeny does not commence until after large numbers of endosperm nuclei have formed. It is of the Caryophyllad type in the Apoc- ynaceae and conforms to the Solanad type in the Asclepiadaceae by Davis’s account, although Asterad, Caryophyllad, Chenopodiad, Onagrad, and Solanad types have been indicated in the former, according to Maheswari Devi (1974). Davis also stated that polyembryony has been reported for members of both families—in one species of the extraregional genus Kopsia Blume (Apocyna- ceae) and commonly in (unspecified members of) Vincetoxicum (author not indicated, see discussion of Gonolobus) (Asclepiadaceae). The Apocynaceae are more diverse than the Asclepiadaceae both in fruit type and in the shapes and sizes of seeds. Follicles are most common throughout the suborder and in fact are the only types represented in the Apocynoideae and Asclepiadaceae. Capsules, berries, and drupes are produced only by certain members of the Plumerioideae. r to be larger when fruits are drupes (e.g., in Ochrosia Juss., Thevetia, V rallesia Ruiz & Pavon) or berries (e.g., in Carissa). Seeds in capsules (e.g., in A/lamanda) and follicles tend to be smaller, are usually flattened, and are sometimes winged (e.g., in 4//amanda), those in capsules and those in follicles of the Plumerioideae lack the apical tufts of long silky trichomes that characterize the Apocynoideae and Asclepiadaceae:; and those in follicles in the Plumerioideae are sometimes arillate (e.g., in Tabernaemontana L.) or have tufts of short hairs at both ends (e.g., in 4/stonia R. Br.). Seeds in follicles seem to be generally more elongate in the Apocynaceae and rounder in the Ascle- piadaceae. Seed coats are multiplicative’ in the Apocynaceae, but only some- times in the Asclepiadaceae. Such seed coats are generally more characteristic of larger seeds. A woody habit occurs in both families of the Apocynineae but is more prevalent in the Apocynaceae, the more primitive and more tropical of the two. A clear majority of the species in this family are trees, erect to scrambling °A term used by Corner for the condition in which cell layers are produced in the integument by periclinal divisions; this usually occurs in addition to the more widespread mechanism by which the number of cells in each layer is increased through anticlinal divisions. 1989] ROSATTI, APOCYNINEAE 315 shrubs, or woody vines; only some are herbaceous, and succulents are rare (except in parts of Africa, in, for example, Adenium Roemer & Schultes and Pachypodium Lindley). Woody members of the Asclepiadaceae constitute a minority and are usually scrambling shrubs or vines (rarely erect shrubs or trees); a larger number of species are erect herbs, and succulents are relatively common Other than differences in the extent to which secondary growth occurs, the Apocynaceae and Asclepiadaceae are quite similar in stem anatomy. In both families the pericycle almost always includes white, sometimes mucilaginous, unlignified cellulosic fibers (slight lignification reported in an extraregional species of Vincetoxicum), either in a continuous ring or as separate strands according to Cronquist, but either isolated or in groups according to Metcalfe & Chalk (1950, 1983) (not detected in material of Rhazya orientalis A. DC., an extraregional member of the Apocynaceae); the cork cambium is superficial (but substantially deeper within the stem in R. orientalis), and wood rays are one to several cells wide. Vessel elements have simple perforations in both families, except that scalariform plates (the more primitive condition) with only a few cross bars (more advanced than many cross bars) have been reported in several genera of Apocynaceae, including an extraregional species of Rau- volfia. Intervascular pitting in both families is alternate; the pits are generally vestured in both families but are evidently smaller in the Apocynaceae. Ac- cording to Cronquist, imperforate tracheary elements (1.e., tracheids) in both families have simple or bordered pits. (The foregoing material is summarized primarily from Cronquist and Metcalfe & Chalk (1950, 1983); see also below.) The most outstanding anatomical features of the Apocynineae are the latic- ifers and the internal phloem present in nearly all species (see primarily Metcalfe & Chalk (1950, 1983)). The former are well developed, nonarticulated (or possibly sometimes articulated), branched or unbranched, and with variously colored (e.g., greenish in Trachelospermum Lem., bright yellow or pale red in some extraregional genera of Apocynaceae, according to Metcalfe & Chalk (1950, 1983)) but usually white contents. They evidently permeate most of the stem in both families; they have been reported, at least in the Apocynaceae, from the primary cortex, pericycle, phloem, pith, and medullary rays. Laticifers in leaves are associated with vascular bundles and frequently (in genera in- cluding Echites P. Br., Catharanthus, Thevetia, and Vinca in the Apocynaceae, and Asclepias in the Asclepiadaceae) extend into the mesophyll and sometimes even to the epidermis (in taxa of Apocynaceae, including Carissa grandiflora (E. Meyer) A. DC. and an extraregional species of Allamanda) as well. Lati- ciferous canals have been reported in the root cortex of Apocynum androsae- mifolium L. and an extraregional species of Thevetia and may also occur in other taxa in both families. Internal phloem (the so-called intraxylary phloem of older literature) has been detected in all examined members of both families (except Pachypodium, a highly specialized, extraregional genus usually assigned to the subfam. Apocynoideae of the Apocynaceae) either as a continuous ring or as isolated bundles around the pith. Datta & Maiti (1971) listed a number of anatomical similarities between the Apocynaceae and Asclepiadaceae (e.g., in the secondary xylem of both: me- 316 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 dium-sized to small vessel members, simple perforation plates, alternate and relatively large pits, bordered pits on fibers; see also above). However, the Apocynaceae can be distinguished from the Asclepiadaceae by generally diffuse- porous wood, with a tendency toward the ring-porous condition in some genera (including . ea. and markedly heterogeneous rays containing latex tubes. Although not explicitly stated, the implication is that in the Asclepiadaceae e wood 1s ring-porous, and the rays are homogeneous and without latex ee throughout the Apocynineae are usually opposite or whorled, simple, and entire: rarely, and only in the Asclepiadaceae, are they either lobed or toothed. True stipules occur in neither family, although colleters® and/or col- leterlike structures are found in the Apocynaceae and, probably to a lesser extent, in the Asclepiadaceae. Stomata are usually paracytic, sometimes an- omocytic, and less frequently actinocytic, anisocytic, brachyparacytic, or cy- clocytic in the Apocynaceae (Chandra er al, 1969, 1972; Dolph & Young; Kapoor & Mitra; Kapoor et a/.; Sharma et a/.; and Trivedi & Upadhyay, 1976, 1977); they are usually paracytic and less often anomocytic or anisocytic in the Asclepiadaceae (see Cronquist). In both families the vascular bundles of the petiole and midrib are evidently bicolateral and arc shaped, and in at least some genera of each there are sometimes small peripheral bundles as well. Nodes are ce throughout the suborder (Metcalfe & Chalk, 1950, 1983). The most common, and therefore presumably ancestral, base chromosome number in the Apocynineae is . = 11, although x = 8, 9, 10, and 12 are also known. These and other cytological data have been of some use systematically and are discussed more thoroughly in the treatments of the families and some of the genera. The abundance of latex and the generally small chromosomes have evidently impeded cytological investigations of many taxa With respect to chemistry, there are probably more similarities than differ- ences between the Apocynaceae and the Asclepiadaceae (see also discussion of the Apocynaceae). A diversity of iridoid compounds, cardiotonic (cardiac) glycosides vs below), and alkaloids has been found in both families, althou the two latte r may be more numerous and widespread in the Apocynaceae a 2. & ! (1964, 1970), Raffauf & Flagler, Willaman & Li, and Willaman & Schubert for alkaloids). Indole, pyrrolizidine, and steroid i aaaieeasend alkaloids occur in the Apocynaceae (see also family discussion), whereas pyridine and phe- ®A term that has not always been used precisely, “colleter” has been applied at one time or another to many kinds of secretory structures of plants, from the simplest few-celled glandular trichomes to much more complex, multicellular appendages derived from both epidermal and subepidermal layers. The term is used here in the latter sense. In the Apocynineae the structures may be homologous to ni sane are often quite conspicuous, are located on the base of the netiole and/or on adjacent areas of the stem, are often root-, finger-, and/or leaf-like, and are reportedly epidermal and subepi- Gounal in derivation, at least in some species of some genera including Carissa, Nerium, Thevetia, and 7 (Jain ef al). Fyell reported ‘ those of Aflamanda a ae Hooker and Thevetia peruviana (Pers.) K. Schum. consist of parenchyma cells il epidermis of clongate, palisadelike cells and a cuticle, and that while those of Vinca minor L. are similar in “a ucture, they were not observed to be secretory and therefore would not qualify as colleters. (S S 1989] ROSATTI, APOCYNINEAE 317 nanthro-indolizidine types are found in—and so-called tylophorine types are reportedly unique to—the Asclepiadaceae; Cronquist is seemingly alone in attributing indole alkaloids to the Asclepiadaceae as well. Crystals of calcium oxalate are often present in both families, while tannins, proanthocyanins, and saponins are variously absent or relatively uncommon throughout the suborder. Cyanogenic compounds are present, but uncom- mon. in both families. Whereas the major component fatty acids in the Apoc- ynineae include linoleic and oleic, palmitic acid is apparently found only in the Apocynaceae (see Alston & Turner). The comparative distribution of phenolic compounds in the two families 1s difficult to assess. For example, although delphinidin, ellagic acid, and myricetin were indicated for neither family by Bate-Smith, Farnsworth (in Taylor & Farnsworth, 1973) listed delphinidin for Vinca minor, and Hrazdina indicated hirsutidin, petunidin, and malvinidin for Catharanthus roseus, as well as cyan- idin for two extraregional genera of Asclepiadaceae. Cardenolides, one of two groups of naturally occurring cardiotonic glycosides, have been reported in at least 12 plant families (see Hoch), including the Apocynaceae and the Asclepiadaceae. Still used as arrow poisons, they are now firmly established for various uses in clinical medicine. The increase in force and the decrease in rate of heart-muscle contractions brought on by carden- olides improves circulation and consequently a number of other aspects of metabolism. As with most drugs, toxic and therapeutic doses are nearly equal. Whereas some reports have indicated that the Apocynaceae are richer in cardiotonic glycosides than the Asclepiadaceae, others suggest that the two are more or less comparable in this regard. According to Kingsbury, about 400 cardiac glycosides had been isolated and characterized by the early 1960's, mostly from the Apocynaceae, Liliaceae, and Scrophulariaceae. Ina later review Singh & Rastogi indicated that these compounds were about equally numerous in ten genera each of Apocynaceae and Asclepiadaceae. Even more recently, Lewis & Elvin-Lewis declared that the compounds are known in more genera of Apocynaceae than any other plant family, although Huber (p. 26) later stated that the “Cerberoideae [included here in the Plumerioideae] and Apocynoideae are distinguished [presumably from other Plumerioideae as recognized here] by the frequent presence of cardiotoxic [i.e., cardiotonic] glycosides, which indicates a strong affinity of these two subfamilies with Periplocaceae [treated here as the Periplocoideae of the Asclepiadaceae] and the tribe Asclepiadeae of Asclepiadaceae.” Regardless of the foregoing, the story of the monarch butterfly (Danaus plexippus) and the cardenolides of Asclepias (Brower ef al.) is perhaps one of the best known in biology (see discussion of that genus). The close relationship between the Apocynaceae and the Asclepiadaceae is not, and to my knowledge has never been, in question. However, the taxonomic history of these families has been relatively complex, at least in part because the plants have been familiar to humans for such a long time. Although the name Apocynum is present in herbals dating back to the fourth century B.c., it is not thought to have been unequivocally and exclusively applied to plants now included in the genus until the seventeenth century. Before that time, practically all plants with milky juice, including those now placed in the Apoc- 318 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 ynaceae, the Asclepiadaceae, and even the Euphorbiaceae, were referred to one monstrous assemblage Taxonomic refinement of this unnatural group was a gradual and sometimes halting process. The “Apocinae” of De Jussieu, on which the family name is based, represented a considerable advance in this direction but nevertheless included as one of four unnamed and unnatural groups a nonlaticiferous one comprising five genera now placed in other families (including the Loganiaceae, see below). Of the three remaining groups, two included genera now assigned to both the Apocynaceae and Asclepiadaceae, although the third was composed only of those now placed in the former. Thus, the system of De Jussieu failed not only to restrict the group to these two families but to distinguish between them as well. Credit for these advancements belongs instead to Robert Brown, who at least tacitly excluded the nonlaticiferous group of De Jussieu and re- moved the “Asclepiadeae” from the rest on the basis of “the singular structure of the stamina,” which at present remains the only consistent difference between the Apocynaceae and the Asclepiadaceae. (For additional taxonomic history, see discussions of the Apocynaceae, Apocynum, and the Asclepiadaceac.) The closest link between the Apocynineae and other Gentianales appears to lie between the Apocynaceae and the putatively more primitive and possibly ancestral Loganiaceae. Probably the least specialized and most diverse family in the order, the Loganiaceae were considered by Takhtajan and by Thorne (1976) to be transitional to the Apocynaceae through Plocosper ma (three species; trees and shrubs; opposite or subwhorled leaves; Mexico and Guatemala). This genus 1s strikingly similar to some Apocynaceae in its elongate, fusiform fruits and apically comose seeds. However, the fruits are capsular, open along two sutures, and usually contain only a single, stalked seed; the ovary is bicarpellate and unilocular, usually with two basal, erect ovules on one parietal placenta and two subapical, pendulous ovules on the other. Carr & Carr implied that Plocosperma, with its apically twice bifid styles, might be considered primitive within the Apocynaceae. Solereder indicated possible connections between Plocosperma and the Apoc- ynaceae (the Hydrophyllaceae were also mentioned in this regard), although he actually placed the genus in the Loganiaceae. However, he was uncertain of its placement, mostly because it lacks internal phloem (in contrast to the Loganiaceae and Apocynaceae) and, presumably, because it has laticifers (lack- ing in the Loganiaceae; present in the Apocynaceae). A unigeneric Plocosper- matoideae was recognized in the Loganiaceae by Takhtajan (tentatively) and Thorne (1976). Stating that the plants combined features of the Apocynaceac, Loganiaceae, Convolvulaceae, and Ehretiaceae, Airy Shaw (in Willis) recog- nized the Plocospermataceae of Hutchinson (1973), who had included the family with the Apocynaceae, Periplocaceae, and Asclepiadaceae in his Apoc- nales. Cronquist included the Plocospermataceae in the Apocynaceae, al- though his description of the latter does not appear to have been adjusted accordingly. Plocosperma 1s here excluded from the Apocynineae As summarized by Rogers, the Apocynaceae and the Loganiaceae both have internal phloem, vestured pits, nuclear endosperm, and colleters but differ in that members of the former are laticiferous (although in the latter, latex is 1989] ROSATTI, APOCY NINEAE B19 present in the fruits of Fagraea Thunb.), generally lack stipules, and frequently have specialized androecia and/or gynoecia. As pointed out by Hutchinson (1973), the Loganiaceae also differ from the Apocynaceae in that the gynoecium is stigmatic at the apex or within apical lobes (vs. at the base of the style-stigma head), the fruits are not follicular, and the embryos are smaller. According to Datta & Maiti (1971), the Apocynaceae have more aspects of wood anatomy in common with the Loganiaceae, Oleaceae, and Salvadoraceae (very small to medium-sized pores and axial parenchyma in all four families) than with the Gentianaceae, Linaceae, Menyanthaceae, and Plocospermata- ceac. Each of these families has been proposed by one systematist or another as ancestral to the Apocynaceae, although the basis for inclusion of the Plo- cospermataceae in this group is unclear because the wood anatomy of Ploco- sperma appears not to have been studied. Pores and fibers of equal diameter (as in the Gentianaceae), exclusively solitary pores (Linaceae), and scalariform or spiral lateral pitting of vessels (Menyanthaceace) are all indicated as lacking in the Apocynaceae (but see Record). Wood characteristics of the Loganiaceae considered to be more specialized than those of the Apocynaceae include den- dritic grouping of vessels, mostly paratracheal parenchyma, and a predomi- nance of exclusively uniseriate rays; additionally, members of the Loganiaceae lack the vasicentric tracheids that have been reported in the Apocynaceae. Datta & Maiti (1971) also indicated that wood of the Apocynaceae is most similar to that of the Salvadoraceae: in cach there are both upright and square cells in the rays, a greater frequency of simple pits on radial walls of fibers, and internal phloem. Paleobotanical evidence seems, for the most part, to be consistent with the phylogenetic relationships between the Loganiaceae, Apocynaceae, and Ascle- piadaceae discussed above. According to Cronquist, pollen assignable to the Apocynaceac has been recovered from Paleocene sediments, while that of the Asclepiadaceae has not been found earlier than the Oligocene. Apparently, fossil pollen assignable to the Loganiaceae has not been found. Macrofossil evidence places the Apocynaceae and Loganiaceae in the Eocene and the As- clepiadaceae no earlier than the Oligocene. (See also Mildenhall; Muller, 1970, 1981.) The economic significance of the Apocynineae lies primarily in the consid- erable medicinal value of cardenolides (discussed above) and of several alka- loids, especially those isolated froin two genera of Apocynaceae, Catharanthus and Vinca (see Taylor & Farnsworth, 1973, 1975). Two in particular from C. roseus (L.) G. Don, vincaleucoblastine and leurocristine, have been extensively applied in chemotherapeutic treatment of various forms of cancer in humans. Often considered to be of “miracle drug” status, leurocristine has been effective against brain tumors, acute leukemia in children, various types of lymphomas including Hodgkin’s disease, and carcinomas of the breast, cervix, and prostate. Experimental evidence has indicated that these two and other related alkaloids may be effective in lowering excessive levels of sugar in the blood and in increasing the flow of urine in humans as well. Many other applications 1n- volving plants in the suborder have had long histories in folk medicine but have not been much evaluated scientifically. 320 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 The latex of these plants has been used to make rubber, while the bark has yielded fibers. The woods are employed in heavy construction, in the building of cabinets, and in carving. Many species are grown for ornamental purposes, including several that in the southeastern United States have either escaped from or persisted beyond cultivation without having become truly naturalized. (For further details, see discussions of families and genera.) REFERENCES: Under references for the Apocynaceae, see CHANDRA et al. (1969, 1972); DATTA & Maiti (1971); DoLPH & YOUNG; FARNSWORTH (in TAYLOR & FARNSWORTH, 1973); FJELL: JAIN et al.; KAPooR & MitRA; Kapoor et al.; MAHESWARI Devi (1971, 1974); RAFFAUF (1964); RAFFAUF & FLAGLER: RAGHUVANSHI “ CHAUHAN; V. S. RAo & GANGULI; RAU (1940), Scnick (1980); SHARMA et al.; TAyLoR & FARNSWORTH (1973, 1975); TRIVEDI & Upapuyay (1976, 1977); and Woopson 1930. 1935b) AcKerRY, P.R., & R. 1. VANE-WriGHT. Milkweed butterflies, their cladistics and biology. ix + 425 pp. London. 1984. [A great many references provided for both Apocynaceae and Asclepiadaceae; systematic index includes plant species; plant species on which larvae feed and to which adults are attracted indicated for each species of butterfly.] Aston, R. E., & B. L. TURNER. Biochemical systematics. xii + 404 pp. Englewood Cliffs, New Jersey. 1963. [Table 13-4 (pp. 206-263) lists alkaloids and/or ‘tcarden- olidglycosides” for many genera of Apocynaceae; discussions of terpenoids and infrafamilial classification of Apocynaceae (pp. 258, 259) and Asclepiadaceae (p. 258). ANDERSON, L. C. Noteworthy plants from north Florida. Ill. Sida 13: 93-100. 1988. [Catharanthus roseus, as an adventive, and Matelea pubiflora, as a native, new to Florida Panhandle. Backer, C. A., & R. C. BAKHUIZEN VAN DEN BRINK, JR. Apocynaceae. Fl. Java 2: 218- 244. 1965a. Asclepiadaceae. [hid. 244-274. 1965b. BaiLey, L.H. Manual of cultivated plants. 1116 pp. New York. 1949. [Bailey Hortorium staffas coauthors; Apocynaceae, 808-813; Asclepiadaceae, 814-817.] _E. Z. Battey, & BAILEy Hortorium Starr. Hortus third. xiv + 1290 pp. New Y York and London. 1976. [Includes many cultivated Apocynaceae and Asclepiada- ic. ] erent H. Apocynacees. Hist. Pl. 10: 146-220. 1889. Asclépiadacées. Ibid. 221-304. 1890. BaRTLING, F. G. Ordines naturales plantarum. 498 pp. / p/. Géttingen. 1830. [Apoc- yneae, 203-205: Peers 201-203.] Bate-Smitn, E. C. The phenolic constituents of plants and their taxonomic significance. I. Dicotyledons. Jour. Linn. Soc. Bot. 58: 95-173. 1962. [Apocynaceae and/or As- clepiadaceac, 153, 154, 168. Benson, L. Plant classification. ed. 2. xxiv + 901 pp. + inside front and back covers. Lexington, Massachusetts, and Toronto. 1979. BENTHAM, G., & J. D. HooKerR. Apocynaceae. Gen. Pl. 2(2): 681-728. 1876. Asclepia- daceae. [hid. 728- 785. [Treatments by BENTHAM; addenda et corrigenda, 1241, 242 —_ BOLKHOVSKIKH .Z., V. Geir, T. MATVEJEVA, & O. ZAKHARYEVA. Chromosome numbers of flowering plants. A. A. Feperov, ed. (Russian and English prefaces.) 926 pp Leningrad. 1969. [Apocynaceae, 45-48: Asclepiadaceac, 59, 60. BREWBAKER, J. L. The distribution and phylogenetic significance of binucleate and trinucleate pollen ae in vane angiosperms. Am. Jour. Bot. 54: 1069-1083. 1967. Brower, L. P., P. B. McEvoy, K. L. WittiAmson, & M. A. FLANNERY. Variation in 1989] ROSATTI, APOCY NINEAE 321 cardiac glycoside content of monarch butterflies from natural populations in eastern North America. Science 177: 426-429. 1972 Brown, R. On the Asclepiadeae, a natural order of plants separated from the Apocineae of Jussieu. Mem. Wernerian Soc. 1: 12-78. 2 foldout keys. 1811. [Preprinted in 1810; Asclepiadaceae, key (on unnumbered pages), followed by 18-58; Apocynaceae, key, -78. CANDOLLE, A. DE. Apocynaceae. Prodr. 8: 317-489. 1844. dao et eo 675- 677. for treatment of Asclepiadaceae in the Brocrornus, see DEc Carr, $.G.M.,&D.J.Carr. The functional s ofsyncarpy. a eel 11: 249-256. 1961. CHITTENDEN, F. J., ed. Dictionary of gardening. ed. 2. 4 vols. 2316 pp. Oxford. 1951. [Many genera of Apocynaceae and eerie included, in aipliabencs’ order. ] CLAPHAM, A. R., T. G. TuTin, & E. F. WARBURG. of the British Isles. ed. 2. xlvii + 1269 pp. Cambridge. 1962. [Probably no ses coca or Asclepiadaceae; only Vinca herbacea, V. major, and V. minor of Apocynaceae se CLEWELL, A. F. Guide to the vascular plants of the Florida Panhandle. + inside front and back covers. Tallahassee. 1985. [Apocynum see en 15 species of Asclepias.] Corner. E. J. H. The seeds of dicotyledons. Vol. 1. xii + 311 pp. Vol. 2. vii + 5 pp. Cambridge, London, New York, and Melbourne. 1976. [Apocynacea ae describ (1: 70-73) and fee (2: 13-16); Asclepiadaceae described (: "14, 75). CorrELL, D. S.. & H. B. Corrett. Aquatic and wetland plants of southwestern United States. xvi + 1777 pp. Water Pollution Control Res. Ser., Environ. Protect. Agency. Washington, D. C. 1972. (Reissued in 2 vols. by Stanford Univ. Press. 1975.) [Apocynaceae, 1331-1338; Asclepiadaceae, 1339-1350; numerous southeastern taxa represented. ] & M. _ JOHNSTON. Manual of the vascular plants of Texas. xv + 1881 pp. enner, Texas 1970. [Apocy eee (including Amsonia, Apocynum, Nerium, Trachelospermum, Vinca), 1211-1218; Asclepiadaceae (including Asclepias, Cy- nanchum, Matelea, Periploca, Chee nae 1219-1241.] Cronauist, A. An integrated system of classification of flowering plants. Frontisp. + xviii + 1262 pp. New York. 1981. [Apocynaceae, 876-879; Asclepiadaceae, 879- 882. DAHLGREN, R. M. T. A revised system of classification of the angiosperms. Bot. Jour. Linn. Soc. 80: 91-124. 1980. _§. ROSENDAL-JENSEN, & B. J. NIELSEN. A revised system of classification of the angiosperms with comments on correlation between chemical and other characters. Pp. 149-204 in D. A. YounG & D. S. SEIGLER, eds., Phytochemistry and angiosperm phylogeny. New York. 1981. [Tropane and pyrrolizidine alkaloids in three genera of Apocynaceae, including Urechites (= Pentalinon), but not in Asclepiadaceae. | Davis, G. L. eo, atic embryology of the angiosperms. x + 528 pp. New York, London, and Sydne DECAISNE, J. een acse DC. ies es 490-665. 1844. [For treatment of Apocynaceae in the Prodromus, see DE CAN E.| DeMETER, K. Vergleichende ha nee Flora 115: 130-176. 1922. [Trans- lators no longer unique to Asclepiadaceae, since structures in Apocynum considered related: other interpretations regarding inflorescences flowers offered.] Dormon, C. Wildflowers of Louisiana. xix + 172 pp. colored pls. Garden New York. 1934. [Apocynaceae, 93, including . yr salicifolia (= A. bernaemontanay,; Asclepiadaceae, 94-96, including Asclepias incarnata, A. ee tuberosa, A. variegata (line drawing), Vincetoxicum carolinense. ! Duke, J. A. On tropical tree seedlings I. Seeds, seedlings, systems, and systematics. Ann. Missouri Bot. Gard. 56: 125-161. 1969. [Apocynaceae, eee ne 158, az2 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 159; first few seedling leaves with green, expanded blades said to be alternate in Apocynaceae alee opposite in examples given), opposite in Asclepiadaceae. ] Duncan, W. Woody vines of the southeastern United States. Sida 3: 1-76. 1967 eet to Psa ra, brief comments, illustrations, maps. Apocynaceae, 34, Trachelo- rin 45,71; ; Asclepiadaceae, 34, Cynanchum scoparium, 46, 71, Perip- joe d graeca, “46, an escape in Knox Co., Tennessee. ] & J Kartesz. Vascular flora of Georgia. An annotated checklist. xii + 147 pp. Athens, tac 1981. [Apocynaceae, 105; Asclepiadaceae, 105, ErcHer, A. W. Bliithendiagramme. Erster Theil. 348 pp. Leipzig. 1875. [Apocynaceae, 251-253: eae ener 253-259] ENDLICHER, S. L. Apocynaceae. Gen. Pl. 8: 577-586. 1838. i ae te Ms 586— 599. [Suppl. 1, 1841: Apocynaceae, 1395-1397: Asclepiadaceae, 1397- Enpress, P. K.. M. JENNy, & M. E. FALLEN. Convergent elaboration ie apocarpous gynoecia in higher advanced dicotyledons (Sapindales, Malvales, Gentianales). Nordic Jour. Bot. 3: 293-300. 1983. [Mortoniella Pittieri Woodson (Apocynaceae) and is lepias physocarpa Schlecter: scanning electron micrographs of carpel-fusion pro- $8. a G. Pollen morphology and plant taxonomy. Angiosperms. Frontisp. + xii + 39 pp. Uppsala. 1952. Eeree M.L. Gray’s manual of botany. ed. 8. Ixiv + 1632 pp. New York (and several other cities). 1950. [Apocynaceae, 1167-1169; Asclepiadaceae, 1169-1177.] Gisss, R. D. Chemotaxonomy of flowering plants. 4 vols. 2372 pp. Montreal and London. 1974. [Vol. 4 includes bibliography, index, and addendum. ] GLeason, H. A. The new Britton and Brown illustrated flora of the northeastern United States and adjacent Canada. Vol. 3. iii + 596 pp. New York and London. 1963. [Apocynaceae, 69-72; Asclepiadaceae, 72-84. ONQuIsT. Manual of vascular plants of northeastern United States and adjacent Canada. li + 810 pp. New York (and several other cities). 1963. [Apocy- 50 naceae, 554-556: Asclepiadaceae, 556-560. ] Goprrey, R. K., & J. W. Wooten. See and wetland plants of southeastern United States. Dicotyledons. x + 933 pp. Athens, Georgia. 1981. [Apocynaceae (Amsonia, Rhabdadenia, Pt ee Urechites [= ee 538-543; Asclepiada- ceae (4Asclepias, Cynanchum, Sarcostemma), 54 GOLDBERG, A. Classification, evolution, and tse : the families of dicotyledons. Smithson. Contr. Bot. 58. iv + 314 pp. Washington, D. C. 1986. [Table 1 indicates ordinal assignments of dicotyledonous families by 12 recent authors; 11 of these included the Apocynaceae and Asclepiadaceae in the Gentianales (one used the name Contortae), HUTCHINSON alone placed them in the Apocynales Goop, R. Features seen eet in the flowering plants. x + 405 pp. New York. 1974. [Corrected version of a work originally published in London, 1956; chapters 7 and 8 are lengthy ae of the Apocynaceae and Asclepiadaceae. ] Gray, A. Synoptical flora of North America. Vol. 2. Part 1. 1878. [Apocynaceae, 79- 85; Asclepiadaceae, 85-106. GREENE, W. F., & H. L. Blomquist. Flowers of the South, native and exotic. xv + 208 pp. 26 pls. Chapel Hill, North Carolina. 1953. [Line drawings ene ne and Asclepiadaceae, plus exotic Apocynaceae (Allamanda cathartica, Carissa gran- diflora, Ervatamia coronaria, Nerium Oleander, Trache jee a aaa Sa Asclepiadaceae (Cryptostegia grandiflora, Hoya carnosa, Stephanotis floribunda), also other references. Guepes, M. Stipules et ligules vraies chez quelques Se sa et Asclépiadacées. Compt. Rend. Acad. Sci. Paris 274: 3218-3221. 1972. [Structures associated with leaf bases in Apocynaceae (e.g., Ervatamia eee [E. coronaria], Vinca major, I. minor) and Asclepiadaceae (e.g., Vincetoxicum officinale) said to include stipules, ligules, and colleters. 1989] ROSATTI, APOCYNINEAE 323 GUIGNARD, L. L’ovule chez les Apocynacées et Asclépiadacées. Mém. Acad. Roy. Sci. Inst. France 55: 1-34. 1917a. [Illustrations representative of genera of Apocynaceae (e.g., Amsonia, Apocynum, Catharanthus, Vinca) and Asclepiadaceae (e.g., Asclepias, Vincetoxicum).| Sur le développement et la structure de l’ovule chez les Apocynacées et les Asclépiadacées. Compt. Rend. Acad. Sci. Paris 165: 981-987. 1917b. HA cuter, H. Provisional i of the natural (phylogenetic) system of flowering plants. New Phytol. 4: 151- 5. Harporne, J. B., & C. ‘ eee Flavone and flavonol glycosides. Pp. 261-311 in J. B. HARBORNE & T. J. Masry, eds., The flavonoids: advances in research. London and New York. 1982. VIE A. M., Jr., T. R. BRADLEY, & C. E. Stevens. Atlas of the Virginia flora. Part 2. Dicotyledons. 148 pp. Farmville, Virginia. 1981. Peseaoee. R. Chemotaxonomie der Pflanzen. Band 3. Dicotyledoneae: Acanthaceae- Cyrillaceae. 743 pp. Basel and Stuttgart. 1964. [Apocynaceae, 124-163; Asclepia- daceae, 199-223. — spre: Y., & K. R. SHIVANNA. The receptive surface of the angiosperm stigma. Ann. Bot. II. 41: ee 1258. 1977. [Strophanthus (Apocynaceae) with uni- vellulas papillae on dry stigmas; Asclepias, Vincetoxicum (Asclepiadaceae) with non- papillate, dry stigmas; Carissa, Mandevilla, Vinca (Apocynaceae) with low to m dium papillae on wet stigmas: -[pocynum, Oncinotus (Apocynaceae), yee (possibly heterogeneous), Strophanthus ee) with nonpapillate, wet stig- mas: bi- and trinucleate pollen, wet and dry stigmas in each family. Heywoop, V. H., ed. Apocynaceae. /n: T. G. — et al., a FI. Europaea 3: 68, 69. 1972. [Neriuwm and Trachomitum by F. MARKGRAF, Vinca by W. T. STEARN.] As- clepiadaceae. [bid. 70-73. [Asclepias, ees 7 ere by F. MARKGRAF.] Hircucock, C. L., A. CRONQuIST, & M. Ownsy. Vascular plants of the Pacific North- west. Part 4: Ericaceae through Campanulaceae. vi + 510 pp. Seattle and London. 1959. [Second printing, 1969; Apocynaceae, 77-82; Asclepiadaceae, 82-85. Hocn, J. H. A survey a cardiac glycosides and genins. South Carolina. 1961.* [Cited by SINGH & RASTOG HRAzDINA, G. ie Pp. 135-188 in J. B. HARBORNE & T. J. MABRY, eds., The flavonoids: advances in research. London and New York. 1982. Huser, H. Apocynaceae. Jn: M. D. DASSANAYAKE & F. R. FosBera, eds., A revised handbook to the flora of Ceylon 4: 25-72. 1983. anes bid 73-124. ee ae . W.H.A. von, A. J. BoNPLAND, & C. S. KUNTH. Apocyneae. Nov. Gen. : 188-233. 1818 [1819 ]. Herein J. The families of flowering plants. Vol. 1. Dicotyledons. ed. 1. xiv + 328 . London. 1926. [Apocynales, 279-282.] . The families of flowering plants. ed. 3. xx + 968 pp. Oxford. 1973. [Apocynales, 469-474. ] JENSEN, S. R., B. J. NIELSEN, & R. DAHLGREN. Iridoid compounds, their oo and systematic importance in the angiosperms. Bot. Not. 128: 148-180. 1975. ee air D. A. Plant embryology: embryology of the Spermatophyta. xviii + 305 pp. Waltham, Massachusetts. 1950. Jouivet, P. Les Eumolpinae (Col. Chrysomelidae) des Apocynaceae et des Asclepia- daceae (Gentianales). Bull. Mens. Soc. Linn. Lyon 51: 214-222. 1982. [Food plants of several leaf beetles, which are often monophagous, belong to these families.] Jones, S. B., JR.. & N. C. Come. List of Georgia plants in the University of Georgia herbarium. 53 pp. Athens, Georgia. 1979. [Apocynaceae, 18: Asclepiadaceae, 19.] Jussieu, A. L. pe. Apocineae. Gen. Pl. 143-151. 1789 Kincssury, J. M. Poisonous plants of the United States and Canada. xii + 626 pp. Englewood Cliffs, New Jersey. 1964. [Apocynaceae, 262-267; Asclepiadaceae, 267- 271.) 324 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 KNOBLOCH, Intergeneric hybridization in flowering plants. Taxon 21: 97-103. 1972. [C Gnabined number of interspecific and intergeneric hybrids reported in lit- erature: Apocynaceac, 4; Asclepiadaceae, 39. KnutuH, P. Handbuch der Bliitenbiologie 2(2): 1-27. 1899. (English translation, Hand- book of flower pollination 3: 1-27. 1909, by J. R. AInsworTH Davis.) [Insect visitors for numerous species and genera. |] Kunze, H. Typologie und Morphogenese des Angiospermen-Staubblattes. (English sum- mary.) Beitr. Biol. Pflanzen 54: 239-304. 1978 [1979]. [Apocynaceae, includin illustrations and photographs of Nerium Oleander, Ochrosia elliptica, Vinca minor, 269-272; oe. including Ceropegia Woodti and Cynanchum vincetox!- cum, 272-2 LAKELA, O., R. _ Lonc, G. FLEMING, & P. GENELLE. Plants of the Tampa Bay area. rev. ed. vi + 192 pp. Sarasota, Florida. 1975. [Apocynaceae, 118; Asclepiadaceae, — LAwRENCE, G. H. M. Taxonomy of vascular plants. xi + 823 pp. New York. 1951. [Apocynaceae, 672, 673; Vinca minor and Allamanda nertifolia illustrated; stigma of Vinca shown in wrong position on style-stigma head. Asclepiadaceae, 673-675; Asclepias curassavica and Stapelia variegata illustrated. LeGcer, V. Recherches sur la Seige voeuuie ie du tégument s¢minal des Apocynacées et des Asclépiadacées. Thése de docteur de Université de Paris. 1913. [Includes descriptions and stration pane genera represented in the Southeast.] LEON [J. S. SAUGET] & ALAIN [E. E. LioGier], HERMANOS. Apocynaceae. FI. Cuba 4: 170-197. 1957. Rel caer Ibid. 198-217 Lewis, W. H., & M. P. F. Etvin-Lewis. Medical botany. Plants affecting man’s health. xvit + 515 pp. New York, epee Sydney, and Toronto. 1977. [Partial list of cardiac glycosides in the Apocynaceae, Li, H.-L. Floristic eines, rier coca Asia and cas North America. Trans. Am. Philos. Soc. Phila. 42: 371-429. 1952. [A (4msonia, but not Trache- lospermum), no ee pen Linptey, J. A natural system of botany. ed. 2. xxvi + 526 pp. London. 1836. [Apo- cynaceae, 299-302; Asclepiadaceae, 302-305. ] Lone, R. W., & O. Lakes. A flora of tropical Florida. xvi1 + 962 ae Coral Gables, Florida. 1971. [Apocynaceae, 697-703: Asclepiadaceae, 703-71 Su J. Acontribution to our knowledge of seedlings. Vol. 2. iv + 646 pp. London. Ig Fruits and seeds also; Apocynaceae, 216-222, descriptions include Echites fae late (illustrated) and species of Plumeria, Tabernaemontana, Thevetia, “As- piadeae,”” 223-233, descriptions include species of Asclepias, Periploca, Sarco- stemma (the last illustrated). MACFARLANE, J. The evolution and distribution of flowering plants. Vol. 1. The Apocynaceae and Asclepiadaceae. x1 + 181 pp. 2 foldout charts. 1933. [A great deal of interesting speculation, although much of the information is outdated. ] MacRoserts, D. T. The vascular plants of Louisiana. Bull. Mus. Life Sci. Louisiana State Univ. 6. 165 pp. Shreveport, Louisiana. 1984, [Apocynaceae, 58, 59: Ascle- ae _ — MAHESWARI Devi, H. Embryological studies in Asclepiadaceae. Proc. Indian Acad. Sci. 60: 52-65. 1964. [Reports of both nuclear and cellular endosperm formation in Calotropis procera need confirmation. Martin, A.C. The comparative internal morphology of seeds. Am. Midl. Nat. 36: 513- 660. 1946. [Apocynaceae, 602, 603; Asclepiadaceae, 614, 61 MetcaLreE, C. R., & L. CHALK. ees Anat. Dicot. 2: 905- 917. 1950. Ascle- piadaceae. hid. 917-925 & Anat. Bic _ ed. 74, 1983. [Numerous references to the Apoc- ynaceae and ‘Asclepiadaceae: e8., eras phloem, 54; laticifers, 72, 73.] ood 1989] ROSATTI, APOCYNINEAE 325 MILDENHALL, D. C. New Zealand Late Cretaceous and Cenozoic plant biogeography: a contribution. Palacogeogr. Palaeoclimatol. Palaeoecol. 31: 197-233. 1980. [Pollen of two species of Parsonsia in Quaternary constitutes first appearance of Apocy- naceae; Asclepiadaceae not mentioned. Morton, J. F. Plants poisonous to people in Florida and other warm areas. ii + 116 pp. Miami. 1971. [Color photographs of Nerium Oleander and Thevetia peruviana, grazing on Cryptostegia madagascariensis Bojer suspected in sudden death of a donkey and a pony in South Florida in 1970.] Folk remedies of the low country. 176 pp. Miami. 1974. [Apocynum canna- un. 27, 28: Asclepias tuberosa, 33, 34. Mutter, J. Palynological evidence on early differentiation of angiosperms. Biol. Rev 45: 417-450. 1970. [Earliest pollen appearances of Apocynaceae include 4/)xia Banks ex R. Br. (Lower Paleocene). Parsonsia R. Br. (Lower Oligocene), A/stonia (lower Upper Miocene) (Asclepiadaceae not mentioned); earliest macrofossils of Apocynaceae are Eocene, those of Asclepiadaceae are Oligocene —. Fossil pollen records of extant angiosperms. Bot. Rev. 47: 1-142. 1981. [Apoc- ynaceae: 4/yxia from Paleocene of Borneo, Rauvolfia from Upper Eocene of Cam- eroon, 4/stonia from Upper Miocene of Borneo, Dyvera Hooker from Upper Miocene of Borneo, and 7rachomitum from Upper Miocene of Spain supported; Parsonsia from Oligocene of New Zealand retracted, that from Upper Eocene of southeastern North America questionable: Nerium from Lower and Middle Eocene of western Europe questionable: ea from Middle Eocene of U. S. A. possible. Asclepiadaceae: Tacazzea Dene. from Oligocene and Lower Miocene of Cameroon supported; Periploca L. and/or Myriopteron Griff. from Miocene of Hungary and Soviet Union possible. ] Munz, P. A. A California flora. vil + 1681 pp. Berkeley, Los Angeles, and London ee Uspoeynaceae eee dimsonia, Apocynum, Vinca), 449-452; Asclepia- aceae Asclepias, Cynanc i Matelea, Sarcostemma), 452-457; see also p. 82, 83 in nee (224 pp.). 19 ee F. Anatomie der ee an Handb. Pflanzenanat. II. Arche- gon. 10. vi + 365 pp. 1926. [Apocynaceae, 268-270; Asclepiadaceae, 270, 271.] NevLinc, L. I., Jk. A catalogue of the published works of Robert E. Woodson, Jr. Ann. Missouri Bot. Gard. 52: 229-233. 1965. [At least 65 references to Apocynaceae and/ or Asclepiadaceac. | No tan, J. R. Developmental morphology of the asclepiadaceous inflorescence. 177 pp. Unpubl. Ph.D. dissertation. Cornell Univ.. Ithaca, New York. 1967.* (Diss. Abstr. B. 28(3): 802. 1967 Ouwt, J. Apocynaceae. Pp. 743, 744 in F. G. Meyer & E. H. KER, eds., Flora of Japan. (English translation.) oS D.C. 1965. a Ibid 744— 751. [Editorial assistance by R. E. Woopson, Jr.] PALTON, J. E., & UDD. ne flora of Paynes ft Basin Se Alachua Sink Hammock, Alachua County, Florida. Castanea 51: 88-110. 1986. [No Apocynaceae; ASC sen ta ennis Walter, uncommon at edge of basin, e ynanchum ae Nutt. occasional at edge of hammock.] ey W.R. Ovular ate ee and the major classification of the dicotyledons. ot. Jour. Linn. Soc. 68: 89-108. soar V. Distribution of aliphatic polyols and cyclitols. i 313-336 in T. Swain, ed., Chemical plant taxonomy. London and New York. 1963. Popepimova, E.G. Apocynaceae. /n: B. K. SHISHKIN, ed., a ora URSS oe pee 18: 645-662. 1952; Flora of the U.S.S.R. (English translation by N. AU) 18: 474-487. 1967. Asclepiadaceae. Flora URSS, 663-718; Flora of the oon 487- 527. [Addenda to English translation include Cynanchum, 561. (eae 326 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 RaAprorp, A. E., H. E. AHLEs, & C. R. BELL. Manual of the vascular flora of the Carolinas. Ixi + 1183 pp. Chapel Hill, North Carolina. 1968. [Apocynaceae (Amsonia, Apoc- yaua, Trac ee Vinca), 845-847, Asclepiadaceae (Asclepias, Cynanchum, ieee 847-858. RAFFAUF a A simple field test for alkaloid-containing plants. Econ. Bot. 16: 171, 172. oo 2. [False positives” obtained for plants in laticiferous families including Apocynaceae and Asclepiadaceae; no clearly positive plants missed by the field assay. A handbook of alkaloids and seaweed Seana plants. x1 + several hundred unnumbered pp. New York, London, Sydney, and Toronto. 1970. [First part, or- ganized alphabetically by plant family. ae about 32 pages for Apocynaceae and less than one for Asclepiadaceae (alkaloids, genera, and references).] RAGHUVANSHI, S. S., & A. K. S. CHAUHAN. Evolution within certain species of Apoc- ynaceae. Pp. 231-235 in P. KACHROo, ed., Advancing frontiers in cytogenetics. Delhi. 1974. [Variety of Tabernaemontana qANCOIG with normal flowers diploid, 21 = 22; variety with double flowers tr plow 2n = 33, in some cases with separate style and style-stigma head for each ovary RecorD, S. J. Classifications of various salonical features of dicotyledonous woods. Trop. Woods 47: 12-27. 1936. [Vessel elements of Apocynaceae and Asclepiadaceae characterized by simple perforation plates on end walls and vestured pits, in Apoc- ynaceae occasionally with scalariform or (e.g., in Vinca) spiral thickenings; wood fibers of Apocynaceae characterized by conspicuous bordered pits; A//amanda, Pe- riploca with ring-porous wood. & R. W. Hess. Timbers of the New World. xv + 640 pp. 7 es New Haven and London. 1943. [Apocynaceae, 56-68; Asclepiadaceae not in J Ric =e T, H. W. Wildflowers of the United States. Vol. 2. The eta States. Part 2. Pp. vi + 323-688. New York. 1966. [Includes color photographs of many of our species of both Apocynaceae and Asclepiadaceae RoperTson, C. Flowers and insects. 221 pp. Carlinville, Ties. 1928. [Visitors listed by plant ae Apocynaceae, 27, 28: Asclepiadaceae, 28-36. ] Rocers, G. The genera . eoeaiaccne in the southeastern United States. Jour. Arnold ee 67: 143-185. 1986. SAFWAT, F. M. The floral oe of Secamone and the evolution = the pollinating apparatus in Asclepiadaceae. Ann. Missouri Bot. Gard. 49: 95-12 Z SCHNARF, K. Vergleichende Embryologie der Angiospermen. vil + ave pp. Berlin. 1931. [Apocynaceae, 199, 200; Asclepiadaceae, 200-203: review of literature; illustrations representative of both families. ] SCHUMANN, K. Apocynaceae. /7: ENGLER & PRANTL, Nat. Pflanzenfam. IV. 2: 109-189. 1895, Onan aae Ibid. 189-306. SEIGLER, D. S. Terpenes and plant phylogeny. Pp. -148 in D. A. YounG & D. S. SEIGLE R, eds., Phytochemistry and angiosperm eee New York. 1981. [Apoc- ynaceae, but not a e SHARP, A. J., R. E. SHANKS, J. K. UNDERWOOD, & E. MCGILLIARD. A preliminary checklist of dicots in Tennessee. Gace ice. 114 pp. 1960. [Apocynaceae and Asclepia- daceae, 74, 75.] SincH, B., & R. P. RAstoci. Cardenolides—glycosides and genins. Phytochemistry 9: 315-331. 1970. [Apocynaceae (¢.g., Carissa, Nerium, Thevetia) and Asclepiadaceae (e.g., Asc es Periploca).] SMALL, J. K. Flora of the southeastern United States. ed. 2. xii + 1394 pp. New York. 1913. eae 933-956, with 19 species and | | genera of Apocynaceae, and 67 species and 12 genera of Asclepiadaceae. | ——. Manual of the southeastern flora. xxii + 1554 pp. New York. 1933. [Asclepia- dales, 1OS7-1078.} SmiTH, E. B. An atlas and annotated list of the vascular plants of Arkansas. 592 pp. 1989] ROSATTI, APOCYNINEAE 327 Published by the author, Fayetteville, Arkansas. 1978 [1979]. [Apocynaceae, 9-12: Asclepiadaceae, 15—20.] SOLEREDER, H. Loganiaceae. [n; ENGLER & PRANTL, Nat. Pflanzenfam. IV. 2: 19-48. 1892; 49, 50. 1895. STANDLEY, P., & L. O. WILLIAMS. Apocynaceae. /n: Fl. Guatemala. Fieldiana Bot. 24: 334-407. 1969. Asclepiadaceae. Ibid. 407-472. [Index to plant names, 473, 474.] STEBBINS, G. L. Flowering plants. Evolution above the species level. xviil + 399 pp. Cambridge, Massachusetts. 1974. [Ovary in Apocynaceae and Asclepiadaceae said to be bilocular with axile placentation—in fact this is only true of one small group within the Apocynaceae as here recognized; both families included in Apocynaceae. } STeEvENS, W. D. A revision of Matelea subg. Dictyanthus (Apocynaceae, sensu lato), 223 pp. Unpubl. Ph.D. dissertation, Michigan State Univ., East Lansing. 1976. (Diss. Abstr. B. 37(2): 587. 1976.) [Higher taxa also discussed. ] STEYERMARK, J. A. Flora of vee Ixxxili + 1725 pp. Ames, Iowa. 1962. [Apocy- naceae (Amsonia, Apocynum, Trachelospermum, Vinca), 1194-1200; Asclepiada- ceae (Asclepias, Cynanchum, Matelea), 1200-1211.] TAKHTAJAN, A. L. Outline of the classification of flowering plants (Magnoliophyta). Bot. R TayLor, L. A. Plants used as curatives by certain southeastern tribes. x1 + 88 pp. Cambridge, se ce 1940. [Asclepias, 52.] THorn_e, R. F. A phylogenetic classification of the Angiospermae. Evol. Biol. 9: 35- 106. 1976. : se aay and angiosperm phylogeny, a summary statement. Pp. 233- 295 in D. A. Younc & D. S. SeiGier, eds., Phytochemistry and angiosperm phy- logeny. New York. 1981. [Recognizes Asclepiadoideae and five other subfamilies in Apocynaceae. | TsIANG, Y., & P.-T. Lt. cae eee Asclepiadaceae. (In Chinese.) Fl. Reipubl. Pop. Sin. 63. xxvi + 621 pp. | Upnor, J. T. C. oo of economic plants. ed. 2. v + 591 pp. New York and Codicote, England. | : Usuer, G. A dictionary a plants used by man. 619 pp. New York. 1974 WacENITz, G. Reihe Gentianales. /; H. MeLcuior, Engler’s Syllabus der Pflanzenfa- milien. ed. 12. 2: 405-424. 1964. oe aa Gentianaceae, Menyanthaceae, Apocynaceae, eee Rubia WILLAMAN, J. J.. & H.-L. Li. Alkaloid- en plants and their contained alkaloids, 1957-1968. Lloydia 33(3A [Suppl.]): 1-286. 1970. [Apocynaceae, 31-56; Asclepia- daceae, 57, 58: species, plant parts, alkaloids, references. ] B. G. ScuusBert. Alkaloid-bearing plants and their contained alkaloids. U. S. Dep. Agr. Tech. Bull. 1234: 1-287. 1961. [Apocynaceae, 25-45; Asclepiadaceae, 47, 48: species, plant parts, alkaloids, references.] Wits, J. C. A dictionary of the flowering plants and ferns. ed. 7. xxu + 1214 + hun pp. London. 1966. [Revised by H. K. Airy SHAw.] eee ae E. Flavones and flavonols. Pp. 189-259 in J. B. HARBORNE & T. J. y, eds., The flavonoids: advances in research. London and New York. 1982. (Only one indicated for Apocynaceae, a pyranoflavonol in Asclepias spriaca. | WUNDERLIN, R. P. Guide to the vascular plants of central Florida. viii + 472 pp. + inside covers (index to families). Tampa and other cities. 1982. [Apocynaceae, 298- 300; Asclepiadaceae, 300-303.] Youna, D. A. The usefulness of eae in angiosperm phylogeny: some selected examples. Pp. 205-232 in D. A. YounG & D. S. SEIGLER, eds., Phytochemistry and angiosperm phylogeny. x + 295 pp. York. 1981. [Flavonoid sulfates in Rhab- dadenia possibly a direct result of growth in mangrove swamps. Youna, D. J., & L. Watson. The classification of dicotyledons: a study of the upper levels of the ae Austral. Jour. Bot. 18: 387-433. 1970. [Data on 83 attributes of 543 gener 328 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 KEY TO THE FAMILIES OF APOCYNINEAE IN THE SOUTHEASTERN UNITED STATES Filaments not connate, unappendaged: anthers free or variously agglutinated to ae other and to style-stigma head; pollen released as individual grains or as tetra HeCariesexteioalis to ovaries or absent: ovaries free or sometimes connate; fruits tice or drupes wl ries free, capsules [or berries] when ovaries connate; seeds in follicles glabrous, each with a tuft of trichomes at ees a at both ends] [or arillate], those in drupes [and berries] glabrous, those in capsules winged. ............ pocynaceae. Filaments connate around gynoecium, ena: abaxially 1 or 2 coronas; anthers connate into an anther head around style-stigma head: pollen released in pollinia, 2 from adjacent anthers united by 2 translator arms and | corpusculum into | pollinarium; nectaries within stigmatic chambers; ovaries free; fruits follicles; seeds each with a tuft of trichomes at apex. 20... ete e eee ey Asclepiadaceae. APOCYNACEAE A. L. de Jussieu, ao Pl. 143. 1789, ““Apocineae,”” nom. cons. (DOGBANE FAMILY) Terrestrial to aquatic (emergent; sometimes in brackish water), erect, as- cending, decumbent, scandent, and/or twining, annual or perennial herbs or subshrubs, or shrubs or trees [or succulents, at least in Pachypodium] with usually white or sometimes nearly colorless latex: axes and leaves glabrous or with various amounts of nonglandular, but sometimes viscid, pubescence [glan- dular trichomes (other than colleters. see below) reported but not confirmed in Rauvolfia tetraphylla]. Leaves simple, usually opposite or sometimes sub- opposite, less frequently alternate, whorled, or subwhorled; blades broad to linear [or panduriform], apices often mucronate, margins entire, usually re- volute, sometimes undulate, bases attenuate to obtuse or truncate, sometimes oblique; petioles distinct, sometimes very short and/or indistinct (especially when leaves linear), without lateral appendages (except in Vinca), bases and/ or adjacent areas of stem often and variously appendaged, usually with colleters (see footnote 6) [or possibly stipular spines in Pachypodium]. Inflorescences terminal or axillary, pedunculate to sessile, 1- to many-flowered, then variously cymose [or racemose]; bracteate or ebracteate; pedicels bracteolate or ebrac- teolate. Flowers perfect, often large, actinomorphic or nearly so; sepals, petals, and stamens 5 [or sometimes 4], carpels 2 [to 8]. Calyx synsepalous but usually divided nearly to the receptacle, persistent in fruit; lobes with or without squa- mellae (small, usually scalelike, glandular or nonglandular projections in axils and/or between bases), commonly imbricate basally to various degrees, var- iously shaped, equal or unequal. Corolla sometimes double (e.g., in Nerium), basal 3-73 united, sometimes with nonglandular pubescence adaxially and/or rarely abaxially, urceolate, campanulate, funnelform, salverform (then tube sometimes widened in | or more places) [or infrequently rotate]; tube or throat sometimes with corona (see filaments); lobes alternating with those of calyx, all overlapping either to right or to left in aestivation, and to a lesser extent at anthesis as well [rarely valvate], erect. spreading, or slightly recurved. Stamens alternating with corolla lobes. Filaments distinct, basally adnate to corolla tube, 1989] ROSATTI, APOCYNINEAE 529 short to very short, terete to flat, straight or curved; corona (i.¢., abaxial ap- pendages) sometimes present, basally adnate to and indistinguishable from corolla tube and/or throat, apically free and laminate. Anthers basifixed and usually somewhat movable on filaments; either distinct from one another and free or only slightly adherent to style-stigma head (in Plumerioideae) or ad- herent to one another and agglutinated to and often forming a cone over style- stigma head (in Apocynoideae), completely fertile’ (in Plumerioideae, except Vinca) or partially sterile (in Apocynoideae, and Vinca), more or less introrsely dehiscent, sometimes with apical appendages (most prominently in Pentali- non). Pollen tricolporate or sometimes bi- or triporate, released as single grains or tetrads (each type sometimes cohering loosely). Nectaries either absent (or appearing so) or adjacent to ovaries externally; 10 (e.g., in Echites, 5 larger, external, and alternating with stamens, and 5 smaller, internal, and opposite stamens), 5 (alternating with stamens and free to variously connate), or 2 (either alternating with carpels and widely spaced or, as in A//amanda, opposite them and forming a shallowly 2-notched annulus). Carpels united to various degrees; ovule-bearing region superior to partially inferior [almost completely inferior, for example, in Plumeria, Himatanthus], sometimes united to form a unilocular ovary with intrusive, parietal placentae (e.g., in Allamanda) [or a bilocular ovary with axile placentae, as in Carissa], but much more commonly free, resulting in 2 distinct, unilocular ovaries each with a marginal placenta, non- ovule-bearing region generally united, either differentiated into a single style and often basally umbraculiform style-stigma head (see footnote 4) or (as in Apocynum) comprising a clavuncle (see FiGuRE Se, 1). Ovules | to many per locule, sometimes pendulous, more or less anatropous, tenuinucellar, and with a single, massive integument: megagametophyte of the Polygonum type. Fruits capsular (e.g., in A/lamanda) [or baccate, as in Carissa] when carpels basally connate, drupaceous (e.g., in Vallesia, [Ochrosia, Rauvolfia, Thevetia}) or fol- licular (in remainder of our genera) when carpels basally free; usually 1 carpel developing in Vallesia [2 developing and fusing during fruit maturation in Thevetia], and 2 developing in remainder of our genera. Seeds 1 to many per developed carpel. variously shaped, those in capsules winged, those in drupes [and berries] naked, and those in follicles glabrous (Amsonia, Catharanthus, Vinca) or each with a tuft of trichomes at apex [or at both ends, as in A/stonia] for arillate, as in Tabernaemontana]; endosperm nuclear, sometimes with starch (e.g., in Apocynum), embryo usually straight, its development of the Cary- ophyllad type. (Emeticaceae Dulac, Plumeriaceae Horan., and Vincaceae S. F. Gray, ‘“Vincae,” are superfluous; excluding Plocospermataceae Hutchinson, including Willughbeiaceae J. G. Agardh, ““Willughbejieae.”) TypE GENUS: Apo- cynum L. A family, here considered in the traditional sense (i.e., distinct from the Asclepiadaceae), of perhaps as many as 200 genera and 2000 species distributed primarily throughout the tropics of the world but with some subtropical and ™Completely fertile” and “partially fertile” or * be sia fertile’ are used here instead of the more commonly used “connective not expanded” and “connective expanded” (or something com- parable). respectively, in part because “ex pi aos parts (and peer es may not always be derived completely from tissues of the connective (se _S. Rao & Ganguli). 330 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 temperate representatives as well. The plants grow in a variety of habitats including deciduous and rain forests, mangrove and other types of swamps, hammocks, thickets, and fields. A diversity of habits characterizes genera rep- resented in the southeastern United States by plants that are native, naturalized, or escaped from or persistent beyond cultivation (see below and at end of this discussion). Alstonia (naturally distributed throughout the Old World tropics) is one of a number of genera in the family that consist mostly of large trees, while others including Carissa ale gi the Old World tropics and subtropics), Taber- naemontana (pantropic), Neriwm (Mediterranean region), Ochrosia (Indone- sian region), Rauvolfia (pantropic), 7hevetia (Mexico to Paraguay), and Vallesia (New World tropics and subtropics) are composed of smaller trees and shrubs. In many genera most of the species are woody vines. Of these, 4//amanda, Angadenia Mueller-Arg., Echites, Rhabdadenia Mueller-Arg., and Pentalinon (Urechites) are confined to tropical and subtropical America; Trachelospermum includes one species indigenous in the Southeast, vane the — are native to southeastern Asia. Principally herbaceous, tem era of annuals or perennials include Amsonia (outliers United States ‘and northern Mexico; one species in Japan and possibly adjacent areas), Catha- ranthus (originally endemic to Madagascar; one species possibly native to India), Vinca (mostly Mediterranean), and Apocynum (northern Mexico to southern Canada). The taxonomic history of the Apocynaceae was intertwined with that of the Asclepiadaceae, at least until Robert Brown first established the ‘“Asclepiadeae”’ on the basis of staminal characters (see discussion of suborder). Brown divided his resulting restricted ““Apocineae” into three unnamed divisions on the bases of seed attachment and pubescence (presence or absence and distribution on the seed), but two genera included in two of these are now lumped under Wrightia R. Br. Lindley, who was evidently the first to use the family name in its present form, accepted the group as defined by Brown but adopted the infrafamilial classification of Bartling. who established three divisions based on fruit and seed characters. Endlicher refined this approach and subdivided the family on the bases of ovary structure and fruit type into three somewhat rearranged assemblages (each a “subordo” or subfamily), one of which was further divided into four groups (each a “tribus’’) differing primarily in seed attachment and pubescence (as above). To characters that had been used pre- viously in grouping genera, De Candolle added ones such as placentation, ovule type, and the presence or absence of squamellae in recognizing seven tribes and several subtribes within the family. Mueller-Argoviensis was evidently the first to incorporate the degree of an- ther fertility and the direction of corolla-lobe overlap in classifying the Apoc- ynaceae. He divided the family into two unnamed groups primarily on the basis of such characters, and each of these groups into tribes based on a number of others that had been previously used; completely fertile anthers and corolla lobes (mostly) overlapped to the left were indicated for four of the six tribes, while partially sterile anthers and corolla lobes overlapped to the right in part defined the other two. Bentham (in Bentham & Hooker) may have been the 1989] ROSATTI, APOCYNINEAE 331 first to suspect the importance of anther agglutination to the style-stigma head, although he actually placed less emphasis on the anthers than did Mueller- Argoviensis in the sense that in two of the three tribes he recognized the anthers were said to be completely fertile and free from the style-stigma head (the corolla lobes usually overlapped to the left), while in the third they were partially sterile and agglutinated to this structure (the corolla lobes usually overlapped the third both fused and free carpels (but see Trivedi & Upadhyay, 1977). In his treatment of South American Apocynaceae, Miers recognized three ‘classes,’ one of which was in part characterized by partially fertile, ““conni- vent” anthers; the condition in the other two classes was not specified. On the basis of a wide range of characters, he further divided the family into 21 tribes, a classification that is for the most part unnatural. Asa Gray, on the other hand, was more in line with the trends begun by Mueller-Argoviensis and Bentham (in Bentham & Hooker); he was the first to recognize the full taxonomic value of the anther characters they introduced. Gray established two tribes in the family in his treatment of North American plants, the Plumerieae Gray,* with completely fertile anthers free from each other and from the style-stigma head (and corolla lobes overlapped to the left in bud), and the Echiteae Gray (as ‘“Echitideae” [= Apocyneae]), with partially sterile anthers adherent to each other and to the style-stigma head (and corolla lobes almost always overlapped to the right). Division of the family along these lines is generally, but not universally, accepted at present. Baillon, seemingly unaware of Gray’s classification, recognized six “‘séries”’ within the Apocynaceae. In five of these the anther ee indicated to be free from the style-stigma head, although they were specifi ald to be completely fertile in only three; in the sixth they were reportedly aie at the base and agglutinated to the style-stigma head. Schumann elevated Gray’s tribes to the level of subfamily, as the Plume- rioideae (see footnote 8) and the Echitoideae K. Schum. (= Apocynoideae). The Plumer ; ; more than two carpels; and the Plumerieae (as the ‘““Plumiereae’’), with apo- This name is based on sagt L., the spelling of which has been variable. Although Linnaeus (Critica Botanica, p. 94. 1737: p. 77 in A. Hort, English translation, 1938) indicated that “P/umieria” commemorated the anes ee co Plumier, “P/imeria’’ was used in both the first edition of Species Plantarum (1: 209. 1753) and the fifth edition of Genera Plantarum (p. 99. 1754), the simultaneous, original places ae wie name. Therefore, it appears that the alteration in spelling to “Plumeria” by Linnacus was intentional and that it should not be considered an ortho- graphic or ty peer rnc error. This being the case, Linnaeus’s original spelling (from a nomenclatural point of view, L.e., “Plumeria’ aoe. be ee ney 73.1, ICBN, 1988), even though several authors have used *Punieria,” as well as “P/lumiera. °Carissa L. (1767), listed as a synonym of .trduina Miller ex L. (1767) by Schumann, is conserved (against Cine Adanson, 1763). The two (Carissa and Arduina) are based on different types according to Jndex Nominum Genericorum, yet there appears to be no reason why the correct name for the tribe proposed by Schumann should not be Arduineae, as indicated by Schumann, even though most authors have taken up Carisseae for the group, following Pichon. 332 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 carpous ovaries and two carpels. Whereas the Arduineae and the Plumerieae were further divided, the Pleiocarpeae were not. > Arduineae included two subtribes: the Melodininae, with bilocular ova- ries, includes | 1 genera (one of which, 4rduina Miller ex L., included Carissa, as sect. Carissa K. Schum., see footnote 9), and the Landolphiinae, with uni- locular ovaries, has nine (including 4//amanda). The Plumerieae included four subtribes, in the following order: the Alstoniinae K. Schum. (many seeds from each carpel, unappendaged calyces, dry fruits), with 21 genera (e.g., Alstonia, Amsonia, Catharanthus, Vinca), the Tabernaemontaninae K. Schum. (many seeds from each carpel, appendaged calyces, fleshy fruits), with six genera (e.g., Tabernaemontanay, the Rauvolfiinae K. Schum. (as the ‘““Rauwolfiinae”’) (two or rarely as many as six seeds from each carpel: placentae thin, not intrusive; seeds not peltately attached), with eight genera (e.g., Rauvolfia, Vallesia), and the Cerberinae K. Schum. (two or rarely as many as six seeds from each carpel: placentae thick, intrusive; seeds peltately attached), with eight genera (e.g., Ochrosia, Thevetia). It is notable that on the basis of evidence from cytology and stem anatomy involving species in seven genera (including 4/stonia, Catharanthus, Taber- naemontana, Rauvolfia, and Thevetia), Datta & Maiti (1971, 1972) concluded that the subtribes of Plumeriecae recognized by Schumann are distinct, and that within this group the Tabernaemontaninae and Rauvolfiinae are the most primitive and the most advanced, respectively; the Cerberinae and the Plu- merunae (= Alstoniinae) were considered intermediate groups, derived from the main line of evolution. Schumann divided the Echitoideae K. Schum. (= Apocynoideae) into two tribes: the Echiteae (as the “‘Echitideae,’” = Apocyneae, anthers not exserted from corolla), with 50 genera (e.g... Echites, Trachelospermum, Rhabdadenia [including Angadenia, as sect. Glandulosae K. Schum.], Pentalinon, Apocynum, Nerium), and the Parsonsieae (anthers exserted from corolla), with 12. Woodson (1930) retained the Plumerioideae (but see below) of Schumann but removed to the Echitoideae all genera of Apocynoideae except Apocynum and the closely related Poacvnum and Trachomitum. Hence, he recognized within the Apocynaceae the Echitoideae (calyces rarely, if ever, without squa- mellae; anthers four-locular at maturity; pollen released as individual grains), the Apocynoideae (calyces always without squamellae; anthers two-locular at maturity; pollen released as tetrads), and the Plumerioideae, in that order. Although other differences between the Echitoideae and the Apocynoideae were indicated by Woodson (1930), only those having to do with pollen and anther locularity appear to be absolute, and the two subfamilies seem to be more closely related to each other than either is to the Plumerioideae. In addition, the Apocynoideae seem to be most advanced within the Apocynaceae, instead of intermediate and more primitive than the Plumerioideae, especially if release of pollen as tetrads is considered to represent a condition transitional to that seen in the Asclepiadaceae. (Pollen grains are aggregated into and re- leased as pollinia in the Asclepiadaceae, and evidently throughout the family they are arranged as tetrads.) Furthermore, 1t seems reasonable to suggest that the character states by which the Apocyneae and Parsonsieae of Schumann 1989] ROSATTI, APOCYNINEAE 333 differ—nonexserted vs. exserted anthers, respectively—are of comparable sig- nificance, since the latter condition is also widespread in the Asclepiadaceae. In view of the foregoing, it would seem most reasonable to retain two subfam- ilies within the Apocynaceae, the Plumerioideae and the Apocynoideae. If development of persistent tetrads and anther exsertion are considered to be comparable advancements toward the Asclepiadaceae, then it is also reasonable to propose three groups, as tribes, within the Apocynoideae. The Echiteae (pollen released as individual grains, anthers held within united part of corolla) constitutes the most primitive assemblage and includes Angadenia, Echites, Nerium, Pentalinon, Rhabdadenia, Trachelospermum, and all the other genera of Schumann’s tribe except the three removed by Woodson. Each of the two other tribes is characterized by one of the advancements. The Apocyneae, with persistent tetrads but nonexserted anthers, includes Apocynum, Poacynum, and Trachomitum, the three genera Woodson removed from Schumann’s Echiteae. The Parsonsieae, with exserted anthers but nonpersistent tetrads, corresponds completely to the tribe of the same name as recognized by Schumann. Although additional research is necessary to assess these proposals more rigorously, this classification of the Apocynoideae 1s adopted here. Woodson (1930) also rearranged the Plumerioideae of Schumann, recogniz- ing within the subfamily four tribes and two subtribes, instead of three tribes and six subtribes, and employing characters in a slightly different way. Thus, he recognized—in order—the Plumerieae (as the “‘Plumiereae’’; ovaries apo- carpous, follicles two), including the Tabernacmontaninae (calyces with squa- mellae, fruits pulpy) and the Alstoniinae (calyces without squamellae, fruits dry); the Pleiocarpeae (ovaries apocarpous, follicles more than two); the Rau- volfieae Woodson (as the ““Rauwolfieae’’; ovaries compound, bilocular, and with axile placentation); and the Arduineae (ovary compound, unilocular, and with parietal placentation). Woodson’s (1930) treatment is confusing to the extent that he did not list included genera. This is especially significant in the case of the Rauvolfieae, since Schumann’s Rauvolfiinae were said to be apo- carpous (although in some genera, including Rauvolfia, the carpels often fuse during fruit maturation [see A. S. Rao]). Woodson (1930) not only redefined several of Schumann’s groups but, as indicated above, presented a nearly opposite view of the direction of evolu- tionary changes within the Apocynaceae. Schumann had considered plants with free, completely fertile anthers (the Plumerioideae) to be more primitive than those with fused (to each other and to the style-stigma head), partially sterile ones (the Apocynoideae), while Woodson (1930) held the opposite view. In addition, Schumann had considered members of the Plumerioideae with com- pound ovaries to be more primitive and those in which the carpels are free to be more advanced, again in contrast to the opinion of Woodson (1930). Debate over the latter issue (summarized below), based primarily on evidence from Allamanda provided by Woodson (1930) and more recently by Fallen (1985), seems to have led to better understanding of the course of evolution within the Plumerioideae and throughout the family. The most recent classification of the Apocynaceae was presented by Pichon in a series of papers appearing over several years (primarily between 1947 and 334 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 1953). According to Wagenitz, who adopted the system for his treatment of the family in Engler’s Sy//abus der Pflanzenfamilien, Pichon recognized three subfamilies, thirteen tribes, and many more subtribes in the family. Among the more significant departures from earlier systems was Pichon’s recognition of the Cerberoideae Pichon and the Allamandeae. Cerbera L., Ochrosia, The- vetia, and five other genera had been included by Schumann in the Cerberinae, a group that corresponds only partially to the Cerberoideae of Pichon. Among other differences, the Cerberoideae were defined by a different set of characters, and such genera as Ochrosia and Kopsia, which were referred by Pichon to the Rauvolfieae, were excluded. As defined by Pichon, the Cerberoideae appear to differ most substantially from the Plumerioideae in relatively minor characters of the anthers (e.g., slightly latrorse vs. clearly introrse dehiscence). The Cer- berinae of Schumann were completely ignored in the system of Woodson (1930), and the subfamily of Pichon is not recognized here. Pichon recognized the Allamandeae as a monogeneric, isolated tribe, the most advanced in the Plumerioideae. Although his reasons for doing this were not entirely clear, his treatment presaged the later conclusions of Fallen (1985; discussed below). Also interesting were Pichon’s inclusion of Pleiocarpa Bentham in the Carisseae (= Arduineae), which was placed first in his system (see below), and his division of the Apocynoideae (as it is circumscribed here) into three tribes (not recog- nized here) differing in construction of the retinaculum. The sequence of genera in most early systems, as well as in Schumann’s, suggested that in the Apocynaceae simple (i.e., free) ovaries were considered to have been derived from compound (i.¢., fused) ones. However, citing as evidence the compound, unilocular ovary of A//lamanda, Woodson (1930) proposed that the free condition is primitive, that the compound, bilocular ovary with axile placentation is intermediate, and that the compound, uniloc- ular ovary with two parietal placentae along the lines of carpel fusion, as seen in A/lamanda, is advanced. He suggested that the parietal placentae in A/la- manda were sterile toward the center of the ovary because they were derived from the evolutionary degeneration of the central part of the septum in a compound, bilocular ovary with axile placentation. That the former type evolved from the latter in the Apocynaceae was also indicated, according to Woodson (1930), by the fact that the ovary in A//amanda is bilocular and the placentae axile during early stages of development. Without reference to Woodson’s earlier discussion of A/lamanda, Woodson & Moore (p. 151) stated that “in frequent instances [taxa unspecified] one finds that a compound ovary is bilocular at both base and apex, where placentation is axile, but unilocular in the middle where placentation is parietal.” It is also notable that Woodson & Moore provided micrographic evidence ofa unilocular ovary with parietal placentae in Lacmellia edulis Karsten, apparently unaware that Schumann had included the genus in his Melodininae, in which the ovaries are supposedly bilocular. These and other considerations suggest that gynoecial evolution within the Apocynaceae could be clarified by careful substantiation of the basic data involved. Fallen (1985) argued that neither of the contrasting evolutionary sequences outlined above for the Apocynaceae (i.e., ovaries simple to compound, or the 1989] ROSATTI, APOCYNINEAE 330 reverse) is entirely correct. She demonstrated that in the ovule-bearing region in flowers of A/lamanda Schottii Pohl, epidermal cells can be recognized where the two carpels meet and are fused, and the carpels in this region are actually free in the earliest stages of their differentiation, suggesting that at least in this case simple (free) ovaries are (more immediately) ancestral. However, in other Apocynaceae with compound ovaries, such cells have evidently not been re- ported, and in addition, the carpels in the ovule-bearing region are fused even in the earliest stages of their development, suggesting that in these groups the ovaries either are primitively compound or are further evolved from their ancestors with simple ovaries. Thus. she considered gynoecial evolution in the Apocynaceae to have proceeded from a primitively compound ovary to one in which the carpels are free in the ovule-bearing region, and from there to an advanced compound condition in 4//amanda. The ovary in Allamanda is not comparable to that seen in other Apocynaceae with compound ovaries, as 1s further evidenced by the fact that, at least in 4. Schottii, the carpels fuse in the ovule-bearing region without ever closing; while such a developmental pattern is common in the Gentianaceae, it has not been reported in other Apocyna- ceae.!° The absence in other genera of Apocynaceae of the gynoecial features summarized above for Allamanda needs to be substantiated. n alternative and at least potentially more reasonable interpretation of the course of gynoecial evolution within the Apocynaceae—one that has not been specifically presented in this debate—is suggested by conditions in the Pleio- carpeae and the Asclepiadaceae. It is to some extent surprising that the former group, with more than two simple (free) ovaries, is placed in an intermediate position in the treatments of both Schumann and Woodson (1930); it seems more consistent with modern concepts to regard such a condition as less spe- cialized than the others seen in the Apocynaceae. In addition, 1t would seem that if the Asclepiadaceae, with two simple ovaries, are considered to have been derived from the Apocynaceae, such a condition should characterize the more advanced groups within the latter. It is notable in this regard that Pichon (1948a) placed Pleiocarpa in the Carisseae, the first (i.e., presumably most primitive) tribe of Apocynaceae treated, and that Corner regarded Lepina Dene., included in the Pleiocarpeae by Schumann, to be one of the more primitive genera of Apocynaceae on the basis of its gynoecium, which consists of three or four carpels that separate widely in the ovule-bearing region during fruit maturation. Rather than proposing—in an attempt to explain an independent origin of compound ovaries in 4//amanda—that evolution of the ovary within the Apo- cynaceae proceeded from compound to simple, and from there to an advanced compound condition in A//amanda, it is more parsimonious to suggest that the main line of evolution within the family was from gynoecia with more than \0An isolated position within the Apocynaceae for -t//amanda is also suggested by its unique fruit type (a capsule), as well as by evidence from chemistry and androecial morphology (see discussion of Allamanda). That it in particular does not belong with other members of Schumann’s Arduineae (i.c.. other Apocynaceae with compound ovaries) is perhaps suggested by cytological and geographic evidence summarized by Van der Laan & Arends (sce discussion of cytology below). 336 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 two simple ovaries, as in the Pleiocarpeae, to those with two simple ovaries, as seen in advanced Plumerioideae, Apocynoideae, and the Asclepiadaceae. The compound, bilocular ovary with axile placentation present in the Melo- dininae of Schumann and the compound, unilocular ovary with parietal pla- centation characterizing Al//lamanda (and perhaps other genera in Schumann Landolphiinae; see Monachino, 1945) then might be viewed as two distinct branches from the main line, separately derived from a gynoecium of two simple ovaries within the Plumerioideae. (Even though the gynoecium of A/- lamanda may have passed through an intermediate condition of bilocularity and axile placentation, that intermediate appears not to have been related to the Melodinineae.) Additional research aimed toward assessment of these pro- posals would not only help clarify relationships within the Apocynaceae but would also illuminate the family’s ancestry as well. (See also discussion of nectaries, below Walker’s (1975a, b, c; 1978) ontogenetic studies on Catharanthus roseus have suggested that in this species and perhaps throughout the family, evolution in the nonovuliferous (apical) region of the gynoecium (uppermost part of the ovaries, at least in C. roseus, in addition to the style and style-stigma head) has proceeded from a free to a fused condition. Remnants of cuticle and the lack of plasmodesmata between adjacent but independently derived cells in the fused portion of the ovaries, as well as other evidence, apparently indicate that carpel fusion was developmentally superimposed onto an apocarpous con- dition. Such a conclusion was also supported by experiments in which the surgical destruction of one carpel primordium prior to carpel fusion only slightly affected development of the other. The single carpel differentiated into an ovary, style, and style-stigma head, although the shapes of these structures were slightly unusual, the pubescence of the style-stigma head was irregular, and the um- braculiform structure at the base of the style-stigma head was not formed on the side normally fused with the other carpel. That the normally fused surface of the developed carpel was abnormal indicated genetically based ‘taccom- modations”’ for fusion and, therefore, perhaps a trend from facultative to ob- ligate apical syncarpy in the evolutionary history of this species and possibly the entire family. The functional significance of the peculiar structure of the gynoecium (basally free but apically united in most Apocynaceae and all Asclepiadaceae) has not been clearly established. The general predominance among angiosperms of gynoecia with variously fused carpels suggests that they are adaptively superior to those in which the carpels are free. Endress proposed that the suggested superiority of fused carpels results from four general advantages conferred by the condition, at least two of which have to do with the apical, nonovuliferous portion of the gynoecium in particular. In one of these, competition between pollen tubes is intensified in the common growth space resulting from such fusion. That is, fertilizations are effected by the most fit pollen tubes from a single, unified competition rather than from each of a number of separate competitions. The second advantage is the pre- sumed ability of pollen on any part of the stigmatic surface to fertilize ovules in any carpel within the flower—clearly a plus if pollen is limited. In fact, 1989] ROSATTI, APOCYNINEAE 337 Walker’s (1978) experiments on Catharanthus roseus demonstrated that pollen tubes are completely uninhibited in growing from one carpel to the other. According to Endress, one of the other advantages of carpel fusion has to do with resource economy and the entire gynoecium, that is, relatively less wall tissue is produced when the carpels are fused. The fourth advantage results from the fact that a more diverse set of fruit types may be evolved in the case of compound ovaries. Despite the foregoing, certain advantages are also conferred by gynoecia with free carpels (Endress ef a/.). In such gynoecia ovaries in which few or no ovules have been fertilized need not needlessly consume resources by participating in fruit maturation, and those destroyed by predation need not jeopardize normal development of the others. Thus, since the foregoing suggests that it is at least conceivable that in some instances apocarpy developed from syncarpy, con- siderations of selective advantage alone cannot resolve the debate concerning the course of gynoecial evolution in the Apocynaceae. In any case, it is rea- sonable to suggest that in most members of the family the advantages of apically fused gynoecia during anthesis have been combined with those of free ovaries during fruit maturation, and that selection has favored maintenance of this combination in the Asclepiadaceae as well. Schick (1980) originally recognized two types of style apices (here, style- stigma heads) in the family, a P/umeria type and an Allamanda-Nerium type, but later (1982) considered those in the series Plumeria, Catharanthus, Vinca, Thevetia, Allamanda, and the Apocynoideae to represent a progression from more primitive to more advanced. In all but the putatively primitive “plu- merioid” genera, the style-stigma head consists of a stigmatic chamber at the base, a pollen depository at the apex, and an adhesive zone between. The adhesive zone, believed to have been stigmatic originally, was thought by Schick (1982) to effect the attachment of pollen to insects and therefore to be func- tionally comparable (analogous) to the translator arms (and probably the cor- pusculum as well) of the Asclepiadaceae. Ultrastructural and biochemical evi- dence also provided by Schick (1982) suggested that the adhesive zone and the translator arms are homologous as well and that the condition in the Periplo- coideae is transitional. Schick’s (1982) account of pollination in Vinca minor is very similar to but less detailed than that outlined here for Apocynum, in the discussion of that genus. Largely on the basis of anatomical evidence from flowers of Apocynum androsaemifolium L. and A. cannabinum L., Woodson (1930) interpreted the nectaries in the genus and throughout the family to represent a whorl of vestigial carpels. According to him, the nectaries and carpels are composed of a similar number and type of glandular cells and differentiate from the receptacle Sl- multaneously and in a comparable manner. In addition, three vascular traces of similar origin and orientation were said to serve each. However, while the nectary traces were shown to arise as a single branch from corolla traces, later dividing into three (one central and two lateral, the lateral ones largely disor- ganized at anthesis), the three traces of each carpel were depicted as having more independent origins. Woodson (1930) mentioned but did not accept an alternate interpretation, at least for the condition in Apocynum, whereby the 338 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 nectaries because of their number (five) and position (alternating with the stamens) would be considered vestigial staminodia. It is notable that Legget (1872b) had suggested earlier that the nectaries of Apocynum androsaemifolium probably represented an “inner whorl of stamens.’ Woodson & Moore later provided evidence from comparative morphology to support the presumed homology between carpels and nectaries in the Apoc- ynaceae. Among flowers produced by a single plant of Pleiocarpa mutica Ben- tham, they reported gradual transitions from those with four fertile carpels to those with two fertile carpels alternating with two sterile “‘carpellodes.”’ Indeed, the nectaries of Catharanthus and Vinca are very similar in appearance to carpels; according to Woodson & Moore they differ from the carpellodes re- ported in P/eiocarpa only to the extent that they are not connected (and, hence, presumably do not contribute) to the style. Based on acceptance of the proposal that the nectaries, which are common and widespread in the Apocynaceae, do in fact represent vestigial carpels, Woodson & Moore concluded that polycarpy probably represents the ancestral condition in the family. However, they did not comment on the systematic position of the Pleiocarpeae per se, a polycarpous tribe of intermediate place- ment in the schemes of both Woodson (1930) and Schumann that is here considered primitive (see also discussion above). Various adaxial appendages of the corolla and calyx have long attracted the attention of systematists. Those of the corolla, which are not known to be glandular, include, for example, the petaloid, laciniate structures (the “‘corona’’) exserted from the throat in Neriu; the deltate, transversely oriented flaps that meet above the anthers, nearly completely sealing the tube at its summit in Thevetia; and the pointed projections near the base of the tube in Apocynum. Although some (for example, those of Apocynum, Trachomitum, and perhaps Poacynum) are opposite the lobes of the corolla (i.e., alternating with the stamens) and not vascularized, the majority are marginally inserted (more or less opposite the stamens) and are served by extensions of the lateral veins in each petal. The latter observations led Woodson & Moore to conclude that all appendages of the corolla in the Apocynaceae are homologous to stipules. Among the genera that include plants native to or at least persisting out of cultivation in the southeastern United States, Thevetia, Nerium, and Apocynum have appendiculate corollas. Unlike the appendages of the corolla, those of the calyx (the squamellae) are usually glandular and relatively inconspicuous. They arise from the very base of the calyx, toward or at the margin or center of each lobe, or both. Although they usually lack vascular tissue, in some extraregional species they have small traces that extend from the lateral veins of each sepal. Therefore, they too were interpreted as being homologous to stipules by Woodson & Moore, instead of to stamens as Woodson (1930) had earlier proposed. On the basis of such an interpretation, squamellae at or near the margins were considered by Woodson & Moore to be primitive, since such a position is similar to that of true stipules, while those located at or near the center of cach sepal, the most common condition in the family, were considered to be advanced; the more or less continuous distribution of squamellae around the base of the calyx was envi- 1989] ROSATTI, APOCY NINEAE 339 sioned to have resulted from what they called (p. 148) ““developmental lacer- ation”’ of either type. Despite the foregoing | only the presence or absence of squamellae has been indicated i in major floristic and monographic accounts of the Apocynaceae, including Woodson’s (1938b) for North America. Al- though this character has been employed mostly at the generic level, it is perhaps of some importance at higher taxonomic levels as well. Squamellae are rare within the Plumerioideae, with the exception that in the Tabernaemontaninae they are universally—or very nearly so—present (see Schumann). Within the Apocynoideae, on the other hand, they are present in at least some species of perhaps |2 genera of Parsonsieae but are sporadic in the Echiteae (of about 50 genera, 30 have all species squamelliferous, four have some, and perhaps 16 have none) and absent in the Apocyneae, as these groups are circumscribed here. Among genera native to or at least persisting outside of cultivation in the southeastern United States, squamellae are present in representatives of Carissa (in which they are otherwise rare), Tabernaemontana, and Thevetia of the Plumerioideae, and Echites, Trachelospermum, Angadenia, Pentalinon, and Nerium of the Echiteae. With respect to vascularization of the calyx, Woodson & Moore recognized four general but taxonomically unimportant types of organization within the Apocynaceae. Although all are unilacunar, there is evidently a fairly gradual transition in the family from a three- to a one-trace condition associated with each sepal. In addition, the four types appear to cross many taxonomic lines, and different combinations of at least two of the types have been reported by Woodson (in Woodson & Moore) from a single inflorescence of Rauvolfia tetraphylla L. Cytological data from the Apocynaceae have been reviewed most recently by Van der Laan & Arends, who provided 520 records (many of them new) for 189 species and 62 genera. The following is based primarily on that account, as well as on a fairly extensive and in part nonoverlapping compilation by Hamel (1984). In general, the systematic value of chromosome morphology appears to be minimal, with the possible exception that in a group of genera corresponding to the Tabernaemontaninae of Schumann, the chromosomes are sometimes especially long (1-3 um, vs. 1-2 in the rest of the family) and then occasionally include a single pair in which the centromeres are terminal or subterminal (instead of submetacentric, the usual condition). Although secondary constric- tions and satellites have not been detected in a majority of the species inves- tigated, one or two per cell have been indicated for species in s1x genera in the Plumerioideae, including ones in the Melodininae and Landolphiinae of the Arduineae and the Alstoniinae, Rauvolfiinae, and Tabernaemontaninae of the Plumerieae, as well as in one genus in the Parsonsieae of the Apocynoideae. Most genera in the Apocynaceae (exceptions include A/stonia, Apocynum, and Ochrosia), including all investigated genera with discontinuous geographic ranges (such as 4msonia, Carissa, Catharanthus, and Nerium), are character- ized by a single base chromosome number; data for outlying species in a few genera with such ranges (e.g., Trachelospermum difforme (Walter) Gray, of 340 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 southeastern North America, and Vinca erecta Regel & Schmalh., of Afghan- istan) are not available. It is evident that infrageneric polyploidy is uncommon in the family, since it has been reported in only eight of the 62 genera investigated, including Allamanda (2n = 18, 36), ae (2n = 22, 42, 44, 48), Carissa (2n = 22, 66), Tabernaemontana (2n = 22, 33, 66), and Rauvolfia (2n = 22, 44, 66, 88; in addition to 20, 24, 55, 68, according to Hamel, 1984). Cytotypes or chro- mosomal races (1.¢., infraspecific polyploidy) occur even less commonly, having been reported in but one species of 4/stonia and four of Rauvol/fia, and probably have arisen exclusively from autopolyploidization. The base chromosome number x = I|1 has been the only one determined in about 60 percent of the 62 genera for which such data are available, including all three of the Pleiocarpeae, all 11 of the Arduineae, and all six of the Taber- naemontaninae; other genera with this number include 4/stonia (along with x = 20 and 21), Amsonia, Rauvolfia, Vallesia, Nerium, and Apocynum (along with x = 8). About 12 percent of the 62 genera have x = 10 (e.g., Ochrosia, Thevetia, Trachelospermum), eight percent have x = 9 (e.g., Plumeria, Alla- manda), and five percent have x = 8 (e.g., Catharanthus, Apocynum in part): one to three percent have x = 6 (only /chites and Pentalinon), 12 (from x = 6 in some cases, from x = 11 in others), 16 (from x. = 8), 18 (from x = 9), 2 (from x = 10; only Cerbera and A/stonia in part), 21 (from the unreported x = 22: only A/stonia in part), or 23 (also from the hypothetical + = 22; only Vinca). The rather widespread occurrence of the base number x = 11, as well as its apparent universality among genera considered here to be most primitive (1.e., the Pleiocarpeae and Arduineae of Schumann) on the bases of morphological and, according to Leeuwenberg (in Van der Laan & Arends, p. 34), chemical evidence, suggests that it 1s the primitive or ancestral number within the family. Evidently no correlation exists between derived life forms and derived chro- mosome numbers within the family. Only three of 35 investigated genera (of a total of about 40) represented in Africa include base chromosome numbers other than 11. Even so, 1t would be premature to propose on that basis an origin for the family on that continent primarily because non-African genera are not nearly as well known cytologi- cally. Nevertheless, of 12 investigated genera restricted to America (of about 55), just two have x = II only (4pocynum has x = 8 and 11); similarly, of s1x strictly Asian genera (of about 50), only one has this number and no other (one has x = 10 and 11). Cytological data as reviewed by Van der Laan & Arends also bear on the problem of the systematic position of A//amanda (discussed above). According to their scheme, the base number +. = 9 was derived from x = 10, which in turn developed from + = 11. All genera with compound ovaries investigated have x = Il and a primarily African distribution, except 4//amanda, which has x = 9 and is restricted to the New World. For these reasons, and because x = 10 has been reported only for genera with simple ovaries, an independent origin for compound ovaries in A//amanda 1s suggested. It is perhaps significant that a base number of . = 9 was reported for only 1989] ROSATTI, APOCYNINEAE 341 two other genera, Himatanthus and Plumeria, each naturally distributed in the American tropics and subtropics. The ovaries in these two genera are simple and approach complete inferiority more closely than any others in the family. Although it is unclear whether or how this condition is related to that in Allamanda, it is notable that Allamanda and Plumeria have in common a number of rather rare iridoids (see discussion of A/lamanda). Nevertheless, Van der Laan & Arends presented an unpublished classification by Leeuwen- berg in which Himatanthus and Plumeria were included in the Plumeriinae, and Allamanda was placed alone in the Allamandeae, far removed from other genera with compound ovaries. Terpenes and various other compounds of terpene origin are well represented and in some cases are systematically important in the Apocynaceae. Iridoid monoterpene glycosides and indole alkaloids characterize the order Gentianales as it is defined here and generally, occurring in all families except the Ascle- piadaceae (but see Cronquist, who attributed indole alkaloids to the Asclepia- daceae as well) and perhaps the Buddlejaceae (e.g., they occur in the Menyan- thaceae, Loganiaceae sensu stricto, Rubiaceae, Gentianaceae, and Apocynaceae; so-called tylophorine alkaloids are reportedly unique to the Asclepiadaceae). Cardiac glycosides occur in the Apocynaceae (see Bisset, 1957, 1961a), As- clepiadaceae, and ten other families of flowering plants (see discussion of sub- order). Monoterpenes (noniridoid types), diterpenes, sesquiterpenes, sesqui- terpene lactones, and several other terpenoids (perhaps including saponins, see below) have evidently not been found in the Apocynaceae or the Asclepiadaceae (see Dahlgren et al. for other compounds absent from these two families). Although present in all groups of land plants (Seigler), triterpenes other than saponins are often taxonomically important and should be investigated in the Apocynaceae. Several were indicated by Gibbs for the family and for the Asclepiadaceae, and one of them (bauerenyl-acetate) may be unique to the Apocynaceae. According to Harborne, Reichstein found triterpenoids to be of value at and below the level of species in the primarily African genus Aco- kanthera G. Don. Unlike common monoterpenes, which are found in the essential oils of a great diversity of plants exclusive of the Apocynaceae and the Asclepiadaceae, iridoid monoterpenes appear to be restricted to about 40 families of flowering plants, including the Apocynaceae but not the Asclepiadaceae (Seigler). Jensen and colleagues recognized ten major groups, most reported from a variety of families, in their compilation of the literature about these compounds. One group included both primitive and unclassified types, five others comprised those with relatively simple structures (“carbocyclic iridoids”’), and the re- maining four included more complex types (““seco-iridoids’’). Iridoid monoterpenes of the primitive or unclassified group indicated b Jensen and colleagues for the Apocynaceae included loganin (Catharanthus, Rhazya Dene., Vinca), desoxy-loganin (Vinca), ketologanin (Vinca), and “iri- doid pyridine alkaloids” (Rawvolfia). Carbocyclic iridoids reported for the fam- ily included aucubin (Thevetia), theviridoside (Cerbera, Thevetia), daphylloside (4lstonia), and plumieride (Al/amanda, Plumeria). The seco-iridoids, all ulti- mately derived from loganin, were divided into a relatively simple group in- 342 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 cluding secologanin (Catharanthus, Rhazya, Vinca) and three more groups each derived from secologanin. One of these included sweroside (Vinca), while another (“complex iridoid alkaloids’) was for the most part composed of compounds biosynthesized by condensations of secologanin and either tryptophane or tryp- tamine to yield indole (and other types of) alkaloids. Indole alkaloids have been extensively investigated, primarily because of their considerable economic and medicinal value. Those characterized by rel- atively simple structures comprise one of two classes recognized by Kisakiirek and co-workers (see below) and are quite widespread, having been detected in the Apocynaceae and nine other families including the Loganiaceae and Ru- biaceae but not the Asclepiadaceae. Indole alkaloids with more complex struc- tures (1.e., those with either tryptamine or trytophan as the indole unit and either a C,- or a C,,-monoterpene moiety derived from secologanin as the other) make up the second class and have been reported from the Apocynaceae, Loganiaceae, and Rubiaceae almost exclusively. Within the Apocynaceae, the two classes of indole alkaloids have evidently been reported only in the Plu- merioideae; steroidal types characterize the Apocynoideae (Kisakiirek er al.). A tremendous amount of literature about complex indole alkaloids in the Apocynaceae, Loganiaceae, and Rubiaceae had accumulated by the end of 1978. It was reviewed by Kisakiirek and colleagues (754 papers cited), who took into account the absolute configurations and biogenesis of these com- pounds in attempting to interpret the data for the first time within a systematic context, using the classification of Pichon to assess relationships within the Apocynaceae.!! Kisakiirek and colleagues recognized eight different skeletal types among the 1200 complex indole alkaloids isolated from the Apocynaceae, Loganiaceae, and Rubiaceae, including those with so-called nonrearranged and rearranged secologanin skeletons. This and other evidence led them to conclude that the Loganiaceae should be considered ancestral to both the Apocynaceae and the Rubiaceae, but more closely related to the former. It was also suggested that of the four tribes of Pichon investigated, the Carisseae (= Arduineae) do not correspond particularly well to any of the groups of genera recognized by Schumann but the Plumerieae, Rauvolfieae, and Tabernaemontaneae (K. Schum.) generally do. The tribe Arduineae was considered by Kisakiirek and co-workers to be closely linked to the Loganiaceae and to be primitive within the Apocynaceae. It is notable in this regard that Pleiocarpa was considered most typical of (most primitive within) this primitive tribe. While this would be consistent with the proposal that the ovary of more than two free carpels characterizing the Pleio- carpeae 1s primitive (see above), it is not consistent with many other classifi- "Extensive and detailed accounts of the occurrence of alkaloids in ees in the Apocynaceae had been previously ane. by, for Biinesel es Bisset (1958, 1961b), Willaman & Li, and Willaman & Schubert. Although these Jered the taxonomic implication to various extents, each was primarily a listing of “compounds bi taxa. Less substantial accounts had been provided by Farnsworth, Ganzinger & Hesse nd co-workers, Raffauf(1970), Raffauf & Flagler, and Taylor & Farnsworth (1973, 1975). 3 1989] ROSATTI, APOCYNINEAE 343 cations, including those of Schumann and of Woodson (1930), in which the Pleiocarpeae were placed in a more intermediate position within the family. Complex indole alkaloids also suggested to Kisaktirek and colleagues that the Plumerieac are more advanced than the Arduineae, and that the tribe may have a profile of skeletal types unique among the four tribes of Pichon that were investigated. Alkaloids of the ibogan skeletal type, evidently characteristic of the Tabernaemontaneae, had been reported for only two genera in the Plu- merieae, 4/stonia and Catharanthus. Both ibogan and bisindole alkaloids have been reported for Catharanthus but not Vinca, thus supporting maintenance of the two as distinct (see also discussion of Catharanthus). The Rauvolfieae were interpreted by Kisakiirek and colleagues as being more advanced than the Plumericae and as lacking ibogan alkaloids (the indication of such compounds for Va//esia apparently was a mistake). The characteristic presence of these alkaloids, as well as a unique emphasis on rearranged seco- loganin skeletons in the Tabernaemontaneae, led them to propose tentative elevation of the group to the rank of subfamily. On the basis of ‘evolutionary advancement parameters” calculated (using both degree of skeletal specialization and level of oxidation) for each of the indole alkaloids considered by Kisakiirek and co-workers, Bolzani and col- leagues later proposed for the Apocynaceae a position closer to the Rubiaceae and farther from the Loganiaceae. Both the calculations of Bolzani and col- leagues and other measures based on morphology suggested a relatively low level of advancement for the Rubiaceae, followed 1n order by the Apocynaceae and the Loganiaceae, also in contrast to the proposals of Kisakiirek and co- workers. The calculations of Bolzani and colleagues also suggested that the Arduineae are most primitive within the Apocynaceae, followed in order by the Plumerieae, Rauvolfieae, and Tabernaemontaneae. Bate-Smith tested leaf material from a large number of species of Apocy- naceac, including many represented in the southeastern United States, for phenolic compounds. He found that members of the family produce leucoan- thocyanins but lack the trihydroxy constituents leucodelphinidin, myrcetin, and ellagic acid. Other phenolics detected in a majority of the species of Apocynaceae screened include quercetin, kaempferol, and caffeic acid. Cyan- idin, an anthocyanin, was reported in only one of nine species of Plumerioideae tested but in five of six in the Apocynoideae. Conversely, whereas p-coumaric, sinapic, and ferulic acids were found in a majority of the species of Plume- rioideae analyzed, they were mostly absent in the Apocynoideae. In the Apoc- ynaceae p-coumaric acid, scopoletin, and possibly luteolin, as well as the an- thocyanins delphinidin, hirsutidin, malvidin, and petunidin, have also been detected (Bate-Smith; Carew & Krueger; Gibbs; Farnsworth, in Taylor & Farns- worth, 1973). Daniel & Sabnis sought phenolics (flavonoids and phenolic acids) in leaf material of 22 species (including a number represented in the Southeast) be- longing to 11 genera of Plumerioideae and eight of Apocynoideae. Kaempferol and p-coumaric acid appeared to be better represented in the former (they were detected in eight vs. two genera and four vs. one, respectively); glycoflavones, leucoanthocyanins, and three phenolic acids (protocatechuric, gentisic, and 344 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 purser seemed more common in the latter (one vs. three, three vs. six, two s. four, one vs. five, and two vs. five genera, respectively). Some compounds were detected only in the Plumerioideae: isorhamnetin in one genus, caffeic acid in another, and a-resorcyclic acid in two others. Several flavonoids were about equally distributed in the Plumerioideae and the Apocynoideae, including quercetin (in eight vs. five genera, respectively), tamarixetin (one vs. one), apigenin (one vs. one), and luteolin (one vs. one), as were the phenolic acids vanillic (ten vs. eight), syringic (11 vs. eight), 2-hydroxy, 6-methoxy (two vs. two), salicylic (two vs. two), melilotic (one vs. one), and o-coumaric (two vs. two). Ferulic and p-hydroxy benzoic acids were detected in five and seven genera of Plumerioideae and five and two of Apocynoideae, respectively. According to information compiled by Gibbs, cyanogenesis is relatively rare in the Apocynaceae. Although it had been reported for six genera in the family, including A/stonia and Nerium, Gibbs’s results were positive for only one, an extraregional member of 4/stonia, and negative or probably negative for nine others, including Catharanthus, Echites, and Rauvolfia. Gibbs also reported that whereas others had reported saponins from a variety of organs in ten genera of Apocynaceae (including 4/stonia and Vinca) and the probable absence of such compounds in 23 others (including 4//amanda, Al- stonia, Amsonia, Apocynum, Carissa, Tabernaemontana, Ochrosia, Rauvolfia, and Thevetia), his own tests using leaf material indicated that they were possibly present in A/lamanda, Nerium, and R. verticillata Baillon but absent or prob- ably so in Amsonia, Carissa, Catharanthus, and R. vomitoria Afz. There appears to be some question regarding the occurrence of tannins in the Apocynaceae. According to Gibbs, some reports indicated their absence from species in 14 genera in the family (including 4//amanda, Alstonia, Och- rosia, Rauvolfia, Thevetia, Vallesia, and Vinca), while others indicated their presence in those of six (including 4//amanda, Apocynum, Carissa, and Trache- lospermum). His own tests for tannins yielded positive results for species of Allamanda, Amsonia, Carissa, and Nerium but were equivocal in the cases of Catharanthus and Vinca. Farnsworth indicated that tannins had been reported in 1. minor, but he and colleagues failed to detect them in V. major (see below). Cronquist stated that the Apocynaceae were not tanniferous. Other tests conducted by Gibbs indicated that compounds such as syringin, Juglone (a naphthaquinone), and aucubin or related substances are probably not produced by members of the Apocynaceae, although he noted that other researchers had reported plumbagin (another naphthaquinone) from an ex- traregional species of Rauvo/fia and various aucubin glycosides from genera including 4/stonia and Thevetia. Farnsworth and colleagues screened leaves and stems of Yugoslavian plants of Vinca major for many compounds. Although they confirmed the presence in the species of alkaloids, saponins, unsaturated sterols, organic acids, and phenols, they were unable to detect tannins, flavonols, or cardiac glycosides. The distribution of polyalcohols, in particular the nonaliphatic types (cycli- tols), was considered within a systematic context by Plouvier. One cyclitol, conduritol, had been reported by the time of that review only from the Ascle- piadaceae (Gonolobus Michx.), while five others, all methyl ethers of inositol, 1989] ROSATTI, APOCYNINEAE 345 had been detected in the Apocynaceae. Although D-bornesitol had been found only in the Rubiaceae and in both subfamilies of the Apocynaceae (in genera including Amsonia and Vinca), L-bornesitol had been detected in the Protea- ceae, Leguminosae, Rhamnaceae, Boraginaceac, and Apocynaceae (only in Apocynum; it has since been reported by Nishibe and co-workers from Ca- tharanthus and a Japanese species of Amsonia), dambonitol had been reported in the Moraceae and both subfamilies of the Apocynaceae (in genera including Nerium, Trachelospermum, and Vinca). The two other inositol methyl ethers recorded for the Apocynaceae evidently have wider distributions: L-quebrach- itol had been found in 11 families of flowering plants, but within the Apocy- naceae only in three (extraregional) genera of the Plumerioideae, and D-pinitol, known from six gymnosperm and 13 angiosperm families, had been detected in the Apocynaceae only in Landolphia Beauv. Information about the fatty acids of seed oils obtained from species of Ca- tharanthus, Nerium, Rauvolfia and four other genera of Apocynaceae was pro- vided by Daulatabad & Ankalgi. Although palmitic, oleic, linoleic, and stearic acids, all very widespread among angiosperms in general, were indicated for the nine species involved, lauric, myristic, arachidic, and behenic, each with a more restricted distribution, were reported only for the three (including C. roseus and R. tetraphylla) they themselves analyzed. Although Gibbs did not claim to list each occurrence of these seed-oil compounds, it is notable that the Apocynaceae escaped mention in his account. Regarding circumscription and ordering of subfamilies, tribes, and subtribes of Apocynaceae, it is primarily Schumann’s system that is adopted here, with a few exceptions. Within the Plumerioideae, the more primitive of the two subfamilies, the extraregional Pleiocarpeae are considered to be most primitive, instead of intermediate. The ordering of two of the four subtribes of Plumericae has also been changed, due to considerations of plants in the Southeast. Despite the fact that much of the evidence summarized above suggests that the Rau- volfiinae are more advanced than the Alstoniinae, in this treatment the order is reversed because Vallesia, the genus of Rauvolfiinae treated here, seems more primitive than Amsonia, Catharanthus, and Vinca, the genera of Alsto- niinae considered. Following Pichon, A//amanda has been placed in its own tribe and positioned as the most advanced group within the Plumerioideae. Three tribes (Echiteae, Parsonsieae, Apocyneae), instead of the two recog- nized by Schumann, have been recognized in the Apocynoideae. Two of these are represented in the Southeast: the Echiteae, by Angadenia, Echites, Penta- linon (Urechites), Rhabdadenia, and Trachelospermum, and the Apocyneae, by Apocynum. In the southeastern United States, the Apocynaceae are represented by in- digenous species in the genera Amsonia, Angadenia, Apocynum, Echites, Pen- talinon, Rhabdadenia, Trachelospermum, and Vallesia, and by naturalized plants in Allamanda, Catharanthus, and Vinca. Whereas several primarily tropical or subtropical genera (Allamanda, Echites, Pentalinon, Rhabdadenia, Vallesia) are represented in our area by native or naturalized species solely in Florida, two species of the temperate genus Vinca (V. minor and V. mayor) that are naturalized throughout much of North America appear not to be so 346 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 established in Florida or perhaps other of the more southern parts of the Southeast. In addition to native and naturalized Apocynaceae, an especially large num- ber of species have escaped from or persisted beyond cultivation, particularly in Florida, without having become truly naturalized. At least Trachelospermum Jasminoides (Lindley) Lem. belongs to a genus represented in the Southeast by native or naturalized plants, while several others (briefly discussed below; in- cluded as bracketed entries in the key) do not. Carissa (subtribe Melodininae), a genus of about 35 species of usually armed shrubs and small trees, is native to warm, dry areas in Africa, Asia, and Aus- tralia. I have seen specimens indicating that in southern Florida (e.g., Dade and Lee counties) C. grandiflora (E. Meyer) A. DC. (C. macrocarpa (Ecklon) A. DC., according to Palmer & Pitman), which is widely grown in this region, sometimes persists beyond cultivation. In addition, according to Judd (pers. comm.) the species fruits abundantly and is a rare escape in this area. Infor- mation on the label of an essentially sterile specimen (Rhoades & West 6/8/ 28, FLAS) perhaps assignable to C. Arduina Lam. or C. bispinosa (L.) Desf. ex Brenan, two species with smaller flowers and fruits than C. grandiflora, ac- cording to Bailey, indicates that the plant was naturalized somewhere in Florida (“Tropie”). Plants of C. Carandas L. (fruits 1.3-2.6 vs. 2.6-5.2 cm long in C. grandiflora), are frequently cultivated in Florida but evidently do not escape. The structures usually referred to as spines appear instead to be thorns (that is, they are homologous to stems and not leaves). They develop from accessory buds, according to Jain et a/., and are modified inflorescence axes fide Palmer & Pitman. They are present on all material I have seen from Florida and are usually bifurcate and axillary. Alstonia (Alstoniinae), a genus of 30-40 species of large trees, is distributed throughout the tropics of the Old World, especially in the western Pacific region. It is represented in our area by 4. macrophylla Wall. ex G. Don, a species that, after its intentional introduction to Sri Lanka for its valuable timber, has become one of the most prominent naturalized elements of the secondary rain forest there (Huber). Specimens collected in Florida indicate that it has escaped from cultivation on numerous occasions. Judd (pers. comm.) observed that cultivated plants in Miami grow rapidly and flower and fruit regularly: he believes that they could easily become established from the wind-dispersed seeds in many disturbed habitats in the area. Considered here in the broad sense to include Ervatamia, Tabernaemontana (Tabernaemontaninae) is a pantropic genus of about 230 species of small trees and shrubs. (Nowicke recognized both 7abernaemontana, with about 140 species in the American and African tropics, and Ervatamia, comprising approximately 90 species distributed from Asia to Australia and New Zealand.) I have seen one specimen (Hansen & Robinson 8473, FLAS) that was considered by the collectors to be a “legitimate escape” in a “pine flatwoods” in Osceola County, Florida. On the basis of a key provided by Nowicke, it is assignable to E. coronaria (Jacq.) Stapf (Tabernaemontana coronaria (Jacq.) Willd., Nerium coronarium Jacq.). Huber considered this species to be conspecific with E. divaricata (L.) Alston in Trimen (= 7abernaemontana divaricata (L.) R. Br.: 1989] ROSATTI, APOCYNINEAE 347 Nerium divaricatum L.), a native of the southern Himalayas now cultivated in tropical areas throughout the world. The plants in cultivation may have either single or double flowers. The species is reportedly heterostylous (Huber). Most species of Rauvolfia (Rauvolfiinae), a large, pantropic genus of perhaps 110 species, are shrubs or trees. Rauvolfia tetraphylla L., a shrubby plant, 1s indigenous to an area from Mexico to northern South America and is reportedly naturalized in India and Australia (Nowicke). It has been attributed to our area as a rather uncommon escape, in—for example— Palm Beach County, Florida. Nowicke has remarked that it is variable throughout its natural range, and that it is often difficult to separate from R. /ittoralis Rusby. My observations on material collected in Florida conform with this assessment. Ochrosia (Cerberinae) includes perhaps 30 species of small to medium-sized trees native to Indonesia and surrounding areas. Because of their nonacuminate leaves and large fruits with relatively narrow wings, variously determined spec- imens collected in Florida are assignable to O. elliptica Labill., according to a key presented by Boiteau. Native to Australia, New Caledonia, and many islands in the Pacific Ocean, the species is cultivated in numerous places, including Hawaii, Puerto Rico, and Florida. The fruits float and remain fertile for long periods in sea water, readily germinating upon contact with land (Boiteau). According to Judd (pers. comm.), the species 1s probably a very rare escape in Florida, and its naturalization there is doubtful. A small genus of probably fewer than ten species of trees, shrubs, and rarely vines, Thevetia (Cerberinae) occurs naturally from Mexico to Paraguay. The- vetia peruviana (Pers.) K. Schum., of this area, is frequently cultivated and is sometimes escaped or naturalized in many tropical and subtropical areas of the world. The species was reported by Small (1913) to be growing naturally in dry soil on Key West, Florida, and Long & Lakela have indicated that in southern Florida, including the Keys, it is naturalized in coastal areas and cultivated more widely. Nevertheless, Judd (pers. comm.) maintains that he has seen no evidence that the species has become established outside of cul- tivation in the area. It is commonly planted for ornament 1n Guatemala, where according to Standley & Williams it is not native but sometimes appears to be naturalized along roadsides and in pastures. In contrast, Woodson (1938b) indicated that it was possibly native throughout Central America. As traditionally considered, Nerium (Echiteac) comprises two or three species of shrubs or small trees native to the Mediterranean region, subtropical Asia, and Japan, although Leeuwenberg recently suggested that there is only one, Nerium Oleander L. This species is commonly planted as an ornamental throughout the warmer parts of the world and is sometimes escaped and nat- uralized as well. Occasionally escaped in tropical America, according to Now- icke, it is naturalized throughout the region fide Woodson (1938b). It is widely naturalized in the warm parts of North America, including Texas, according to Correll & Johnston. Although Judd (pers. comm.) indicated that it 1s a rare escape in southern Florida, Long & Lakela did not think that it is established there. Evidently, plants may persist for a long time after cultivation and may even propagate or at least spread to some degree by vegetative means. However, sexual reproduction by plants out of cultivation does not appear to occur 348 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 anywhere in the Southeast. Many of the plants in our area and elsewhere have double flowers and are sterile. The latex of this plant 1s lethally poisonous, even in small quantities. Many other species of Apocynaceae appear to be confined to cultivation in the southeastern United States. Some of these belong to genera not otherwise represented in the area, such as Acokanthera, Beaumontia Wall., Mandevilla Lindley, and Plumeria. REFERENCES: Under references for the Apocynineae, see BAILEY; BAILLON; BARTLING; BATE-SMITH; BENTHAM (in BENTHAM & Hooker); BROWN; DE CANDOLLE; CORNER; CORRELL & aesey CRONQUIST; Foe et al., ENDLICHER; ENDRESS ef al.: me Sieh HuBer; JENSEN et a/l., Lewis & ELvin-Lewis; LINDLEY; LONG & LAKELA; PLOUVIE RAFFAUE (1970): ea CHAUHAN; ROGERS; SCHUMANN; SEIGLER; Siaut (1913), STANDLEY & WILLIAMS: WAGENITZ; WILLAMAN & Li; and WILLAMAN & SCHUBERT. Under references for Catharanthus, see CAREW & KRUEGER. ALLORGE, L. oe de la aes ee Allamandées aux Echitoidées (Apocynacées). Adansonia, II. 15: 273-276. 1975. [Placement of Allamandeae in Apocynoideae ee oe of presence of ee a proposal rejected by FALLEN (1985).] Morphologie et biologie florale des Apocynacées. Applications taxonomiques. These. Paris. 1976.* Morphologie, systématique, chimio-taxonomie et biogéographie des Tabernae- montancidées (Apocynacées) américaines. Thése. Paris. 1983. [Treatment of Er- vatamia, 473-486, including F. coronaria (= Tabernaemontana divaricata).] ———. Monographie des Apocynacées-Tabernaemontanoidées américaines. Morpho- logic, systématique, chimio-taxonomie. Mém. Mus. Natl. Hist. Nat. II. B. 30: 1- 213. 14 an J On, & C. SAsTRE. Morphologie et chimiotaxonomie des Apocynacées. C ae phviogsnuqus . biogéographiques. (English summary.) Compt. Rend. Soc. Biogéogr. 57: -126. 1980. [Allamanda placed in its own tribe, in most advanced position in ie (One ee (the fourth of five subfamilies recognized); phytochemical data suggest relationship between Apocynaceae and Loganiaceae, Rubiaceae, and Nyssaceae. | ANDERSSON, A. Studien tiber die Embryologie der Familien oo Oleaceae und Apocynaceae. eae niv. Arsskr. II. Sect. 2. 27(7): 1 2. pls. 1-4. 1931. [Apoc- ynaceaec, 75-101, pls. 3, 4; illustrations of several Leek species. ] Bisset, N. G. The cee glycosides of the Apocynaceae. Part I. Indon. Jour. Nat. Sci. 109: 173-211. 1953. Part I. /hid. 111: 76-117. 1 Cardiac glycosides. Part I. Apocynaceae: the genera Carissa L., Vallaris Burm. = Beaemaniia Wall., pearl A. DC., and others. Ann. Bogor. 2: 193-210 . Apocynaceae: a preliminary paper-chromatographic study of the slycosides od Thevetia ee (Pers.) K. Schum. /bid. 4: 145-152. 196la. [Several kinds from many parts of the plants; extracts from seeds of yellow-flowered plants see equal to those of orange-flowered ones. . The occurrence of alkaloids in the Apocynaceae. /bid. 3: 105-236. 1958. peor system ee taxonomic interpretation mostly confined to level of genus and below various uses by native peoples discussed; many species of the southeastern United States included.] Part II]. A review of recent developments. /bid. 4: 65-161. 1961b. Borreau, P. Flore de la Nouvelle Calédonie et dépendances. 10. Apocynacées. 302 pp. Paris. 1981. fie the papas of L. ALLoRGE.] BoLzANnl, V. D.S., M. F. D. G. F. D. Sitva, A. I. D. Rocua, & O. R. Gorruies. Indole — 1989] ROSATTI, APOCYNINEAE 349 alkaloids as systematic markers of the Apocynaceae. Biochem. Syst. Ecol. 12(2): 159- a 1984. Campo, M. van, S. Nitsson, & A. J. M. LEEUWENBERG. Palynotaxonomic studies in Fabermaernontan L. sensu lato Taree Grana 18: 5-14. 1979. [Light mi- crographs of pollen of Ervatamia coronaria (= Tabernaemontana divaricata).] CANDOLLE, A. DE. Apocynaceae. DC. Prodr. 8: 317-489. 1844. [Addenda et corrigenda, -677. CHANDRA, V.. S. L. Kapoor, P. C. SHARMA, & L. D. KAPoor. Epidermal and venation studies in Apocynaceae—I. Bull. Bot. Sury. India 11: 286-289. 1969. [Stomata mostly paracytic in six arborescent species in four genera, including A/stonia; also anomocytic in 4. macrophylla.) S. L. KApoor, & L. D. Kapoor. Epidermal and venation studies in Apocynaceae—IV. Bull. Bot ri India 14: 76-82. 1972. [Stomata mostly paracytic in various genera, including 4/lamanda, Carissa, Catharanthus, Trachelospermum, also anomocytic (4//amanda, ee ane Trachelospermum), anisocytic (Alla- manda, Catharanthus), actinocytic (Catharanthus), and cyclocytic (Carissa). CHAUHAN, A. K. S., & . RAGHUVANSHI. Cytotaxonomy and induced polyploidy in Apocynaceae. Biol. Meu §: 163-209. p/s. 1-3. 1981. [Natural triploids in Taber- naemontana divaricata discussed in detail; base number x = 11 in family. DanteL, M., & S. D. SABNIS. Chemotaxonomical studies on Apocynaceae. Indian Jour. 978. Datta, P. - is R. K. Marti. Relationships of Plumiereae (Apocynaceae) I. Based on karyology. Bull. Bot. Soc. Bengal 26: 9-19. 1972. II. Based on wood microscopy. Ibid. 25: 55- 61. 1971. [Species of various cece including A/stonia, Catharanthus, Ervatamia, a ea described in detail. DAULATABAD, ,&R. F. eae Studies on Apocynacea [sic] seed oils. Fette Seifen pees 87: 247-249. 1985. Die, J. vAN. A comparative ae - the particle fractions from Apocynaceae latices. Ann. Bogor. 2: 1-124. 1955. [Systematic value confined to a few cases at the species level and below.] DovpH, G. E., & J. YOUNG. mane variation in five genera of the Apocynaceae. Proc. Indiana ‘Acad. Sci. 89: 94, 95. 1980. [Stomata confined to abaxial epidermis, either paracytic or ae ce guard cells sometimes sunken in 23 North Amer- ican species of various genera, including [chites. Epcar, J. A. Plant-butterfly coevolution: pyrrolizidine alkaloids and cardenolides. (Ab- stract.) Proc. 13th Int. Bot. Congr. 125. 1981. [Species of the Danainae and Ith- omiinae (Nymphalidae) obtain Dee camer alkaloids (for use in defense and as precursors of courtship pheromones) from Apocynaceae (Parsonsieae) and several other families; larvae of more advanced species of Danainae obtain cardenolides from Apocynaceae and Asclepiadaceae.] Enpress, P. K. Syncarp py and alternative sa of escaping disadvantages of apocarpy in primitive angiosperms. Taxon 31: 48-5 82 EzcurrA, C. Revision de las Apocynaceas de - Argentina. Darwiniana 23: 367-474. 1981. [Sixteen genera, including Rauwvolfia, Ricbaadens. Thevetia, Vallesia, Vinca; illustrations FaLLen, M. E. A systematic revision of Anechites (Apocynaceae). Brittonia 35: 222- 231. 1983a. [Treated as a monotypic genus of the Caribbean region; ae. not known from the Southeast; placed in Cerbereae, Thevetia among its closest rela- tives xonomic revision of Condyloc ae een Ann. Missouri Bot. Gard. 70: 149-169. 1983b. [PICHON’s systen supported below the level of tribe (e.g., subtribes said to be too narrowly pera ot head and structure of anthers considered very important in family; relationship between Condylocarpon and Rau- volfia suggested. ] 350 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 The 2 € ie of the se ae its structure and significance. (Abstract.) Am. Jour. 71(5, part 2): 28, 1984. [Floral evolution in family primarily involved pane of distinct een and functional regions of style head and its coadaptation with androecium, culminating in “gynostegium” of the Apoc- ynoideae : The: gynoccial development and systematic position of A//lamanda (Apocyna- ceae). Am. Jour. Bot. 72: 572-579. 1985. FARNSWORTH, N. R. The re aaa of the periwinkles: Vinca and Catharanthus. he 24: 105-138. BS ER, & F. J. sae Studies on Vinca major ener at ae ceo Jour rm. Sci, 51: 217-224. 1962. — I. Anatomy of the xeromorphic leaves of. Hoe meets Thevetia peruviana nd Vinca minor (Apocynaceac). Nordic Jour. Bot. 3: 383-392. 1983. [Branched and unbranched, nonarticulated laticifers; at least in first two species, ek ap- pendages secretory when young, as well as anatomically similar to colleters of Ru- biaceae; scanning electron micrographs of oxalate crystals. eS D., & M. Hesse. A chemotaxonomic study of the subfamily Plumerioideae “the Apocynaceae. ores 39: 326-349. 1976 [1977]. [Detailed listing of plu- merane alkaloids by species.] Hame.L, M.-C. Contribution pie oeniane a la connaissance botanique des Apocy- nacees. I. Les données morphologiques et biologiques. Revue Gén. Bot. 90: 221- 300. 1983. II. Les one caryologiques. /bid. 91: 171-188. 1984. [Not cited by VAN DER LAAN & ARENDS; Includes a number of additional counts.] HARBORNE, J. B. Biochemical systematics, the use of chemistry in plant classification. Pp. 545-588 in L. REINHOLD & Y. Liwscuitz, eds., Progress in phytochemistry. nol Jain, S., S. BANSAL, & A. PILLAt. Stem-node-leaf continuum in Apocynaceae. Feddes Repert. 95: 183-188. 1984. [Unilacunar, one-trace leaf nodes in species of Carissa, Nerium, Thevetia, and others; armaments of Carissa said to be accessory buds on basis of vascular anatomy; similarity to Asclepiadaceae in nodal vasculature, petiolar appendages, laticifers, and intraxylary phloem noted. oor, 8. L., . Mirra. Epidermal and venation studies in Apocynaceae— VI. Bull. Bot. Surv. India 21: 68-80. i La aa Dareoyue in various genera, including Carissa, Trachelospermum, and Vinca, also cylic, anisocytic, anomocytic, and cyclocytic in Carissa: Claas Vinca shown to differ in leaf venation; review of taxonomic problems in Ca , P. C. SHARMA, V. CHANDRA, & L. D. KAP poor. Epidermal and venation studies in Apocynaceae II. Bull. Bot. Surv. India 11: 372-376. 1969. [Stomata paracytic in Allamanda, Alstonia, Carissa, Ervatamia, Thevetia, and others; also anomocytic in Sad Carissa, Thevetia, and others. KISAKUREK, M. V., A. J. M. LEEUWENBERG, & M. Hesse. A chemotaxonomic investi- ie of the plant families of Apocynaceae, Saas and Rubiaceae by their indole alkaloid content. Pp. 211-376 in S. W. PELL rR, ed., Alkaloids: chemical and biological perspectives. Vol. 1. New York and eee cities. 1983. [Eight alkaloid subgroups recognized. ] N, F. DER, & J. C. ARENDS. Cytotaxonomy of the Apocynaceae. Genetica 68: 3-35. 1985. LEEUWENBERG, A. J. M. Series of revisions of Apocynaceae, part III. Notes of Nerium L. and ign L. Meded. Landb. Wageningen 83(7): 57-60. 1983 [1984]. [Also see Adansonia I]. 16: 383-392. 1976.* LeGGeT, W. H. pe ocvnum, no. |. Bull. Torrey Bot. Club 3: 46. 1872a. No. 2. Ibid. 49, 50. 1872b. No. 3. Ibid. 53-55. 1872c. No. 4. [hid. 4: 1, 2. 1873a. No. 5. Ibid. 23. 1873b. Le Men, J., & H. Pourrar. Présence de l’acide ursolique dans les feuilles de Vinca 1989] ROSATTI, APOCYNINEAE a5 minor L., Nerium Oleander L. et Salvia officinalis L. Ann. Pharm. Frang. 10: 349- 351. 1952. [See Gipss, who indicated compound for Vinca but not for Nerium. | LiproLp, H. Die Gattung Plumeria L. (Apocynaceae) auf Kuba. ae Repert. 90: 193- 215. 1979. [Twelve endemic species; ?. rubra only in cultivation.] . Die Gattungen Thevetia L., Cerbera L. und C oe ee Rafin, (Apocynaceae). Thid. 91: 45-55. 1980. [Thevetia considered monotypic, 7. peruviana placed in Maueswarli Devi, H. Embryology of Apocynaceae—I. Plumiereae. Jour. Indian Bot. Soc. 50(1): 74-85. 1971. [Catharanthus roseus, Rauvolfia tetraphylla.| 2. Arduineae (Carissa spinarum Linn.). Pl. Sci. Lucknow 6: 24-29. 1974. [Illustrations include pollen.] MEYER, S. Studies in the family Apocynaceae. Calcutta Univ. Jour. Dep. Sci. 1: 131- 158. pls. 27-25. 1938. [Embryology of four extraregional species in genera including Alstonia, Rauvolfia, and Thevetia.] Miers, J. On the Apocynaceae of South America. 11 + 291 pp. 35 p/s. Edinburgh. 1878. [Helpful illustrations of / don nia (pl. 27), Echites (pl. 29), Rhabdadenia ( pl. 15.A), Thevetia 4.4), and others.] Monan, J. S.S., & J. A. eae Leaf architecture of Apocynaceae. Proc. Indian Acad. Sci. ae Sci. 91(3): 189-200. 1982. [Several of the taxa treated represented in the southeastern U. S. MOoNACHINO, J. revision of \facoubea and the American species of Landolphia. Lloydia 8: 291-317. 1945 ——. A résumé of the American Carisseae (Apocynaceae). [bid. 9: 293-309. 1946. . Arevision of the genus A/stonia (Apocynaceae). Pacific Sci. 3: 133-182. 1949. MUELLER-ARGOVIENSIS, J. Apocynaceae. /n: C. F. P. von Martius, Fl. Brasil. 6(1): I- 196. pls. 1-54. 1860. [Excellent illustrations of Allamanda (pls. 1, 2), Catharanthus (pl. 25), Echites (pls. 47, 50, 53), Rauvolfia (pls. 9, 10), Rhabdadenia (pl. 52) Vallesia (pl. 12), and others.] Nirsson, S., & M. vAN Campo. Pollen morphological studies in the Apocynaceae with special reference to fine structure. Proc. 4th Int. Conf. Palynol. Lucknow 1976/77. * NisHise, S., S. HisApA, & I. INAGAKI. Cyclitols of Vinca rosea and Amsonia elliptica. se 1177, 1178. 1973. [L-(+ )-bornesitol in each, the latter a Japanese species. ] NISHINO, 'E Corolla tube formation in six species of Apocynaceae. Bot. Mag. Tokyo 95: 1-17. 1982. lene ie a, COT rare TOSEUS, Vinca major, and eee species a Nerium and Trach NowlIckeE, J. ea OR, ot We ODSON, JR., & R. W. SCHERY, eds., Panama. ha Missouri Bot. noe 57: 59-130. 1970. PALMER, E., & N. PITMAN. Trees of southern Africa. Vol. 3. Pp. i-vi, 1499-2235. Cape Town, 1972 {197 3]. PicHon, M. Classification des fa I. Carissées et Ambélaniées. Mém. Mus. Natl. Hist. Nat. 24: 111-181. 1948a. II. Genre “Rauvolfia.” Bull. Soc. Bot. France 94: 31-39. 1947. IV. Genre 4/stonia et genres voisins. Bull. Mus. Hist. Nat. Paris, II. 19: 294-301. 1947b. V. Cerbéroidées. Not. Syst. Paris 13: 212-229. 1948b. VI. Genre Tabernaemontana. iid 230-254. 1948c. IX. Rauvolfiées, Alstoniées, Alla- mandées et Tabernémontanoidées. Mém. Mus. Natl. Hist. Nat. 27: 153-251. 1949a. XV. Genres Trachelospermum, Baissea et Oncinotis. Bull. Mus. Hist. Nat. Paris, I]. 20: 190-197. 1948d. XXII. Les espéces du genre Vinca. Ibid. 23: 439-444. 1951. XXII [sic]. Supplement aux Landolphiinae. Ba 20: 549-553. 1948e. XIX. Le rétinacle des Echitoidées. Bull. Soc. Bot. France 95: 211-216. 1948f. XXV. Echi- toidées. Mém. Mus. Natl. ae Nat. B. I: l- 143, 1950a. PS SEE Supplément a Plumérioidées. [bid. 145-173. 1950b. XX XV nographi Landolphiées. Mém. Inst. Franc. Afr. Noire 35: 3-437. 1953. 302 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 . Lesg ee nodales des Apocynacées et leurs modifications. Bull. Mus. Hist. Nat. Paris, IH. 21: 467-473. 1949b. RAFFAUF, R. F. ae chemotaxonomic considerations in the Apocynaceae. Lloydia 27: 286-298. 1964. [Alkaloid distribution with respect to PICHON’s classification; al- kaloid differences between Catharanthus and Vinca. FLAGLER. Alkaloids of the Apocynaceae. Econ. Bot. 14: 37-55. 1960. [Helpful discussion of systematic importance; evidence reportedly supporting re- latedness of Catharanthus and Vinca, Ochrosia and Rauvolfia, and genera of Ta- bernaemontaninae RAGHUVANSHI, S. S., & A. K. S. CHAUHAN. Evolution within certain species of Apoc- ynaceac. Pp. 231-235 in P. KAcHRoo, ed., Advancing frontiers in cytogenetics. Delhi. 1974. [Single-flowered variety of Tabernaemontana an diploid, 2n = 22; double- flowered variety ses 2n = 33, in some cases with separate style and style-suugma head for each ov Rao, A. S. A revision of Rauvolfia with particular reference to the American species. Ann. Missourl Bot. Gard. 43: 253-354. p/. 3. 1956. [About 80 species in nine sections; 34 species in four series and two sections in the Western Hemisphere; many topics discussed. ] Rao, V. S., & A. GANGuLI. Studies in the floral anatomy of the Apocynaceae. Jour. Indian Bot. Soc. 42: 419-435. 1963. [Twelve species investigated; terminal anther a Sega se partially composed of tissues from the connective; nectaries con- firmed as carpellary in genera including Catharanthus and Vinca, possibly recep- tacular in ee with annular nectaries (e.g., A//lamanda, Alstonia, Ervatamia).] Rau, M. A. Studies in the Apocynaceae. Jour. Indian Bot. Soc. 19: 33-44. 1940. [Em- bryology of extraregional species of six genera, including Carissa. ] Roy Tapapar, N. N. Cytotaxonomic ae in Apocynaceae and delineation of the cua 7 olutionary tendencies operating within the family. Caryologia 17: 10 138. 1964. [Tentative Boasegey of wae included; one branch within the Alsto- niinae ee of Amsonia, Lochnera (= Catharanthus), and Vinca; within Plu- merioideae, Plumerieae od Arduineae as side branches, Pleiocarpeae on main branch; Landolphiinae (including A//amanda) and Melodininae as two side branches of Arduineae. | Scuick, B. Untersuchungen tiber die Biotechnik der Apocynaceenblite I. Morphologie und Funktion des Narbenkopfes. (English summary.) Flora 170: 394-432. 1980 pees of style-stigma heads of various taxa, including 4//amanda cathartica, nsonia Tabernaemontana, Apocynum cannabinum, Catharanthus roseus, Nerium nee Ochrosia elliptica, Thevetia peruviana, Pere eee aN as and Vinca major; also, some light and transmission electron micrographs.] II. Bau und Funktion des Bestéubungsapparates. (English summary.) /bid. 172: 347-371. 1982. [Pollination of Vinca minor: illustrations ane DAOLS graphs of longitudinal sections through flowers of various taxa, including 4A//amanda cathartica, Thevetia peruviana, and I. minor. ] SHARMA, P. C., V. CHANDRA, S. L. KApoor, & L. D. KApoor. Epidermal and venation studies in ea | Bull. Bot. Surv. India 12: 92-96. 1970. [Stomata para- cytic and anomocytic in Rauvolfia tetraphylla, anomocytic and anisocytic in Catharanthus roseus (but see TRIVEDI & UPpADHYAY, 1976, 1977 SPETA, F. Weitere eae liber Proteinkérper in Zellkerner und ihre taxo- nomische Bedeutung. (English summary.) Pl. Syst. Evol. 132: 1-26. 1979. [In plants of sar - Vinca studied, spindles evidently thinner in former.] TAYLOR, W. — eds. The Vinca alkaloids: botany, Coase and caer xx + 357 pp. New York. 1973. [Contents include: STEARN, A synopsis of the genus }7nca including its taxonomic and nomenclatural history. 19-94: FARNSWoRTH, N. R., The phytochemistry of Vinca species, 95-180; TROJANEK, J.. M. Novacek, & F. Stary, Chemotaxonomy of Vinca species, 245-278. ] 1989] ROSATTI, APOCY NINEAE 353 & eds. The Catharanthus alkaloids: botany, chemistry, age and clinical use. ix + 323 pp. New York. 1975. ate include: STE W. A synopsis of the genus Cicadas (Apocynaceae), 9-44; SvoBoDa, G. H, & D. A. BLAKE, The phytochemistry and pharmacology of Catharanthus roseus (L.) G Don, 45-83.] Trivept, B. S.,& N. UpADHYAY. Ontogeny of stomata of some members of Apocynaceae. Geophytology 6: 92-97. 1976. [Paracytic, anomocytic, and anisocytic stomata in Catharanthus roseus, Tabernaemontana divaricata, and an extraregional species of Rauvolfia.) & . onan ies aah in Apocynaceae: epidermal structures. [bid. 7: 29-37. 1977. Tribes of BEN m & Hooker characterized by “distinct cuticular patterns’; data on stomata and ie Seren cells) and trichomes summarized for 27 species in 20 genera; stomata exclusively paracytic (e€.g., in es ca- thartica, Rauvolfia a. both paracytic and anomoc ytic (e.g., in Alstonia macrophylla, Thevetia peruviana, extraregional species of Carissa, Tae mun), exclusively anomocytic (only in sa eee divaricata), OY paracytic, anomocytic, and anisocytic (only in Catharanthus roseus), glandular trichomes ab- sent from all investigated taxa except R. tetraphylla. WaLkeR, D. B. Postgenital carpel fusion in Catharanthus roseus s(Ap ). L. Light and scanning electron microscopic study of gynoecial ontogeny. Am. Jour. Bot. 62: 457-467. 1975a. II. Fine structure of the epidermis before fusion. Protoplasma 86: 29-41. 1975b. III. Fine structure of the epidermis during and after eat Thid. 43- 63. 1975c. IV. Significance of the fusion. Am. Jour. Bot. 65: 119-121. Wana, D.-N., & Y.-N. ZHAO. New Late Cretaceous pollen genera and species in the Jiang mes Basin of Hubei. (In Chinese; English summary.) Acta Bot. Sinica 21: 320- 327.19 aa with possible affinities in the Apocynaceae; illustrations. ] WILKINSON, os . Apocynaceae. Pp. 224, 225 in V. H. HEywoop, ed., Flowering plants of the world. New York. 1978. Woopson, R.E., Jn. Studies in the Apocynaceae. I. A critical study of the Apocynoideae (with special reference to the genus 4pocynum). Ann. Missouri Bot. Gar 212. pls. 1-20. 1930. [Systematic history, morphology, and phylogeny of the family, as well as taxonomic treatments of Apocynum (83-156), Trachomitum (156-164), and Poacynum (164-168).] I. i revision of the genus se eae Ibid. 15: 341- 378. pls. 47-49. 1928a. [A genus of Central America and northern South America, allied most closely with I pate HLA Pooh of the genus 4msonia Ibid. 15: 379-435. 1928b. IITA. A new species of Amsonia from the South-Central States. Ibid. 16: 407-410. 1929. lV. The American genera of Echitoideae. /hid. 20: 605-790. 1933. [Key to genera, 624, 625.] 1V. The American ear of Echitoideae (continued). Ibid. 22: 153-306. 1935a. [Trachelospermum, 232-238.] 1V. The Amer- ican genera of Echitoideae (concluded). Ibid. 23: 169-438. pls. 1-7. 1936. [Anga- denia, 191-198: Urechites (= Pentalinon), 198-205; Rhabdadenia, 205-211; Echites, 217-252.] VII. An evaluation of the genera Plumeria L. and Himatanthus Willd. hid. 25: 189-224. 1938a. [Reasons for separation discussed; “carpels” in each said to be “strikingly subinferior”’; P. rubra indicated only from cultivation in the South- east, and there only in southern Florida. ] Observations on the inflorescence of Apocynaceae (with special reference to the American genera of Echitoideae). bid. 22: ae ‘ 5b. [Dichasial cyme most common and primitive within family, evolutiona athways to other types sug- gested: observations on various genera, including amen Catharanthus, Echites, and Vinca. Apocynaceae. N. Am. Fl. 29: 103-192. 1938b. [Asclepiadaceae never com- pleted.] J. A. Moore. The vascular anatomy and comparative morphology of apoc- ynaceous flowers. Bull. Torrey Bot. Club. 65: 135-166. pis. 3-5. 1938. 354 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 KEY TO THE GENERA OF APOCYNACEAE IN THE SOUTHEASTERN UNITED STATES General characters: trees, shrubs, vines, or herbs with latex; leaves simple, opposite often with colleters, inflorescences cymose or sometimes flowers solitary; flowers perfect, actinomorphic or nearly so, 5-merous except carpels 2; sepals nearly free, often with squamellae; corollas variously svmpetalous, sometimes appendaged adaxially, lobes ofien overlapping; stamens epipetalous, alternating with corolla lobes, filaments free and short, anthers free (Plumertoideae) or variously agglutinated to each other and to style-stigma head (Apocynoideae), pollen released as as ie ds (Apocyneae) or individual grains, nectaries external to ovaries or abso usually 5, free to completely connate; carpels free or connate in ovule-bearing region, connate in non-ovule-bearing region into either style and style-stigma rie or clavuncle; ovaries generally superior, simple, unilocular, and with marginal placentae or sometimes compound, unilocular [or bilocular] and with parietal [or axile] placentae; ovules 1 to many per carpel; fruits follicles, sometimes dr ag capsules, [or berries]; seeds in follicles glabrous, each with a tuft of trichomes at apex at both ends] [or arillate], those in drupes [or berries] glabrous, those in capsules nee A. Anthers free or only slightly adherent to style-stigma heads; corolla lobes usually overlapping to left, at least in aestivation (but usually to right in A/stonia macroph\ lla and Ochrosia elliptica); fruits follicular (seeds then without trichomes, or with tri- chomes at both ends, or arillate), drupaceous, capsular, or baccate. B. Ovary |, compound, in fruit remaining so; styles not cleft at base. C. Ovaries perenne fruits capsular, spiny; plants unarmed: — usually whorled or subwhorled. 200.000.0000... 00.000 Allamanda. C. Ovaries oe gare baccate, not spiny: plants armed; Cee opposite. eect ar lslay aye ene ence tones belo eon Aye oes eee arissa ii ae ‘a.! B. Ovaries 2, simple, in fruit remaining so or fusing; styles cleft at ba D. Plants shrubs or trees; fruits follicular (seeds then with trichomes i both ends or arillate) or drupaceous. E. Leaves alternate. Corollas narrower and shorter than leaves, salverform (but tube c stricted near middle or widened only subapically); calyces srr eae nectaries aeeoh fruits drupaceous, usually 1-seeded, usu- ally only | ovary maturing. f F. Corollas ee than ad nearly as long as leaves, funnelform: calyces with i asec aaae nectaries annular, shallowly 5-lobed; fruits drupa- ceous, 2- to 4-seeded, both ovaries maturing and fusing together. . Gp Mesh Sa ioe Bitte eS ne tore a heat cates hag Thevetia peruviana. 12] E. Leaves opposite or whorled. G. Leaves opposite; calyces with squamellae; fruits follicular (seeds ar- illate, without trichomes). ....... [Vabernaemontana divaricata.'"] Leaves opposite or whorled; calyces without squamellae; fruits follic- ular (seeds then not arillate, but with trichomes at both ends) or dru- paceous (seeds then without trichomes). Leaves all more or less the same size and shape; ovules/seeds attached at middle; fruits follicular or drupaceous. Pedicels several times longer than calyces; leaf apices usually acuminate; fruits follicular (ovaries remainin [- rm G. Nee g free). .....00, Se Wad Se he aed ete bod eee eas eats stonia macrophylla.'?] I. Pedicels about as long as calyces: leaf apices usually acute to obtuse; fruits drupaceous (ovaries remaining free during fruit "Escaped from or persistent beyond cultivation but clearly reproducing neither sexually nor asex- ually in the southeastern United States. For further consideration see end of family discussion. 1989] ROSATTI, APOCYNINEAE 355 development) although each may resemble 2 that are fused, or sometimes becoming fused at base). .[Ochrosia elliptica.'? H. Leaves clearly of different sizes and shapes: ovules/seeds_ attached at | end; fruits drupaceous (ovaries becoming fused during fruit development). ...................... Rauvolfia tetraphylla.'?) D. Plants herbs or subshiriibs: fruits follicular (seeds without trichomes). J. Leaves alternate, sometimes pees iat or subwhorled:; nectaries evi- dently pace or connate into a shallow ring. ............ 2. Amsonia. J. Leaves opposite. sometimes oe a K. Corollas ee a. mouths of tubes ee or cae widened subapically; anthers completely fertile. .......... : eee K. eee funnelform, mouths of tu bes not — or tubes ened subapically: anthers incompletely fertile. ....... TNCQ. A. pane ee to style-stigma heads or to clavuncles; corolla lobes overlapping a at least in aestivation; fruits follicular (each seed with a tuft of trichomes at L "Pats shrubs or small trees: leaves usually whorled or subwhorled, rarely opposite; nectaries absent; corollas often double; ovaries free but somewhat agglutinated. os ie ea aes BAe duke Abn naeeaie cohen es [Nerium Oleander.'?] . Plants herbs, vines, or scrambling to erect subshrubs; leaves usually opposite, sometimes subopposite, rarely alternate; nectaries 5, a or basally connate; co- rollas normal (not double): ovaries free and not agglutinated. Plants vines or scrambling to erect subshrubs; se funnelform to salver- form, unappendaged adaxially; non-ovule-bearing region of gynoecium differ- entiated into style and style-stigma head; pollen released as individual grains. lants vines; corollas salverform, about as long as or longer than leaves, tubes usually widened at or near middle; follicle pairs (from single flowers) diverging usually ca. 180° (but ranging from ca. 160 to 200°). .......... Pee a ieee tated 2 decline ee etch encanta eases 6. Echites. Plants vines or scrambling to erect subshrubs; corollas funnelform, about as long as or longer than leaves (in 7rachelospermum clearly shorter than leaves, funnelform to sometimes approaching salverform but then tubes widened apically); follicle pairs (from single flowers) diverging ca. 90° or tm 2 less O 1 A: g ( 4 ging 1 ) : leaves generally 1-3 cm long, margins revolute and ia meeting BO ARIAI x, 5 Annie a Aerator! oe ip areeatentd con condo Bian de aes gadenia. O. Nectaries not exceeding ovar 5, leaves generally 4-9 cm nae pare See but not meeting abaxial ly. . Anthers with long, spirally lapi placentae ea (seeds each encased ina ee deciduous, BoE. haped CANS) a3 eph teas ah a ecseat hance has goa eee hep 8. Pentalinon. Re re without apical a placentae not chaffy (seeds not each encased in a scale Q. Inflorescences many- ieee calyx lobes usually aristate and keeled, with squamellae; corollas funnelform or sometimes ap- proaching salverform (but then tubes widened apically), clearly shorter than leaves: leaf apices acute to subcaudate. ........ Be die ee ts gpa ts et agian, eke kes eee ae 9. Trachelospermum. Inflorescences |- or 2-, sometimes up to 6-flowered; calyx lobes mucronate, not keel ed, without squamellae; ieee ae form, about as long as or longer than leaves; leaf a CLONA tes 225: tis ones itd passin tas eats 2 10. Riabdadenia. M. Plants herbs; corollas campanulate, short-cylindrical, or urceolate, appen aged within; non-ovule-bearing region of gynoecium a sessile, rhomboid cla- vuncle; pollen released as tetrads. .................00.0-. 11. Apocynum. 0 356 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Subfam. salar a Schumann, in Engler & Prantl, Nat. Pflanzenfam. V. 2: 122. 1895, **Plumieroideae.”’ Tribe PLUMERIEAE Schumann, ibid., “‘Plumiereae.” Subtribe Rauvolfiinae Schumann, ibid. 1. Vallesia Ruiz & Pavon, Fl. Peruv. Chil. Prodr. 28. 1794. Terrestrial shrubs or small trees; axes and leaves glabrous. Leaves alternate; blades lanceolate, narrowly ovate, or oblanceolate, apices rarely mucronate, then only slightly so, margins usually revolute, sometimes slightly undulate, bases obtuse to attenuate, often oblique; petioles without lateral appendages, colleters present on bases and adjacent areas of stem. Inflorescences terminal but soon appearing lateral (opposite a leaf) by sympodial growth, pedunculate, few-flowered, bracteate; pedicels much longer than calyces, ebracteolate. Calyx Y-Y as long as corolla or less, glabrous; lobes triangular, bases usually not imbricate, squamellae lacking. Corolla salverform (but tube widened basally and subapically [or only subapically]), shorter than leaves, with nonglandular pubescence (adaxially, at lobe bases); lobes overlapping to left in aestivation, spreading, elliptic to oblong (apices acute), entire to somewhat undulate and/ or revolute, somewhat oblique; tube without adaxial appendages but thickened apically. Stamens inserted at base of subapically widened portion of corolla tube, glabrous; filaments short, slender, terete, straight or slightly curved; an- thers not agglutinated to style-stigma head, cordate (apices and basal lobes obtuse), completely fertile, introrsely dehiscent, long axis mostly parallel to that of filament. Nectaries lacking. Carpels completely free in ovule-bearing region, weakly united above; ovaries 2, simple, distinct, unilocular, with mar- ginal placentae and several ovules; style in transverse section bilobed; style- stigma head ellipsoid, not umbraculiform, with 2 obtuse lobes apically, gla- brous. Fruits drupaceous, usually only | from each flower, oblique, clavate, obpyriform, or ovoid, the surface more or less translucent, whitish. Seeds | for 2] per fruit, usually fusiform, without trichomes. LECTOTYPE SPECIES: V. glabra (Cav.) Link (Rauvolfia glabra Cav.). Britton & Millspaugh, Bahama FI. 339. 1920, stated “two known species, the following typical’ in reference to 1. glabra, Woodson, N. Am. FI. 29: 138. 1938, also indicated this species. (Named for Francisco Valles, died 1592, physician to Philip II of Spain.)— TEAR-SHRUB. A small genus of two (Chittenden; Long & Lakela; Schumann) to aragtl et al.; Standley & Williams; Woodson, 1938b) or perhaps ten (Airy Sha Willis; Tomlinson) species of shrubs or small trees with alternate leaves, aul flowers, and drupaceous fruits. The plants occur from Argentina to Mexico and in the southern parts of California and Florida according to some (e.g., Airy Shaw, in Willis; Schumann; Woodson, 1938b) but are restricted to tropical and subtropical America according to others (e.g., Bailey ef a/.; Chittenden; Long & Lakela; Standley & Williams; Tomlinson). The genus is evidently not represented in Panama (see Nowicke). Vallesia somewhat resembles and may be most closely related to Rauvolfia 1989] ROSATTI, APOCYNINEAE 3o7 L. The latter is a pantropic genus of perhaps as many as | 10 species (Nowicke): it is represented in the Southeast by R. tetraphylla, a species native to the American tropics and subtropics that has occasionally escaped from cultivation but is evidently not fully naturalized in Florida (see family discussion). Both of these genera are apocarpous and have drupaceous fruits, but in Va/lesia the leaves are alternate (vs. opposite or whorled), the flowers lack nectaries, and each fruit is the product of a single carpel (vs. both carpels). Va/lesia has been allied with Rauvolfia and eight other genera in the Rauvolfiinae, one of four subtribes comprising the Plumerieae. This tribe also includes eight genera in the Cerberinae. Two of these, 7hevetia and Ochrosia, with bicarpellate and unicarpellate drupes, respectively, have a species each that is cultivated and escaped, but probably not naturalized, in Florida (see family discussion). Plants of Vallesia have much smaller flowers than those of 7. peruviana and can be readily distinguished from those of O. e/liptica by their alternate (vs. opposite or whorled) leaves. Vallesia includes one or perhaps two species native to southern Florida and the American tropics. Vallesia antillana Woodson, chromosome number ev- idently unreported, has long been confused with V. glabra (Cav.) Link, 2n = 22, but according to Woodson (1937, 1938b) it differs in its longer (10-12 vs. 5-7 mm), more deeply lobed corollas and wider (1-3 vs. 0.6-2 cm) leaves. Woodson (1938b) indicated only I. antil/ana for the southeastern United States (southern Florida) and reported V. g/abra from California to Veracruz and Oaxaca, as well as from Pacific coastal South America; Tomlinson agreed that V. glabra had been incorrectly attributed to Florida. Nevertheless, Long & Lakela recognized both species in the southern part of the state, using differences in pedicel length (in addition to those outlined above) to distinguish them. However, whereas Long & Lakela indicated that in V. antillana and V. glabra the pedicels were less than or greater than 5 mm long, respectively, Woodson (1938b) had indicated almost the opposite (4-5 vs. 2-4). According to Long & Lakela, both Vallesia antillana and V. glabra occur in coastal hammocks in southern Florida and throughout much of tropical America, although in Florida V’. glabra was indicated only from the Keys. I do not consider the five specimens I have seen from the Southeast (all from the Florida Keys) to represent more than one species, although the situation remains confused. Four of them have pedicels over 5 mm long (Vallesia glabra): two of these have corollas at least 10 mm long (V. antillana), while the other two have retained no corollas at all. The fifth specimen has pedicels less than 5 mm long (V. antillana) but corollas 6-9.5 mm long (mostly V. glabra). According to Tomlinson, the fruits of 1”. antillana develop quickly and are one-seeded; the common name (tear-shrub) refers to their pear shape and whitish, almost translucent quality. At least in southern Florida, flowers and fruits are usually present throughout the year (Tomlinson). It is notable that, although the plants are evergreen, the leaves tend to disappear from herbarium sheets with time. According to information summarized by Gibbs, leucoanthocyanins and tannins have yet to be detected in the genus, despite specific testing, although alkaloids appear to be numerous. 358 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 The fruits of Vallesia glabra are sometimes eaten by humans, and a juice from the plants is reportedly used in ophthalmia (Uphof). REFERENCES: Under references for the Apocynineae, see Airy SHAW (in WILLIS); BAILEY ef al.; CHITTENDEN; GIBBS; LONG & LAKELA; SCHUMANN; STANDLEY & WILLIAMS; and UPHOF. Under references for the Apocynaceae, see NowicKEe and Woopson (1938b). ToMLINSON, P. B. The biology of trees native to tropical Florida. x + 480 pp. Allston, assachusetts. 1980. [V. antillana, 96-98.] Woopson, R. E., Jk. New or otherwise noteworthy Apocynaceae of tropical America Ann. Missouri Bot. Gard. 24: 11-16. 1937. [Protologue for V. antillana, with discussion. Subtribe Alstoniinae cuca * in era - — Nat. Pflanzenfam. IV. 2: 1305 2. Amsonia Walter, Fl. Caroliniana, 98. 1788. Terrestrial, erect, perennial herbs; axes and leaves glabrous or with non- glandular pubescence. Leaves alternate to sometimes subopposite and/or sub- whorled; blades ovate to linear, apices acute to acuminate, margins occasionally revolute, bases obtuse to attenuate, often oblique; petioles without lateral ap- pendages, colleters lacking on bases and adjacent areas of stem. Inflorescences usually terminal, sometimes axillary, pedunculate, several- to many-flowered, minutely bracteate; pedicels usually longer than calyces, rarely bracteolate. Calyx “s as long as corolla or less, glabrous; lobes triangular and sometimes attenuate, bases not imbricate to sometimes slightly so, squamellae lacking. Corolla salverform (but tube usually widened apically), shorter than leaves, with nonglandular pubescence (adaxially, along tube and at lobe bases; abax- ially, along tube); lobes overlapping to left in aestivation, usually spreading, sometimes erect, lanceolate to elliptic or narrowly spatulate, entire to undulate, generally not oblique; tube without adaxial appendages. Stamens inserted near top of corolla tube, glabrous: filaments short, terete, slightly curved; anthers not agglutinated to style-stigma head, each sagittate (apex and basal lobes mostly obtuse), completely fertile, introrsely dehiscent, long axis mostly parallel to that of filament. Nectaries evidently lacking or connate into a shallow ring. Carpels completely free in ovule-bearing region, united above; ovaries 2, sim- ple, distinct, unilocular, with marginal placentae and numerous ovules; style terete; style-stigma head short-cylindrical, basally umbraculiform, sometimes with 2 obtuse lobes apically, glabrous. Fruits follicular, usually 2 from each flower (the pair divergent less than 90°), narrowly cylindrical to slightly monili- form; surface striate, glabrous; dehiscence by segmentation and/or disintegra- tion, usually along spiralled lines. 7 numerous, narrowly cylindrical, with- out trichomes. LECTOTYPE SPECIES: eT Walter; see Britton & Brown, Illus. Fl. No. U.S. & C ance ed. 2. 3: 20. 1913. (Probably named for Charles Amson, circa 1760, physician of Gloucester Co., Virginia, and friend of John Clayton [see Woodson, 1928b].)—BLUESTAR. 1989] ROSATTI, APOCYNINEAE 359 One of the few genera of Apocynaceae in temperate to subtemperate regions, comprising perhaps as few as five highly variable or as man 0 more narrowly defined species in the southern United States and northern Mexico, and one in Japan (and possibly Korea and China as well; see Ohwi). As originally circumscribed by Walter, the genus comprised only Amsonia ciliata Walter and Amsonia Tabernaemontana Walter (Tabernaemontana Amsonia L.), the latter of which had been included by Linnaeus (Sp. Pl. 2: 301. 1762) as the sole herbaceous, Temperate Zone member of Tabernaemontana. Amsonia was progressively expanded by the addition of both new and transferred taxa (sum- marized by Woodson, 1928b), perhaps the most interesting of which is the Japanese A. elliptica (Thunb.) Roemer & Schultes (Tabernaemontana elliptica Thunb.). Members of Amsonia are perennial, mostly erect, somewhat woody herbs with alternate leaves, medium-sized flowers with blue corollas (hence the common name bluestar), and continuous or articulated follicles. Amsonia, with its free, unappendaged anthers, seeds without trichomes, and corolla lobes that overlap to the left in bud, is well placed within the Plume- rioideae. It is, furthermore, assignable to the Plumerieae because of its two separate ovaries, and within that to subtribe Alstoniinae on the basis of its many ovules, calyces without squamellae, and dry (follicular) fruits. (The fruits may be follicular in the Tabernaemontaninae, but in such cases the seeds have a fleshy aril.) Amsonia is perhaps most closely related to Haplophyton A. DC., a monotypic genus with subopposite leaves that is distributed from Arizona and New Mexico southward to Guatemala (reports from southern California and Cuba are evidently unconfirmed; see Woodson, 1928b, 1938b). Haplo- phyton cimicidum A. DC. differs from most P including Amsonia in its apically comose seeds. Woodson (1928b) idered Rhazya (two species, Greece to India) to be another very close relative of Amsonia, from which it differs 2a in haves nectaries. Bee » and geographically distinct sections were recognized pie ie by Schumann: sect. Euamsonia K. Schum. (= sect. AMSONIA) cua of corolla tubes not constricted, apices of stigmas depressed-capitate or truncate) of southeastern North America and Japan, and sect. SPHINCTO- SIPHON K. Schum. (summits of corolla tubes constricted, apices of stigmas with two distinct lobes) of the southwestern United States and northern Mexico. Woodson (1928b) elevated these sections to subgenera, noting that in addition to the differences mentioned by Schumann. the inflorescence bracts were much more conspicuous in subg. SPHINCTOSIPHON (K. Schum.) Woodson. He re- moved four species from the latter group to subg. ARTICULARIA Woodson (which has articulated instead of continuous follicles, and seeds that are rounded or pointed instead of truncate at the ends) and divided the remainder into sects. MICRANTHAE Woodson and LONGIFLORAE Woodson, the latter with larger flowers, fruits, and seeds. McLaughlin maintained subg. ARTICULARIA but el- evated Woodson’s two sections to subg. SPHINCTOSIPHON Sensu stricto and subg. LONGIFLORAE (Woodson) McLaughlin (as subg. “Longiflora’’), respectively, on the basis of morphological and ecological evidence (see McLaughlin). There appears to be a hierarchy of relationships within Amsonia that 1s reflected at least partially in the systems of Schumann and Woodson (1928b) 360 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Figure |. Amsonia. a-l, 4. Tabernaemontana: a, distal part cess stem, x 1/2 b, rhizome with bases a 2 hee above-ground stems, x '/; c, flower bud, corolla cee overlapping to left, x 3; d, flower, salverform corolla with apically widened tube, x 3: e, flower cut ioneitadinally. corolla with nonglandular pubescence, anthers free both ae each other and from style-stigma head, gynoecium free in ovule-bearing region, united above into style and style-stigma head, x 5; f, apical part of a (terete) and style- “stigma head (umbraculiform Seite slightly bilobed apically), 12; g, base of flower cut in transverse section through ovaries (5 calyx lobes; serail tube; shallow, annular nectary; 2 distinct, unilocular ovaries, each with a marginal placenta and nu- merous ovules), x 12: h, 3 lobes of persistent calyx, annular nectary, 2 immature follicles after fall of style and corolla, x 12; 1, mature follicles (both carpels typically developing), x Vo. j, seed, x 3; k, mature seed cut longitudinally (seed coat hatched, endosperm stippled, embryo unmarked), x 3: 1, embryo, but not in that of McLaughlin, who treated the four groups apparent in the genus as subgenera. The most substantial discontinuity among these plants, morphologically and geographically, seems to lie between eastern (in south- eastern North America and Japan) and western (in the southwestern United 1989] ROSATTI, APOCYNINEAE 361 States and northern Mexico) populations. Thus, although I agree with Schu- mann that two major groups exist within the genus, I would favor their rec- ognition at the subgeneric level, as subg. AMSONIA and subg. SPHINCTOSIPHON, respectively, in order to accommodate as sections and subsections the addi- tional groups recognized at various levels in the latter by Woodson (1928b) and McLaughlin. Although both recognized subg. ARTICULARIA as a distinct subgenus, Woodson (1928b) indicated that it was more closely related to subg. SPHINCTOSIPHON than to subg. AMSONIA, suggesting that inclusion of the group as sect. ARTICULARIA within subg. SPHINCTOSIPHON would be more indicative of relationships. This proposal would require that the two groups recognized as sections of subg. SPHINCTOSIPHON by Woodson be demoted to subsections." Possibly two extremely variable, or perhaps as many as seven or even eight species of Amsonia, all assignable to subg. AMSONIA, occur in the southeastern United States, although the genus is not represented in southern peninsular Florida. Among those characterized by externally glabrous corollas are A. rigida Shuttlew. ex Small (including 4. g/aberrima Woodson, according to Godfrey & Wooten), 4. ciliata, and A. Hubrichtii Woodson, three species for which chromosome numbers have evidently not been reported. Amsonia rigida is a species of wet or moist, relatively open habitats on the Coastal Plain from southern Georgia and northern Florida to southeastern Texas (but probably only as far west as central Louisiana, if 4. glaberrima is excluded). Amsonia ciliata is a variable species found in relatively dry, open places, often in sandy or limy soil, throughout most of the Southeast and adjacent areas. It differs from A. rigida sensu lato in its sessile to subsessile (vs. distinctly petiolate) upper leaves. Many specimens obtained throughout most of the range of 4. ci/iata have linear to nearly filiform leaves (up to 50 times as long as wide by my measurements) and on that basis have been recognized as var. tenuifolia (Raf.) Woodson (var. filifolia Woodson). Plants in the eastern parts of Oklahoma and Texas are assignable to var. texana (Gray) Coulter, which allegedly (Small, 1913) differs from var. ciliata in its longer calyx lobes (2 vs. | mm) and corolla tubes (9-11 vs. 6-8 mm), as well as in its glabrous (vs. pubescent) stems and branches. Such plants may be transitional to subg. SPHINCTOSIPHON Amsonia Hubrichtit is slesely related to the so-called Ozark phase of A. ciliata var. filifolia (= var. tenuifolia), according to Woodson (1943), but the plants differ in their larger size, more lustrous foliage, obscurely (vs. distinctly) pe- dunculate inflorescences held among (vs. well above) the leaves, and smaller, narrower corollas. They reportedly occur in rocky or gravelly creek bottoms in the Ouachita Mountains of western Arkansas and southeastern Oklahoma, although the species was not included by Godfrey & Wooten (coverage includes Arkansas) or Correll & Correll (coverage includes Oklahoma). Other species of 4msonia 1n the Southeast have externally pubescent corollas. Amsonia Tabernaemontana, 2n = 16, 22, 32, differs from A. /udoviciana Vail, 2n = 22, in its glabrous (vs. pubescent, at least distally) follicles, glabrous '3Such proposals regarding these taxa of the southwestern United States and northern Mexico would require further justification and nomenclatural changes beyond the scope of the present treatment. 362 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 calyces, and abaxially glabrous or glabrescent (vs. tomentose) leaves. The for- mer often occurs in sandy soil, in mesic woods and on floodplains, river banks, and lake shores throughout most of southeastern North America with the exception of peninsular Florida and possibly coastal areas. It has reportedly escaped from cultivation in several northeastern states (see Woodson, 1928b). Amsonia ludoviciana, on the other hand, is restricted to Mississippi, Louisiana, and Arkansas, where it usually grows in moist, open woods, according to several accounts, although Lemke reported it only from Louisiana, Mississippi, and South Carolina. Amsonia illustris Woodson, 21 = 22, occurs in wet sand or gravel in or near streams and ditches as well as in swamps and wet meadows from eastern Kansas and southern Missouri to eastern Texas and western Arkansas. It allegedly differs from 4. Tabernaemontana in its lustrous (vs. dull) adaxial leaf surfaces, its sparsely pubescent (vs. glabrous) calyces, and its drooping (vs. erect) mature follicles (a condition not seen elsewhere in the genus, according to Woodson, 1929) that are somewhat constricted between the seeds (but by no means articulated as in sect. ARTICULARIA, according to Woodson, 1929). This group of southeastern species with externally pubescent corollas appears also to include 4msonia repens Shinners, chromosome number evidently un- reported, which according to Correll & Johnston differs from A. Tabernae- montana and A. i/lustris in its usually geniculate (vs. straight) stem bases and more xeric habitats (in prairies and along railroad tracks vs. along streams and near lakes or in wet soil of swamps, meadows, ditches, etc., respectively). Compared to 4. Tabernaemontana, A. repens also has smaller (less than or equal to 6 by 3.cm vs. 6-15 by 5 cm), thicker, more lustrous leaves, and pilose (vs. glabrous) calyces. That 4. repens was reported for neither Arkansas (see Smith) nor Louisiana (see MacRoberts) is consistent with its apparent restric- tion to eastern Texas (Correll & Johnston). Correll & Johnston remarked that Amsonia Tabernaemontana, A. tllustris, and A. repens, the three species with externally pubescent corollas, ae with A. glaberrima (and therefore probably 4. rigida), form an extremely close and possibly conspecific alliance, and that the pubescence characters by which they are largely defined are not entirely reliable. Thus, with the exception of those assignable to 4. ciliata (or A. Hubrichtii) and possibly A. ludoviciana, which is also based on pubescence characters, all plants of Amsonia (i.e., subg. AMSONIA) that occur or might occur in the Southeast may belong to one extremely variable species, 4. Tabernaemontana. Nevertheless, after reconsideration of morpho- logical evidence, Mathé & Mathé continued to recognize most of these species. As McLaughlin pointed out, it is notable that while Woodson’s treatments (1928b, 1938b) of the species in subg. AMSONIA are rather consistent, those dealing with species of subg. SPHINCTOSIPHON are not, suggesting that species in the latter subgenus are even less distinct than those in the former. In fact, McLaughlin found that the pubescence characters used by Woodson (1928b), for example, to separate some of the species in sect. ARTICULARIA were not rehable. Several indole alkaloids have been reported in Amsonia (see primarily Kisa- kiirek ef al.). Tests conducted by Gibbs using leaf material of 4. angustifolia 1989] ROSATTI, APOCY NINEAE 363 (Aiton) Michx. (= A. oe were reportedly negative for saponins but gave strong reactions for tannin Species of Amsonia are grown as ornamentals. The plants are propagated in the spring by division or by seed, and in the summer by cuttings (Bailey ef a/.; Chittenden). At least some species are hardy and have been cultivated in several northeastern states, where 4. Tabernaemontana has been reported as an escape (Woodson, 1928b) REFERENCES: Under references for the Apocynineae, see BaILey et a/.; CHITTENDEN; CORRELL & CORRELL; CORRELL & JOHNSTON; GIBBS; GODFREY & WooTEN; MACROBERTS; OHWI; SCHUMANN; SMALL (1913); and SMITH. Under references for the Apocynaceae, see KISAKUREK ef a/. and Woopson (1928b, 1929, 1938b). Lemke, D.E. Recent collections and a Hanae y ues of Amsonia ludoviciana Vail (Apo ynaceae). Sida 12: 343-346. In Louisiana, fewer than 1000 plants anges from a number of small ean in the longleaf-pine region, perhaps 200 plants in cultivation.] MatHE, I., & I. MATHE, JR. The morphological characteristics of the taxa belonging to the subgenus Euamsonia. (In Hungarian; English summary.) Bot. Kézlem. 69: 145- 150. 1982. [Includes 4. ilustris, A. ludoviciana, A. rigida, and A. Tabernaemontana (vars. oe salicifolia).] ay ene n, S. P. A revision of the southwestern species of Amsonia (Apocynaceae). eee Bot. Gard. 69: 336-350. 1982 Se 1 K. Amsonia ludoviciana. Addisonia 17: 23, 24. pl. 556. 1932. [Color plate. ] . Amsonia rigida. Ibid. 20: 25, 26. pl. 653. 1937. [Color plate.] WILDEMAN, E. DE. Amsonia Tabernaemontana Walt. Ic. Selec. Horti Thenensis 2: 121. pl. 69. 1901. [Helpful illustration: stigma shown to be shallowly bilobed apically.] Woopson, R. E., Jk. A new 4msonia from the Ozarks of Arkansas. Rhodora 45: 328, 329. 1943. [4. Hubrichtii Woodson. ] oa 3. Catharanthus G. Don, Gen. Hist. 4: 71, 95. 1838.'4 Terrestrial, erect, ascending, or decumbent, annual or perennial herbs, often woody toward the base; axes and leaves glabrous or with nonglandular but “There has been SS. disagreement regarding the correct name for this genus. Reichenbach (Conspec. Regni Veg. 1: _ 1828) was the first to propose that Vinca rosea L. is generically distinct from the other species : aa and published the monotypic Lochnera to accommodate it. Unfor- tunately, the name was not accompanicd by a description, although Dwyer argued that by listing V. rosea L., and only this species. in a position subordinate to Lochnera, an indirect reference to a ag and effectively ieee eee (. e., the species description of Linnaeus, Syst. Nat. 759) was provided. However. Article 41.2, ICBN, 1988, specifies that in order for the name ofa genus to be validly publis a it must i accompanied cither by a description or “by a reference (direct or indirect) to a previously and effectively published description or diagnosis of a genus or subdivision of a genus,” the latter referring only to * ‘taxa ofar. bane between genus ae sree! es” (Art. 4, Note 1, ICBN, 1988). In addition, Article 34.1 specifically dly published “by the mere mention of the ae taxa included in the taxon concerned.” ‘Thus, Lochnera Reichenb., a nomen nudum, is illegitima Endlicher’s description of Lochnera in Fake 1838 (Gen. Pl. 583. 1838), would have established Lochnera Reichenb. ex Endl. as the correct name for the genus, except that current evidence suggests 364 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 sometimes viscid pubescence. Leaves opposite to rarely subopposite; blades variously ovate to obovate, apices abruptly and narrowly mucronate, margins sometimes slightly revolute, sometimes undulate, bases obtuse to attenuate, often oblique; petioles without lateral appendages, colleters present on bases and adjacent areas of stem. Inflorescences axillary, essentially epedunculate, l- to less often 4-flowered, ebracteate (but associated leaves often reduced); pedicels usually shorter than calyces, ebracteolate. Calyx %[—] as long as corolla or less, with nonglandular pubescence but sometimes viscid; lobes narrowly triangular and attenuate, bases not imbricate, squamellae lacking. Corolla sal- verform (but tube widened subapically), as long as or longer than leaves, with nonglandular pubescence but sometimes viscid (adaxially, near juncture of tube and lobes; abaxially, sometimes along tube [and/or lobe margins]); lobes usually overlapping to left in aestivation, spreading, obovate, entire (but mucronate), somewhat oblique; tube without adaxial appendages but thickened apically. Stamens inserted in upper 4 of widened portion of corolla tube, mostly gla- brous; filaments very short, narrowly terete, straight or slightly curved; anthers not agglutinated to style-stigma head, each narrowly cordate (apex acute or obtuse, basal lobes mostly obtuse), completely fertile, introrsely and somewhat laterally dehiscent, long axis mostly perpendicular to that of upper part of filament. Nectaries 2, alternating with carpels and widely spaced, elongate and usually exceeding ovaries. Carpels completely free in ovule-bearing region, united above; ovaries 2, simple, distinct, unilocular, with marginal placentae and numerous ovules; style terete; style-stigma head cylindrical, basally um- braculiform, with 2 small or obscure lobes apically, with a tuft of hairs sub- apically and at base of cylindrical part. Fruits follicular, | or 2 from each flower (nearly sessile in leaf axils) (the pair divergent less than 90°), short and usually somewhat moniliform; surface striate, with nonglandular pubescence; dehis- cence tardy, primarily by an adaxial (and sometimes also an abaxial) suture, Seeds numerous, ovoid, without trichomes. (Lochnera Reichenb., 1828, nom. illegit.) LECTOTYPE SPECIES: C. roseus (L.) G. Don (Vinca rosea L.); Woodson (N. Am. Fl. 29: 124. 1938) indicated 1. rosea as the type species of Lochnera. (Name from Greek, katharos, pure, and anthos, flower, according to Don, in reference to the “neatness and beauty” of the flowers.) A genus of three (Standley & Williams) to six (Nowicke) or seven (Stearn, 1975) species of trailing or erect herbs (often woody at the base) with opposite, deciduous leaves, conspicuous flowers, and follicular fruits. With the exception of Catharanthus pusillus (Murray) G. Don of India, all of the species were originally endemic to Madagascar. Catharanthus roseus (L.) G. Don (Lochnera rosea (L.) Reichenb.; Vinca rosea L.), the Madagascar periwinkle, is commonly an earlier date for the competing Catharanthus G. Don. According to Stearn (1966, 1975), the fourth volume of Don’s General History, which included on page 95 his description of Catharanthus, was listed as being “now first published” in an issue of Bent’s Month/y Literary Advertiser dated 10 April 1838. In fact, since parts of the ba volume were issued in advance of its completion, an even earlier date is likely: a reference to page 23 of this 908-page volume was published in the Botanical Register in June, 1835. For more eee accounts of this controversy, see Dwyer, Lawrence, and Stearn (1966, 1975). 1989] ROSATTI, APOCYNINEAE 365 and widely cultivated, escaped, and naturalized in tropical and subtropical regions throughout the world, including, in the Southeast, most of peninsular Florida and the Keys (Judd, pers. comm.); it is adventive but 1s not naturalized in the Florida Panhandle (Anderson), in the Carolinas, and in southern Cali- fornia but has evidently not been reported out of cultivation elsewhere in the continental United States or Canada. The species is completely naturalized throughout most of Central America but is especially abundant in sand under coconut palms along the Atlantic Coast (Standley & Williams). Apparently, no other species of Catharanthus has been reported out of cultivation in the Southeast. Catharanthus conforms to the general characters defining the Plumerioideae and (as Lochnera) was assigned by Schumann to the Plumerieae because of its free ovaries of two carpels. Within that tribe it was included with Amsonia, Vinca, and 18 other genera in the Alstoniinae, on the basis of its numerou ovules per carpel, calyces lacking squamellae, and dry, follicular fruits. Catharanthus is most closely related to Vinca, from which 11 is readily distin- guished by its salverform (vs. funnelform) corollas. These two genera differ conspicuously from Amsonia in having nectaries and mostly opposite (vs. alternate) leaves. The segregation of Catharanthus from Vinca has been widely accepted since Reichenbach first proposed it, some of the notable exceptions being Bailey (but not Bailey ef a/.), Bentham & Hooker, and De Candolle. The genera are clearly distinct morphologically, as is indicated by the numerous differences listed by, for example, Lawrence and Pichon (1949). Although Catharanthus appears to be more advanced than Vinca in vegetative morphology, it may be less ad- vanced in certain floral characters. It is therefore interesting that within the native ranges and in this hemisphere, species of Catharanthus clearly have more tropical as well as more southern distributions than do those of Vinca. This is illustrated by the fact that while C. roseus is a common, naturalized weed in Florida, especially in the south, . minor and V. major are naturalized nowhere in the state (Judd, pers. comm.). Plants of Catharanthus are annual or perennial herbs that do not persist as long as those of Vinca, which usually are more woody. In addition, whereas the leaves of the former are in all cases deciduous, those of Vinca are more likely to be persistent and evergreen, although those of V. major are reportedly deciduous (Woodson, 1938b). The anthers of Catharanthus are completely fertile and free from the style-stigma head, as in most Plumerioideae, while those of Vinca are only partially fertile and converge above (but are not fused with) it. The latter conditions appear to be transitional to those seen in the Apocynoideae, and for this reason Vinca is placed after Catharanthus here. According to Farnsworth, there are few if any anatomical or micromor- phological (e.g., pubescence, cuticle) differences between Catharanthus and Vinca, and those that have been reported are ontogenetic and/or environmental in origin. Nevertheless, Cross & Johnson reported, at least in diploid shoots, tunicas of two and three cell layers in C. roseus and V. minor, respectively, although in each species derivatives of the three outer layers of the shoot apex form the leaf primordia and cortical tissues. Anatomical and ontogenetic in- 366 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 FiGure 2. Catharanthus. a-m, C. roseus: a, distal part of stem with axillary ae buds, and developing fruits (both carpels from each flower developed), x '/2; b, flow bud, corolla lobes overlapping to left (the usual condition, but overlapped to the at in‘‘a’’), x 3;c, upper part of flower cut longitudinally (salverform corolla with subapically widened tube and nonglandular pubescence, anthers free from each other and from style- stigma head, each completely occupied by 2 locules), x 6; d, anther before dehiscence, side view, with short, curved filament, x 10; e, f, anther before dehiscence, abaxial and adaxial sides, x 10; g, apical part of style (terete) and style-stigma head (umbraculiform basally, tufts of nonglandular trichomes at base of cylindrical part and subapically, concealing 2 small or obscure lobes apically), x 12 h, bas ase of lower, after al of 2 calyx lobes (foreground) and corolla, 2 g g 2g 1989] ROSATTI, APOCYNINEAE 367 vestigations of carpel fusion involving the gynoecium above the ovule-bearing region in flowers of C. roseus (Walker, 1975a—c; 1978) are summarized in the family discussion Although the literature regarding alkaloid chemistry 1s enormous (see pri- marily family discussion), Farnsworth contended that a majority of it is actually of little comparative value because of inaccuracies in plant identification, as well as general failures to provide vouchers and to control variations in plant collecting, in the plant parts used, and in the methods used to isolate the compounds. Nevertheless, Paris & Moyse provided paper ae nae ag evidence that distinguished several species of Catharanthus (including C. r seus) from a number of others belonging to Vinca (including V. major, y. minor), findings that Farnsworth later substantiated by both one- and two- dimensional thin-layer chromatography. Janot and colleagues reported so-called iboga-type alkaloids (one of the three major classes they recognized) in Catharanthus (C. roseus) but not Vinca. Farnsworth and co-workers reported that of the 21 alkaloids isolated from species of Catharanthus, only one, aku- ammine, had been detected in species of Vinca as well; however, Aynilian and colleagues later indicated that this number had grown to eight. Farnsworth also compiled a long list of nonalkaloid constituents that had been reported for Catharanthus and Vinca. Among some of the compounds reported in one genus and not the other were formic acid, kaempferol, malvidin, palmitic acid, quercetin, and stearic acid in Catharanthus (from C. roseus), and caffeic acid, p-coumaric acid, fructose, gentisic acid, sucrose, and vanillic acid in Vinca (from V. minor). Nonalkaloids reported for species in both genera included choline, o-protocatechuic acid, tannins, and ursolic acid; saponins had been verified for neither genus (but see below). Daniel & Sabnis later reported from the leaves of C. roseus five phenolic acids (2-hydroxy, 6-methoxy benzoic; p-coumaric; protocatechuic: syringic; and vanillic) that had not been indicated for the species by Farnsworth and substantiated the presence of two flavonoids (kaempferol and quercetin): three additional phenolic acids (ferulic, a-resorcylic, and salicylic) were also reported for the extraregional C. pusillus. According to information compiled by Gibbs, compounds including cyano- genic glycosides and leucoanthocyanins have been confirmed for neither genus, while saponins may be present in Vinca. Tests for tannins have evidently been inconclusive. De Candolle recognized three sections in }7nca, two of which included species that are now assigned to Catharanthus. Pinchon (1949) recognized these two ith 2 ovaries ask eround « and foreground), x 6; 1, base of flower cul in transverse eis through ovaries, showing 5 calyx lobes, corolla nae 2 distinct nectaries (left and right), and 2 distinct, uniloc a ovaries, each with peat placenta and numerous ovules, x 12; j, 3 lobes of persistent calyx, aaa of 1 nectary (between middle and left calyx lobes), 2 nearly mature follicles, x 2; k, immature ae opened along line 2 carpel fusion, with seeds on marginal ieee x 2; 1, mature seed, showing hilum, 12; m, mature seed cut lon ee through hilum on sent (seed coat hatched, ern stippled, embryo unshaded), 368 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 as well as a third to accommodate an additional species, C. scitu/us (Pichon) Pichon. Thus, as recognized by Stearn (1975), the genus is divisible into three sections: sect. CATHARANTHUS (stamens inserted near the top of the corolla tube; follicles greater than 20 mm long; corolla tubes greater than | cm long: plants perennial), with six species; sect. CUPAVEELA (A. DC.) Pichon'> (stamens inserted near the top of the corolla tube; foilicles greater than 20 mm long: corolla tubes less than 1 cm long; plants annuals), with C. pusillus; and sect. ANDROYELLA Pichon (stamens inserted near the middle of the corolla tube: follicles not more than 12 mm long), with C. scitulus. Diploid chromosome numbers have been determined for four species of sect. CATHARANTHUS, including C. roseus, as well as for the monotypic sect. Cupa- VEELA. In all cases 1t is 2” = 16 (Stearn, 1975), except that 27 = 24 has been reported in C. roseus (see Bolkhovskikh et a/.). In addition, colchicine-induced tetraploids of C. roseus have been produced on several occasions, usually in hopes of increasing the levels of economically important alkaloids. Relative to the corresponding diploids, such plants demonstrate overall increased growth rates and have broader leaves, larger stomata, larger, more deeply colored flowers, and larger pollen grains, at least among the viable ones (Janaki Ammal & Bezbaruah). According to data provided by Krishnan and colleagues, alkaloid content does not seem to be appreciably affected by such treatments. Tetra- ploids obtained by Cross & Johnson had larger cells, shorter, more massive stems, thicker, greener leaves, and larger flowers than the corresponding dip- s. The vegetative differences in morphology were attributed to differences in n cell shape in the apical meristem. In addition to its ornamental value, Catharanthus roseus has considerable importance as a source of medicinal compounds (see discussion of suborder) REFERENCES: Under references for the Apocynineae, see ANDERSON; BAILEY; BAILEY et a/.: BENTHAM & HOOKER; BOLKHOVSKIKH ef a/.; DE CANDOLLE; G1BBs; SCHUMANN; and STANDLEY & Under references for the Apocynaceae, see DANIEL & SABNIS; FARNSWORTH; FARNSWORTH et al., NOWICKE; PICHON (1949a); WALKER (1975a—c; 1978), and Woopson (1938b) Under references for inca, see AYNILIAN et al. and JANOT et al. Boke, N. H. Development of the adult shoot apex and floral initiation in Vinca rosea I. Le Jour. Bot. 34: 433-439. 1947. [Inflorescence suggests alternate-leaved ances- tor.] Development of the eaten in Vinca rosea L. Ibid. 35: 413-423. 1948. [Petals completely free in early ontogeny: sepals and petals resemble bud scales more than foliage leaves in See and ontogeny; polypetalous ancestor suggested.] Development of the stamens and carpels in Vinca rosea L. Ibid. 36: 535-547, “The original spelling by De Candolle, “Cupa-vleea,” is evidently an error of orthography or typography. According to Stearn (1975), the epithet is derived from the Malayan kapavila, used for the only species in the section, Catharanthus pusillus, and reportedly rendered as eas veela by Rheede in 1689. The hyphen is here removed in accordance with Article 73.9, ICBN, 1989] ROSATTI, APOCYNINEAE 369 1949. [Mode of initiation and general morphology indicate carpellodes (nectaries) are reduced carpels, suggesting whorl of four carpels in ancestor. Carew, D. P., & R. J. KRUEGER. Anthocyanidins of Catharanthus roseus callus cultures. Phytochemistry 15: 442. 1976. [Anthocyanins, identified as glycosides of hirsutidin, malvidin, and petunidin, matched those found in corollas of intact plants.] Cross, G. L., & T. J. JoHNson. Structural features of the shoot apices of diploid and colchicine- induced. tetraploid strains of Vinca rosea L. Bull. Torrey Bot. Club 68: 618-635. Dwyer, J. D. oe taxonomy of the genera inca, Lochnera, and Catharanthus. Lloydia 27: 282-285. 1964. Guépes, M., & J.-P. Gourret. Architecture des phyllomes végétatifs et périanthaires: cas de Catharanthus roseus (L.) G. Don (Vinca rosea L.). (English summary.) Flora 162: 309-334. 1973. [Illustrations of vasculature of leaves, sepals, and petals, all considered homologous. ] Ho.pen, D. G., & R. D. PARKER. Seed germination in Catharanthus pusillus (Murr.) G. Don and C. ean (Bak.) Pich. ec Hortscience 20: 186. 1985. [Procedures worked out for C. roseus ineffect JANAKI AMMAL, E. K., & H. P. BEZBARUAH. rauced tetraploid . Catharanthus roseus L.) G. Don. Proc. Indian Acad. Sci. B. 57: 339-342. pl. 15. 1963. JouHnson, T. J. Development of the foliage leaves of Vinca rosea. Unpubl. Master’s degree thesis, ae Oklahoma. 1946.* KRISHNAN, R., M. V. CHANDRAVADANA, G. N. MOHANKUMAR, & P. R. RAMACHANDER. Effect of induced autotetraploidy on alkaloid content and root weight in Catha- ranthus roseus (L.) G. Don. Herba Hungar. 24: 43-52. 1985 Lawrence, G. H. M. Vinca and Catharanthus. Baileya 7: 113-119. 1959. Paris, R.-R., & H H. Moyse. Examen en électrophorése sur papier des alkaloides totaux de diverses espéces de pervenches. Compt. Rend. Acad. Sci. Paris 245: 1265-1268. 1957 ParKer, R. D., & J. ViTTI. heals ae cee in the genus Catharanthus (L.) Pich. (Abstract.) Hortscience 20: 186. 1985. [Six species; traits essentially controlled by a single locus include prostrate et recessive), spreading (dominant), and dwarf habits (recessive); corolla pigmentation controlled by a single locus in some species, by au wo or more interacting loci in others. STERN, W. T.. Catharanthus roseus, the correct name for the Madagascar periwinkle. ee 29: 196-200. 1966. . Asynopsis of the genus ye ies \ ). Pp. 9-44 in W. I. TAYLOR & N. R. Farnsworth, eds., The Catharanthus alkaloids: botany, chemistry, phar- pee and clinical use. cc York. 1975. [A very detailed account. Vitti, J. D., & R. D. ParKerR. Inheritance of floral patterning in Catharanthus G. Don. (Abstract. ) Hortscience 20: 186. 1985. [Hybrids between C. roseus and C. longifolius (Pichon) Pichon, both of sect. CATHARANTHUS, evidently obtained. ] 4, Vinca Linnaeus, Sp. Pl. 1: 209. 1753; Gen. Pl. ed. 5. 98. 1754. Terrestrial, [erect,] ascending, or decumbent, perennial herbs, often woody toward base, or subshrubs; axes usually glabrous, leaves with nonglandular pubescence. Leaves opposite; blades subdeltoid, ovate, lanceolate, or elliptic, apices obtuse to acute, rarely mucronate, margins flat or revolute, bases truncate to attenuate, often oblique; petioles often with 2 small, lateral, alternate or subopposite appendages. colleters lacking on bases and adjacent areas of stem. Flowers solitary, axillary; peduncles usually longer than calyces, ebracteolate. Calyx “2-4 as long as corolla, with nonglandular pubescence; lobes narrowly triangular to nearly linear, bases not imbricate, squamellae lacking. Corolla 370 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 funnelform, as long as or longer than leaves, with nonglandular pubescence (adaxially, near juncture of tube and throat); lobes overlapping to left in aes- tivation, spreading, usually dolabriform, entire, oblique; tube and throat with- out adaxial appendages but throat thickened and/or with ridges apically (at juncture with lobes). Stamens inserted at summit of corolla tube, with non- glandular pubescence; filaments very short, terete, geniculate;'® anthers not agglutinated to style-stigma head (but curved over it), each narrowly obov (apex oe to subcaudate, bases attenuate), incompletely fertile ee adaxially 2 locules at about middle), introrsely dehiscent, long axis mostly perpendicular to that of filament. Nectaries 2, alternating with carpels and widely spaced, usually not exceeding ovaries. Carpels completely free in ovule- bearing region, united above; ovaries 2, simple, distinct, unilocular, with mar- ginal placentae and numerous ovules; style terete, gradually widened apically; style-stigma head slightly conical, basally umbraculiform, lobes not evident apically, with a tuft of hairs apically and at base of conical part. Fruits follicular, usually 2 from each flower (the pair divergent less than 90°), widened around seeds; surface striate, glabrous; dehiscence tardy, by an adaxial suture. Seeds usually single, ovoid and deeply grooved on funicular side, without trichomes. LECTOTYPE SPECIES: V. major L.; see Britton & Brown, Illus. Fl. No. U.S. & Canada, ed. 2. 3: 20. 1913. Vinca minor L. was later selected by M. L. Green (in Hitchcock & Green, Prop. Brit. Bot. 136. 1929) because it was considered to be the better known of the two original species, the other being V. mayor, and because it, and evidently not V. major, was included where the generic name was applied in Hortus Cliffortianus and Hortus Upsaliensis. Vinca minor was also selected by Woodson (N. Am. FI. 29: 125. 1938). (Vinca indicated by Linnaeus, Philosophia Botanica, 150. 1751, as a name derived from Pervinca of Tournefort, which is evidently also reflected in colloquial Italian (pervinca), French (pervenche), German (Berwinkel, Barwinkel), and English. According to Stearn, the name vincapervinca, which was used in reference to these plants by Pliny in the first century A.p., may have been derived from Latin vincio, to bind or wind about, in reference to the slender and flexible shoots used in wreath making, and per, through, because the leaves of some species remain green throughout the year.)— PERIWINKLE. A genus of about six or seven species (Stearn) of subshrubs or suffrutescent herbs with opposite leaves, conspicuous flowers, and follicular fruits. All but one of these are native to an area from France, Spain, Portugal, Morocco, and Algeria to the Caucasus region and northern Iran; Vinca erecta Regel & Schmalh. is geographically and perhaps taxonomically isolated in Afghanistan and neigh- boring areas (see below). Two of the species commonly cultivated throughout the world, TV. #inor L. and V. mayor L., are escaped and naturalized in the Western Hemisphere, including parts of the southeastern United States. A third species, 1. herbacea Waldst. & Kit., is also cultivated in North America and An equally likely interpretation 1s that the filaments are straight and that the stamens distal from the points of geniculation are entirely antherial. Anatomical or other evidence evaluating these alternatives appears to be lacking. 1989] ROSATTI, APOCY NINEAE 371 at least has the potential to escape in our area; Fernald reported it out of cultivation in Massachusetts Vinca is most closely related to Catharanthus, and many aspects of its chem- istry, anatomy, morphology, geography, and systematics are considered in the discussion of that genus Plants belonging to Vinca minor, 2n = 46 (32 and 92 also reported), are usually smaller in all respects and less erect than those of V. major, 2n = 92 (16 also reported). In addition, the leaves of V. minor are coriaceous and evergreen (vs. membranaceous and deciduous or semi-evergreen), tapered (vs. rounded, truncate, or subcordate) at the base, and glabrous (vs. ciliate) at the margin; the calyx lobes are also glabrous instead of ciliate. In V. herbacea, 2n = 46 (92 also reported), and V’. erecta the secondary veins diverge from the midrib of the leaf at angles of 5—30°, while in all other species this angle is 40- 50°. In addition, the leaves of V. herbacea are distinguished from those of other species of Vinca by their narrowly elliptic to lanceolate (vs. wider) shape and more subtle venation. The natural distribution of Vinca minor and V. major (and probably other species as well) has been obscured, perhaps permanently, because of a long history of cultivation in the Old World, where plantings in graveyards and probably other places date back at least to the fourteenth century (see Stearn). Nevertheless, in 1753 Linnaeus indicated a more northern distribution for V. minor (“habitat in Germania, Anglia, Gallia’) than for V. major (“habitat in Gallia Narbonensi, Hispania’), a difference that currently persists in Europe and is also reflected in this hemisphere. The hardiest and most commonly cultivated species in the genus, Vinca minor is escaped and naturalized in much of the North Temperate Zone, including an area in North America bounded by Nova Scotia, Minnesota, Georgia, Arkansas, and Kansas (Steyermark). It occurs out of cultivation but rarely fruits in Texas (Correll & Johnston) and has been listed as escaped but evidently not naturalized in Louisiana (MacRoberts) and eastern Tennessee (Sharp et a/.). According to Radford and colleagues, it occurs in scattered localities in North Carolina (but not South Carolina), where 1t reportedly spreads from roadsides and waste places to fields and woods. The species was not reported for Guatemala by Standley & Williams, and although Nowicke in- dicated that it is cultivated in Panama. she only conjectured about its escape fHeke: Both Vinca minor and V. major are reportedly naturalized in Georgia (Jones & Coile), although neither is so established in Florida (Judd, pers. comm.). While V. minor is a frequent escape in the northeastern United States, V. major is only occasionally found outside of cultivation, in the southern part of the country to as far north as Virginia (Gleason & Cronquist). Vinca major evi- dently does not grow in Panama (see Nowicke) but is thoroughly naturalized in Guatemala, where in the western highlands it blooms profusely ea the cold dry season when flowers in general are scarce (Standley & Willia Vinca herbacea is also grown in North America and is reportedly aiicned outside of cultivation in Massachusetts (Fernald). Nevertheless, I have seen no evidence that it escapes from gardens in the Southeast. The species is native 372 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 FiGureE 3. Vinea. a-m, I. minor: a, distal part of stem with axillary flower (corolla lobes ‘dolabsiform), x 1: b, flower bud, corolla lobes overlapping to left, but otherwise as in ‘“‘n” (seed coat hatched, endosperm stippled, embryo unmarked), x 10 396 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 is only accomplished after those relatively rare occasions on which trapped insects exert enough energy in attempting to escape that the anther cone 1s disrupted. Despite Woodson’s (1930) belief that pollination in Apocynum is not under- stood, at least two accounts (Knuth; Legget, 1872a—c, 1873a, b), based primarily on observations of 4. androsaemifolium, seem to have some merit. The fol- lowing is based upon a combination of these and my own observations. Nectar guides in the form of pink stripes extend from the sinuses between corolla lobes to the base of the corolla tube, highlighting five openings to spaces at the base of the flower, between the nectaries. Each opening ts defined by the base of one anther and the corolla tube from one of its inwardly projecting appendages to the next. The two auricles at the base of each anther are curved toward each other. They and the appendages of the corolla are shaped and oriented in such a way that while an insect part such as a proboscis or a leg could freely pass through the opening in a downward direction to reach the nectar, 1t would likely be funneled during withdrawal by one half of one ap- pendage and the nearest auricle into the seam between that anther and the next, which becomes increasingly narrow from its base to its summit. That the insect parts in question thus become tightly wedged between the anthers is indicated by the frequency with which they are left behind by departing insects. A visitor that leaves intact would probably have pulled its inwedged part along the seam between adjacent anthers until the widest part of the clavuncle was reached (the point at which the clavuncle and the stamens are agglutinated). Thus, scraping from three sides ultimately would have been effected during the insect’s attempts to free the part and most, if not all, of any pollen that might have been on it would have been deposited on or near the clavuncle immediately beneath the widest part, suggesting that this is the most effective location for the stigmatic surface. The struggle for freedom would usually be sufficient to force apart the anther cone, allowing pollen in the space above the clavuncle to be released, deposited on the visitor, and transferred to another flower. Itis perhaps significant that the nectaries are located beneath the corollar appendages, where they are somewhat protected from the activities described above. Several fruit and seed characters were thought by Woodson (1930) to be taxonomically important. He considered erect follicles to be more primitive than pendulous ones because they occur only in Apocynum pumilum (Gray) Greene, in which the inflorescence/infructescence in general was considered primitive. He also used follicle length and shape (straight or curved), as well as seed and coma length, to separate taxa. Woodson’s (1930) key divided the species of Apocynum into four groups based on leaf orientation (ascending, spreading, or drooping) and the relative lengths of calyces and corollas. Although the correspondence between these divisions and four sections proposed by Béguinot & Belosersky is close, Wood- son (1930, p. 70) contended that recognition of such infrageneric taxa was both unnecessary and artificial because “‘the species are so confluent as to make sectional lines extremely obscure.” It is therefore not entirely clear why he chose to recognize seven species instead of one, except that he was a strong 1989] ROSATTI, APOCYNINEAE oof proponent of the idea that hybridization is responsible for most of the inter- gradation between species. One of the four species and most of the numerous varieties recognized by Woodson (1930) in eastern North America, as well as 12 of the 21 species and varieties he recognized in the entire genus, were thought by him to have resulted from hybridization. Numerous examples of fruit-bearing, putatively parental populations adjacent to and/or intermingled with sterile, presumably hybrid populations were cited in support of this view. For example, although Apoc- yaum androsaemifolium, A. cannabinum, A. sibiricum (cited as A. hyperici- folium Aiton), and A. medium were all abundant in a field near South Bend, Indiana, during the summer of 1928, only plants assignable to the first three species later developed fruits; 4. medium, then presumed to bea hybrid between A. androsaemifolium and A cannabinum, was completely sterile in this sense. Woodson (1930) also cited a locality in Colorado at which fruiting clones of 4. androsaemifolium var. incanum A. DC. were intermingled with nonfruiting individuals of 4. medium var. floribundum (Greene) Woodson. One small plant with the leaves of the former and flowers quantitatively intermediate between the two was discovered in 1929. The plant was named 4. androsaemifolium var. intermedium Woodson even though, evidently, no other individuals had been discovered. Whether or not that single plant ever set fruit was not men- tioned. Despite the failure of many plants of presumed hybrid origin to set fruit, those of, for example, the common and typical eastern variety of A. medium frequently do (although those of the more restricted western varieties evidently do not). Woodson (1930) was unable to support his presumptions of hybridity with observations of meiosis in pollen mother cells because of the small size of the anthers involved and because of complications posed by the latex. However, meiotic irregularities were often indicated by individual pollen tetrads in which some of the grains were abortive and relatively deficient in nuclear material, while others included more than the usual complement. He reported that, although all taxa have some sterile pollen, percentages were lowest in those not considered to be of hybrid origin. In what has become one of the truly classic studies in plant biology, Anderson experimentally evaluated what he termed (p. 159) the “rather extreme position” of Woodson concerning the frequency of hybridization in Apocynum. Seeds from specimens determined by Woodson to belong to 4. androsaemifolium, 4. cannabinum, and A. medium were germinated, and the resulting plants were grown for about a year and a half. Pollen fertility among progeny of the pre- sumed parental species was generally higher than that among offspring prop- agated from the supposed hybrid: the average percentage among seven seedlings propagated from one individual assigned to 4. cannabinum was 95.1, with values ranging from 91 to 98; for 4. androsaemifolium the figures for 11 seed- lings from one individual were 80.9 and 42 to 99, respectively; and for A. medium the overall figures for 25 seedlings of five individuals were 64.9 and 4 to 99. On the basis of identifications provided by Woodson, who was unaware of the experiment and therefore the origin of the specimens, both of the pu- tatively parental species bred true in the sense that all of the progeny were 398 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 assigned to the same species as the plants from which they were propagated; however, both glabrous and pubescent varieties were said to have been seg- regated from each. Of 25 specimens representing the progeny of what he had eee as A. medium, Woodson (1930) indicated that only 18 could still be ned to that taxon (14 definitely, four questionably), five to A. androsae- mifolium (three definitely), and two to 4. cannabinum (one definitely). It is also notable that all of the specimens Woodson considered to be of questionable identity had been propagated from A. medium. Anderson concluded that although 4pocynum medium is a hybrid between A. androsaemifolium and A. cannabinum, its taxonomic status, at least in an evolutionary context, is not comparable to that of the parental species. Indi- viduals assignable to this entity appear to comprise a spectrum of cases, each one or more generations removed from the original cross (i.e., an original cross) between the species, in which some are sterile, others display high levels of heterozygosity, and still others breed nearly true. He also implied that while hybridization is a frequent occurrence throughout the genus, it is responsible for fewer taxonomic problems in western than in eastern North America. He considered increased variation in species that were originally more distinct to be one of the most significant effects of introgression in the genus. The conclusions of Woodson (1930) and Anderson regarding hybridization in Apocynum appear to be generally accepted at present. However, it is notable that they have remained unsubstantiated by, for example, chemical and cy- tological evidence, despite considerable advances in and increased taxonomic uses of the techniques involved. Since publication of Anderson’s work in 1936. chromosome numbers have been reported for 4. androsaemifolium (2n = 16, 22), A. cannabinum (2n = 16, 22), and A. sibiricum var. cordigerum (Greene) Fern. (2” = 22) (see Van der Laan & Arends) Gleason (New Britton & Brown Illus. Fl. No. U. S. & Canada 3: 72. 1963) and a majority of his contemporaries, according to him, thought that plants assignable to Apocynum medium represented hybrids between A. androsae- mifolium and either of two other species he considered well marked, 4. can- nabinum and A. sibiricum. Gleason indicated that a plant should be referred to . medium if it did not agree with at least six of the seven elements he used to distinguish either 4. androsaemifolium or, collectively, A. cannabinum and A, sibiricum. Gleason also thought that the extreme variation in 4. cannabinum was due to hybridization with A. sibiricum, and perhaps with another, evidently unnamed, densely pubescent species as wel While it is clear that Apocynum androsaemifolium, A. cannabinum, and various hybrids between them (including those assignable to A. medium, as well as to varieties of this and the parental species) are represented in the southeastern United States, the occurrence there of 4. sibiricum is questionable. The purer forms of 4. androsaemifolium (corolla pink or striped with pink, the tube at least as long as the calyx, the lobes spreading; leaves drooping) and A. cannabinum (corolla whitish or greenish, the tube about as long as the calyx, but sometimes shorter or longer, the lobes erect or nearly so; leaves ascending or only slightly spreading), as well as plants influenced by the genomes of other species, are all widespread in North America and the Southeast. 1989] ROSATTI, APOCYNINEAE 399 Plants assignable to Apocynum androsaemifolium appear not to occur in Florida (see, for example, Clewell; Correll & Johnston; Small, 1913, 1933) and generally to be scarcer in the more southern parts of our area. According to Hitchcock and colleagues (p. 79), the species occurs throughout much of Canada and the United States, except for ‘‘the s.e.”’ Individuals of this species with pubescent abaxial leaf surfaces and perhaps others with glabrous ones may occur in the Southeast, although the latter are more representative of popu- lations in the western United States.'? Apocynum cannabinum is represented in our area by two pubescence-based variants that, perhaps like the foregoing, may be of questionable significance. What has been called var. pubescens (Mitch.) . DC. (many parts of the plants hairy) is mostly restricted to eastern North Asnerics, while the typical element (plants glabrous throughout or only sparsely hairy in places) is not. Plants assignable to A. medium may be more common in the eastern two-thirds of the continent, although they occur as far west as California (Munz) and the Pacific Northwest (Hitchcock et al.). Apocynum sibiricum (A. hypericifolium), at least in its purer forms, differs most conspicuously from other members of the genus that may be found in the Southeast in its sessile or nearly sessile leaves. Variants with deeply cordate leaf blades, the bases of which often clasp the stem, and corollas about as long as broad have been referred to var. cordigerum (Greene) Fern., and those with very narrow leaves and corollas longer than broad would qualify as var. sa- lignum (Greene) Fern. Evidently, the typical element includes plants that are glabrous and have oblong to narrowly ovate leaves and corollas about as long as broad; pubescent plants otherwise fitting this description have been referred to var. Farwellii (Greene) Woodson. While most authors have not indicated sibiricum for our area in particular, some (e.g., Correll & Correll; Correll & Johnston) have stated that it occurs throughout most of North America. Al- though Steyermark indicated that in Missouri it occurs as far south as Texas County, the southernmost border of which is within 50 miles of Arkansas, and Harvill and colleagues reported it for Isle of Wight County in Virginia, I have seen no specimens from the Southeast that are assignable to this species. Species of Apocynum occupy a variety of habitats from sea level to the subalpine zone, including saline and fresh-water beaches and shores, cultivated and fallow fields, open woods, roadsides, railsides, prairies, and dunes. In the Carolinas 4. androsaemifolium appears to be confined to the mountains o North Carolina (Radford et a/.). At least in Missouri, plants of = stbiricum reportedly occur in moister habitats than do those of other sp (Steyermark), “Determination of the typical element of Apocrnum androsae deca was addressed by Woodson (1932). He considered the variant of eastern North America to be typical, primarily because Linnaev (Sp. PL. 1: 213. 1753) wrote in the protologue “habitat in Virginia, Canada.” Although Linnacus a indicated at that time “foliis ovatis utrinque glabris” and most eastern plants are pubescent, the wichontes are generally evident only with the aid of a hand lens, and according to Woodson (p. 31) ‘glabrous” at a st was a SUN term of not too ‘en cpa Thus, specimens considered by Li vy have bee 1. If Woodson’s — correct, the correct name for the glabrous variety of w western Sot ied ica would be var. glabruim Macoun, and most if not all plants in the east would be assignable to var. ee Further resolution of this problem appears to be lacking. 400 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 although in Ohio individuals assignable to this species were considered to be xerophytic (Schaffner). Other ecological differences among species of Apocynum are not obvious. Apocynum cannabinum is often a serious weed in, for example, crop fields of corn, sorghum, and soybean. According to Schultz & Burnside, yields from irrigated plantings of corn were reduced very little by the presence of this species, but those from unirrigated corn and sorghum were reduced by as much as 15 and 30 percent, respectively. Apocynum has had a long and varied history of use by humans, most of which has been reviewed by Woodson (1930). The fibrous bark of several species, but most notably that of 4. cannabinum, Indian hemp, has been and continues to be a source of thread and cord for the construction of such articles as clothing, sails, fishnets, and rope. Studies during the late nineteenth and early twentieth centuries suggested that the fiber could be profitably used to make paper, although nothing much seems to have been accomplished along these lines since then. Dried and usually pulverized roots and rhizomes of A. cannabinum have a stimulating effect on heart function (undoubtedly due to e presence of cardiac glycosides; see discussion of the suborder) and were used by early settlers in the treatment of dropsy and as a diurectic (Uphof), while those of most species enhance the movement of a variety of bodily substances in humans. According to Chittenden, the roots of 4. androsaemi- folium and A. cannabinum are poisonous and differ in their medicinal effects. Reports published in the early part of this century indicate that latex from species of Apocynum (and Asclepias) was a better potential source of rubber than was that of any of the many other species investigated and that the production of both latex and fiber was optimal in plants grown on arid land generally unfit for most other crops. Fields in the western United States were, and perhaps still are, allowed to become dominated by species of Apocynum because of the nectar they provide for honeybees. REFERENCES: Under references for the Apocynineac, see CHITTENDEN; CLEWELL; hen & CORRELL; CorRELL & JOHNSTON; DEMETER; HARVILL et al.; HITCHCOCK et al.; KNUTH; MARKGRAF (in HeEywoop),; Munz; Onwt; PoBEDIMOVA; RADFORD ef al.; SCHUMANN: ae (1913, 1933): STEYERMARK; TSIANG & Lr; and UpnHor. Under references for the Apocynaceae, see VAN DER LAAN & ARENDS; LEGGET (1872a- 1873a, b); ScHick (1980, 1982); and Woopson (1930). ANDERSON, E. An Seige study of eros in the genus Apocynum. Ann. Missouri Bot. Gard. 23: 159-168. pl. / 36. BALLARD, C. W. eae and rhatmacogtion of the genus Apocynum. Proc. Int Congr. Pl. Sci. 2: 1406-1412. 1926 ee 9]. [According to Woopson (1930), chat acters of the clavuncle incorrectly us BEGUINOT, A., & N. BELOSERSKY. Revisione regs del genere . pea Linn. Atti Accad. Lincei Mem. Cl. Sci. Fis. V. 9: 595-734. pls. 1-12, figs l E._uiotr, C. L. Qu ees analysis of pika (Ochotona ea hay oe in central Idaho. Northwest Sci. 54: 207-209. 1980. [4. androsaemifolium comprised 11 per- cent by weight of the food piles this mammal stored for winter, second to Smilacina stellata (L.) Desf. (70.5 percent). — 1989] ROSATTI, APOCY NINEAE 401 Fowter, W. B., & A. R. TIEDEMANN. Phenological relationships of ee betulifolia Pall. and. pe apn Srenriny | vet L. Northwest Sci. 54: 17-25. 1980. [In eastern Washington, transplants of 4. slaves clara showed Deed ee pheno- logical differences saeas with eleva Frye, T. C., & E. B. BLopcetr. A contribution to the life history of Apocynum andro- sia Bot. Gaz. 40: 49-53. pl. 2. 1905. [Helpful illustrations of anther clavuncle connections and of gradation from bilateral to the less-common aac te ioe many embryological details. ] SCHAFFNER, J. H. Xerophytic adaptations of Apocynum hypericifolium. Ohio Nat. 10: 184, 185. 1910. [At time of writing. the species was known in Ohio only from Cedar Point, now mostly develope ScHULTz, M. E., & O. C. BuRNsIDE. Distribution, competition, and phenology of hemp dogbane (Apocynum cannabinum) in a oes Weed Sci. 27: 565-570. 1979. [Emergence from root crowns at 17-19°C; flower buds formed four to seven weeks later; seeds first viable ten weeks after eee Woopson, R. E., Jr. The ea and nomenclature of Apocynum androsaemifolium L. Rhodora 34: 30, 31. 1932 [Asclepiadaceae to follow] CRINS, TAMARICACEAE 403 THE TAMARICACEAE IN THE SOUTHEASTERN UNITED STATES'! WILLIAM J. CRINS? TAMARICACEAE Link, Enum. Hort. Berol. 1: 291. 1821, ‘““Tamariscinae,” nom. cons. (TAMARISK FAMILY) Small trees or shrubs [subshrubs] of temperate and subtropical deserts, steppes, and shorelines, on sands, clays, and saline soils. Wood ring-porous or semi- ring-porous; sieve-element plastids of the S-type. Branches slender, persistent or deciduous. Root systems variable, ranging from taproots with relatively short lateral roots to extensively branched, horizontally spreading roots (branches when buried often also producing adventitious roots). Leaves scalelike or small, alternate, simple, entire, usually sessile [rarely subpetiolate], exstipulate, sub- imbricate, [amplexicaul in some species of Tamarix], often fleshy or punctate, with salt-secreting glands, usually glaucescent, with sclereids consistently pres- ent at or near vein endings; stomata anomocytic, on both surfaces of leaves repared for the Generic Fl fthe Sout! United States, a long-term project made possible by grants from the Natio nai Science aes and at this writing supported by BSR-8716834 (Norton G. Miller, principal investigator). under which this account was prepared, and BSR-8717333 (Carroll E. Wood. Jr., principal investigator). This treatment, 129th in the series, follows the format established in the first paper (Jour. Arnold Arb. 39: 296-346. 1958) and continued to the present. The area covered by the Generic Flora includes North and ak Carolina, Georgia, Florida, nessee, Alabama, Mississippi. Arkansas, and Louisiana. The descriptions are ba sed primarily on the plants of this area, with information about extraregional members of a family or genus in brackets |]. The reference I did not verify is marked with an asterisk. I have enjoyed working with Norton Miller and Carroll Wood on the Generic Flora project, a I thank them for their interest and advice. Ihsan A. Al-Shehbaz and Barbara Nimblett have nln frequently. Stephen A. Spongberg and Elizabeth B. Schmidt improved a final manuscript with thei editorial precision and expertise. Thanks are extended to the staffs of the New York State Mee (especially Alta Beach, Senior Librarian) and the Botany Libraries of oe ard University (especially Geraldine Kaye. Librarian) for providing many references. 1 thank the curators of the following herbaria, who have sent eee or provided access to collections and ho spitality during my visits: A. ALU, DAO, F, FLAS, FSU, GH, LSU, LTU, NCSC, NCU, NO, NY, NYS, UARK, UNA, and USF. ae vcr on of Tamarix was prepared by Karen Stoutsenberge from dissections by Kenneth R. Robertson and Carroll Wood from material collected 2 _ E. Umber (6/4, GH; along the Rio Grande west of Mountainair, Socorro Co.. New Mexico), R. L. Wilbur (7/73, Gu; Carteret Co., North Carolina). and D. S. Correll and H. B. Correll (9599, Gu: . ameron Parish, Louisiana). This treatment is published as contribution number 604 from the oe York a Science Service. Biological Survey. New York State Museum, The State Education Department, Albany, New York © President and Fellows of Harvard College, 1989. Journal of the Arnold Arboretum 70: 403-425. July, 1989. 404 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 [only on abaxial surface in Myricaria germanica]. Inflorescences spicate or racemiform, often aggregated in terminal panicles [or flowers solitary in Hol- olachna and Reaumuria]. Flowers perfect [2 species of Tamarix dioecious], regular, small, hypogynous, 4- or 5- [6-]merous; pedicels subtended by small, scalelike bracts. Sepals 4 or 5 [6], free or slightly connate, persistent, imbricate. Petals 4 or 5 [6], alternating with and larger than sepals, free, sometimes persistent [with a pair of ligular appendages near base in Reaumuria], inserted at base of a fleshy or membranaceous disc, imbricate. Stamens 4 or 5 [to 10 in Tamarix, to 75 in Reaumuria], usually haplostemonous or obdiplostemon- ous, usually free [monadelphous 1n Myricaria, fasciculate in 5 bundles in Reau- muria], filaments filiform: anthers tetrasporangiate, 2-locular, versatile, ex- trorse in Tamarix [introrse in Myricaria], longitudinally dehiscent, slightly longer than wide; pollen (2-) 3- (4-)colpate, simple or in tetrads, binucleate (in Tamarix) [trinucleate in Myricaria]. Gynoecium of (2) 3 or 4 (5) carpels united in a compound, superior, unilocular ovary; placentation parietal, placentae sometimes intruded, making the ovary appear plurilocular near base and apex, or reduced, making placentation appear basal: styles 3[-5], distinct or connate at base; stigmas apical or decurrent adaxially [sessile in Myricaria]; ovules 2 to many on each placenta, with short funiculi, erect, anatropous, bitegmic, crassinucellar. Fruit a loculicidal, pyramidal capsule, the 3[-5] valves placen- tiferous at base, coriaceous; seeds long-pilose over entire surface [Reaumurieae] or at apex only, with unicellular hairs, which may be prolonged into a con- spicuous hygroscopic coma (Tamariceae), micropyle formed by endostome; endosperm with development nuclear [cellular in Tamarix tetrandra], degen- erating quickly in Tamariceae [persistent in Reaumurieae]; cotyledons 2, ob- long, obtuse at apex, flat or plano-convex, the radicle short, conical, inferior: megagametophyte (embryo sac) tetrasporic, but postmeiotic development vari- able, usually Fritillaria or Adoxa type, embryogenesis of Solanad type, with broad, massive suspensor formed, polyembryony occasional. Base chromo- some number 12. (Including Reaumuriaceae Ehrenberg.) Type GENUS: Tamarix Linnaeus. A relatively small family of five genera (ffololachna Ehrenb., Myricaria Desv., Myrtama Ovez. & Kinz., Reaumuria L., and Tamarix L.) and about 100 species, occurring primarily in central Asia, the Middle East, northern Africa, and the Mediterranean region of Europe. Several species of Tamarix have been introduced into North America, and six of these have become naturalized in the Southeast.’ The Tamaricaceae are one of numerous families that have been placed in the catchall order Parietales. Other early suggestions of relationship included the Guttiferales (Endlicher; Grisebach; Hallier, 1911), Rutiflorae (Caruel), and ‘Myricaria germanica (L.) Desv. has been cultivated in the Southeast. The following key will allow its identification in the event that it is discovered as an escape Filaments free, inserted on a disc: anthers extrorse; our species with 4 or 5 stamens. ... Tamarix. Filaments fused into a membranaceous tube (monadelphous); anthers introrse; stamens 10. ...... Sascha Aan iio at ae atta, Bs aah tepar eee ant ean ahd hace dict de ee ener lor Myricaria. 1989] CRINS, TAMARICACEAE 405 Centrospermae or its segregates (Adanson: Bessey; De Candolle; Hallier, 1903; Jussieu; Saint-Hilaire; Vesque). In most recent ordinal classifications, the families of the Parietales have been redistributed, and the Tamaricaceae are now generally placed in the small order Tamaricales, along with the Frankeniaceae (Thorne, 1976) and/or Fouquieri- aceae (Barkley; Behnke: Bentham & Hooker; Boivin; Hutchinson, 1964; Takh- tajan, 1980), or in the larger order Violales (Cronquist, 1981, 1988; Gibbs: Lindley, 1853; Melchior; Thorne, 1983). These hypotheses, based largely on reproductive features, are also supported by the anatomically similar salt glands of the Frankeniaceae and Tamaricaceae (Solereder) and by the production of D-pinitol by members of both families (Plouvier). A recent novel phylogenetic scheme has the Tamaricaceae in the order Primulales (Goldberg), but this placement is virtually unjustified, except on the basis of such ubiquitous char- acter states as simple perforation plates in xylem vessels, lack of stipules, perfect flowers, superior, unilocular ovaries, and the absence of alkaloids. None of the ordinal characters listed by Goldberg would refute the placement of the Tam- aricaceae in the Violales or Tamaricales, and in fact, his description of Prim- ulales requires several qualifiers because of the unusual (for Primulales) char- acter states found in the Tamaricaceae. A cladistic analysis utilizing existing data, or the examination of new lines of evidence, may be required to shed further light on the relationships of this family and, indeed, all of the families of the catchall Parietales. The basis for the uncertainty in ordinal placement of the Tamaricaceae 1s the inclusion of two groups of genera (comprising the two tribes Reaumurieae Bentham & Hooker (Hololachna, Reaumuria) and Tamariceae (Myricaria, Myrtama, Tamarix)*) within the family, which at times (e.g., Lindley, 1833, 1853; Maout & Decaisne) have been interpreted as being unrelated. This hy- pothesized polyphyly is probably based on a misunderstanding of certain char- acter states and their ontogeny. The tribes have been differentiated on the basis of inflorescence structure, distribution of pubescence on the seed, and presence or absence of endosperm in the seed. Developmental studies of the seed have shown that members of the tribe Tamariceae produce cellular (or initially nuclear in Jamarix tetrandra Pallas) endosperm, which disappears as the em- bryo grows to occupy all of the space in the embryo sac (Davis; Puri; Sharma). Thus, the Tamariceae merely show reduction in the development of the en- dosperm, a trend that correlates with other advanced character states exhibited by the tribe, including tendencies toward elaboration of the hygroscopic hairs on the seed, localization of hairs on the seed, reduction in flower size, increasing complexity of inflorescence architecture, elaboration of the disc, and tetramery in some members. Members of the Tamaricaceae are generally shrubs or small trees of deserts, occur along mountain streams. The branches are covered with small, often scalelike leaves that are punctate with salt-secreting glands. The four- or five- “Several authors (¢.g.. Bentham & Hooker) have included a third tribe, Fouquierieae, which 1s generally accorded family status. Its ordinal placement is as unsettled as that of the Tamaricaceac. 406 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 [six-]merous, white, pink, or even reddish flowers are quite small but are aggregated into panicles of racemes or spikes that provide spectacular displays. The capsules contain numerous small seeds that may be hairy over the entire surface (tribe Reaumurieae) or have long hairs at the chalazal end forming a conspicuous coma that aids in dispersal (tribe Tamariceae). Fossils of the Tamaricaceae are very few. Tamaricoxylon africanum (Krausel) Boureau, based on fossil wood, is identical in anatomy to modern 7amarix (Boureau). The fossils are from northern African Quaternary deposits and thus shed no light on the age of the family. An attribution of Oligocene age to one of these specimens is considered doubtful by Boureau. REFERENCES: Apanson, M. Familles des plantes. 2 vols. 640 pp. Paris. 1763-1764. [Tamarix, 2: 266- 269, placed in “XX XVII. Famille. Les Amarantes, Amaranthi”’ —1.¢., the Amaran- thaceac.] AGARDH, J. G. Theoria systematis plantarum. Ixxx + 400 pp. Lund. 1858. [Tamaris- cineae, 342, 343; Reaumuriaceae, BAILLON, H. Tamaricacées. Hist. Pl. 9: 236- 245. 1888. [Tamarix (Myricaria) germanica L., 7. tetrandra, and Reaumuria hypericoides Willd. illustrated.] BARKLEY, F. A. Clave descriptiva de los ordenes de las Anthophyta. Revista Fac. Nac. Agron. Medellin Univ. Antioquia 9(33): 27-69. 1949. [Tamaricaceae, Fouquiteri- aceae, and Malesherbiaceae in Tamaricales.] BATTAGLIA, E. Contributo all’embriologia delle Tamaricaceae. Nuovo Giorn. Bot. Ital. Il. 48: 575-612. 1941. [Variability in embryo-sac development in Tamarix gallica L. and T. africana Poiret.] ene a R. The genus Tamarix. ix + 209 pp. Jerusalem. 1978. [Worldwide revision.] BEHN H.-D. Sieve-element plastids of Fouquieria, Frankenia (Tamaricales), and Rha becile ndron Megas ee sometimes allied with Centrospermae (Caryophyl- lales). Taxon 25: 265-268. 1976. [Transmission electron micrographs of S-type sleve- oo aie ae icarla germanica and Tamarix juniperina Bunge, 266, figs. 3 BELYAEVA, _ S. Embryonic processes in —— Kotschyi Bunge (Tamaricaceae). (In Russian; English summary.) Izv. Akad. Nauk Turkmensk. SSR, Ser. Biol. Nauk 1983(3): 3-11. 1983. [Variability in ie ee development Benson, L. Plant classification. xiv + 688 pp. Boston. 1957. [Tamaricales, Tamarica- ceae, 149, 150.] BENTHAM, G., & J. D. Hooker. Tamariscineae. Gen. Pl. 1: 159-162. 1867. ] Bossier, E. Flora orientalis. Vol. : Thalamiflorae. xxxiv + 1017 pp. Basel. 1867. nee ae 758-779, two oe recognized. Borvin, B. Les familles de trachéophytes. Bull. Soc. Bot. France 103: 490-505. 1956. [Tamaricales contains Tamaricaceae and Fouquieriaceae.] BOLKHOVSKIKH, Z., V. GriF, T. MATVEJEVA, & O. ZAKHARYEVA. Chromosome numbers of flowering plants. A. A. Feperov, ed. (Russian and English prefaces.) 926 pp. Leningrad. 1969. [All members of the family 2” = 24] Boureau, E. Sur l’anatomie comparée et les pie ad see fossiles de ne ricaceae, découverts en Somalie fran¢aise et en Mauritanie. Bull. Mus. Hist. Nat. Paris. II. 23: 462-469. on [Tamar ag described as new fossil genus Ps ac- commodate African specim BREWBAKER, J. L. The distribution and phylogenetic significance of binucleate and trinucleate pollen grains in the angiosperms. Am. Jour. Bot. 54: 1069-1083. 1967. 1989] CRINS, TAMARICACEAE 407 BRUNNER, C. Beitrage zur vergleichenden Anatomie der Tamaricaceen. Jahrb. Hamburg. Wiss. Anst. 26: 89-162. 1909. [Detailed anatomical descriptions of many taxa a.] Brunswik, H. Uber das Vorkommen von Gipskrystallen baa den Hamancaceae,, Sitz- ber. Akad. Wiss. Wien 129: 115-136. 1920. [ um in all genera of Tamaricaceae, but not in Fouquieria. ] CANDOLLE, A. P. DE. Tamariscineae. Prodr. Syst. Nat. 3: 95-98. ture ee] Corner, E. J. H. The seeds of dicotyledons. Vol. 1. x + 311 pp. Cambridge, England. 1976. [Tamaricaceae, 265. Cronouist, A. An integrated system of classification of flowering plants. 1262 pp. New York. 1981. [Violales, including Tamaricaceae, 404-406. _ The evolution and classification of flowering plants. vii + 555 pp. New York. 1988. [Violales, 339-345.] Davis, G. L. Systematic erbnyo pe) of the angiosperms. viii + 528 pp. New York. 1966. [Tamaricaceae, 255, 256. DesvaAux, N. A. Sur la nouvelle famille de plantes fondée sur le genre Tamarix. Ann. Sci. Nat. 4: 344-350. 1825. [Myricaria described, 349; detailed descriptions of family EHRENBERG, C. G. Observations sur la famille des Tamariscinées, et sur la manne du tamarisque du mont Sinai. Ann. Sci. Nat. 12: 68-78. 1827a. [Hololachna and sub- genera of Tamarix oe manna obtained from 7. gallica L. var. mannifera Ehrenb.] Ueber die Manna-Tamariske nebst allgemeinen Bemerkungen tber die Tam- ariscineen. Linnaea 2: 241-282. 1827b. [Diagnostic characters for genera and sub- genera; discussion of derivation of generic name Tamarix.] E1cuer, A. W. Bliithendiagramme. Vol. 2. xx + 575 pp . 1878. Senne ie 243, 244, fig. 97; floral diagrams of Myricaria germanica, Tamarix tetran ENDLICHER, S. Genera plantarum secundum ordines naturales disposita. es + 1483 pp. ‘iene. 1836-1840. [Reaumuriaceae, 1037, 1038; Tamariscineae, 1038, 1039.] ERDTMAN, G. Pollen morphology and plant taxonomy. Angiosperms. 539 pp. Stock- holm. 1952. (Corrected reprint with addendum [pp. 541- 553]. New York. 1971.) [Tamaricaceae, 426, fig. 2444; descriptions of pollen of Myricaria germanica, Reau- muria hirtella Jaub. & Spach, Tamarix gallica, T. tetrandra.| Faun, A. Xylem structure and annual rhythm of development in trees and shrubs of ae cag 1. Tamarix aphylla, T. jordanis var. negevensis, T. gallica var. maris- m i. Trop. Woods 109: 81-94. 1958. [Detailed descriptions of wood anatomy.] meres A. Cytologische und entwicklungsgeschichtliche Studien an Myricaria ger- manica Desv. Sv. Vet.-Akad. Handl. 48(7): 1-62. 1912. [7 = 12; detailed descriptions of cell divisions, embryogenesis, pollen-grain formation, and fruit developmen nt.] Gipss, R. D. Chemotaxonomy of flowering plants. 4 vols. xx + 2372 pp. Montreal and London. 1974. [Tamaricaceae, 3: 1838, 1839.] Go pserG, A. Classification, evolution and phylogeny of the families of panes Smithson. Contr. Bot. 58. 314 pp. 1986. [Tamaricaceae, 173, fig. 86 (includes il- lustrations of Tamarix x articulata Vahl (= T. aphylla), T. senegalensis DC., Reau- muria hypericoides, Hololachne (sic) songarica (Pall.) Ehrenb., and Myricaria ger- manica).} aaa ai S. G. Tamaricaceae. (In Russian.) Jn: B. K. SCHISCHKIN & E. FI. URSS 15: 276-327. 1949. [Many new eae taxa ne in co 408 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 GRIseBACH, A. Grundriss der systematischen Botanik fiir akademische Vorlesungen. Gottingen. 180 pp. 1854. hay cane weed 91; Tamariscineae, 91.] Gupta, A. K., & Y. S. sna The leaf epidermal structures in Tamaricaceae. Acta Bot. Indica 12: 200-204. 1984 Hatuer, H. Uber die Te iia: bei Engler’s Rosalen, Parietalen, Myrtifloren und in anderen Ordnungen der Dikotylen. Abh. Naturw. Ver. Hamburg 18: 1-98. een [Tamaricaceae, 43, 49.] . Lorigine et le systeme phylétique des angiospermes. Exposés a l’aide de leur arbre eh Arch. Néerl. Sci. Exact. Nat. IIB. 1: 146-234. 1911. [Tamari- caceae, 216.] Hurcuinson, J. The genera of flowering plants (angiosperms). Dicotyledones. 2 vols. eer 1964-1967. [Tamaricales, 1: 28; Tamaricaceae, 2: 321-324.] Jouri, B. M., & D. Kak. The embryology of Tamarix Linn. Phytomorphology 4: 230- 247. cn Josu1, A. C., & L. B. KasALE. A note on the structure and development of the embryo- sac, ovule, and fruit of Tamarix dioica Roxb. Ann. Bot. IL. 50: 421-425. 1936. Jussieu, A. L. be. Genera plantarum secundum ordines naturales disposita. 498 pp. Paris. 1789. [Tamarix, 313, placed in Ordo IV. Portulaceae, les Portulacées. | KNABEN, G. Chromosome numbers of Scandinavian arctic-alpine plant species. I. Blyttia 8: 129-155. 1950. [AMfvricaria germanica from Norway, 2n = 24.] Linptey, J. Nixus plantarum. 28 pp. London. 1833. [Reaumurieae in Cistales, 13: ip eee in Silenales, 14.] e vegetable kingdom. ed. 3. Ixviii + 908 pp. London. 1853. [Tamaricaceae, 341: ne 407.] Link, H. F. Enumeratio plantarum horti regii botanici berolinensis altera. I. viii + 458 pp. Berlin. 1821. [Tamariscinae, 291.] Luspock, J. A contribution to our knowledge of seedlings. Vol. 1. vii + 608 pp. New York. 1892. [Tamariscineae, 229, 230.] MAHESHWARI, P. The angiosperm embryo sac. Bot. Rev. 14: 1-56. 1948. Maout, E. Le, & J. DecaAIsNE. Traité général de botanique descriptive et analytique. vill + 746 pp. Paris. 1868. [Tamaricaceae, 432, 433: Tamarix indica Willd. illus- trated.] Me tcuior, H. Tamaricaceae. oo ae 334 in H. Mevcuior, A. Engler’s Syllabus der Paanvenfamilicn: ed. 12. . Berlin. 1964. METCALFE, C. R., & L. oe Gee Anat. Dicot. 1: 157-159. 1950. Murty, Y. S. St udies in the order Parietales. IV. Mes anatomy of the flower of Tamaricaceae. Jour. Indian Bot. Soc. 33: 226-238. 1954. NiepeNzu, F. Tamaricaceae. Jn; A. ENGLER & K. alae Nat. Pflanzenfam. II]. 6: 289-298. 1895. ——. Tamaricaceae. /n; A. ENGLER & K. PRANTL, Nat. Pflanzenfam. ed. 2. 21: 282- 289, 1925, OVCHINNIKOV, P. N., & T. K. Kinzikaeva. Myrtama Ovez. et Kinz.. gen. nov genus of the tamarisk family Sy ae Link). (In Russian.) Dokl. Alera. Nauk Tadzhik. SSR. 20(7): 54-58. d On a new ae . ie woody flora of Pamir-Alai— Myrtama Ovez. et Kinz. (fam. Tamaricaceae a (In Russian.) Izv. Akad. Nauk Tadzhik. SSR. Otd. Biol. Nauk 1983(3): 77-79. 1983. PAROLI, V. eC ae sulla embriologia di Tamarix gallica L. Atti Accad. Lincei Rend. Sci. Fis. Mat. Nat. VI. 26: 352-354. 1937. . Contributo ein studio embriologico delle Tamaricacee. Ann. Bot. Roma 22: |- 18. 1940. [Adoxa-type embryo-sac development in Tamarix gallica J PLouvigr, V. Recherche des L-inositol, L-québrachitol et D-pinitol dans quelques sara botaniques. Présence de l’acide shikimique dans le Mammea americana L. (Gut 1989] CRINS, TAMARICACEAE 409 tiféres). Compt. Rend. Acad. Sci. Paris 258: 2921-2924. 1964. [Frankeniaceae and Tamaricaceae are among only 13 families known to produce D-pinitol.] PritzeEL, E. Der systematische Wert der Samenanatomie, insbesondere des Endosperms, bei den Parietales. Bot. Jahrb. 24: 348-394. 1898. [Tamaricaceae, 368, 369. Putte, A. A. Compendium van de terminologie, nomenclatuur en systematiek der zaadplanten. ed. 3. viii + 376 pp. Utrecht. 1952. [Tamaricaceae, 222, included in Cistales; floral diagram of Myricaria.] Puri, V. Studies in the order Parictales. I. A contribution to the embryology of Tamarix chinensis Lour. Beih. Bot. Centralbl. 59A: 335-349. 1939. [Fritillaria-type embryo- sac development.] QAISER, M. aya in the seed morphology of the family Tamaricaceae from Pakistan. Bot. Jour. Linn. Soc. 94: 469-484. 1987. et ie taxa examined; scanning electron eon of seed surface and con S. I. Att. Tamaricaria, a new genus of Tamaricaceae. Blumea 24: 151-155. 1978. [Synonymous with Myrtama. ] Rao, T. A., & S. CHKRABORTI. On the vein termini idioblasts in Reauwmuria Hasselq. ex L. (Tamaricaceae). Curr. Sci. ene 51: 907, 908. 1982. [Ilustrated.] RENDLE, A. B. The classification of flowering plants. Vol. 2. Dicotyledons. ed. 2. x1x + 640 pp. ee England. 1952. [Tamaricaceae, 202, fig. 97, placed in Parietales.] SAINT-HILAIRE, A. DE. Sur les plantes auxquelles on attribue un placenta central libre, et revue des familles auxquelles ces eee appartiennent. III. De la famille des Portulacées. Mem. Mus. Hist. Nat. Paris 2: 195-208. 1815. [Tamarix, 205-208.] SCHNEIDER, C. K. Tamaricaceac. tence: eters der Laubholzkunde 2(8): 340- 345. oe [Illustrations of Tamarix chinensis, T. pentandra Pall., T. juniperina, T. africana, T. parviflora DC., - tetrandra, Myricaria germanica, Reaumuria hyper- eres Se iahna SONLALICE SCHNIZLEIN, A. Iconographia eee naturalium ee vegetabilis. Vol. 3. Ordines 162-228 (Umbelliferae—Coriarieae). Bonn. 1843-1870. [Tamarix africana, pl. 221, Reaumuria vermiculata, pl. 220; Myricaria germanica, pl. 221.] SHARMA, Y. ne Samiclbecnesis: and embryogeny of Tamarix ericoides Rottl. Ann. Bot. IT. : 861 -870. 1939. [n = 12; Fritillaria-type embryo-sac development; first report ; potyembryony in the ernie a a H. Systematische Anatomie der Dicotyledonen. xi1 + 984 pp. Stuttgart. 18 eee 129-132. Soo, os Die modernen Grundsatze der Phylogenie 1m neuen System der Blitenpflanzen. Acta Biol. Acad. Sci. Hungar. 4: 257-306. 1953. [Tamaricaceae, 289, in Parietales, suborder Tamariscineae. | TAKHTAJAN, A. Flowering plants: origin and dispersal. x + 310 pp. Washington. 1969. ee Tae 106, 219, 243.] utline of the classification of flowering plants (Magnoliophyta). Bot. Rev. 46: 295-359. 1980. [Tamaricales, 273, 274; derived from Violales, highly advanced. ] THORNE, R. F. A phylogenetic classification of the Angiospermae. Evol. Biol. 9: 35- osed new realignments in the angiosperms. Nordic Jour. Bot. 3: 85-117. 1983. a abies and Frankeniaceae in suborder Tamaricineae nae Violales:] VESQUE, J. Contributions 4 l’histologie systématique de la feuille des Caryophyllinées, précédées de remarques complémentaires sur importance des caractéres anatom- iques en ee ue Ann. Sci. Nat. Bot. VI. 15: 105-148. 1883. [Tam- ariscinées, 137-14 WILLDENow, C. L. aieate der Gattung Tamarix. Abh. K6nigl. Akad. Wiss. Berlin Physik Klasse 1812-1813: 76-86. 1814. [Early monograph; 16 species.] ZABBAN, B. Come aumenta il numero dei cromosomi nei nuclei inferior: del sacco embrionale di Myricaria germanica Desv. Atti Accad. Lincei Rend. Sci. Fis. Mat. Nat. VI. 21: 208-211. 1935. [v7 = 12; embryo-sac development described.] 410 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Tamarix Linnaeus, Sp. Pl. 1: 270. 1753; Gen. Pl. ed. 5. 131. 1754. Shrubs or small trees of shorelines [deserts, steppes] on sands and clays, often in areas of high salinity, with extensive root systems, often also with adven- titious roots from buried branches. Leaves herbaceous, small, scalelike, entire, alternate, exstipulate, glabrous or papillose [or hairy], sessile with narrow base or somewhat amplexicaul [or vaginate], with salt-secreting glands. Inflorescence of simple or compound racemes, often paniculately branched, occurring on branches of previous and/or current year’s growth. Flowers small, perfect [1m- perfect], 4- or 5-merous, actinomorphic. Bracts shorter than, equal to, or longer than their subtended pedicels, single [2 or 3 in Tamarix rosea Bunge], scalelike, entire, glabrous or papillose [or hairy], sessile with narrow base or somewhat amplexicaul [vaginate] (not necessarily paralleling leaf morphology). Sepals 4 or 5 (6), connate for a short distance near base, entire to denticulate [incised], glabrous or papillose [or hairy], imbricate. Petals 4 or 5 (6), white, pink, or red, ovate or elliptic to obovate, entire to emarginate, equilateral or not, per- sistent or deciduous after anthesis, contorted in bud. Androecium diplostem- onous, the outer whorl of 4 or 5 antesepalous stamens, the inner whorl abortive and developed into a nectariferous hypogynous disc with antepetalous lobes [or of 4-10 antepetalous stamens]; stamens free, inserted onto the central disc; anthers dithecal, usually cordate, apiculate or not. Pollen usually tricolpate (rarely tetracolpate), prolate to spheroidal, [9.5—]1 1.5—23.5[-26] um wide, [12.5-] 14—28.5[-30] um long, walls with fine [to coarse] reticulate sculpturing, lumina circular, polygonal or irregularly shaped, clearly defined to the edge of colpi or not. Gynoecium pyramidal or bottle shaped, consisting of 3[—5] carpels, each with a parietal placenta; styles and stigmas 3[-—5]. Fruit a loculicidal, many- seeded capsule. Seeds 0.5—0.7 mm long, elongate-obovoid, straight, comose at chalazal end, otherwise glabrous [or hairy near apex]; hairs of coma unicellular, hygroscopic, with wavy thickenings near base; endosperm none. Base chro- mosome number 12. Lectotype species: 7. gallica L.; see Britton, North Amer- ican Trees, 702. 1908. (The etymology of Tamarix is uncertain, but the name is probably derived from the Tambre (Tamariz) River in Spain, although the Tamaro River in Nepal and the Hebrew word tamaruk have also been suggested as possible sources.) — TAMARISK, SALTCEDAR. A genus of about 54 species, of the Middle East, the Mediterranean area, central Asia, and southern Africa. Ten species have been introduced into North America, and six of these are naturalized to varying degrees in the Southeast: Tamarix gallica L., T. ramosissima Ledeb., 7. canariensis Willd., T. africana Poiret, 7. tetragyna Ehrenb., and 7. parviflora DC. The first southeastern records of tamarisks as escapes from cultivation appear to date from 1897 (Horton, 1964). The infrageneric taxonomy of Tamarix has undergone y revisions. Bunge divided the genus into sections on the basis of seasonality of flowering and used numbers of floral parts, raceme and vegetative morphology, filament- insertion patterns, petal persistence, and capsule and style morphology as the basis for further subdivisions. However, Baum (1964) and others have shown 1989] CRINS, TAMARICACEAE 411 that seasonality of flowering is unreliable and that many species have both vernal and aestival (or continuous) anthesis. One of the most useful sets of characters derives from disc morphology. The three sections recognized by Baum (1978) are characterized by features of the disc, as well as by raceme width, petal length, and stamen number and position. Section TAMARIX gen- erally has racemes less than 5 mm broad, petals 1-2.25 mm long, and five antesepalous stamens. Series within this section differ in vegetative features such as the presence or absence of papillae and leaf morphology. Species of sect. OLIGADENIA (Ehrenb.) Endl. have broader vernal racemes, some tetran- drous members, and discs with nectariferous lobes. The series within this section differ in bract length relative to pedicel length, petal shape and length, and raceme morphology. Species of sect. POLYADENIA (Ehrenb.) Baum also have broad racemes but have more stamens and discs that lack nectariferous lobes. Its two subdivisions (ser. ARABICAE Baum, ser. PLEIANDRAE Bunge) differ in the number of antepetalous stamens arising from the disc. The taxa in North America belong to sects. TAMARIX and OLIGADENIA. Collectors should be aware that most tamarisks are virtually unidentifiable in the vegetative state. Sect. TAMARIX ser. TAMARIX? 1S represented in North America by 7. gallica, French tamarisk (filaments five, arising directly from the lobes of the disc, raceme rachis glabrous, racemes arising mostly from green (new) wood, sepals entire or subentire, petals oval, caducous), and 7. ramosissima (T. pentandra Pallas, pro parte, nom. illegit., fide Baum, 1978; 7. odessana Stev. ex Bunge) (filaments five, inserted under the disc between its lobes, sepals denticulate, petals obovate, persistent). Tamarix gallica is the most widely recorded species in North America, including our area (Chapman: Jones & Coile; Long & Lakela; MacRoberts; Radford ef a/.; Small, 1913, 1933). However, most of these rec- ords are incorrect; the species is actually quite rare in North America. In the Southeast it is naturalized in Georgia, Louisiana, and North Carolina. Its nat- ural range includes France, Spain, Switzerland, Italy, and Sicily. Tamarix ra- mosissima, a halophyte commonly cultivated in Arizona and California, has become naturalized in the Southwest. Its natural range spans central Eurasia from eastern Turkey, the Ukraine, and eastern Iraq eastward to Mongolia, Tibet, central China, and North Korea. In our area it is known from Arkansas, Georgia, Louisiana, Mississippi, North Carolina, and South Carolina. Sect. TAMARIX ser. LEPTOSTACHYAE (Bunge) Baum is represented in the nat- uralized flora of North America by Tamarix canariensis Willd. (filaments in- serted at tips of lobes of disc, with bases dilated, sepals denticulate, rachis of racemes usually papillose). Although Baum (1967) noted that 7. canariensis had only recently been introduced into Arizona, an examination of specimens from many southeastern herbaria indicates that it has been naturalized in the Southeast since at least 1903. It has been found in North Carolina, South Carolina, Georgia, and Louisiana. The native range of 7. canariensis is the s TAMARIX was called ser. Gallicae by Baum (1978), but since it contains the type of the genus, it must be called ser. TAMARIX. 412 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 —_ Pom Ny ‘or y \ Ye? Wey er Figure |. Tamarix. a—p, 7. ramosissima: a, tip of branch with inflorescence, x 1/2; b, detail of vegetative twig with 2 developing lateral buds, x 6; c, detail of inflorescence, « 5: d, young flower with anthers ready to dehisce, x 12; e, flower after anthers have fallen, the stigmas receptive, x 12; f, flower in vertical section, showing attachment of filaments to disc and of ovules to 2 parietal, basal placentae, x 20; g, gynoecium and androecium removed from flower (upper portions of 2 stamens removed), showing attachment of stamens between lobes of disc, < 12; h, detail of disc and bases of filaments, 1989] CRINS, TAMARICACEAE 413 western Mediterranean Basin (Sicily, extreme southern France, Sardinia, Spain, Morocco, Algeria, Tunisia, and the Canary Islands). Similar to T. africana and T. gallica, it differs from the former in its smaller flowers (petals less than 2 mm long, racemes not over 5 mm wide) and its obovate (vs. ovate) petals, and from the latter in its denticulate sepals, papillose raceme rachises, and some- what expanded filament bases. There appears to be considerable morphological overlap between 7. canariensis and T. gallica, however, and the taxonomic and evolutionary relationships of these two taxa require reevaluation. Section OLIGADENIA ser. ANISANDRAE Bunge is represented in the Southeast by Tamarix africana (stamens five, arising directly from disc lobes, petals relatively large, to 3 mm long) and 7. fetragyna (antesepalous stamens 4 or 5, arising from lobes of disc, one to several smaller antepetalous stamens also present along edge of disc; sepals denticulate; raceme rachis densely papillose; leaves tending to be amplexicaul; vernal racemes more than 5 cm long, loosely flowered). Tamarix africana is naturalized in Louisiana and South Carolina. Its indigenous range includes southern Europe (France and Spain to Italy and Yugoslavia) and northern Africa (Canary Islands and Morocco to Tunisia). Similar to 7. gallica in disc morphology, it has dilated filament bases. Tamarix fetragyna, native to the Middle East (Egypt, Israel, Jordan, Lebanon, Syria, and Cyprus), is reported here for the first time as a naturalized member of the North American flora. It is established on Cumberland Island in Camden Co., Georgia (W. H. Duncan 29934, 9 April 1982; DAo). This species differs from all the others that have become naturalized in North America in having an- tepetalous stamens in addition to the antesepalous ones. It has been the subject of physiological investigations on salt tolerance (see below). Section OLIGADENIA ser. ARBUSCULAE Baum contains Tamarix parviflora (T. tetrandra of American authors, not Pallas), small-flowered tamarisk, the only naturalized North American species that is consistently tetramerous and te- trandrous. Its sepals are denticulate and ridged, and its filaments arise directly from the lobes of the disc. Naturalized in Florida, Louisiana, Mississippi, and North Carolina, as well as throughout the Southwest, its natural range includes southern Europe from Corsica to Turkey, and perhaps also Spain and northern Algeria. In addition, Baum (1967) acc aes Tamarix aphylla (L.) Karsten (sect. being naturalized or cultivated in North America. Tamarix aphylla is known to be naturalized only in the southwestern United States; it is native to North gynoecium removed, x 20; i, mature, unopened fruit, with persistent calyx lobes, petals, and filaments, x 10; j, capsule with the 3 valves beginning to separate apically, x 10; k, fully open See 2 seeds still inside. x 10: 1, nearly mature capsule with | valve removed to show seeds attached to panciah Pasa) placenta, x 12; m, valve of capsule after shedding of seeds, placenta visible below, x 12; n, wind-dispersed seed with coma, 12: 0, seed body, x 25; p, embryo, eiiented as in seed body in “‘o,” x 25. q, 7 canartensis: detail of disc, showing insertion of filaments on lobes of disc, < 20. 414 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 and East Africa, the Middle East, and western Asia eastward to central India. Tamarix aralensis has not yet been found as an escape in the Southeast, but it has been grown as an ornamental shrub in Orange Co., North Carolina (W. C. Coker s.n., 2 June 1942; ncu). It is indigenous to the southern Soviet republics of Kazakhstan, Turkmenistan, Tadzhikistan, and Uzbekistan, as well as to Iran and northern Afghanistan. It differs from its close relative, 7. canariensis, in having filaments that arise from between the lobes of the disc rather than directly from the lobes. Tamarix chinensis (1. pentandra of American authors, in part, not Pallas; 7. juniperina Bunge) is widely cultivated in North America. It has been grown in North Carolina (N. Henry s.n., 5 July 1932; Ncu), but contrary to Baum’s (1967) findings, I have located no authentic record of escape from cultivation in the Southeast. The natural range of 7. chinensis includes Mongolia, The People’s Republic of China, and Japan. It is not a halophyte, unlike most other species of the genus, although it is a xerophyte. Although very similar to 7. ramosissima, T. chinensis has subentire sepals, and in the vernal flowers some of the filaments are inserted between the lobes of the disc on its margin rather than below it. However, in aestival flowers the filament- insertion pattern breaks down. It is difficult to see how these taxa can be recognized as different species, let alone members of different sections of the genus! Both 7. chinensis and 7. ramosissima have gone under the name 7. pentandra in North America, where they are considered to be serious pests along water courses in the deserts of the Southwest. An extensive body of literature on their spread, physiological ecology, and control has developed (see below). One additional species, 7. mascatensis Bunge (sect. TAMARIX Ser. TAMARIX) has been cultivated in Florida and Texas but is not yet known to have escaped. There are two centers of diversity for the genus. One of these, the Indo- Turanian region, 1s also said to be its center of origin, with several of the species there exhibiting primitive character states (poorly developed nectariferous disc, large flowers, both as in Tamarix ericoides Rottler). A secondary center of diversity exists in the Middle East. Although most species of Tamarix are desert or shoreline plants, Vasilchenko & Vasileva argued that the genus may have originated in mesic habitats along mountain streams on what were prob- ably islands in the Tethys Sea during the Paleogene. Their argument is based on the current distribution of 7. arceuthoides Bunge, which, however, is not one of the more primitive taxa, at least according to Baum (1978). More detailed analyses of the biogeography of the region, and of that time period, will be needed to test their hypothesis. Brotherson & Winkel have outlined several aspects of the life history of Tamarix ramosissima that contribute to its success as a colonizer. These include the ability to produce seeds throughout the growing season and under a wide range of conditions, cross-pollination by wind, self-compatibility when cross- pollination is not effected, high seed output, long- and short-range dispersal capabilities, vigorous vegetative reproduction, allelochemical potential (pre- sumably by way of salt secretion), and wide environmental tolerances. Seed germination can take place at any time during the growing season as long as 1989] CRINS, TAMARICACEAE 415 sufficient moisture is available. The primary root elongates very quickly after germination. here is contradictory evidence regarding the mode of pollination in Tam- arix. Although Brotherson & Winkel suggested that pollination is mediated by wind, floral morphology indicates that entomophily must also occur in some species. The staminal disc secretes nectar, a nutritional reward for potential pollinators. However, within Tamarix there has been a trend toward the pro- duction of smaller grains of pollen in smaller flowers, which has been interpreted as a tendency toward anemophily (Baum ef a/., 1971). No experimental in- vestigations of incompatibility systems have been undertaken, and ae patibility may vary from species to species. It has been reported for 7. + mosissima and for Myricaria germanica, but cultivated material of 7. ee in New York has failed to set seed (R. G. Means, New York State Museum, pers. comm., 1988). Vegetative modes of reproduction are also important in the spread and persistence of Tamarix species. Several species are known to produce adven- titious roots and vigorous new shoots from stem tissue that becomes buried. Periodic burning, clearing, or flooding may lead to the development of im- penetrable thickets of Tamarix (Everitt). Thus, although reproduction from seed may be important, the significance of layering and reproduction from vegetative fragments that have been dispersed along water courses should not be underestimated in the rapid spread and colonization of extensive areas in the Southwest by 7. ramosissima and other species. Horton (1964) reviewed the history of the introduction and spread of tam- arisks in the United States, and several authors have dealt with the problem of the ‘‘weedy” species since then. Tamarix “gallica” had become naturalized along the Gulf Coast of Texas by 1877. Its migration into Kansas is well documented. It moved east from Colorado during the 1920’s on floods of the Arkansas River and had spread halfway across Kansas by 1939 (Gates). Tam- arix “pentandra” had become naturalized in North Carolina by 1897 and was first recorded as an escape in Arizona in 1901. Not present in Utah before 1925, it spread rapidly along the major river systems and lake shores between 1935 and 1955. More recently, it has begun to invade mountain streams and reservoirs (Christensen, 1962). It is now a common riparian tree in Utah, Nevada, Arizona, New Mexico, Colorado, and Texas (in Texas, 450,000 acres had been colonized by 1965) and is also known from southern Montana, Idaho, Nebraska, and South Dakota. Tamarix “tetrandra” (T. parviflora), introduced into the Southwest largely to prevent erosion resulting from overgrazing by livestock, became established in Texas by 1888 and in North Carolina by 1897 (Horton, 1964). This species has also been reported as an escape in Portugal. Numerous other studies document the spread and effects of tamarisks along river systems and in irrigable land in the Southwest. The weediness of tamarisks, especially in dry areas, stems from their ability to become established by rapid seed germination and quick growth of very extensive root systems where water resources are limited. Once established, the plants may cause reductions in the width of stream channels, which leads 416 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 to flooding, as well as to excessive use of limited water supplies (although varying opinions exist on this point; see, for example, Waisel, 1960b). Tamarix chinensis has the capacity for stomatal closure under conditions of high wind or temperature, thereby diminishing its evapotranspiration rate. In T. aphylla diurnal stomatal closure may be mediated by an endogenous cir- cadian rhythm (Hagemeyer & Waisel). Other xeromorphic features of the leaves probably also minimize water loss. In 7. chinensis stomata appear to respond directly to changes in the leaf-to-air absolute humidity gradient, leading to more efficient water use during the hottest part of the day, when the evaporative demand Is greatest. Tamarisks replace native vegetation and are less attractive to most species in the southwestern avifauna. Only the mourning dove (Zenaida macroura (L.)) and the white-winged dove (Z. asiatica (L.)) find tamarisk stands to be suitable nesting sites, according to a 1977 report by B. W. Anderson and colleagues, although ladder-backed woodpeckers (Picoides scalaris (Wagler)) have since been recorded as nesting in Tamarix chinensis in the Southwest. The salt glands of several species of Tamarix have been studied extensively, with the aim of determining mechanisms by which these xerophytes survive in soils with high salt concentrations. Located on both leaf surfaces and young branches, the glands consist of two highly vacuolated and six less vacuolated cells, surrounded by epidermal cells with thickened walls. Numerous plas- modesmata connect the collecting cells with adjacent mesophyll cells. Although Mace believed that there was no association between the salt glands and vas- cular bundles, later studies of apoplastic transport (e.g., C. J. Campbell & Thomson) indicate that solutes can be shunted from the xylem to the salt gland under conditions of high salinity. The types of salts secreted by these glands vary, depending on the root environment. Both monovalent and divalent cat- ions, including Na‘, K*, Ca**, and Mg'’*, are present in the solutions. There is no apparent selectivity between cations of the same valence, but divalent cations are prevalent. Anion constitution also varies, with Cl, NO, , HCO, , and SO, being the major ones. In one investigation HCO, comprised 60 percent of the anion pool (Berry, 1970), and in another SO, was predominant (Hem). Salt secretions are sometimes so extensive that columns or cuboidal crystals of salts develop above the gland pores, and concentrations of salts in the secretions may be 20 times greater than those in the local ground water. Tamarix tetragyna (and probably other species) has additional physiological adaptations to soil salinity. Tolerance to salinity stress is at least partially mediated by very large amounts of free proline in the roots. The only investigations of organic natural products in the Tamaricaceae have involved Tamarix species. Long-chain alkanes and esters have been partially characterized in 7. gallica. The alkanes generally contain 18-35 carbons, and the esters have 40-52. Various hydroxylation substitutions occur along these chains, but the level of hydroxylation is very low. The pool of waxy constituents generally contains mixtures of isomers. In 7. “‘pentandra’”’ the epicuticular waxes are variable in constitution, with the variability apparently unrelated to ecological factors (Wilkinson, 1980). Several complex or unusual phenolics, 1989] CRINS, TAMARICACEAE 417 including tannins, a furanofuran, a phenolic glyceride ae a 3,4-ben- zocoumarin (a rare class of natural products), a new 4-methylcoumarin, a new digalloylglucose (nilocitin), and several aldehydes, have been isolated from Tamarix species. Various flavonoids, including unmethylated, mono-, and dimethyl ethers of the flavonols quercetin (rhamnetin, isorhamnetin, tamarix- etin, tamarixin) and kaempferol (kaempferide, rhamnocitrin), as aglycones, glycosides (glucosides, rhamnosides, and glucuronides), and sulfates, have also been found in such tissues as leaves, flowers, and roots. The leaves of all five species of Tamarix analyzed contain D- pinitol (Plouvier). In various species the organic acids trans-2-hydroxy-4 acid, ellagic acid and its 3,3’-dimethyl ether, gallic acid, and isoferulic acid have been isolated. The systematic significance of most of these compounds has yet to be determined. There is some evidence of host-specificity among certain gall-forming insects associated with tamarisks. For example, most species of Psectrosema (Diptera: Cedidomyiidae) utilize only a single species of Tamarix. One species is known to use two hosts; two other species use three. Host specificity is maintained in mixed stands of tamarisks. When more than one species of Psectrosema utilizes the same host species, the insects attack different parts of the tree, minimizing competition (Habib; Harris: Kieffer). There is also some suggestion that other insects, such as mirids (Heteroptera) and aphids (Homoptera), are monopha- gous or oligophagous with regard to Tamarix species (Habib ef al.; Zocchi). Girdling of branches by cicadas (Homoptera) has been noted in Arizona, but vigorous regrowth from below the girdled areas indicates that no measure of biological control is achieved by these insects. Despite the detrimental effects Tamarix species have had on natural eco- systems in North America, some of the same species, as well as others, are economically important in a positive sense. Tamarix aphylla and other species are sources of tannins, which are produced in healthy branch, bark, fruit, and leaf tissue, as well as in galls induced by various insects, including moths, scale insects, and gall midges. Tamarisks have also been used for afforestation of dunes and other “unproductive” sandy areas (Decaux; Trabut, 1926; Vil- bouchevitch). The wood serves as a source of fuel, charcoal, and construction material, and its ash yields various inorganic salts. Decaux noted that a fine powder of Tamarix ash acted as a successful contact insecticide against cat- erpillars. In Manchuria and Tibet young branch tips of tamarisk have been used to make tea, and in northern Europe tamarisk has served as a substitute for hops in the making of beer. Tamarix “‘pentandra” has been shown to accumulate trace elements such as boron, cadmium, copper, lead, lithium, manganese, molybdenum, nickel, and zinc (Dreesen & Wangen; Shiber), both as surface deposits on leaves and internally, which may mean that the plants could serve as indicators of environmental contamination. Medicinally, species of Tamarix have been used in the treatment of infections, hemorrhages, en- teritis, rickets, gout, and rheumatism, and as astringents, antiseptics, and di- uretics (Vilbouchevitch). Some tamarisks are also prized for the honey pro- duced from their nectar. The biblical ‘“‘manna” is a gummy exudate composed mainly of carbohydrates secreted by coccids (scale insects) that feed on the 418 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 branches of 7. mannifera (Ehrenb.) Bunge (Maleki & Djazayeri). Several species, including those naturalized in North America, are also grown as ornamentals (Bailey et al.). REFERENCES: ost of the family references are also pertinent, but see especially BAUM (1978) and PLOUVIER In reference to the genus. ANDERSON, B . A. Hiccins, & R. D. OHMART. Avian use of saltcedar communities in the ares Cslorada River valley. Pp. 128-136 in R. R. JOHNSON & D. A. JONES, tech. coords., pea ee preservation and management Sea habitat: a sym- ae U.S. Dep. Agr. Forest Serv. Gen. Tech. Rep. RM-43. 1977. ANDERSON, J. E. Factors controlling transpiration and photosy Bete in Tamarix chi- nensts Lour. Ecology 63: 48-56. 1982. ANDHIWAL, C. K., & K. Kishore. Hentriacontan-!2-ol and other waxy constituents of Tamarix gallica Linn. Jour. Indian Chem. Soc. 61: 820-822. 1984. ARENDT, G. Beitrage zur Kenntnis der Gattung 7amarix. Inaugural-Dissertation, Fried- rich-Wilhelms- nec zu Berlin. 56 pp. Borna-Leipzig. 1926. [Worldwide re- vision; discs of several species, including 7. africana and T. parviflora, illustrated. ] AtAEV, E. A. Notes on Tamarix Meyeri Boiss. and its can with the saline soils of ihe burnt planes of central Kopetdaga. (In Russian.) Izv. Akad. Nauk Turkmensk. SSR, Ser. Biol. Nauk 1977(6): 64-66. 1977. ameter descriptions; ionic com- position of soil.] BAHL, C. P., M. R. PARTHASARATHY, & T. R. SESHADRI. Chemical components of Tamarix dioica. Indian Jour. Chem. 5: 171, 172. 1967. [Flavonoids, carbohydrates, organic acids. BaiLey, L. H., E. Z. BAitey, & STAFF OF THE BAILEY HorTORIUM. Hortus third. A concise dictionary of plants cultivated in the United States and Canada. xiv + 1290 . New York. 1976. [Tamarix, ees H.H.,M.A. M. Naw WWAR, J. BuDpRus & M. LinscuHerp. Niloticol, a phenolic glyceride and two phenolic aldehydes from the roots of Tamarix nilotica. Phyto- chemistry 26: 1837, 1838. Bar-Nun, N., & A. POLJAKOFF-MAYBER. Some aspects of protein metabolism in 7am- arix tetragyna roots grown in a saline substrate. Austral. Jour. Pl. Physiol. 1: 237- 246. 1974 > Salinity stress and the content of proline in roots of Piswm sativum and Tamarix tetragyvna. Ann. Bot. II. 41: 173-179. 1977. BATE-SMITH, E. C., & C. R. MeTCALFE. Leuco-anthocyanins. 3. The nature and system- atic a iets of tannins in dicotyledonous plants. Jour. Linn. Soc. Bot. 55: 669— 705. 1957. [Tamaricaceae, 694.] Baum, B. On the vernales-aestivales character in Tamarix and its diagnostic value. Israel Jour. Bot. 13: 30-35. 1964. [Seasonality of flowering not a valuable taxonomic character. ] —. Introduced and eee tamarisks in the United States and Canada (Tam- aricaceae). Baileya 15: 19-25. 1967. [Eight species reported for North America; key; illustrations of dises and Bere insertion. ] 1. J. Bassett, & C. W. Crompton. Pollen morphology and its relationships to taxonomy and distribution of Tamarix, series Vaginantes. Osterr. Bot. Zeitschr. 118: 182-188. 1970. [Scanning electron micrographs of pollen surfaces; key to pollen grains of five species.] J& Pollen morphology of 7anarix species and its relationship to ‘the Peace of the genus. Pollen Spores 13: 495-521. 1971. [Key and descrip- tions of pollen grains of most species; scanning electron micrographs. ] Berry, W. - ae ames of salt secretion in Tamarix. (Abstract.) Pl. Physiol. 44(Suppl.): 21. 1989] CRINS, TAMARICACEAE 419 —. Characteristics of salts secreted by Tamarix aphylla. Am. Jour. Bot. 57: 1226- ae 1970. HOMSON. eee of salt secreted by salt glands of Tamarix ears ‘Canad. Jour. Bot. 45: 1774, 1775. 1967. BosaBALipis, A. M. Origin, ultrastr oe estimation of the possible manners of growth a chyma ee of Tamarix aphylla L. Jour. Submicroscop. Cytol. 19: 423-432. 1987. & . THomson. Light microscopical studies on salt gland development in Tamarix api Ann. Bot. II. 54: 169-174. 1984a & Ultrastructural differentiation of an unusual structure lining the an- ticlinal walls of the inner secretory cells in Tamarix salt glands. Bot. Gaz. 145: 427- 435. 1984b. Ultrastructural development ayes secretion in the salt glands of 7am- arix aphylla L. Jour. Ultrastructure Res. 92: 55-62. 1985. Bowpen, W. M. Diploidy, polyploidy, and ane aa relationships in the flow- ering plants. Am. Jour. Bot. 27: 357-371. 1940. [7. gallica, T. hispida Willd., T. odessana, T. parviflora, n = 12; T. parviflora, T. pentandra, 2n = 24.] A list of chromosome numbers in higher plants. I]. Menispermaceae to Ver- benaceae. Ibid. 32: 191-201. 1945. [Same counts as above for Tamarix.] Bowser, C. W. Introduction and spread of undesirable tamarisks in the Pacific south- western section of the United States and comments concerning the plants’ influence upon the indigenous vegetation. (Abstract.) Am. Geophys. Union Trans. 38: 415 416. 1957. BroTHERSON, J. D., J. G. CARMAN, & L. A. SzysKA. Stem-diameter age relationships of Tamarix ramosissima in central Utah. Jour. Range Managem. 37: 362-364. 1984. & D. FieLp. Tamarix: ey of a successful weed. Rangelands 9: 110-112. 1987. [Popular review of the effects in the western United States. ] & WINKEL. Habitat piers of saltcedar (Tamarix ramosissima) 1n central Utah. Great Basin Nat. 46: 535-541. 1986. Brus, T. First nesting of a New World woodpecker in — (Tamarix chinensis). Southwest. Nat. 28: 113. 1983. [Ladder-backed woodpec Bunce, A. Tentamen generis Tamaricum. 81 pp. Dorpat. 1852. {Wor rldwide revision. ] CAMPBELL, C. J., & J. E. STRONG. Salt ae eee in Tamarix pentandra (Tama- ricaceae). Southwest. Nat. 9: 232-238. | CAMPBELL, N., & W. W. THOMSON. eee localization in the leaf of Tamarix. Pro- toplasma 83: 1-14. 1975. [7. aphylla.] CARLETON, M. A. Adaptation of the tamarisk for dry lands. Science 39: 692-694. 1914. [7. gallica and T. parviflora mentioned. CARMAN, J. G., & J. D. BRoTHERSON. Comparisons of sites infested and not infested with saltcedar (Tamarix pentandra) and Russian olive (E/laeagnus angustifolia). Weed Sci. 30: 360-364. 1982. [Salt concentration alone insufficient to classify salt- cedar sites; seasonal moisture availability also important.] CHAKRABARTY, G., S. R. Gupta, & T. R. SesHADRI. Polyphenols of Tamarix Troupti and 7. aphylla. Indian Jour. Chem. 3: 171-174. 1965. [Flavonoids, organic acids. ] CHAPMAN, A. W. Flora of the southern United States. ed. 2. xxxvuil + 698 pp. New York. 1883. [T. gallica in South Carolina, 620 CHRISTENSEN, E. M. The rate of naturalization of 7amarix in Utah. Am. Midl. Nat. 68: eee 1962. [7. pentandra.] yering in Tamarix induced by algae. [bid. 70: 250. 1963. [7. pentandra.] nem D. C., P. E. Martin, & R. M. HAGAN. Evapotranspiration from riparian vegetation: water relations and irrecoverable losses for saltcedar. Jour. Soil Water Conserv. 37: 233-236. 1982. [Factors affecting evapotranspiration rates in 7. chi- nensis in the Southwest. Decaux, M. L’avenir du Tamarix articulata en Tunisie, Algérie et Maroc: utilité de ses galles, moeurs et l’insecte qui les produit et de ses parasites. Bull. Soc. Natl. 420 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Acclim. France, IV. 5 janvier 1893: 1-10. 1895. [Economic uses; life histories of insects associated with galls.] Decker, J. P. Salt secretion by Tamarix pentandra Pall. Forest Sci. 7: 214-217. 1961. [Photographs of salt glands and salt crystals.] Dreesen, D. R., & L. E. WANGEN. Elemental composition of saltcedar (Tamarix chi- nensis) impacted by effluents from a coal-fired power plant. Jour. Environ. Qual. 16. 1981. Duncan, W. H., & M. B. Duncan. Trees of the southeastern United States. x1 + 322 pp. Athens, Georgia. 1988. [Tamarix, 47, 48; key; follows set ear en tee Ev Ansari, M. A., M. A. M. Nawwar, A. Ev Dern, A. EL SHERBEINY, & H. I. Ev Sissi. A sulphated kaempferol 7,4’-dimethyl ether and a quercetin aes from the flowers of Tamarix aphylla. Phytochemistry 15: 231, 232. 1976. [Multiple sulfation.] Ex Sissi, H. I., M. A. M. Nawwar, & N. A. M. SALen. Plant constituents of Tamarix nilotica leaves (Tamaricaceae). Experientia 29: 1064, 1065. 1973. [Flavonoids, hy- drolyzable tannins. ] Everitt, B. L. Ecology of saltcedar—a plea for research. Environ. Geol. 3: 77-84. 1980. [T. chinensis (including 7. ramosissima).] Faun, A., & N. Arnon. The living wood fibres of Tamarix aphylla and the changes occurring in them in transition from sapwood to heartwood. New Phytol. 62: 99- 104. 1963 FerRAzzi, P. Tamarice. Apic. Mod. 74: 143-145. 1983. [Apicultural value of Italian tamarisks. | sARY, H. L. Root distribution of five-stamen tamarisk, seepwillow, and arrowweed. Forest Sci. 9: 311-314. 1963. [7. pentandra.] & J. S. Horton. Some sprouting ee of five-stamen tamarisk. U. S. Dep. Agr. Forest Serv. Res. Note RM-39. 7 pp. 1965. [7. pentandra Gates, F. C. Trends of tree migration in Kansas. ae Kansas Acad. Sci. 42: 127- 132. 1939. eee of T. gallica in Kansas, 130, map 10.] GATEWOOoD, J. S., T. W. Rosinson, B. R. Cotsy, J. D. Hem, & L. C. — Use of water by ‘bottom- land US saga in ace Safford valley, Arizona. U. S. Geol. _— Surv. Water-Supply Pap. 1103. xiv + iene) . 1950. [Saltcedar te for more than 75 Pals of the water use in rire 7 Gay, L. W., & L. J. FrirscHen. An energy aes analysis of water use by saltcedar. Water Seen Res. 15: 1589-1592. 1979. [7. chinensis.] GERLING, D., J. KUGLER, & A. Lupo. The gall insects and mites that occur on Tamarix spp. in Israel and the Sinai. Boll. Lab. Entomol. Agrar. “Filippo Silvestri” Portici 33: 53-79. 1976. GINZBURG, C. ‘Organization of the adventitious root apex in Tamarix aphylla. Am. Jour. Bot. 54: 4-8. 1967. GOLDSMITH, F. B., & N. SMART. Age, spacing and growth rate of 7amarix as an indication of lake boundary fluctuations at Sebkhet Kelbia, Tunisia. Jour. Arid Environ. 5: 43- 51. 1982. [T. africana.] Grar, W. L. Fluvial adjustments to the spread of tamarisk in the Colorado Plateau region. Geol. Soc. Am. Bull. 89: 1491-1501. 1978. [Historical photographs of col- onization by 7. chinensis and concurrent changes in stream morphology.] Grozier, R. U. Growth of salt cedar oe gallica) in the Pecos River near the New Mexico—Texas boundary. U.S. Geol. Surv. Prof. Pap. 525-B: 175, 176. 1965. Gupta, S. R., & T. R. SESHADRI. Survey of anthocyanins. Part VI. Colouring matter of Tamarix Troupti. Constitution of the aglycone and its synthesis. Jour ) London 157: 3063-3065. 1954. [Flavonoids; chemical synthesis and analytical chemistry. ] Haars, R. The bionomics of Psectrosema spp. (Diptera: Cecidomyiidae) reared from 1989] CRINS, TAMARICACEAE 42] galls on Tamarix spp. in Pakistan. Bull. Entomol. Res. 73: 457-463. 1983. [Host ranges; distribution; use of different parts of same tree by different species of gall- formers. _S. A. HASAN, SHUKAR-UD-DIN, & R. AHMAD. Final report. Insect enemies at- tacking tamarisk Tamarix spp. in Pakistan. June 1975—June 1980. cay valicanal sane of Biological Control, Pakistan Station, Rawalpindi. 138 p HAGEMEYER, J., & Y. WAISEL. An endogenous circadian rhythm of arene in eres aphylla. Physiol. Pl. 70: 133-138. 1987. Harporne, J. B. Flavonoid bisulphates and their co-occurrences with ellagic acid in the Bixaceae, Frankeniaceae and related families. Phytochemistry 14: 1331-1337. 1975. [Flavonol bisulphates are characteristic of Tamaricaceae and co-occur with ellagic acid in this family.] Harris, K. M. Descriptions of Psectrosema spp. (Diptera: Cecidomytidae) ra from galls on Tamarix spp. in Pakistan, including four new species. Bull. Entomol. Res. 73: 447-456. 1983. [Illustrations of galls, p/. 6; many gall-formers host- specific ] Hay, J. Salt cedar and salinity on the upper Rio Grande. Pp. 288-300 in M. T. FARVAR & J Mitton, eds., The ae technology: ecology and centers develop- ment. Garden City, New York. 1972. [History of introduction in this region; 7. pentandra introduced to prevent erosion caused by overgrazing Hem, J. D. Composition of saline residues on leaves and stems of saltcedar (Tamarix oe Pallas). U. S. Geol. Surv. Prof. Pap. 491-C. 9 1967. ae vist, H., & F. Grazi. Studies on variation in embryo sac development. Bot cere 141-166. 1964. [Tamarix odessana (= T. Be nN T. parv viflora.] Pane J.S. Inflorescence development in Tamarix pentandra Pallas (Tamaricaceae). Southwest. Nat. 2: 135-139. 1957. [Continuous flowering from early spring until autumn; early racemes on old wood, later racemes on new shoots. _ Notes on the introduction of deciduous tamarisk. U. S. Dep. Agr. Forest Serv. Res. Note RM-16. 7 pp. 1964. [Reviews early introductions of T. gallica, T. pen- tandra, T. tetrandra, and T. (= Myricaria) germanica in the U.S. A The development and perpetuation of the permanent tamarisk type in the phreatophyte zone of the southwest. Pp. 124-127 in R. R. JoHNsoNn & D. A. JONEs, tech. coords., Importance, preservation and management of riparian habitat: a sym- posium. U.S. Dep. Agr. Forest Serv. Gen. Tech. Rep. RM-43. 1977. [Brief review of life history traits of 7. chinensis (including 7. ramosissima).] & J. E. FLoop. Taxonomic notes on Tamarix pentandra in Arizona. Southwest. Nat. 7: 23-28. 1962. [Arizona specimens are most similar to 7. ramosissima sensu a oe Hunt, D.R. Typification of eee aphylla L. Kew Bull. 16: 481, 482. 1963. [Basionym of Tamarix aphylla (L. Hunter, W. C., B. W. pases & R. D. SES Sune avian community com- oe of gail Hew se ert riparian systems. Pp. es 134 in R. R. JoHNson, C. D. , D. R. Patton, P. F. FFoLiiott, & R. ae en ee Tieeaes per and their management: Rea con- flicting uses. U. S. Dep. Agr. Forest Serv. Gen. Tech. Rep. RM-120. 1985. [Lower species richness and total density of birds in Tamarix habitats (ee in natural HyickaMa, T. E. A. vAN. Water use by saltcedar as measured by the water budget method. U. S. Geol. Surv. Prof. Pap. 491-E. 30 pp. 1974. [Depth of ground water or water table affected average water use by 7. pentandra.] Weather and evapotranspiration studies in a saltcedar thicket, Arizona. Ibid. 491-F. 78 pp. 1980. [T. chinensis reacts to high wind speeds and high temperatures by stomatal closure; riparian vegetation does not always use water at the max1 imum potential rate.] 422 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 IsHAK, M. S., H. I. Ex Sissi, M. A. M. NAwwar, & A. E. A. EL SHERBEINY. Tannins cee ene of a galls of Tamarix aphylla, Part I. Planta Med. 21: 246-253. , A. E. A. SHERBEINY, & M. A. M. NAwwar. Tannins and polyphenolics of the galls of Tamarix aphylla. Part IU. Ibid. 374- a 1972b. IsRAILl, A. H., R. C. SHARMA, & A. = Kripwal. Chemical examination of Tamarix gallica Linn. Indian Jour. Chem. 3: 48. 1965. [Ellagic “acid and derivatives. ] Jones, S. B., & N.C. Come. The distribution of the vascular plants of Georgia. 230 pp. Athens, Georgia. 1988. [7. gallica in Chatham and Glynn counties. Katie, A., & A. POLJAKOFF-MAYBER. Malic dehydrogenase from 7amarix roots. Effects of sodium chloride im vivo and in vitro. Pl. Physiol. 55: 155-162. 1975. [Sodium chloride postulated to induce conformational changes in malic dehydrogenase in 7. tetragyna. | & ct of salinity on respiratory pathways in root tips of Tamarix tetragyna. Ibid. 57: 167-170. 1976. [Salinity affects root metabolism in tamarisk differently than in Pisum sativum L.] KHAN, N. A., M. Kamit, & M. ILyas. Constituents of Tamarix dioica Roxb. (Tama- riscineae). Indian Jour. Chem. 18B: 390. 1979. [First isolation of trans-2-hydroxy- 4-methoxycinnamic acid as an aglycone from a plant source KreFFerR, J. J. Les Cécidomyies du 7Varmarix. Marcellia 11: 169- 172. 1912. [Newly described gall-formers and host records.] KLerInkopr, G. E., & A. WALLACE. Physiological basis for salt tolerance in Tamarix ramosissima. Pl. Sci. Lett. 3: 157-163. 1974. KnutH, P. Handbook of ree pollination. Vol. 2. vii + 703 pp. Oxford. 1908. [Tamariscineae, 202: insect visitors to ne germanica include flies and lepi- dopterans; self-pollination possible if cross-pollination not effected.] Lesreton, P., & M.-P. BoucHEz. Recherches chimiotaxinomiques sur les plantes vas- culaires. V. Distribution des composés polyphénoliques chez les Pariétales. Phyto- chemistry 6: 1601-1608. 1967. [Flavonols, organic acids.] Lona, R. W., & O. LAKELA. A foe of tropical Florida. xvil + 962 pp. Coral Gables, Florida. ‘1971, [T. gallica, ~~ 610.] Lupo, A., & D. GeRLING. Bionomics of the Tamarix spindle-gall moth Amblypalpis olivierella Rag. (Lepidoptera: Gelechiidac) and its natural enemies. Boll. Lab. Ento- mol. Agrar. “Filippo Silvestri” Portici 41: 71-90. 1984. Mace, A. C. Osmotic water stress: mesophyll Setar deficit and transpiration rates of tamarisk. Southwest. Nat. 16: 117-120. 1. [7 pentandra.] MacRoperts, D.T. A documented checklist ae Ae of the vascular flora of Louisiana. Part III. Dicotyledoneae. Fagaceae to Zygophyllaceae. 756 pp. Shreveport, Louisi- ana. 1988. [7. pees and T. gallica mapped for Louisiana, 731, 732.] MA ekI, M., & M. DyAzAyert. Chemical characteristics and composition of ‘‘gaz”” gum from Tamarix mannifera. Starke 24: 296-299, 1972. [Gum is 90 percent carbo- hydrate, four percent protein.] MartIs, B. pe, M. C. Lor, & M. B. Poco. Contribution to a better understanding of the genus in Portugal. Candollea 41: 293-298. 1986. [Key to Portuguese species. ] McATEE, W. L. rther notes on tamarisk. Science 39: 906. 1914. Mc cack E. ptr in oo ornamental plants. 3. The tamarisks. Jour. Calif. Hort. Soc. 12: 76-83. | MerKEL, D. L., & H. H. to PKINS. Life history of salt Sa ial L.). Trans. Kansas Acad. Sci. 60: 360-369. 1957. [Detailed accou Nawwar, M. A. M., J. Bupprus, & H. BAvER. Dimeric sterol constituents from the roots of Tamarix nilotica. Phytochemistry 21: 1755-1758. 1982. [A furanofuran, a new ellagic-acid derivative, and other organic acids. ] _ A. E. A. Ev SHERBeINy, & M. A. EL Ansari. Plant constituents of Tamarix 1989] CRINS, TAMARICACEAE 423 aphylla flowers eas Experientia 31: 1118. 1975. [Flavonoid aglycones, glycosides, and sulfates.] M.A. Sa 3.4,8,9,10-pentahydroxy-dibenzo [b,d] pyran-6-one from Tamarix open Phytochemistry 23: 2966, 2967. 1984. [Rare natural product.] J.B us, H. BAuer, & M. LinscHetp. Polyphenolic constituents of the flowers of Tamarix nilotica: the structure of nilocitin, a new digalloylglucose. Tetrahedron Lett. 1984: 4 4. MeL NSCHEID. Flavonoids of the flowers of Tamarix nilotica. Phytochemistry 23: ae 2349. 1984. NIEDENZU, F. De genere Tamarice. Index lectionum in lyceo regio hosiano Brunsber- gensi. I. Pp. 3-11. Brunsberg. 1895. [Key to species of Tamarix of the world, based on post-BUNGE specimens. ] Parmar, V. S., J. S. RATHORE, S. SinGH, A. K. JAin, & S. R. Gupta. Troupin, a 4-methylcoumarin from Tamarix Troupii. Phytochemistry 24: 871, 872. 1985. Perkin, A. G., & P. J. Woop. Yellow colouring principles contained in various tannin matters. Part V. Pistacia Lentiscus, P. Terebinthus, Tamaris [sic] africana, T. gallica, Ailanthus glandulosa, Ficus Carica. Jour. Chem. Soc. London 73: 374-386. 1898. [Ellagic and gallic acids yield yellow dyes.] PLatt-ALoiA, K. A., R. D. Buss, & W. W. THomson. Lipid-lipid interactions and membrane fusion in plant salt glands. Pp. 160-172 in W. W. THomson, J. B. Mupp, . Grass, eds., Biosynthesis and function of plant lipids. Proc. 6th Ann. Symp. Botany (January 13-15, 1983), Univ. Calif., Riverside. 1983. [7. aphylla.] QaiserR, M. The genus Tamarix Linn. (Tamaricaceae) in Pakistan. Pakistan Jour. Bot. 13: 107-158. 1981. [Twenty-six species (five new), Some new combinations; stylized illustrations of leaf shapes, 108, and disc morphology, 109. RApFoRD, A. E., H. E. AHLES, & C. R. BELL. Guide to the vascular flora of the Carolinas. 383 pp. oe Hill, North Carolina. 1964. Raja, M. I ‘‘Tamerici.”? Pro Natura 21(74): 7-9. oe [Ecology, distribution, and morphology of the seven Italian species of Tamarix. ] Reyno ps, H. G., & R. R. ALEXANDER. Pe ena ee eos stamen tamarisk. In: C. S. ScHOPMEYER, tech. coord., Seeds of woody plants in the United States. U.S. Dep. Agr. Agr. Handb. 450: 794, 795. 1974. Tiree oa and seedling. ] Roserts, G. Plotting a better future for lambs: a practical guide to providing shade. Queensl. Agr. Jour. 110(1): 25, 26. 1984. [Tamarix articulata (= T. aphylla) rec- ommended as a source of shade in lambing areas. ] Rosinson, T. W. Introduction, spread and areal extent of saltcedar eens - western states. U.S. Geol. Surv. Prof. Pap. 491-A. 12 pp. 1965. [Rehash of H 1964, plus data on acreages invaded and effects of eae on stream Sac SALEH, N. A. M., H. I. Ev Sissi, & M. A. M. NAawwar. A rhamnetin glucuronide trisulphate from . leaves of Tamarix aphylla. Phytochemistry 14: 312, 313.1975. SAVCHENKO, V. A. Ecology of flowering of tamarisks, salt-spikes and sea-blites of the floodplain forests of southern Tadzhikistan. (In Russian.) Izv. Akad. Nauk Tadzhik. SSR, Otd. Biol. Nauk 1975(1): 18-24. 1975. ees see and 7. ramosissima. | SCHIMAN-CZEIKA, H. Tamaricaceae. Fl. Iranica 4: 1-17. ao oo species 0 Tamarix in liane T. aphylla, fig. 8: T. ramosissima, figs SHIBER, J. G. Trace metals in Tamarix pentandra and Malva ee from Beirut. Revue Roum. Biol. Sér. Biol. Végét. 26: 55-60. 1981. [Cadmium, copper, chromium, iron, lead, nickel, zinc.] SHIMONY, C., & A. FAHN. Light- and electron- ee ees studies on ae structure of salt glands of Tamarix aphylla L. Jour. Linn. Soc. Bot. 60: 283-288. SHINNERS, L. H. Salt cedars (Tamarix, Ta crce of the Soviet vane by S. G. GorscCHKOVA. Translation and commentary. Southwest. Nat. 2: 48-73. 1957. [Key 424 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 to North danse material; preliminary taxonomic conclusions regarding North American spec SMALL, J. K. Flora ote ait United States. ed. 2. x11 + 1394 pp. New York. 1913. Lion grees 783, of the aes flora. xxii + 1554 pp. New York. 1933. [Tamani- caceae, Re Smirnov, I. A. Agrotechniques of growing planting material of Tamarix in the desert zone. (In Russian.) Lesn. Khoz. 12: 28-30. 1981. [7. hispida. Situ, E. B. An atlas and annotated list of the vascular plants of Arkansas. ed. 2. 1v + 489 pp. Fayetteville, Arkansas. 1988. [7. chinensis common along Arkansas River; T. gallica not verified.] STEVENS, O. A., & W. N. CLUTE. ela habit of Tamarix. Am. Bot. 30: 34. 1924. [Notes on flowering i in New Mex1 Swenson, J. E., P. HENDRICKS, & A. FARJON. Arrival and occurrence of Tamarix chi- nensis (tamarisk) along the Yellowstone River in Treasure and Rosebud counties, Montana. Proc. Montana Acad. Sci. 41: 67-70. 1982. THomson, W. W., W. L. Berry, & L. L. Liu. Localization and secretion of salt by the salt elands of Tamarix aphylla. Proc. Natl. Acad. Sei. U.S. A. 63: 310-317. 1969. & L. L. Liv. Ultrastructural features of the salt gland of Tamarix aphylla L. Planta 73: 201- 220. 1967. —.,K.A. Piatt, & N. CAMPBELL. The use of lanthanum to delineate the apoplastic continuum in plants. Cytobios 8: 57-62. 1973. [Tamarix aphylla. PLATT-ALolA. The ultrastructure of the plasmodesmata of the salt glands of Tamarix as revealed by transmission and freeze-fracture electron microscopy. Protoplasma 125: 13-23. 1985. [7- aph via.) TrasutT, L. Le takaout, galle de |’‘ettl” ou “tlaia” (Jamarix articulata Vahl). Bull. Stat. Rech. Forest. Nord Afrique 1: 171-182. 1917. [Illustrations of galls on flowering and oo branches. ] tlaia, Tamarix articulata Vahl. [bid. 336-349. 1926. [Uses.] Uruor, J. C. T. Dictionary ofeconomic plants. ed. 2. 591 pp. New York. 1968. [Tamarix, 510. Urkin, L. M. The structure of tamarixin. Chem. Nat. Prod. 2: 128-130. 1966. VASILCHENKO, I. T., & L. 1. Vasiteva. Was tamarisk a desert tree? (In Russian, English summary.) Biull. Moskov. Obshch. Isp. Pri. Biol. Il. 81: 139-143. 1976. [Mountains of central Asia (former islands in the Tethys Sea) as place of origin of the genus.] VILBOUCHEVITCH, J. Les Tamarix et leurs applications: leur valeur au point de vue du reboisement. Bull. Soc. Natl. Acclim. France. 5 septembre 1890: 1-18. 1890. [Eco- nomic and medicinal uses.] WaIsEL, Y. Ecological studies on Tamarix aphylla (L.) Karst. I. Distribution and re- production. Phyton Buenos Aires 15: 7-17. 1960a. Hl. The water economy. /bid. 19-28. 1960b. Watts, J. G. Adaptation of indigenous insects to aes pentandra. Folia Entomol. Mex. 25-26: 51, 52. 1973. [Native insects feeding on tamarisk are generalists.] Wess, P.B. Tamarix gallica of Linnaeus. Jour. Bot. iendon 24: 1-11. 1841. [Discussion of the origin of the staminal disc.] WETTSTEIN, R. R. von. Handbuch der systematische Botanik. ed. 2. vil + 914 pp. Leipzig and Vienna. 1911. els ice 583, illustrations an tetrandra and 7. pauciovulata J. Gay ex Coss., fig. 4 WIEDEMANN, H. T., & B. T. Cross. i inundation for control of saltcedar along the periphery of lakes. (Abstract.) Proc. Southern Weed Sci. Soc. 31: 229. 1978. [Ta marix gallica.) WILKINSON, R. E. Seasonal development of anatomical structures of saltcedar foliage. Bot. Gaz. 127: 231-234. 1966. [7T. pentandra. | 1989] CRINS, TAMARICACEAE 425 —. Water stress in salt cedar. /bid. 133: 73-77. 1972. _ Ecotypic variation of Tamarix pentandra epicuticular wax and possible rela- tionship with herbicide sensitivity. Weed Sci. 28: 110-113. 1980. [Alleged ecotypes represent different provenances; no statistical treatment of data, and alternative explanations of variation not considered.] Zual, S.-H., C.-G. WANG, & X.-Z. Gao. Morphological and anatomical observations of clasping leaves of Tamarix L. (In Chinese; English abstract.) Acta Bot. Sinica 25: 519-525. 1983. [Thirteen species of eae - China; xeromorphic features of leaves discussed in relation to survival in dese Zoccul, R. Contributo alla conoscenza dell’ Re ae delle tamerici in Italia. Redia, III. 52: 31-129. 1971. [Feeding habits of 92 species of insects on various species of Tamarix.] ZonARY, M. The genus Tamarix in Israel. Trop. Woods 104: 24-60. 1956. [Numerous illustrations; discussion of disc morphology. & um. On the androecium of Tamarix flower and its evolutionary trends. Israel Jour. Bot. 14: 101-111. 1965. [Illustrations of stamen insertion and lobing patterns of disc; scheme of evolutionary trends of these features.] Ecological studies on east Mediterranean dune plants. Bull. Res. Council Israel 1(4): 38-52. 1952. [Adaptations of tamarisks to shifting sand levels. ] AL-SHEHBAZ, DRABA 427 NEW OR NOTEWORTHY DRABA (BRASSICACEAE) FROM SOUTH AMERICA IHSAN A. AL-SHEHBAZ! Four Colombian species, Draba bovacana, D. cocuyana, D. pseudocheiran- thoides, and D. ritacuvana, are described as new. Two new names, D. argen- tifolia and D. ecuadoriana, are proposed to replace the later homonyms D. argentea O. E. Schulz and D. affinis J. D. Hooker, respectively. Draba L. is the largest genus of the Brassicaceae, with about 350 species distributed primarily in the Northern Hemisphere (Al-Shehbaz, 1987). There are at least 75 species in South America that grow at high elevations from Colombia and Venezuela southward along the Andes to Patagonia During the past several years, | have been working on a revision of the South American species of Draba. In this continent the genus is more highly diversified in habit and morphology of its species than elsewhere in the world. There are many undescribed drabas in South America, and the nomenclature of several species 1s problematic. In the present paper, four new species are described and two new names are propose Draba argentifolia Al-Shehbaz, nom. nov. Based on Draba argentea O. E. Schulz, oo IV. 105(Heft 89): 134. 1927, non D. argentea (All.) Lam. Encycl. 2: 329. 1786. A close examination of Draba argentifolia and D. pickeringii A. Gray sup- ports Schulz’s (1927) view that the two species are indeed very closely related. However, the latter species has slender styles ca. | mm long, fruiting pedicels 4-6 mm long, petals 4-5 mm long, fruits 4-6 mm long, and valves primarily covered with simple trichomes. In D. argentifolia all these structures are clearly shorter than those of D. pickeringii, and the fruit valves are covered with furcate and stellate trichomes. Draba boyacana Al-Shehbaz, sp. nov. FIGURE |. Herba perenna pusilla. Caulis erectis vel ascendens, 2.5—9 cm longa, pilis crispis simplicibus vel stipitato-furcatis. Folia basalia rosulata, obovata vel spathulata, acuta, 2-dentata, ciliata, sessilia, superne hispida, pilis rigidibus subappressis simplicibus usque ad 1.4 mm longis et brevioribus stipitato-fur- ‘Arnold Arboretum, Harvard University, 22 Divinity Avenue, Cambridge. Massachusetts 02138. © President and Fellows of Harvard College, 1989. Journal of the Arnold Arboretum 70: 427-437. July, 1989. 428 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 —' ; | RL. : < ~ Figure 1. Holotype of Draba boyacana. Scale bar = | cm. catis, inferne glabrescentia, 7-23 mm longa, 4-8 mm lata; folia caulina 2-5, sessilia, oblonga vel ovata, 5-8 mm longa, 1.5-3.5 mm lata. Inflorescentia corymbosa, ebracteata; pedicelli fructiferi divaricati, 3-5 mm longi. Sepala ovata, erecta, nonsaccata, 3-3.5 mm longa, 2—2.5 mm lata. Petala alba, obovata, cuneata, 5-6 mm longa, 2—2.5 mm lata. Siliqua elliptica vel sublanceolata, compressa, glabra, 7.5—8.5 mm longa, 3—3.5 mm lata; stylus 0.6-0.9 mm longus. Semina matura ignota. Small perennial herbs. Stems erect to ascending, | or few from base, un- branched, 2.5—9 cm long, crisped pubescent with mixture of simple trichomes and shorter, stalked, furcate or substellate ones. Basal leaves forming rosettes, sessile, obovate to spatulate, 7-23 mm long, 4-8 mm wide, acute at apex, cuneate at base, 2-toothed, ciliate, the upper surface hispid with mixture of long, simple, subappressed trichomes to 1.4 mm long and fewer much shorter, stalked furcate ones, the lower surface glabrescent. Cauline leaves 2 to 5, sessile, oblong to ovate, 5-8 mm long, |.5-3.5 mm wide, entire, pubescent on both surfaces (or only on the upper) with mixture of simple and stalked furcate or substellate trichomes. Inflorescences ebracteate, corymbose racemes, slightly elongated in fruit; fruiting pedicels divaricate, 3-5 mm long, pubescent. Sepals erect, ovate, nonsaccate, 3—-3.5 mm long, 2—2.5 mm wide, membranaceous at 1989] AL-SHEHBAZ, DRABA 429 margin, pubescent. Petals obovate, cuneate, not clawed, 5-6 mm long, 2-2.5 mm wide, white. Filaments erect, dilated at base, 2.2-2.5 mm long, white; anthers ovate, ca. 0.6 mm long. Nectar glands confluent, low, subtending bases of all filaments. Fruits elliptic to sublanceolate, flattened parallel to septum, 7.5-8.5 mm long, 3-3.5 mm wide, glabrous; styles slender, 0.6-0.9 mm long. Mature seeds not seen. Tyre. Colombia, Depto. Boyaca, Sierra Nevada del Cocuy, in paramo below Ritacuva glacier, 4300 m alt., 5 Aug. 1957, P. J. Grubb, B. A. B. Curry, & A. Ferndndez- Pérez 303 (holotype, us!). ADDITIONAL SPECIMEN EXAMINED. Colombia. Depro. BoyAcA: Sierra Nevada del Cocuy, Alto Ritacuva, Barclay & Juajibioy 7485 (mo). Draba boyacana, which grows on peat at the base of Espeletia Mutis plants or on moraines above the Espeletia stands (4300-4500 m alt.), has no close relatives among the northern South American drabas. Of the 11 Mexican and Guatemalan species (Rollins, 1984), only D. helleriana Greene bears some superficial resemblance to D. hovacana. However, the latter has white instead of yellow flowers and is quite different in leaf morphology and pubescence, as well as in habit and inflorescence. Perhaps the nearest relative of D. bovacana is D. magellanica Lam., a species widely distributed in Patagonia (Boelcke, 1984). From that, D. boyacana is easily distinguished in having glabrous rather than pubescent fruits, few- instead of many-flowered inflorescences, confluent instead of no median nectaries, and primarily simple or furcate as opposed to stellate leaf pubescence. Draba cocuyana Al-Shehbaz, sp. nov. FIGURE 2. Herba perenna suffruticosa caudicibus simplicibus vel ramosis. Caules lig- nosi, 4-8(-20) cm longi, inferne petiolis bruneis, foliorum emortuorum ad- pressis imbricatis indurati, superne dense foliati. Folia caulina integra, oblonga vel ovata, imbricata, glabrescentia, ciliata, 3-8(-10) mm longa, 2-3 mm lata. Racemi corymbosi ebracteati; pedicelli fructiferi divaricati, 4-9(—-1 1) mm longi. Sepala oblonga, subsaccata, 4-5 mm longa. Petala obovata, flava, 6-7 mm longa, ca. 3.5 mm lata. Siliqua elliptica vel subovata, compressa, glabra, (6-)8—-12 mm longa, (3.5-)4.5-5.5 mm lata; stylus 1.5—2.7 mm longus; semina brunnea, compressa, ovata, reticulata, 1.6-1.8 mm longa, 1.2-1.4 mm lata. Suffrutescent perennial herbs; caudex short, simple or branched. Stems sev- eral from caudex, unbranched, 4—8(—20) cm long, leafless below and covered with persistent, imbricate, appressed, brown, petiolar remains of previous years, leafy above. Leaves oblong to ovate, 3-8(-10) mm long, 2-3 mm wide, obtuse to subacute at apex, abruptly ending in a thick, brown, persistent petiolar base, entire, ciliate, the upper surface glabrous, the lower sparsely pubescent with dendritic trichomes but soon glabrescent. Inflorescences ebracteate corymbose racemes, slightly elongated in fruit; rachis sparsely to densely pubescent with furcate to dendritic trichomes; fruiting pedicels divaricate, usually straight, 4— 9(-11) mm long, pubescent. Sepals erect, oblong, subsaccate, 4-5 mm long, ca. i) 430 JOURNAL OF THE ARNOLD ARBORETUM [voL. 7 FiGure 2. Holotype of Draba cocuyana. Scale bar = 1 cm. 2.5 mm wide, yellow, sparsely pubescent. Petals broadly obovate, 6-7 mm long, ca. 3.5 mm wide, bright yellow. Filaments erect, ca. 4.5 mm long; anthers oblong, ca. | mm long. Nectar glands confluent, subtending bases of all fila- ments. Fruits elliptic to subovate, flattened parallel to septum, (6-)8—12 mm long, (3.5-)4.5-5.5 mm wide, glabrous; septum complete, membranaceous; styles slender, 1.5-2.7 mm long. Seeds 4 to 7 per locule, broadly ovate, com- pressed, 1.6-1.8 mm long, |.2-1.4 mm wide, brown, reticulate. Type. Colombia, Depto. a cae Sierra Nevada del Cocuy, Valle de los Co- rallitos, in peaty rock crevices, ca. 4000 m alt., 31 July 1957, P. J. Grubb, B. A. B. Curry, & A. Fernandes: Pérez 188 (holotype, us!; isotype, msc!). ADDITIONAL SPECIMENS EXAMINED. Colombia. Depto. Boyaca: Sierra Nevada del Cocu W face of Boqueron de Cusiri, Grubb, Curry, & Ferndndez- Pérez 831 (us); Sierra Noid: del Cocuy, Quebrada Bocatoma, Cleef & Florschiitz 5644 (us) Draba cocuyana 1s most closely related to D. litamo L. Uribe, resembling it in many features of habit, leaves, flowers, and fruits. Both species are known in Colombia as “‘litamo real” and are extensively collected for alleged medicinal purposes. A note on the holotype reads: ‘ta potion made of whole plant in boiling water used to relieve pain and prolong life.” Garcia-Barriga (1974) stated that D. /itamo is greatly demanded in Colombia for medicinal properties and is used even against cancer. The differences between the two species can hardly be recognized by those who collect them for such use. Therefore, both species may well be endangered. Draba cocuyana 1s easily distinguished from D. /itamo in having ebracteate inflorescences, glabrous fruits, and usually dendritic trichomes on the inflo- 1989] AL-SHEHBAZ, DRABA 43] rescence rachis, pedicels, and lower leaf surfaces (see FIGURE 3a). Draba litamo has bracteate inflorescences, scabrous fruits with simple trichomes, and usually simple or furcate trichomes on the rachis, pedicels, and lower leaf surfaces (see FIGURE Both Draba cocuyana and D. litamo belong to a group of four species that Schulz (1927, 1929, 1936) recognized as sect. Dolichostylis (Turez.) O. E. Schulz. This section includes the Colombian D. schultzei O. E. Schulz and the Vene- zuelan D. empetroides Brandt, Gilg, & O. E. Schulz, D. /indenii (W. J. Hooker) Linden ex Sprague, and D. arbuscula J. D. Hooker. The last species also grows in the Sierra Nevada de Santa Marta, Colombia. Unlike most of Schulz’s (1936) other sections of Draba, sect. Dolichostylis is well defined and evidently mono- phyletic. Its six species can be distinguished as follows. KEY TO THE SPECIES OF SECTION DOLICHOSTYLIS A. Styles (5-)8-12(-17) mm long; pedicels (1-)1.5-2.5(-3.5) cm long; leaves glabrous. D. lindenit. A. Styles 0.75-3 mm long: pedicels 0.3-1(-1.4) cm long; leaves variously pubescent. B. Leaves narrowly linear, conspicuously ciliate; styles 2-3 mm long. ........... Sep agt ded anetd eh iow hoki ead erase onde neat eae hein Wen e aaha arco ea schultzei. B. Leaves ovate to narrowly oblong, nonciliate to ciliate; styles usually than 2 mm lon C. Fruits with long-stalked stellate trichomes. C. Fruits glabrous or scabrous with simple or furcate trichomes. D. Leaves canescent, tomentose, almost exclusively iti aedacaoe eee D. empetroides. ichome Gee genes ee en ett Saae uae aa eh ae Ee oes arbusc a D. Leaves neither canescent nor tomentose, with simple, are or dendritic trichomes. E. Inflorescences a fruits glabrous; leaves and pedicels with den- AritiG trichOMES:. ccc ee ep esas Glas Sn Eee ocuyana. Inflorescences o acteate; _ scabrous; leaves and pedicels with both simple and furcate trichomes. ................0.00005- D. litamo. ™m Draba ecuadoriana Al-Shehbaz, nom. nov. Based on Draba affinis J. D. Hooker, Fl. Antarct. 2: 235. 1845; non D. affinis Host, FI. Austr. 2: 238. 1831: non D. affinis Hoppe in Sturm, Deutschl. Fl. I, Abt. 14(Heft 60). 1833: non D. affinis Ledeb. Fl. Ross. 1: 148. 1842 As shown above, the Ecuadorian Draha affinis J. D. Hooker is a later hom- onym of three binomials that represent minor variants of the highly polymor- phic European D. aizoides L. and D. alpina L. Draba pseudocheiranthoides Al-Shehbaz, sp. nov. FIGURE 4, Planta perennis, caudicibus crassis, petiolis foliorum emortuorum persisten- tibus; caulis erectis, 15—46 cm longis. Folia basalia rosulata, anguste oblan- ceolata, remote serrulata, ciliata, 3-8 cm longa, 4-6 mm lata, pilis simplicibus et furcatibus. Inflorescentia bracteata, multiflora; pedicelli fructiferi divaricati, (5—)8—12(-17) mm longi. Sepala oblonga, 2-3 mm longa. Petala flava, obovata, subemarginata, 5.5-7 mm longa, 2—2.5(-3) mm lata; filamenta 2—-2.5 mm longa; 432 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 . WESOAd 3 WAN wey \ @eia ig een =SiAn av 6 oy Adi WEL, erat Only the Asclepiadoideae are represented by plants native in the Southeast. Matelea and Gonolobus (recognized separately for reasons outlined in the dis- anc resented by Morrenia odorata, naturalized in Florida, and possibly by Calo- tropis procera (see footnote 23). The Periplocoideae are represented by Cryp- tostegia grandiflora and Periploca graeca, each sometimes escaped from cultivation in Florida. Despite its frequent and perhaps usual inclusion in the Asclepiadoideae, Secamone is similar in a number of characters to putatively more primitive taxa within the suborder, justifying placement of the genus in its own subfamily (Safwat). Anthers are four-locular in the Apocynaceae, Periplocoideae, and Secamone, but two-locular in the remainder of the Asclepiadoideae. However, Safwat showed that the condition in Secamone is not due to the formation of a false partition in each of two essentially two-locular anthers, as Demeter had reported for Thevetia neriifolia Juss. ex Steudel (= JT. peruviana (Pers.) K. Schum.) (Apocynaceae; Plumerioideae), but rather that four groups of arche- *5Sundell provided an annotated list of available and rejected names for taxa between the levels of family and genus in the Asclepiadaceae. According to him, the correct names and author citations for Endlicher’s subtribes ‘““Hoyeae” and “‘Stapelieae’’” would be Hoyinae G. Don and Stapelinae G. Don, respectively. Although Sundell did not include them, Endlicher’s names should evidently be rejected in favor of Don’s because of priority. According to Stafleu and Cowan (Taxonomic Literature, ed. 2, 1976), the pages in Endlicher’s Genera Plantarum on which his subtribal names appeared were eiblshed in August 1838, whereas those in the fourth volume of Don’s A General History of the Dichlamydeous Plants on which his subtribes ‘“‘Hoyaceae” and “‘Stapelieae” appear (according to Sundell) were published no later ne April 1838, and possibly as aes as 1837. It is essentially for this same reason that Don is cited instead of Endlicher as the r of the Gonolobeae. Not surprisingly, there appear to be no available names between the a a subtribe and genus for the possible nonsucculent and succulent groups within the ee linae. According to Sundell, the correct author citation for the Asclepiadoideae as a subfamily of the Apocynaceae (i.e., the Asclepiadaceae as here defined) is Meisner, Pl. Vasc. Gen. 1: 257, 266; 2: 174 1838. However, according to Stafleu and Cowan (Taxonomic Literature, ed. 2, 1976), these pages were all published in April 1840. In any case, the ‘“‘Asclepiadoideae”’ would have been established within the Asclepiadaceae as an autonym with the publication in August 1838 of the Periplocoideae R. me ex Endl. and the Secamonoideae Endl. (See also footnote 23.) should be noted that Swarupanandan presented a very complex case against some of the names cael listed as available. According to this argument, all of the names Sundell indicated as being based on Pergularia L. were illegitimate because they were actually based on Pergularia R. Br., considered a later homonym and itself therefore illegitimate. For example, although Endlicher cited Linnaeus as the author of Pergularia and excluded the ane listed under the penis by Brow n, he tea. Marsdenieae Sere (1868), and Se Dene. (1 a in favor of ee Benthait (1873), Marsdenieae N. . Br. 07); ane Slepelicee Bentham (1868), eee, because they all included Pergularia and t p with the earlier Pergularieae Endl. (1838), Swarupanan- dan considered them eae because the Pergularieae Endl. is not. If Swarupanandan were to be considered correct, the tribe with erect pollinia recognized here (i.e., the penuitieae Endl.) would be in need of a legitimate name 452 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 sporial cells are evident in the hypodermis during the earliest stages of internal differentiation. Demeter had also noted that in the Plumerioideae (Apocynaceae) the four locules are essentially equal in pollen production, but that in the Apocynoideae the abaxial two are partially sterile and perhaps otherwise less developed. He therefore proposed that the bilocular condition in the Asclepiadoideae repre- sented the culmination of a trend involving progressive suppression of the abaxial locules; that is, that the two adaxial locules in the Apocynaceae, Periplo- coideae, and Secamonoideae are homologous to the two locules in the Ascle- piadoideae. Evidence cited by Safwat supports this (however, see Frye, 1901: Richharia). According to Safwat (p. 110), the two adaxial groups of archesporial cells in the four-locular anther appear “in exactly the same position” as the two in the two-locular condition. In addition, he showed that although the staminal vascular bundle is between the adaxial and abaxial locules in the four- locular anther, 1t 1s adjacent to the abaxial epidermis in the two-locular con- dition. According to Safwat, the two meiotic divisions of pollen mother cells are simultaneous in most Apocynaceae, most Periplocoideae, and Secamone, but successive in all cases investigated in the Asclepiadoideae (e.g., Asclepias, Cy- nanchum, Gonolobus, Matelea, and Sarcostemma). The resulting tetrads are usually tetrahedral or isobilateral in the Apocynaceae; generally tetrahedral or isobilateral but sometimes linear (or T-shaped; see Maheswari Devi) in the Periplocoideae; rhomboid and T-shaped in Secamone (within pollinia) but perhaps linear as well; and linear (or rarely T-shaped; see Maheswari Devi) in the Asclepiadoideae (within pollinia). The style-stigma head in Secamone also appears to retain certain putatively primitive features according to Safwat, although in my view the supporting generalizations he cited need information from additional taxa. According to Safwat, its entire surface 1s secretory in the Apocynaceae, although in Apocy- num, an advanced genus, the secretions 1n areas alternating with the stamens are loosely organized into five amorphous bodies. At the extreme opposite, in most Apocynaceae—that is, in all Asclepiadoideae investigated but not in Secamone—secretions comprising both nectar and complex elements of the pollinaria (1.e., the corpuscula and the translator arms) are limited to well- defined areas within the stigmatic furrows that alternate with the stamens. The style-stigma heads in the Periplocoideae and in Secamone appear to be intermediate between the extremes described above, but in different ways. In the Periplocoideae, as in the Asclepiadoideae, secretions in each of five rela- tively well-defined areas form structures involved in pollen transfer. In the Periplocoideae these include an adhesive disc that is attached to the base of an internally adhesive, inverted, usually elongate cone or partial cone into which individual pollen tetrads fall and are held after anther dehiscence. On the other hand, in Secamone, as 1n the Apocynaceae, the entire surface of the style-stigma head is secretory, but unlike the condition in Apocynum, the se- cretions in areas alternating with the stamens are tightly organized into struc- tures thought to be homologous with but reduced from those described for the Periplocoideae. Pollen tetrads in Secamone are held within pollinia, as in the 1989] ROSATTI, APOCYNINEAE 453 Asclepiadoideae, and the inverted or partial cones present in the Periplocoideae do not develop (presumably because they would serve no purpose); instead the hardened secretions in Secamone become attached to the pollinia after anther dehiscence in much the same way that the translator arms become attached to the pollinia in the Asclepiadoideae. Safwat also presented evidence suggesting that the hardened secretions in Secamone are homologous only to the corpuscula in the Asclepiadoideae, and that the translator arms of the latter were separately (and perhaps subsequently) evolved. Indeed, I have observed (see illustrations in Schumann) that the structures in Secamone are similar in appearance and orientation to corpuscula (i.e., they resemble inverted V’s or U’s), and that they and the pollinia to which they are attached are very similar to the pollinaria with erect pollinia char- acterizing the Pergularieae. In view of the foregoing, I suggest that intercalation of translator arms be- tween the bases of the hardened secretions of the style-stigma head and the bases of the pollinia in Secamone (and reduction in the number of pollinia involved from four to two) could have resulted in the condition in the Per- gularieae and that within the Asclepiadoideae the erect pollinia of the Pergu- larieae are most primitive. By similar reasoning, the horizontal pollinia of the Gonolobeae are more primitive than the pendulous ones in the Asclepiadeae. The tribal sequence employed here is intended to reflect these possibilities. Safwat concluded that his investigations suggested recognition of the Apocy- naceae in the broad sense, and within that group the subfamilies Plumerioideae, Apocynoideae, Periplocoideae, Secamonoideae, and Asclepiadoideae. Among recent workers, some (e.g., Stevens) have adopted this arrangement, while others (e.g., Pichon, 1948b) have chosen to recognize an additional subfamily, the Cerberoideae Pichon, near the Plumerioideae and the Apocynoideae (see discussion of the Apocynaceae). However, the case for recognizing the Seca- monoideae at the subfamilial level seems more compelling than that for the Cerberoideae. Therefore, while only the Plumerioideae and Apocynoideae are recognized in my treatment of the Apocynaceae, the Periplocoideae, Seca- monoideae, and Asclepiadoideae are all accepted here in the Asclepiadaceae, although only the first and third are represented by plants that in the south- eastern United States are native, naturalized, or escaped from or persistent beyond cultivation. As indicated above, the Periplocoideae are generally considered to be more primitive than the Asclepiadoideae, especially with regard to the structures involved in the transfer of male gametes from one flower to another. As men- tioned above, secretions from each of five areas on the style-stigma head in the Periplocoideae solidify to form an adhesive disc connected to the base of an internally adhesive, inverted cone or partial cone, into which individual pollen tetrads fall from adjacent anthers and are held after anther dehiscence; this entire unit often becomes attached by the adhesive disc to an insect visitor. In the Asclepiadoideae, secretions from each of five areas on the style-stigma head harden to form a corpusculum and, in addition, two connected translator arms that each become attached after anther dehiscence to one pollinium (the tetrads of one anther locule encased in tapetal secretions) from each of two 454 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 adjacent anthers; this entire structure (i.e., pollinarium) frequently becomes attached by the corpusculum to an insect visitor. (See description of pollination in discussion of Asclepias.) The mechanism of pollen transfer in the Periplocoideae is less effective than that in the Asclepiadoideae, due in part to the greater effectiveness in the latter of the corpusculum. Whereas insect parts become attached to the disc just by adhesion, they become wedged—often irreversibly—in an abaxial groove in the corpusculum. In addition, although pollen tetrads are merely adherent to the structures described for the Periplocoideae, in the Asclepiadoideae they are completely enclosed in the pollinia, which are firmly attached to the corpus- culum by the translator arms. The Periplocoideae also differ from the Asclepiadoideae in the extent to which the anthers are associated with one another and with the style-stigma head. Although sometimes adherent to one another, they are not connate into an anther head. In addition, they are appressed against, but not agglutinated to, the style-stigma head. Anthers in the Periplocoideae also lack the wings that characterize the Asclepiadoideae, the adaxial, hyaline flaps present in some Pergularieae and in the Asclepiadeae, and the abaxial appendages of Gonolobus (and the extraregional Fischeria DC.). Many aspects of the chemistry, anatomy, embryology, and morphology of the Asclepiadaceae are summarized in the discussion of the suborder. Of taxo- nomic relevance within the family is Huber’s observation that a wide distri- bution of cardiac glycosides in the Asclepiadeae indicates a “primary position” of the tribe, as well as a close affinity with the Periplocoideae (and the Apoc- ynaceae, especially the Apocynoideae). In other tribes of Asclepiadaceae, these compounds are “gradually” replaced by more specialized types of picric sub- stances. Clearly, a primitive position for the Asclepiadeae within the family is contrary to what has been suggested by morphological evidence (discussed above). Cytological data regarding the Asclepiadaceae (and Apocynaceae) that had accumulated up to about 1976 were evaluated and summarized to the level of tribe by Stevens. As in the Apocynaceae, the most common and presumably ancestral base chromosome number in the family is x = 11; in fact, about 90 percent of the genera and 78 percent of the species for which chromosome counts are available include the sporophytic number 2” = 22, and only spo- rophytic numbers based on x = || have been reported for the Secamonoideae (Secamone,; 2n = 22), Pergularieae (20 genera; 2n = 22, 33, 44, and 66), and Gonolobeae (Matelea; 2n = 22). Other base numbers include x = 8, 9, 10, and possibly 12 and 23, each of which could have been derived from x = 11. Although polyploidy in the Periplocoideae, Secamonoideae, and Gonolobeae may be only apparently absent (i.e., due to insufficient sampling), it may be less common throughout the family than in the Apocynaceae (see also Albers). In general, the Asclepiadaceae are not well known cytologically. For example, of about 300 species thought by Stevens to comprise the Gonolobeae, only a single, extraregional member of Mate/ea is known in this regard (see above). According to data compiled by Bolkhovskikh and colleagues, sporophytic num- bers have been reported for, among genera represented in one way or another 1989] ROSATTI, APOCYNINEAE 455 in the Southeast, Asclepias (2n = 22; also 20 and 24), Calotropis (2n = 22, 44, and 48), Cryptostegia (2n = 24), Cynanchum (2n = 22; also 18 and possibly 44), Periploca (2n = 22 and 24), and Sarcostemma (2n = 22 and possibly 20). The Asclepiadaceae are usually considered to be of relatively little economic importance. The seed trichomes are silky but too brittle to be of much use as a textile (Standley & Williams). However, in the southern United States they have been used as a substitute for down, especially in making pillows and cushions (Bramwell). Several uses for these plants have been proposed and to some extent investigated. Particularly during World War II, species of Asclepias were evaluated as potential crop plants and sources of alkaloids for oral con- traceptives, of latex for the production of rubber, and of stem fibers as sub- stitutes for hemp in rope and paper making (see primarily Whiting; Woodson, 1954). In the year following its construction in 1943, a United States Govern- ment ““milkweed floss- and seed-extracting plant” in Petoskey, Michigan, using material of (evidently) several species obtained from uncultivated plants in 26 states, provided the military with two million pounds of seed trichomes for use instead of kapok as a stuffing in life-saving equipment (Berkman). Upon restoration of peace, interest in the cultivation of these plants waned, primarily because of their toxicity to livestock (however, see Van Emon & Seiber). Re- cently there have been evaluations of Asc/epias for these and related purposes (e.g., Buchanan ef al.; Campbell; Van Emon & Seiber), as well as a historical account of the many uses of A. syriaca (Gaertner). Many species of Asclepiadaceae, especially those belonging to Asclepias, Cryptostegia, Hoya, and Stapelia (among the genera discussed above), are used in horticulture. Cryptostegia grandiflora has reportedly been grown as a po- tential source of natural rubber (Lawrence), although the expenditure of sub- stantial amounts of time and energy on this species during World War II in the Dominican Republic resulted in only one small sheet of rubber (C. E. Wood, Jr., pers. comm.). Many species are toxic to livestock and humans, and the latex of some (especially of Matelea, according to Lawrence) is used as arrow poison in Central and South America. Nevertheless, young, tender fol- licles of several viny species of Matelea and other Asclepiadaceae are eaten either fresh or cooked in this same area (Spellman, 1975b; Standley & Wil- liams). REFERENCES: nder sige for the Apocynineae, see BAILLON; BARTLING; BENTHAM & HOOKER; BOLKHOVSKIKH ef a/.; BROWN (1811); DECAISNE; DEMETER; ENDLICHER; HUBER: LAw- RENCE; LINDLEY; ne STANDLEY & WILLIAMS; and WAGE NITZ. Under references for the Apocynaceae, see PICHON (1948b). Under references for Asclepias, see WOODSON (1954). ALBERS, F. Karyological studies in Bean ene (Abstract.) Proc. Int. Bot. Con 13: ae 1981. [Survey of 60 genera; “‘overwhelming majority” of species diploid 2n = 22), as high as 2n = 132 i in some succulents. ] Baum, H. Die Stellung der Samenanlagen am Karpell bei Asclepias syriaca, Cynanchum vincetoxicum, und ee ee. Osterr. Bot. Zeitschr. 95: 251-256. 1949.* 456 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 [According to SAFWAT, showed placentae to be adaxial, submarginal projections and not infolded margins of carpels.] BERKMAN, B. Milkweed—a war strategic material and a potential industrial crop for sub-marginal lands in the United States. Econ. Bot. 3: 223-239. 1949. [Detailed account of re use of many species. Bookman, S.S. The floral morphology of Asclepias speciosa (Asclepiadaceae) in relation to es, and a clarification in terminology for the genus. Am. Jour. Bot. 68: 675-679. 1981. BRAMWELL, D. Asclepiadaceae. Pp. 225, 226 in V. H. HEywoon, ed., Flowering plants of the world. New York. 1978. Brower, L. P. Plant poisons in a ene food chain and implications for mimicry theory. Pp. 69-82 in K. L. CHAmBers, ed., Biochemical coevolution. Corvallis, Oregon. 1970. [Species of Asclepias, anes Gomphocarpus, and Gonolobus.] Browicz, K. The genus Periploca L., a monograph. Arb. Kornickie 11: 5-104. 1966. [Two sections, three series, 11 species recognized; fossil, chemical, cytological, an- atomical, and other data summarized. Brown, R. On the organs and mode of fecundation in Orchideae and Asclepiadeae. Trans. Linn. Soc. London 16: 685-745. pls. 34-36. 1833. [Asclepiadaceae, 715-733, 744, 745, pls. 34-36.] BUCHANAN, R. A., I. M. Cutt, F. H. Orry, & C. R. RUSSELL. Hydrocarbon- and rubber- producing crops. Evaluation of U.S. plant species. Econ. eu 32: 131-145. 1978. [Deserving of further consideration as crop species, among 100 evaluated, were Asclepias incarnata and A. syriaca (for rubber); highest al rating for A. syriaca (various uses).] CAMPBELL, T. A. Chemical and agronomic evaluation of common milkweed, Asclepias syriaca. Econ. Bot. 37: 174-180. 1983. [Goal (dry weight) of 6 percent oil and 7 percent polyphenol “readily” attainable through selective breeding; that of 10 percent polymeric nyctorsTpon more cine ult ae Dan Dicko-ZAFIMAHOVA, L. Ul t es parois des pollinies de Calotropis procera (Ait.) Aut. f. ae (English pana Adansonia 17: 455-463. lege and SEMs. & J.-C. AuDRAN. Etude ontogénique de la pollinie de Calotropis procera (As- clepiadaceae), apports decyl de la microscopie électronique. (English sum- mary and figure captions.) Grana 20: 81-99. 1981. [A continuous tectum of spo- ropollenin formed around the pollinium J Decalisn_g, J. Asclepiadeae. DC. Prodr. 8: 490-665. 1844 DraPALik, D. J. A hoe study of the genus Matelea i in the southeastern United States. 225 pp. Unpubl. Ph.D. dissertation, Univ. North Carolina, Chapel Hill. 1969. (Diss. Abstr. B. 31(5): 2504. 1970.) [Includes M. gonocarpa (= Gonolobus gonocar- pus).] E1sikowiTcH, D. Morpho-ecological eieee on the serra of Calotropis procera (Asclepiadaceae) in Israel. Pl. Syst. Evol. 152: 185-194. 986. [Concentration of nectar within stigmatic chamber oe germination of ie ia.] EL-Gazzar, A., & M. K. Hamza. Morphology of the twin pollinia of Asclepiadaceae. Pollen Spores 15: 459-470. 1973. [Thirty-two species in 20 genera, including As- clepias and Cynanchum, individual grains invariably inaperturate and without ex- ternal exine sculpturing; Schumann’s Pern of Hoya and Stapelia (pollinia hor- izontal) in Tylophoreae (pollinia erect) q ; many other morphological details. ] Emon, J. vAN, & J. N. SE1BER. Chemical constituents and energy content of two milk- weeds, Asclepias speciosa and A. curassavica. Econ. Bot. 39: 47-55. 1985. [For use as fuel (values compared to various fossil sources) and/or “chemical feedstock”; latex of A. speciosa (western U. S.), with much lower levels of cardiac glycosides, much less (if at all) toxic to livestock.] Fasris, H. A. Asclepiadaceae. Pp. 38-66 in A. L. CABRERA, ed., Fl. Prov. Buenos Aires. Part V. 1965. 1989] ROSATTI, APOCYNINEAE 457 Fournier, E. Asclepiadaceae. Jn: C. F. P. von Martius, FI. Brasil. 6(4): 189-332. pis. 50-98. 1885. [Many excellent illustrations.] Frye, T. C. Development of the pollen in some Asclepiadaceae. Bot. Gaz. 32: 325-331. pl. 13. 1901. [Several species of Asc/epias, most represented in the Southeast; bispo- rangiate anthers of Asclepiadoideae not due to elie aee ue reduction because only two oe of archesporial cells evident from earliest sta rphological study of certain Asclepiadaceae. [bid. 34: 389-413. pls. 13- 15. 1902. seienl species of Asclepias, most represented in the Southeast; umbels said to be terminal; many anatomical and embryological details. ] GAERTNER, E. E. The history and use of milkweed (Asclepias syriaca L.). Econ. Bot. 33: 119-123. 1979 [1980]. [Brief review of various uses. GILMARTIN, A. J. Variations within populations and classification. II. Patterns of vari- ation within Asclepiadaceae and Umbelliferae. Taxon 29: 199-212. 1980. [Ascle- piadaceae and Umbelliferae similar to each other but not to Bromeliaceae in ranges of intragroup variation at several taxonomic levels; similarities attributed in part to comparable frequencies of umbellate inflorescences . Morphological variation within five angiosperm families: Asclepiadaceae, Bro- meliaceae, Melastomataceae, Piperaceae, and Rubiaceae. Syst. Bot. 6: 331-345. 1981. [Asclepiadaceae sampled in western North America, the other four in Colom- bia; variation within populations of Asclepiadaceae and Melastomataceae compa- rable, greater than that for other three. Goop, R. An atlas of the Asclepiadaceae. New Oe on ee ma noe ae maps for the Pert iadoide pn as well as for several tribes (e.g., Asclepiadeae, Gonolobeae) aa ee (e.g., clepias, Sarcostemmay, high percentage of narrowly distributed and endemic es in family; greatest species diversity in southern Africa.] Hux.ey, C. Symbiosis between ants and epiphytes. Biol. Rev. Cambridge Phil. Soc. 55: 321-340. 1980. [Three groups of species in Dischidia represent three levels of increasingly sophisticated modifications, especially regarding leaves; in all, adven- titious roots absorb nutrients from organic debris accumulated by ants in spaces enclosed by individual leaves (greater density of stomata on adaxial surface maxi- mizes intake of CO, from decomposition and minimizes water los Jounson, M. F. Studies in the flora of Virginia: Asclepiadaceae. Castanea 48: 259-271. 1983. [Keys to three genera and 18 species (Asclepias, 13; Cynanchum, one; Matelea, four), all represented in the Southeast. ] KuHaToon, S., & S. I. Avi. On the localized pollinium germination in Calotropis procera (Willd.) R. Br. (Asclepiadaceae). (Abstract.) Pakistan Jour. Bot. 14(Special Issue): 45. 1982. [Among dissected pollinia, germination along incision in pollinial wall suggests wall is responsible for localized germination. KonzALova, M. Asclepiadaceae pollen in the Cheb Basin, Tertiary Czechoslovakia. Vestn. Ustred. Ustavu. Geol. 56(2): 79-85. 1981.* Kunze, H. Typology and morphogenesis of the angiosperm stamen. Beitr. Biol. Pflanzen 54: 239- 304. 1978 [1979].* Morphogenese und Synorganisation des Bestaubungsapparates einiger Ascle- piadaceen. (English summary.) /bid. 56: 133-170. 1981. a developmental sequences include those for stamens and coronas in Asclepi Maueswari Devi, H. Embryological studies in Asclepiadaceae. Pree. Indian Acad. Sci. 60: 52-65. 1964. [Reports of both nuclear and cellular endosperm formation in Calotropis procera need confirmation. ] Murphy, H. revision of the genus Fischeria (Asclepiadaceae). Syst. Bot. 11: 229- 241. 1986. [Considered to be monophyletic and distinct from Matelea, which is probably paraphyletic; see also SPELLMAN, 1975a.] Newron, L. E. Terminology of structures associated with pollinia of the Asclepiadaceae. Taxon 33: 619-621. 1984. NIsHINO, E. Corolla tube formation in the Tubiflorae and Gentianales. Bot. Mag. Tokyo 458 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 96: 223-243. 1983. [In corolla-tube formation Apocynaceae differ from other Gen- tianales investigated, including Cynanchum. ] No tan, J. R. Developmental morphology of the asclepiadaceous inflorescence. 177 pp. ets Ph.D. dissertation, Cornell Univ., Ithaca, New York. 1967.* (Diss. Abstr. B. 28(3): 802. 1967.) PEREIRA, J. F. Estudos em Asclepiadaceae, XI. Chave para determinacao dos géneros de Asclepiadaceae brasileiras e mais cultivadas no Brasil. (English summary.) Bol. Mus. Bot. Munic. Curitiba 42: 1-16. p/s. J-/]. 1980. [Key to genera represented by native and/or cultivated species. ] Perry, L. M. Gonolobus within the Gray’s Manual range. Rhodora 40: 281-287. pl. 494, 1938. [G. gonocarpos and several other species here included in Matelea.] RAGHAVENDRA, A. S., & V. 8. R. Das. The occurrence of C,-photosynthesis: a supple- mentary list of C, plants reported during late 1974—-mid 1977. Photosynthetica 12: 200-208. 1978. [Stapelia sp. listed.] RAMAKRISHNA, T. M., & D. A.. GovinpAppa. Studies of areoles in Asclepiadaceae and Periplocaceae. Curr. Sci. Bangalore 52: 1 145-1148. 1983. [Four main types described from 35 species; although said to be of no taxonomic value, only types 1 and 2 in Periplocoideae and Secamonoideae, all four in Asclepiadoideae.] Rao, V. S., & A. GANGULI. The floral anatomy of some Asclepiadaceae. Proc. Indian Acad. Sci. 57: 15-44. 1963. [Ovary semi-inferior in all species studied (two in Periplocoideae, ten in Asclepiadoideae).] RICHHARIA, R. H. The number of microsporangia in each stamen in Asclepiadaceae. Curr. Sci. Bangalore 2: 340-342. 1934. [Four in Periplocoideae, with tetrahedral or—in one genus—linear tetrads:; two in Asclepiadoideae, with linear tetrads; evi- dence of neither suppression nor fusion of sporangia in the latter ROBERTSON, C. Insect relations of certain asclepiads. I. Bot. Gaz. 12: 207-216. pl. 12. 1887a. IL. bid. 244-250. 1887b. [Details of pollination for nine southeastern species of ee ] Flowers and insects, Asclepiadaceae to Scrophulariaceae. Trans. Acad. Sci. St. tou 5: 569 598. 1891. [Insect visitors to seven southeastern species of Asclepias listed. ] RoeskeE, C. N., J. N. Serper, L. P. BRower, & C. M. Morritr. Milkweed cardenolides and their comparative processing by monarch butterflies (Danaus plexippus L.). Pp. 93-167 in J. W. WALLACE & R. L. Sete eds., Biochemical interaction between plants and insects. New York and London. 1976 SAFWAT, F. M. The floral morphology of Seeamon and the pan of the pollinating apparatus in Asclepiadaceae. Ann. Missouri Bot. Gard. 49: 95-129. 1962. SATA, S., S.C. GOYAL, & A. PILLAI. eae anatomy of some ee Proc. Indian Acad. Sci. Pl. Sci. 95: 147-152. 1985. [Unilacunar, two-trace cotyledonary node in species studied, including 4sclepias currasavica and Calotropis procera.] SCHAARSCHMIDT, F. The significance of lianas in Eocene forests of central Europe. (Abstract.) Proc. Int. Bot. Congr. 13: 197. 1981. [Leaves of lianous Asclepiadaceae presumably growing near border of Eocene lake.] SHINNERS, L. The species of Mate/ea (including Gonolobus) in north central Texas (Asclepiadaceae). Field Lab. 18: 73-78. 1950. [Includes key to genera of Asclepia- ss in region as Asclenindacese other than Asclepias. Sida 1: 358-367. 1964. ern ae = L. A revision of the genus Fischeria (Asclepiadaceae). 189 pp. Unpubl. Ph.D. dissertation, St. Louis Univ., St. Louis, Missouri. 1975a.* (Diss. Abstr. B. 36(6): 2599. 1975.) [See also MuRPHY.] . Asclepiadaceae. In: R. E. Woopson, Jr., & R. W. SCHERY, eds., Fl. Panama. Ann. Missouri Bot. Gard. 62: 103-156. 1975b. 1989] ROSATTI, APOCY NINEAE 459 & C.R.GuNN. Morrenia odorata and Araujia sericofera (Asclepiadaceae): weeds in citrus groves. Castanea 41: 139-148. 1976. SREEDEVI, P., & A. N. NAMBoopirRI. The germination of pollinium and the organization of germ furrow in some members of Asclepiadaceae. Canad. Jour. Bot. 60: 166- 172. 1982. [Position of the germination zone (furrow), in all cases discernible after acetolysis, ““genus specific” for Asclepias, Calotropis, and 12 other genera (eight of which not indicated).] Stevens, W. D. A revision of Matelea subg. me acon (Apocynaceae, sensu lato). 223 pp. Unpubl. Ph.D. dissertation, Michigan State Univ., East Lansing. 1976. (Diss. Abstr. B. 37(2): 587. 1976.) [Higher ae i discuss ed.] SUNDELL, E. The Seer tribal, and subtribal nomenclature of the Asclepiadaceae. Taxon 29: 257-265. SWARUPANANDAN, K. i illegitimate Pergularia and the tribal names Ceropegieae, Marsdenieae and Stapelieae (Asclepiadaceae). Taxon 32: 468-471. 1983. Wanntorp, H.-E. Ca/lotropis gigantea (Asclepiadaceae) and Xylocopa tenuiscapa (Hy- menoptera, Apidae). Studies in flower morphology and pollination biology. Sv. Bot. Tidskr. 68: 25-32. 1974. seas structure thought to be specifically adapted for pollination by carpenter bees WarbD, D. B., ed. Plants. Vol. 5 in cP. C. H. PrircHARD, series ed., Rare and endangered biota of Florida. xxix + 175 pp. Gainesville, Florida. 1979. [Asclepias Curtissli, by . WarRD, threatened; Matelea alabamensis, by D. J. DRAPALIK, endangered; M. Baldwyniana, ee . J. DRAPALIK, rare. Wuirte, A. C., & B. L. SLOANE. The Stapelieae. 3 vols. 1186 pp. Pasadena, California. 1937. [Each volume with separately paginated index; 20 genera of succulents treated, 1233 figures and 39 plates include photographs and/or illustrations of nearly all species. Wuitinc, A. G. A summary of the literature on milkweeds (Asclepias spp.) and their utilization. U. S. Dep. Agr. Bibliogr. Bull. 2: 1-41. 1943. [Extensive review of topics including the production of fiber, rubber, paper, cellulose, and oil from several ies. Woopson, R. E., Jk. The North American oe Perspective of the genera. Ann. Missouri Bot. Gard. 28: 193-244. KEY TO THE GENERA OF ASCLEPIADACEAE IN THE SOUTHEASTERN UNITED STATES General characters: subshrubs, vines, or herbs with latex; leaves simple, opposite, some- times with colleters; inflorescences terminal but often appearing interpetiolar, cvmose but often appearing umbellate; flowers perfect, actinomorphic or nearly so, 5-merous except carpels 2; sepals nearly free, often with squamellae, corollas variously ee often appendaged adaxially (see filaments), lobes often overlapping; stamens cy ee US, he ternating with corolla lobes; filaments connate, forming a column around o styles, bearing abaxially 1 or 2 coronas; outer corona (when eed ae Gace basally adnate to corolla, often apically free; inner corona (when present) comprising 5 free or connate, laminate, conduplicate, vesicular, or rarely digitate segments, each some- times appendaged adaxially; anthers connate into anther head around style-stigma head, subtended by inner corona, ed corona, or corolla; pollen released in pollinia, 2 from adjacent anthers united by 2 translator arms and I corpusculum into I pollinarium; nectaries within stigmatic se ee alternating with anthers on Style-stigma head; carpels free in ovule-bearing and stylar regions, united apically into style-stigma head; ovaries superior to subinferior, simple, unilocular, with marginal placentae; ovules usually nu- merous in each carpel; fruits follicles; seeds each with a tuft of trichomes at apex. 460 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 A. Pollen assis as tetrads; corpuscula absent B. In scences several- to many-flowere red: corolla rotate, diameter at apex 2-3 cm: “follicles long, narrow, subterete, divergent less than 45° at base, usually erect. orolla funnelform, diameter at apex 5—8 cm; follicles short, stout, angled, divergent ca. 180° at base, usually horizontal. ........... joate ab deb Gaeevets aenaain de Ween yeep nreted wanes, & ‘ryptostegia grandiflora.*®] A. Pollen released in pollinia; corpuscula present. rpuscula exceeding style-stigma hea translator arms more or less horizontal, pollinia horizontal. Anthers lacking abaxial appendages; outer corona absent; inner corona lam- inate to fleshy, mostly erect, free from corolla, with 15 nesta 5 shallow, obtuse, single, and alternating with 5 longer or cute pairs), often with 5 adaxial, adnate, ligulate appendages (opposite each cee lobe), exceeding corpuscula (corona in M. alabamensis intermediate between inner and outer in position, orientation, fusion to corolla, and a logy). =< = 3 o SS D. Each anther with an abaxial, fleshy to laminate appendage; outer corona fleshy, mostly ae adnate to corolla but free apically (forming a mostly spread- ing annulus), entire or with 10 lobes (each very shallow and obtuse), unap- pendaged adaxially, not exceeding corpuscula; inner corona ee heen ds Geelane 4. danse Ws eee eb a ek Gd a Ae onolobus. C. Corpuscula not exceeding or (in Morrenia) exceeding style-stigma ee panaefor arms more or less pendulous, pollinia pendulou E. Inner corona of 5 distinct, vesicular segments; ‘outer corona laminate, spread- ing, adnate to corolla but free apically (forming an erect or ascending annulus), es eee ih ore Sawa eed ee een se eee ey ek Sarcostemma. E. Inner corona of 5 distinct or basally connate, often conduplicate (and some times apically enclosed) but not vesicular segments; outer corona absent. F. Segments of inner corona either digitate and distinct (in C. Northropiae) or laminate, flat, and distinct (but basally adnate to corolla in C. Blodgettii) or basally connate (in C. scoparium and Morrenia), without adaxial, adnate appendages; stems twining. x, Segments of inner corona distinct or (in C. scoparium) connate to level well beneath that of style-stigma head apex (then corolla lobes less than 2 mm long, leaves nna) glabrous adaxially; anther head im- mediately subtended by base of inner corona (although eae - nate to corolla in C. Blodgettii); bases of leaf blades attenuate, Cate, Or COrdate.. coe oa ogy yeep ened Cees dwawes 4. Gdn. Segments of inner corona connate to or above level of style-stigma head apex (and corolla lobes greater than 8 mm long, leaves broad), with nonglandular pubescence adaxially; anther head held well above base of inner corona; bases of leaf blades hastate or less often cordate. Hi ative Re toh tye aerate’ aid auntie an Gescn te onkcd ld duch tue arbe ict domain ae ah 3. Morrenia. pes of inner corona fleshy, conduplicate, and distinct, usually with adaxia Dp y opened; stems pros- trate to erect, not twining. dk iden See AGH ee Hebe ee 6. Asclepias. oo o scaped from or persistent beyond cultivation but clearly not reproducing either sexually or seeenally in the southeastern United States. For further discussion, see family discussion 1989] ROSATTI, APOCYNINEAE 461 Subfam. ASCLEPIADOIDEAE [R. Brown ex Endlicher, Gen. Pl. 589. 1838, ‘“‘Asclepiadeae verae.”’ Tribe GONOLOBEAE G. Don, Gen. Hist. Dichlamydeous Pl. 4: 107, 136. [18377] 1838. 1. Matelea Aublet, Hist. Pl. Guiane 1: 277; 3: pl. 109, fig. 1. 1775. Terrestrial, scandent and/or twining or (in M. cynanchoides, M. pubiflora) prostrate to suberect, perennial herbs (usually woody toward base) or subshrubs with white latex; axes and leaves with often dense nonglandular and (usually much shorter) glandular pubescence. Leaves opposite; blades ovate, elliptic, lanceolate, or sometimes orbiculate, especially near base of stem, apices acute, acuminate, or sometimes obtuse, rarely emarginate, margins sometimes revo- lute, bases cordate or obtusely sagittate; petioles distinct, without lateral ap- pendages, bases and/or adjacent areas of stems usually without colleters. In- florescences pedunculate or (in M. cynanchoides and M. pubiflora) sessile, 2- (often in M. cynanchoides and M. pubiflora) to several-flowered, cymose but appearing umbellate, racemose, corymbose, or paniculate, sometimes with both flowers and pedicel scars present simultaneously, bracteate; pedicels longer or (often in M. cynanchoides and M. pubiflora) shorter than calyces, ebracteolate. Calyx about 74 as long as corolla or less, free nearly to base, usually with both nonglandular and glandular pubescence abaxially, glabrous adaxially; lobes acute, bases imbricate or sometimes valvate, usually with minute squamellae (usually | near each sinus). Corolla much shorter than leaves, free nearly to base to united throughout the basal *4, rotate to campanulate, often glabrous or (in M. pubiflora) with nonglandular pubescence adaxially, with nonglandular and sometimes glandular pubescence or sometimes nearly glabrous abaxially; lobes valvate or sometimes imbricate, erect, ascending, spreading, or reflexed. Outer corona absent [or, reportedly, sometimes present]. Inner corona exceed- ing or (in M. alabamensis) exceeded by anther head, sessile, cyathiform (seg- ments connate about to level of style-stigma head) and more or less erect (mostly concealing anther head and style-stigma head) and free from corolla or (in M. alabamensis) discoid and spreading, free from but in contact with corolla [and/ or outer corona]; lobes 5, laminate, each with (or, in M. cynanchoides, often without) 2 lateral, shorter to longer projections, and with or without a single, adaxial, adnate, shorter or sometimes equally long appendage (especially the latter giving the appearance of a lobe apex tangentially split into adaxial and abaxial flaps or tongues). Anther head short-obconic, immediately subtended by base of inner corona; anther wing slots directed outward but mostly down- ward. Anthers each with an apical, adaxial, hyaline flap inflexed over style- stigma head, unappendaged abaxially. Style-stigma head short-obconic, with flat or depressed, distinctly 5-sided apex. Corpuscula clearly exserted from anther head; translator arms mostly horizontal; pollinia horizontal, each with sterile, hyaline tissue adjacent to attached end. Fruits follicular [unknown for many species], | (or, reportedly, rarely 2) from each flower, lance-ovoid, nar- rowly fusiform to fusiform-ovoid, acute to attenuate apically, dehiscent along 462 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 U Matelea. a-j, /. carolinensis: a, Petes with none and terminal inflores- cence appearing interpetiolar by sympodial growth, x |; b, inflorescence, with pedicel scars and corolla lobes all overlapping to left (ey seen ‘in only Ss, uppermost and rightmost), x 1; c, flower bud, with corolla lobes all overlapping iA eae x 3; d flower at anthesis, ar corolla lobes all see to right, x 2; e, flow above without calyx and corolla (5 principal lobes of corona [erect, i.e., perpendic ae to page] opposite 5 adaxial ligules [arching over style- ~suigma head; solid outline, 2 clearly 2-lobe andS5a apica al. adaxial. I style -su igma re pedal alternating with 5 pairs of lateral projections [more or less erect] and 5 corpuscul corona: 2 panopal lobes, of >, and 2 oo uae of 5, entire, 1 of each in section, 0 a princip ne oS lateral projections exceeding principal lobes), areas of fusion between segments I etween paired lateral aed leftmost and rightmost) and between segments 1989] ROSATTI, APOCYNINEAE 463 adaxial suture, erect on erect pedicels to pendulous on pendulous ones, some- times erect on deflexed pedicels(?); surface with many or, rarely, a few tubercles [or shallowly winged lontigudinally], usually with short nonglandular and glan- ular pubescence. Seeds many, ovate-lenticular, each with a tuft of trichomes at apex; seed-coat surface papillate or smooth, lateral wings crenate-undulate or entire. (Including Cyclodon Small, 1933; Edisonia Small, 1933; and Odon- tostephana E. J. Alex. in Small, 1933. Excluding, among others, Gonolobus Michx., 1803.) Type species: M. palustris Aublet; although Aublet originally included more than one species in Matelea, pl. 109, fig. 1, an illustration of M. palustris, was cited after the name of the genus. (Name of unknown etymolo- gy.)— CLIMBING MILKWEED, SPINY-POD. Recognized here as distinct from Gonolobus Michx. (for reasons summarized in the discussion of that genus, and below), Matelea includes a total of 180 to 200 species distributed from the central and southern parts of the United States into South America, according to Spellman (1975b), who treated nine species in Panama. Matelea had been considered by most botanists (e.g., Schumann) to comprise less than ten species in tropical South America until Woodson expanded it to encompass North and South American plants to which many other generic names had been applied. In his revision of subg. DicTyANTHUS (Dcne.) Woodson, Stevens indicated that even after the recognition of a few genera of Gonolobeae in pea to those favored by Woodson, nearly 200 species of Matelea would remain. Standley & Williams ne Woodson’s circumscription of Matelea, but not oahu certain misgivings. They thought that species in Mexico and Central America that had been placed in the genus in fact represent several genera, and that (p. 450) some of them are “‘less closely related . . . than certain other” species of Matelea are to members of Gonolobus and Fischeria. Woodson recognized both Matelea and Fischeria but considered them to be closely re- lated. Murphy also maintained them as distinct, stating that Mateleais probably paraphyletic and that uniquely derived character states within Fischeria suggest it is monophyletic. Woodson retained formal taxonomic status for some of the genera he placed in Matelea merely because the names were familiar. Mostly on the basis of corollas, coronas, and pollinia, he defined a total of 16 subgenera and six sections, at least some of which were “‘not entirely satisfactory,” according to Stevens (p. 71 d column (i.e igul tleft and right near center, Sa Sorsaer tre oot stigmatic eee not ce x 12; h, androecium without corona and g um without ovule-bearing region (2 corpuscula, partly exceeding anther head, pes pei aad at summit of 2 outwardly projecting slots Doha anthers and 2 stigmatic chambers [column unmarked; openings in column to stigmat scenes stippled eran. x 12; i, structures shown in “h,”’ seen from below (comic not visible; 5 0 utwardly projecting each of 2 adjacent anthers [sterile area near attachment of translator arms unmarked], 2 translator arms, and corpusculum with s ce ce eS aoved for attachment to insects; Des and translator arms more or less horizontal before removal from anthers]), x 464 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Despite recognition ofa large number of infrageneric taxa, Woodson included all but one of our species, in addition to a number of others, in subg. CHTHAMA- LIA (Dene.) Woodson, which he did not further divide and which he considered to be characteristic of the genus in the southern United States and northern Mexico; Matelea alabamensis (Vail) Woodson was placed in subg. Eumatelea (= subg. MATELEA), in the “Reticulatae” (name not validly published). These subgenera are distinguished in Woodson’s key (pp. 220-222) primarily by characters of the pollinia (“‘subquadrate- or oblong-reniform, with a narrow hyaline margin” in subg. CHTHAMALIA; “subtriangular-pyriform, with a con- spicuous hyaline indentation or margin” in subg. MATELEA and three others). In a biosystematic study of Matelea in an area comparable to and perhaps equaling ours, Drapalik recognized nine species, one of which, M. gonocarpa (Walter) Shinners, was included by Woodson—and is included here—in Gono- lobus (as G. gonocarpus (Walter) Perry). Drapalik agreed more closely with Small’s treatment (1933) of these plants, in the sense that among them he recognized four groups of species instead of the three favored by Woodson, although he did not think, as Small did, that they should be recognized at the generic level. Thus, he considered M. gonocarpa, M. alabamensis (Cyclodon alabamense (Vail) Small), and M. pubiflora(Dene.) Woodson (Edisonia pubiflo- ra (Dene.) Small) to be, at least morphologically, unrelated to one another or 0 a group of six closely related species Small (1933) had included in Odon- tostephana E. J. Alex. Comprising what Drapalik called the “Odontostephana group” were M. Baldwyniana (Sweet) Woodson, M. carolinensis (Jacq.) Wood- son, M. decipiens (E. J. Alex.) Woodson, M. flavidula (Chapman) Woodson, M. floridana (Vail) Woodson, and M. obliqua (Jacq.) Woodson (in which Dra- palik included M. Shortii (Gray) Woodson). The species concepts of Drapalik appear to be sound and are adopted here. I recognize Matelea as distinct from and less specialized than Gonolobus because of the absence in the former and presence in the latter of abaxial anther appendages. Despite this fact and Drapalik’s contention that there are four morphologically unrelated groups involved (even though he included them all in the same genus), plants in our area suggest to me a close link between Matelea and Gonolobus. ere appear to be four principal types of androecial appendages in the Asclepiadaceae, at least as the family is represented in the Southeast: an outer, discoid, spreading corona that is adnate to the corolla and sometimes shallowly lobed (i.e., the “faucal corona” or “faucal annulus” of many authors, charac- teristic of, for example, Gonolobus gonocarpus); an inner, erect or ascending corona with five principal, distinct or, less often, connate segments (e.g., the hoods of Asclepias, the vesicular segments of Sarcostemma, the five-lobed, cuplike structures characterizing the “Odontostephana group” in Matelea); a whorl of five structures adnate to the adaxial surfaces of the inner corona (e.g., the horns of Asc/epias, the ligules of Matelea); and abaxial anther appendages (as seen in Gonolobus). There are, in addition, structures in M. alabamensis and M. pubiflora that do not entirely fit into any of the foregoing categories and are therefore of particular interest. That the anther appendages characterizing Gonolobus represent abaxial ad- 1989] ROSATTI, APOCYNINEAE 465 nations to the stamens of structures homologous to the adaxial ligules of Ma- telea (also considered in the discussion of Gonolobus) was rejected by Woodson because of the widespread occurrence in Gonolobus of both. Nevertheless, in G. gonocarpus the anther appendages are accompanied only by an outer corona; there is evidence neither of an inner corona nor of ligules. With regard to androccial appendages in general, Matelea alabamensis ap- pears to be intermediate in three major respects between Gonolobus gonocarpus and members of the ““Odontostephana group,” which have an inner, cuplike corona with ligules but lack an outer corona. Matelea alabamensis approaches G. gonocarpus in having a corona that is discoid, instead of cuplike; that is, more like an outer and less like an inner corona. Furthermore, to the extent to which it is adnate to the corolla, the corona of M. alabamensis is intermediate between that of G. gonocarpus, which is adnate above the point of filament insertion and is therefore a true outer corona, and those of species in the ‘“‘Odontostephana group,” which are not adnate above this point and are there- fore true inner coronas. Moreover, the corona of M. alabamensis bears ligules that are, relative to those of the ‘““Odontostephana group,” less extensively adnate to the corona and more closely associated with the androecium, a disposition that has implications regarding the possibly homologous relation- ship between the ligules of Matelea and the anther appendages of Gonolobus (but see below). Other morphological evidence also suggests a relationship between Matelea alabamensis and Gonolobus gonocarpus. As indicated in photomicrographs provided by Perry (p/. 494) and drawings included by Drapalik (fig. 7), the pollinaria of these two species are similar in size and morphology, yet they are much larger and of considerably different shape than those of M. Baldwyniana, M. carolinensis, M. decipiens, M. obliqua, and M. pubiflora, which are among themselves also fairly uniform in these features. In addition, the apex of the style-stigma head tends to be larger and more clearly pentagonal in M. ala- bamensis and G. gonocarpus than in our other species of Matelea. It is notable that Matelea alabamensis is reportedly quite rare and has not been much collected. Outside of his own specimens, only those obtained by Harbison (on which Vail based the protologue) and by Thorne were known to Drapalik, who indicated after intensive searching that the plants are probably restricted to a few populations near the Apalachicola River in Florida, where the species is considered endangered (Drapalik in Ward), and along the same river in southeastern Alabama and southwestern Georgia (in these places called the Chattahoochee). Matelea pubiflora is another species Drapalik considered to be isolated from the others he treated, primarily because of its prostrate habit, sessile inflores- cences, and campanulate corollas, however, it too suggests a link between Matelea and Gonolobus. The corona in M. pubiflora is similar to that of species in the “Odontostephana group” in being more or less erect, connate around the anther head, not subtended by an outer corona, and not adnate to the corolla. However, it differs in having small, inconspicuous lateral projections on each of the five principal lobes instead of very prominent ones, thus ap- proaching the condition of the discoid, essentially entire outer corona of G. 466 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 gonocarpus (and M. alabamensis). The corona of M. pubiflora is also similar at of G. gonocarpus (and M. alabamensis) in that it is not entirely erect and is in contact with the corolla. However, the contact in M. pubiflora is at least as much due to the campanulate shape of the corolla as to the suberect orientation of the corona; the two structures clearly are not adnate. Other similarities between Matelea pubiflora and Gonolobus gonocarpus are evident. Whereas ligules are associated with the inner corona in several species of Matelea (including M. alabamensis), such structures evidently occur neither on the inner corona of M. pubiflora nor on the outer one of G. gonocarpus. Thus, it is perhaps significant that according to illustrations provided by Dra- palik (fig. 6), the lateral seed-coat wings of M. pubiflora and G. gonocarpus are crenate and undulate, while those of the other two species depicted, M. ala- bamensis and M. Baldwyniana, are nearly entire and flat According to Drapalik, Matelea pubiflora is most closely allied to species comprising prostrate plants that occur farther west than the area covered in his investigations. In fact, one of these, 4. cynanchoides (Engelm.) Woodson, has been reported since Drapalik’s work from the southeastern United States as we define the region, from Miller County in southwestern Arkansas (Smith), and indeed appears to be related to M. pubiflora. Both species are characterized by a prostrate habit, sessile inflorescences, and campanulate corollas, in ad- dition to very small or sometimes nonexistent lateral projections on each of the five principal lobes of a suberect inner corona. That the inner corona bears very prominent Pee in ne cynanchoides but is without such structures in M. pubiflora (see al e) suggests either that the resemblance between these two species 1s a aipeaeial or that the taxonomic significance of this character lies at or below the level of species and has been overestimated (see also below). Matelea pubiflora and M. cynanchoides also have smaller leaves than our other species, probably reflecting the drier habitats in which they occur. Dra- palik noted that the former is found in more xeric conditions than the other species he treated, and M. cynanchoides is known from sandy areas mostly in open woodlands (Correll & Johnston). Whereas M. pubiflora is rare and en- demic to the Coastal Plain in Georgia and Florida (Drapalik; see also Anderson), M. cynanchoides is frequent in northern and central Texas (Shinners, 1964) and is also known from Oklahoma (Correll & Johnston) and, as indicated above, southwestern Arkansas. Our six remaining species of Matelea, the ‘““Odontostephana group” of Dra- palik, are indeed closely related. This is an assemblage of scandent and twining vines with large, cordate leaves; essentially rotate corollas; no outer coronas: cyathiform, mostly erect inner coronas with five principal lobes, each with lateral projections and sometimes adaxial (or possibly abaxial, see below) ligules as well; relatively small, cylindrical style-stigma heads; and tuberculate follicles. Efforts to reveal the true nature of the ligule sometimes present on the inner corona of members of the ““Odontostephana group” probably would consid- erably enhance our understanding of the relationships within Matelea and perhaps between this genus and Gonolobus as well. According to Drapalik (p. 19), in the six species of this group (and in M. pubiflora) “the corona lobes may also have additional small adaxial and abaxial lips, protrusions, or in- 1989] ROSATTI, APOCYNINEAE 467 dentations.”’ Clearly, if these structures were in some cases truly abaxial, it would be more difficult to argue that they are homologous to the anther ap- pendages in Gonolobus than if they were in all cases truly adaxial. In fact, my own observations suggest that the latter interpretation is more tenable. In most cases 1n the ““Odontostephana group,” the ligules clearly arise from the adaxial surfaces of the inner corona and are exceeded by its principal lobes. However, in many instances the principal lobes instead appear to be apically split into equal adaxial and abaxial halves (see FIGURE 6g); in such cases the former could easily be interpreted as an adaxial ligule that equals the principal lobe. In some instances, and especially frequently in Matelea floridana, the smaller, shorter structure (1.e., the apparent ligule) is the outermost, suggesting an abaxial adnation on the principal lobe. In fact, in M. floridana (and perhaps other species) the outer structure appears to be the appendage (1.e., ligule) when the corona is viewed from the adaxial side but is just as easily seen as the principal lobe when the corona is observed from the abaxial side. In view of the foregoing, it seems most reasonable to interpret this as yet another adaxial ligule, but one longer than the principal lobe. Even if the ligules were in all cases truly adaxial, their taxonomic importance (and the likelihood that they are homologous to the anther appendages of Gonolobus) is greatly diminished by the degree to which their presence, absence, and size relative to the principal lobes are variable both among and within our six species of the ‘““Odontostephana group.”’ For example, each species and in some cases individual flowers have both ligulate and nonligulate principal lobes. In addition, whereas the ligules were shown to exceed the principal lobes in perhaps 13 of 17 illustrations of Matelea floridana provided by Drapalik (fig. 2), my own examination of three herbarium specimens evidently belonging to this species (as determined by Drapalik and by me, using his key, but see below) revealed this condition on only one. Although in 12 of 13 drawings of M7. carolinensis (Drapalik’s fig. 2) the ligules were exceeded by the principal lobes, among four herbarium specimens I saw they much exceeded them (1.e., the principal lobes appeared as abaxial ridges) on one and were about equal to them (i.e., aS in FiGURE 6g) on two others. It seems likely that with enough observations each of the conditions described above would be represented in each of the six species. The foregoing examples involving closely related species (1.e., Matelea cy- nanchoides and M. pubiflora on one hand, and those in the ‘““Odontostephana group” on the other) suggest that the ligule is of limited or no taxonomic value in Matelea. Indeed, although Drapalik documented with illustrations much of the variation he observed, the character was used in neither the key nor the formal descriptions he provided. In contrast, the abaxial anther appendages invariably absent in Matelea and present in Gonolobus appear to represent a much more stable character, one by which these genera can consistently be distinguished. The differences in stability between the ligules of Matelea and the abaxial anther appendages in Gonolobus imply genetic differences, which in turn suggest that the structures are not strictly homologous. The closely related species in the “Odontostephana group” differ mainly in aspects of the corolla and corona that are often difficult to discern on living 468 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 plants and are nearly impossible to observe in herbarium material. These include, among others, color of corolla and corona, size and shape of corolla lobes, and shape and size of the lateral projections relative to the principal lobes of the inner corona. As Drapalik documented extensively, the shapes and relative sizes of these structures, particularly the lateral projections, are fre- quently and irreversibly distorted by drying. He therefore stated that the correct assignment to species of herbarium specimens is frequently impossible and that liquid-preserved flowers should be included by collectors when possible. I have frequently observed in herbarium material both color distortion and the loss of significant portions of the lateral projections, the latter evidently due to shriveling in life and/or upon drying. Matelea floridana supposedly differs from the other species in the ““Odon- tostephana group” in that the principal lobes of the inner corona usually exceed the lateral projections. According to Drapalik, the species is rare and endemic to the Coastal Plain in northern Florida, although the possibility that it may occur in southeastern Georgia should be investigated because of the presence there of open or dense oak-hickory (Quercus-Carya) mixed hardwood forests. Wunderlin and colleagues have recently reported the species from central Flor- ida. Drapalik reported that M7. floridana occurs in shadier conditions than do other members of this group, and the later flowering observed among plants of this species in nature was maintained in common-garden experiments. The lateral projections usually equal or exceed the principal lobes in the remaining five species in the ““Odontostephana group.” Two of these supposedly differ from the rest in having rotate-reflexed or rotate instead of rotate-as- cending corollas; Matelea carolinensis and M. flavidula are very similar but can be distinguished on the basis of corolla-lobe color (usually dark maroon vs. some shade of green or rarely light maroon, respectively) and the relative lengths of the principal corona lobes and the lateral projections (the lateral projections always longer in M. carolinensis, see FiGure 6f, g; the structures usually about equally long in M. flavidula). Both species are evidently infrequent and occur in open or second-growth oak-hickory forests, but M. carolinensis appears to have a wider geographic range (see, however, Wunderlin ef ai.). Whereas this species occurs in various floristic provinces from Delaware to Texas, M. flavidula is restricted to the Coastal Plain in South Carolina, Georgia, Florida, and Alabama. Matelea Baldwyniana”’ usually differs from the other two species with rotate- ascending coronas (M. obliqua and M. decipiens) in its white corollas; flowers with cream-colored corollas also occur in M. Baldwyniana and, rarely, M. obliqua, but the lateral projections of the corona lobes are usually much longer in M. Baldwyniana. The corollas of M. obliqua are otherwise normally rose colored, light maroon, or infrequently dark maroon or green, while those of M. decipiens are always dark maroon; in addition, the corolla lobes of M. obliqua are usually four to six times as long as wide, whereas those of MV. decipiens are nearly always stouter. “According to Drapalik, application of 1 | modern usage; type material has not been found. and the species is not now known from the type locality, near Savannah, Georgia. 1989] ROSATTI, APOCYNINEAE 469 Matelea Baldwyniana is rather common in northwestern Arkansas and ad- jacent parts of Missouri, particularly in places close to springs, but it is rare east of the Mississippi River; in addition to Arkansas, the species is known in our area only from Alabama and Florida. Matelea obliqua and M. decipiens are infrequent over large areas in the eastern United States: 4. obliqua ap- proximately from Georgia north to Missouri and Pennsylvania; M. decipiens from Georgia north to Maryland and west to Texas. Drapalik reported that for most species of Matelea included in his investi- gation, vegetative growth 1n nature was greatest in either exposed or partially shaded conditions, but that in garden experiments partial shade was optimal for M. carolinensis, M. decipiens, and M. obliqua. Flowering was not observed among the few individuals (species not indicated, probably indeterminable) found in places with northern exposures. It is thought to occur under dense shade only in M. alabamensis and M. floridana (as well as G. gonocarpus), which flower under very exposed conditions as well. All species of Mate/lea that Drapalik studied are evidently best adapted to well-drained soils, and none is found in areas subject to inundation. Matelea pubiflora and M. cynanchoides occur in drier conditions than the other species; according to Drapalik, M. pubiflora (he did not study M. cynanchoides) has especially large taproots and is restricted to communities dominated by turkey oak (Quercus laevis Walter) and long-leaf pine (Pinus palustris Miller). With the exception of M. cynanchoides, our species of Matelea are usually found in or at the edges of forests dominated by oaks, hickories, and other hardwoods. Drapalik also reported the pH of soil samples obtained within 15 cm of the surface at each of his study sites. All of the samples associated with Matelea Baldwyniana and M. obliqua and most of those collected with M. carolinensis and M. decipiens were basic to circumneutral. In contrast, a majority of those obtained for M. alabamensis, M. flavidula, M. floridana, and M. pubiflora were acidic, although Drapalik reported growing plants of all four species from seed to flowering in pH-neutral substrates. Drapalik reported natural pollination involving plants both in and out of cultivation for each of the species included in his study except Matelea pubi- flora; observation during daylight hours revealed no insect visitors to flowers of this species, although numerous fruits were produced by plants in cultivation. Visitors with attached pollinaria were in all cases small flies belonging to the Anthomyiidae, Chloropidae, Milichiidae, and Phoridae. Indeed, the flowers were said to emit the kinds of odors (1.e., unpleasant, at least to most humans) and in many cases to have the colors (i.e., reds, maroons) to which such insects are attracted. The flies have been observed to effect pollination while feeding on nectar produced by the flowers. The details of pollination as described by Drapalik are very much like those presented here for Asclepias, with a few notable exceptions. Although nectar is produced in the stigmatic chambers in both genera, in Mavelea it accumulates at the base of the inner corona immediately outside the base of the outwardly projecting slot, instead of in the hoods at either side. In many but not all cases involving species of Matelea, pollinator parts appear not to be wedged into the corpusculum, as they evidently always are in Asc/epias, but instead to adhere 470 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 to it even though corpuscular secretions have not been noted. In Matelea, in contrast to Asclepias, the anther wings are much smaller and more rigid and the pollinia are larger than the slot opening, so that pollinia are not and need not be inserted completely but merely placed against the slot. Nevertheless, as in Asclepias, the convex side must be adjacent to the opening for successful pollen-tube growth. In nature all of our species of Matelea appear to be reproductively isolated, with the exception of M. carolinensis and M. decipiens of the ““Odontostephana roup.” Drapalik reported populations of these two species and various inter- mediates between them throughout Georgia and the Carolinas. He specifically mentioned genetically unmixed populations of M. carolinensis in Tennessee, Maryland, and Virginia, and of M. decipiens in Louisiana and Missouri, but did not address the issue for the remaining geographic areas in which these species occur. Drapalik attributed the infrequency of interspecific hybridization to various factors, including temporal, geographic, ecological, and mechanical isolation. Evidence suggested that pollinator specificity was of limited impor- tance in this regard. Breeding experiments conducted by Drapalik in cultivation using the six species belonging to the ““Odontostephana group” indicated that each is self- incompatible (a gametophytic system was assumed) but revealed little about the relationships among them. Artificial cross pollinations involving six of the 15 possible species pairs were successful in the sense that mature fruits and viable seeds were produced. Although fruit set among plants in cultivation was greater than that observed in nature (6.10 vs. 0.22 fruits per plant), the overall percentage of successful interspecific pollinations among them was low (6.84 percent). This observation, together with the fact that successful artificial crosses between Matelea carolinensis and M. decipiens were no more frequent than those between species not known to hybridize in nature, diminishes the reli- ability of these results. Nevertheless, it is tempting and perhaps reasonable to suggest that additional attempts might eventually result in successful crosses between all pairs of species in this group. This is especially so because the percentage of successful pollinations between species was equal to that observed between individuals from different populations of the same species, indicating that interfertility between and within species might be approximately equal. A lower rate of success resulted from pollinations within single populations of a given species, presumably due to self-incompatibility and a greater genetic similarity among the individuals. As expected, an even lower success rate was obtained from self-pollinations. Compared to the parents, the F, hybrid plants resulting from artificial pol- linations were intermediate in flower color and morphology (Drapalik). In addition, the hybrids were as fertile as the parents, as determined by pollen stainability, the number of seeds per fruit, and seed viability. Drapalik provided helpful illustrations of coronas representing the parents and progeny involved in several of the interspecific crosses. Unfortunately, this documentation does not elucidate the genetic basis of ligule variation. Matelea carolinensis is cultivated in North America with sufficient frequency 1989] ROSATTI, APOCYNINEAE 47] to have been treated by Bailey and colleagues. Drapalik provided instructions for growing all species found in the southeastern United States, both from seed and by transplanting mature plants. REFERENCES: Under references for the Apocynineae, see ANDERSON; BAILEY ef ie CORRELL & JOHNSTON; SCHUMANN; SMALL (1933); SmitH; and STANDLEY & WILLIAM Under references for the Asclepiadaceae, see DRAPALIK; MURPHY; PERRY; SHINNERS (1964); SPELLMAN (1975b),; STEVENS; WARD; and WoopDsoNn. VaiL, A. M. Studies in the Asclepiadaceae.— VII. A new species of Vincetoxicum from Alabama. Bull. Torrey Bot. Club 30: 178, 179. pis. 9, 10. 1903. [= M. alabamensis.] WUNDERLIN, R. P., B. F. HANSEN, & D. W. HALL. The vascular flora of central Florida: taxonomic and nomenclatural changes, additional taxa, II. Sida 13: 83-91. 1988. [According to annotations by Drapa.ik, M. floridana, but not the previously re- ported M. carolinensis, present. . Gonolobus Michaux, Fl. Bor. Am. 1: 119. 1803. No Terrestrial, scandent and/or twining [or sometimes prostrate], perennial herbs (usually woody toward base) or subshrubs with white latex; axes and leaves nearly glabrous or with usually sparse nonglandular and sometimes (usually much shorter) glandular pubescence. Leaves opposite; blades ovate, elliptic, or sometimes suborbiculate especially near base of stem, apices acute, acuminate, or subcuspidate, sometimes obtuse, rarely emarginate, margins not revolute, bases cordate or obtusely sagittate [sometimes obtuse or cuneate]; petioles distinct, without lateral appendages, bases and/or adjacent areas of stems usu- ally without colleters. Inflorescences pedunculate [or subsessile], few- to several- flowered, cymose but appearing umbellate, racemose, or corymbose, bracteate; pedicels longer than calyces, ebracteolate. Calyx about '4 as long as corolla or less, free nearly to base, sometimes with nonglandular pubescence abaxially, glabrous adaxially; lobes acute, bases valvate, with minute squamellae (usually 1 near each sinus). Corolla much shorter than leaves, free nearly to base to united throughout the basal 4, rotate [to subcampanulate], usually with at least some degree of nonglandular pubescence adaxially, glabrous abaxially; lobes valvate, spreading to reflexed [or sometimes erect to ascending]. Outer corona discoid, fleshy, crenate-undulate, basally adnate to the corolla tube but apically free, forming a mostly spreading [or erect] annulus. Inner corona absent [or reportedly exceeded by anther head, substipitate, essentially discoid and spread- ing, free from but in contact with outer corona and/or corolla; lobes 5, laminate but sometimes revolute, unappendaged]. Anther head short-obconic, imme- diately subtended by base of outer [or inner] corona; anther wing slots directed outward but mostly downward. Anthers each with an apical, adaxial, hyaline flap inflexed over style-stigma head, and an apical, abaxial, fleshy, appendage projecting outward. Style-stigma head short-obconic, with flat or depressed, distinctly 5-sided apex. Corpuscula clearly exserted from anther head; translator arms mostly horizontal; pollinia horizontal, each with sterile, hyaline tissue adjacent to attached end. Fruits follicular [unknown for many species], | from 472 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 each flower, lance-ovoid, narrowly fusiform to fusiform-ovoid, acute to atten- uate apically, dehiscent along adaxial suture, erect on erect pedicels to pen- dulous on pendulous ones, sometimes erect on deflexed pedicels; surface winged longitudinally [or sometimes reportedly with tubercles], usually with sparse, nonglandular pubescence. Seeds many, ovate-lenticular, each with a tuft of trichomes at apex; seed-coat surface papillate or smooth, lateral wings crenate- undulate. (Vincetoxicum Walter, 1788, nom. illeg., not Von Wolf, 1776, not Moench, 1794.) LecroTyPE SPECIES: G. gonocarpus (Walter) Perry*® (Vincetox- icum gonocarpum Walter); Perry (Rhodora 40: 283. 1938) selected G. macro- phyllus Michx. as the “‘standard-species,”’ a name, as she correctly recognized, invalidated by Michaux’s citation in synonymy of the earlier V. gonocarpum. (Name from Greek, gonia, angle, and /obos, pod, evidently in reference to the longitudinally winged (i.e., angled in transverse section) follicles of the first (and the type) but not the other two species originally included by Michaux.)— ANGLE-POD. Recognized here as distinct from Mate/ea (for reasons put forth in the dis- cussion of that genus, and below), Gonolobus includes perhaps 100 species (Spellman, 1975b). Although most of these occur in Mexico and Guatemala, Gonolobus is distributed from the southeastern United States and nearby areas, where it 1s represented by a single species, G. gonocarpus (Walter) Perry, to Panama, where fewer than ten species are known The correct name and appropriate circumscription of this genus have been unclear as well as controversial. As proposed by Michaux in 1803, Gonolobus included three species. Vincetoxicum gonocarpum Walter (as V. gonocarpos Walter) and V. acanthocarpum Walter (as V. acanthocarpos Walter), comprising the totality of Vincetoxicum Walter (Fl. Caroliniana, 104. 1788) as originally delimited, were listed as synonyms of the first two, G. macrophyllus Michx. and G. hirsutus Michx., respectively (thus invalidating these two names of Michaux), while the third, G. /aevis Michx., was newly described Although it is generally conceded that the material on which Gonolobus laevis was based actually represents more than one species, interpretations regarding it are confusing and/or in some cases contradictory. Gray (p. 75) partially described G. /aevis with the phrase “‘folliculis laevibus 5-angulatis,” suggesting that he chose portions agreeing with Michaux’s description of Gonolobus (‘‘fol- liculi plerumque costati seu angulosi’’) but contradicting the latter’s concept of G. laevis. Although Michaux had explicitly indicated that the fruits of G. mac- rophyllus and G. hirsutus are angled and muricate, respectively, he had de- scribed (p. 75) those of G. /aevis merely with the phrase “‘folliculis laevibus,”’ implying that they are smooth and not angled. According to Woodson, the material selected by Gray is probably conspecific with G. gonocarpus. Unlike Gray, Vail chose material contradicting Michaux’s generic concept but agreeing with his description of Gonolobus laevis. She maintained that such ices ause SCOR OnE is nes ioe the Sorted! epuihet 4 1S ROnOCANpUS, even though gonocarpos was Simila rly Vincetoxicum is neuter, so the Spiers in aah case should be nite even though gonocarpos had been proposed initially by Walter and, again, widely used thereafter 1989] ROSATTI, APOCYNINEAE 473 Tag 4 Mt, at a fi iN A ‘ e \ vas va AS allt Gi Lye ANI (ti Sey \i a i aX aS Por aae fie ite i ih ri) We vy ! | KS FIG Matelea and Gonolobus. a-f, M. carolinensis (from flowering material collected eatin a, oe a Agee os 2 flowers of | inflorescence, surfaces tuber- cu Ya: e follicle, cut transversely (pericarp and laterally winged seeds eae area ae eee pes tufts of trichomes stippled evenly; marginal placenta near center), x 2; c, placenta, after release of seeds, x 1; d, seed, with apical tuft of trichomes, x '/; e, seed, apical tuft of trichomes removed, seed coat winged laterally, x 3; f, embryo, somewhat expanded from soaking in water (oriented as in “‘e’’), x 3. g, G. gonocarpus: mature follicle, surface winged longitudinally, x 2 material was much more abundant than that selected by Gray but nevertheless assignable to Enslenia albida Nutt. (= Cynanchum laeve (Michx.) Pers.). Her assessment accords better with ‘““The guide for the determination of types” (ICBN, 1983: element T.4.e) and is now generally accepted, although it is dificult to determine “current usage” in this case. Vail logically returned to Vincetoxicum Walter the first two species included in Gonolobus by Michaux (see above), leaving in the latter only G. /aevis. Vail’s interpretation of this species was adopted by, for example, Small (1933), but erry, who accepted Gray’s interpretation but nevertheless reestablished the original concept of Gonolobus by selecting Michaux’s first species, G. mac- 474 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 rophyllus (= G. gonocarpus), as the type of the genus. This designation is especially significant because of nomenclatural problems with Walter’s Vin- cetoxicum. Both the name and the description of Gonolobus provided by Michaux sug- gest that he envisioned a group of plants with angled (i.e., longitudinally winged) follicles, so it seems inconsistent for him to have indicated such a condition in only the first of the three species descriptions he included. In doing this he might have been simply following Walter, who had done essentially the same both in naming and in describing the two species comprising his Vincetoxicum viz., V. gonocarpos and V. acanthocarpos, the epithets meaning angled and thorny fruit, respectively) but did not address the issue in his generic descrip- on. The second species included in Gonolobus by Michaux, G. hirsutus (Vince- toxicum acanthocarpos), was placed by Alexander (in Small, 1933), together with six other species of the southeastern United States in Odontostephana E. J. Alex., in synonymy under O. carolinensis (Jacq.) E. J. Alex. Although the descriptions (p. 1076) indicated that the follicles in Vincetoxicum (here Gonolo- bus, for reasons given below) and Odontostephana were “unarmed, wing-ridged toward the apex” and “armed with fleshy spines,”’ respectively, the genera were separated in the key (p. 1065) on the basis of corona structure, primarily “‘disk- like or saucer-shaped” in the former and ‘“‘cup-shaped or incurved at the tip” in the latter (Small, 1933). Odontostephana was included 1 in an expanded Ma- telea Aublet by Woodson, and Michaux’s second s} generally known as M. carolinensis (Jacq.) Woodson (see discussion of Matelea). The correct generic name for the first species listed by Michaux (i.e., the one with longitudinally winged follicles) remained in doubt for several reasons. Although Vincetoxicum Walter (1788) preceded Gonolobus Michx. (1803), it would have been eliminated as a nomen rejiciendum if a proposal to conserve Vincetoxicum Moench (1794) had been adopted (see Perry). Although the latter did not occur, Vincetoxicum Walter is to be considered a later homonym and therefore illegitimate on the basis of Ross’s argument that Vincetoxicum Von Wolf (1776) had in fact been validly published, so Gonolobus now appears to be correct. As indicated above, there has been in addition to the nomenclatural confusion a considerable amount of taxonomic disagreement regarding Matelea and Go- nolobus. It is perhaps surprising that the genera usually have been combined by botanists dealing with the relatively few species in North America (e.g., Clewell; Correll & Johnston; Duncan & Kartesz; Jones & Coile; Radford et al.; Shinners, 1950, 1964; Smith; Steyermark) but retained as separate by those treating plants in Central America (e.g., Spellman, 1975b; Standley & Wil- liams), where the two genera are better represented and evidently less distinct. Woodson chose to maintain both Gonolobus and Matelea despite drastically reducing (from 97 to nine) the number of asclepiadaceous genera represented in the New World outside of South America. Although he stated that differences between the two in corona structure did not exist, he maintained that the fleshy, abaxial appendages of the anthers universally present in Gonolobus do not occur in Matelea and thus based his separation primarily on this character. He 1989] ROSATTI, APOCY NINEAE 475 considered the possibility that the anther appendages of Gonolobus represented adnations to the stamens of structures corresponding to the ligules often adnate to or at least associated with the adaxial side of the corona in Matelea, but he rejected the idea because such structures “almost invariably” occur in Gonolo- bus in addition to the anther appendages. It is notable that ligules nevertheless do not occur in G. gonocarpus (see also discussion of Matelea). Woodson cited fruit and pubescence characters as further support for his generic separation. For the relatively few species of each genus for which fruits were known, he stated that those of Gonolobus are usually winged, while those of Matelea are usually muricate (i.e., angled and thorny, respectively, in the sense of Walter). He also pointed out that a mixture of long, nonglandular and short, glandular trichomes is unique to most members of Matelea (and possibly the closely related if not congeneric Fischeria DC.) and never found in Gonolo- US. Shinners (1950), who believed that Matelea and Gonolobus are not suffi- ciently distinct for separation at the generic level, provided a rebuttal to Wood- son’s position. He reported that the follicles of MM. producta (Torrey) Woodson (western Texas to southern Arizona and northern Mexico) are not at all muri- cate (they are smooth but evidently not angled or winged, according to Correll & Johnston), and that those of M. reticulata (Gray) Woodson (central, southern, and western Texas; northeastern Mexico) are only sparingly so. Indeed, I have seen a specimen of Mate/ea obtained in Florida (Curtiss 5948; GH), probably belonging to M. flavidula (Chapman) Woodson (based on the key provided by Drapalik) that bears a single follicle with only a few short projections; although it does not appear to have been angled before pressing, an intermediacy between the genera in fruit is nevertheless suggested. My observations of both glandular and nonglandular trichomes on speci- mens of Gonolobus gonocarpus are consistent with those reported by Shinners (1950) for this species and for the probably conspecific G. suberosus (L.) R. Br. as well (see below), except that it seems to me that pubescence in general is not as substantial in this species as it is in our representatives of Matelea. Shinners (1950) also cited other evidence suggesting that Gonolobus and Ma- telea are not distinct, at least in Texas. Shinners (1950) regarded the presence of anther appendages in Gonolobus and their absence in Mate/ea to be insufficient grounds for separating the two genera, citing Woodson’s (p. 239) observation that the structures in the former were “rather poorly developed as a rule in the temperate representatives.” However, although those of G. gonocarpus evidently are not as elaborate as those of some more tropical species, they are nevertheless quite apparent and unlike any structures present in Matelea. Drapalik studied nine of these species in the southeastern United States, including Gonolobus gonocarpus (as Matelea gonocarpa (Walter) Shinners) and eight other species here included in Matelea. He believed (p. 80) that all nine should be included in the same genus because of “many similarities” among them, despite the fact that the follicles of G. gonocarpus and the others are “quite different.” His concession that glandular trichomes are relatively scarce on plants of G. gonocarpus is consistent with my observations (as indicated 476 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 above), and I agree that trichome density should not be used as a generic distinction. However, I do not agree with his implication (p. 80) that the anther appendage present throughout Gonolobus and completely lacking in Matelea also fails in this regard merely because it is “simply a sterile enlargement . . . that protrudes outward.” Characters of the androecium, primarily a variety of sterile enlargements of the filaments (i.e., coronas), have served as the mor- phological bases for many and perhaps most genera of Asclepiadaceae, in- cluding all those in the Southeast. Despite the taxonomic decision to recognize a single genus, Drapalik stated that Gonolobus gonocarpus was not morphologically related to the eight species of Matelea. He also reported several other differences between them, involving aspects of physiology, phenology, and ecology. He found greater susceptibility to infections by fungi and rusts in G. gonocarpus than in the six species of the “Odontostephana group” (see discussion of Marelea), among cultivated (but not wild) plants, and reported a fragrance (presumably floral) in G. gonocarpus unlike any in the other eight Drapalik also determined that in garden plots individuals of Gonolobus gon- ocarpus were always the first to resume growth in the spring, yet they flowered later than adjacent plants belonging to the ‘“Odontostephana group” of Matelea. Although he reported some overlap in flowering between G. gonocarpus and members of the ““Odontostephana group,” he found that of six possible hybrids (among 99 flowering individuals grown from seeds resulting from natural pol- linations of the cultivated plants), none showed evidence of parentage by G. gonocarpus. According to Drapalik, when plants of G. gonocarpus and species of Matelea occur together in nature, those of the former usually are found in the moister conditions and flower a bit later (except that M. floridana evidently flowers at about the same time or earlier). At least in the Southeast, plants of G. gonocarpus occur in a wider range of substrate moisture and pH than do individual species of Matelea and, as in the case of only two of our species of Matelea (M. alabamensis and M. floridana), flower under conditions ranging from densely shaded to sunny. This wider ecological amplitude would account, at least in part, for the fact that G. gonocarpus appears to be more abundant in the Southeast than any of the species of Matelea. While Drapalik’s position in combining Matelea and Gonocarpus is certainly not without merit or support from other botanists, it resulted from study of only nine of the perhaps 300 species in question. Despite the existence of southeastern species of Matelea (M. alabamensis and M. pubiflora) that I be- lieve are in several ways intermediate between the genera (see discussion of Matelea), my studies support the decisions of those familiar with all of the plants involved (e.g., Spellman, 1975b; Standley & Williams; Stevens, pers. comm.; Woodson) to recognize both Matelea and Gonolobus. G pnolobiis gonocarpus (as Vincetoxicum gonocarpos Walter) and G. suber- osus (L.) R. Br. (as V. suberosum (L.) Britton) were both recognized in the southeastern United States by Small (1933), but according to Drapalik, the Linnaean epithet has been misapplied and the plants are certainly conspecific. The species have been separated on the basis of corolla-lobe size and pubes- cence, those in G. gonocarpus being more than twice as long as the calyx lobes (vs. twice as long or less) and glabrous (vs. pubescent) adaxially. However, on 1989] ROSATTI, APOCYNINEAE 477 the basis of these characters, Drapalik found both species and “‘every concelv- able intermediate” in several populations and grew both from seeds of a single fruit obtained from a plant in Tennessee. Drapalik emphasized that any en- deavors to interpret the considerable morphological variation in this species taxonomically should be preceded by adequate field (and, presumably, other biosystematic) study. oodson recognized three subgenera”? in Gonolobus, one of which, the ex- traregional subg. ‘“‘Pseudolachnostoma,” has erect or ascending (vs. spreading or reflexed) corolla lobes and lacks the outer, discoid, annular corona present in the other two. Anthers in the extraregional subg. ““Pterolobus”’ are supposedly partially exserted from the style-stigma head, while those in subg. ““Eugonol- obus”’ are not. Gonolobus gon urs throughout the southeastern United States with the exception of southern Florida, in all states bordering our area, and in Illinois and Indiana. The plants often occur in low, moist, rich but well-drained areas with hardwood trees in swamps and on floodplains. Gonolobus gonocarpus is of sufficient horticultural interest in North America to have been treated by Bailey and colleagues, and Drapalik provided instruc- tions for growing the plants both from seed and by transplantation. According to Standley & Williams, the tender young fruits of Central American species of Gonolobus (and related genera) are commonly eaten as vegetables and are sold for this purpose in markets in Guatemala City. REFERENCES; more references for the eg ean see BAILEY ef al.; CLEWELL; CORRELL & JOHNSTON; DUNCAN & KARTESZ; JONES & COILE; RADFORD et al.; SMALL (1933); SmitH; STANDLEY & WILLIAMS: and STEYERMARK. Under references for the ik Rimes see DRAPALIK; PERRY; SHINNERS (1950, 1964); SPELLMAN (1975b); and Woops Gray, A. Contributions to the botany of North America. Proc. Am. Acad. Arts Sci. 12: 51-84. pls. I, 2. 1877. [Gonolobus, 74-79.] Ross, R. The generic names published by N. M. von Wolf. Acta Bot. Neerl. 15: 147- 161. 1966. Vat, A. M. Studies in the Asclepiadaceae.—IV. II. The types of Gonolobus Michaux and descriptions of new species in Vincetoxicum Walter. Bull. Torrey Bot. Club 26: 425-431. 1899 Tribe ASCLEPIADEAE [R. Brown ex A. P. de Candolle, Bot. Gallicum 1: 323. 1828.*] 3. Morrenia Lindley, Bot. Regist. 24(Misc.): 71. 1838. Terrestrial, scandent and/or twining subshrubs with white latex; axes and leaves with often dense nonglandular pubescence. Leaves opposite; blades ovate, elliptic, or lanceolate, apices acute or acuminate, margins sometimes undulate, The names of these subgenera were not validly ees as by Woodson, since they were accom- panied in 1941 neither by a Latin description or diagnosis nor by a reference to a previously and effectively published Latin description or diagnosis (1.¢ " they were not published in accordance with Art. 36.1 [ICBN, 1988], the effective date of which is 1 January 1935). Therefore, they are enclosed in quotation marks here 478 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 rarely revolute, bases hastate or less often cordate; petioles distinct, without lateral appendages, bases nee adjacent nee of stems usually with a few colleters. I w-flowered, cymose but appearing racemose or sometimes subumbellate, eee pedicels longer to shorter than calyces, ebracteolate. Calyx ca. 2-5 as long as corolla, free nearly to base, glabrous or with nonglandular pubescence abaxially, glabrous adaxially; lobes acute, bases valvate, usually with 2 or 3 squamellae near each sinus. Corolla much shorter than leaves, free nearly to base, salverform (but with lobes much longer than tube, and therefore appearing nearly rotate), glabrous adaxially and abaxially; lobes valvate, spreading to reflexed, the margins revolute. Outer corona absent. Inner corona exceeding anther head, stipitate, cyathiform (seg- ments connate about to level of or above style-stigma head), erect (concealing anther head and style-stigma head), free from corolla; lobes 5, each with 2 lateral, shorter projections near sinuses between segments, and either emar- ginate, with (as viewed from the adaxial side) a mucronate convexity at base of emargination, or entire, with acute, marginally incurved apices. Anther head cylindrical, held well above [or sometimes immediately subtended by] base of inner corona; anther wing slots directed outward. Anthers each with an apical, adaxial, hyaline flap inflexed over style-stigma head, unappendaged abaxially. Style-stigma head short-obconic, with conical, bilobed apex. Corpuscula clearly exserted from anther head; translator arms mostly pendulous; pollinia pen- dulous, each completely fertile. Fruits follicular, | from each flower, lance- ovoid to ovoid, obtuse or truncate and often mucronate apically, dehiscent along adaxial suture, mostly pendulous on pendulous pedicels; surface smooth or sometimes shallowly winged longitudinally [or rugose], usually with short nonglandular and glandular pubescence. Seeds many, elliptical-lenticular, each with a tuft of trichomes at apex; seed-coat surface rugose, lateral wings crenate- undulate or entire. Type species: M. odorata (Hooker & Arnott) Lindley (Cy- nanchum odoratum Hooker & Arnott), the only species initially included in the genus by Lindley. (Named for Charles Morren, 1833-1886, professor of botany and director of the botanical garden at Liége, Belgium.)—MILKWEED VINE, STRANGLER VINE. A genus of probably two species. Morrenia odorata (Hooker & Arnott) Lind- ley is native to Paraguay, southern Brazil, and northern and central Argentina but has escaped from cultivation and become naturalized in Florida and per- haps elsewhere (see below), whereas M. brachystephana Griseb. is evidently restricted to northern and central Argentina. Compared to the latter, MM. odorata has larger flowers (e.g., corollas 10-15 vs. 5-6 mm long), stipitate (vs. sessile) anther heads, and smooth (vs. rugose) follicles (see, for example, Fabris; Schu- mann). Details are uncertain regarding the introduction to Florida of Morrenia odor- ata (see primarily Tucker ef a/.), a species favored by horticulturists for the vanillalike fragrance of its flowers (Ryan & Knorr). An individual was report- edly cultivated in Pasco County in 1939, and plants were first reported from citrus groves, clearly the preferred habitat in the adventive range of the species, 1989] ROSATTI, APOCYNINEAE 479 in Orlando in 1957. The species is now widespread in central peninsular Florida but is most abundant in Lake, Orange, and Seminole counties. In addition to citrus groves, where they are both numerous and problematic, the plants have been infrequently observed in other man-made and/or disturbed habitats. They have been found in natural habitats as well, albeit rarely. Morrenia odorata is a perennial, vigorously growing vine that in Florida often kills support plants, particularly citrus trees, through both shading and girdling (hence, “strangler vine’’); parasitism evidently does not occur (Ryan & Knorr), although individuals compete for water and nutrients (Spellman & Gunn). Infestations by M. odorata also reduce the efficiency of irrigation, spray- ing, and harvesting in citrus groves. Control of Morrenia odorata has been difficult, and further spread of the weed seems likely for several reasons (see primarily Tucker et a/.). The plants produce an abundance of viable, wind-dispersed seeds (about 1000 per fruit) at a young age (two years or less) and also propagate vegetatively from root fragments resulting from tillage. They have been known to survive freezing temperatures and have demonstrated high degrees of tolerance to the chemical herbicides registered for use in citrus groves. Biological control of Morrenia odorata is complicated by the presence of latex, which renders the plants (and most members of the suborder) toxic or at least unpalatable to most animals. Woodhead and co-workers reported that populations were reduced by more than 90 percent within one or two years after citrus plants were somehow “treated” in 1978, 1979, and 1980 with a fungus, Phytophthora palmivora (E. J. Butler) E. J. Butler. Mitchell & Kann- wischer-Mitchell determined that applications under citrus trees of 1.5 chlamy- dospores per square centimeter of soil resulted in the death of over 95 percent of the plants of M. odorata. Nevertheless, the species evidently continues to be a serious problem. Singh & Achhireddy studied germination and seedling growth to gain a better understanding of the biology and therefore the possibilities for control of Mor- renia odorata. They found that germination rates were markedly reduced above and below pH 7—to zero percent after seven days at pH 8.0 and pH 5.5; on either side of a 20/25°C (12-hour) thermoperiod, although seedling growth was highest at about 30°C; and at soil depths greater than 2.5 and less than 0.5 cm. Both germination rates and seedling growth decreased with decreasing water potential. Allelopathic effects on Morrenia odorata have been suspected of Lantana Camara L. (Verbenaceae), native to tropical America and the Coastal Plain in the Southeast, but also a serious weed in citrus groves in Florida. In fact, Achhireddy & Singh found that incorporation into the soil of dried material of L. Camara significantly inhibited vegetative growth in M. odorata (e.g., 50 percent seedling mortality within 15 days of germination in soil containing, in terms of dry weight, one percent root material _ Camara), although it had no apparent effect on final germination percent Morrenia odorata is one of several species of eee reported to be used in South America to stimulate lactation in cattle and humans. The latex 480 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 is used in the manufacture of cheese, while people in the Chaco region of Argentina eat the follicles either raw or cooked (see Spellman & Gunn). REFERENCES: Under references for the Apocynineae, see SCHUMANN. Under references for the Asclepiadaceae, see FABRIS; SPELLMAN & GUNN; and SUNDELL. ACHHIREDDY, N. R., & M. SincH. Allelopathic effects of lantana (Lantana Camara) on milkweedvine (Morrenia odorata). Weed Sci. 32: 757-761. 1984. MiTcuHeELL, D. J., & M. E. KANNWISCHER- -MITCHELL. Relationships of inoculum density of Phytophthora species to disease ee in various hosts. Pp. 259-269 in D. C. ERwin, S. BARTNICKI-GARCIA, & P. H. o, eds., Phytophthora, its biology, tax- onomy, ecology, and pathology. St. a Nee 1983. Ryan, G. F., & L. C. Knorr. Milkweed vine: a spreading menace in citrus groves. Citrus Industry 48(5): 8, 10, 30. 1967. SinGH, M., & N. R. Acuuireppy. Germination ecology of milkweedvine (Morrenia odorata. Weed Sci. 32: 781-785. 1984. Tucker, D. P. H., R. L. PHituips, & J. L. JAcKson, Jk. The milkweed vine... a major citrus pest in Flonda citrus groves. Citrus Industry 53(1): 6, 8, 9. 1972. WoovHEAD, S. H., W. H. Ripinas, C. L. ScHoutties, R. K. CLark, & N.E. EL-GHOLL. Field efficacy of Phytophthora palmivora for control of milkweed vine. Phytopatholo- gy 71: 913. 19 4. Cynanchum Linnaeus, Sp. Pl. 1: 212. 1753; Gen. Pl. ed. 5. 101. 1754. Terrestrial to semiaquatic (sometimes at edges of salt marshes), scandent, perennial herbs (usually woody toward base) [or erect to prostrate subshrubs] with white latex; axes and leaves glabrous or with various degrees of nonglan- dular pubescence. Leaves opposite, sometimes caducous and the plants leafless or nearly so; blades ovate, elliptic, lanceolate, linear, or filiform, apices acute, acuminate, or mucronate, margins often revolute, bases attenuate, truncate, or cordate; petioles distinct, sometimes very short and/or indistinct (especially with linear or filiform leaves), without lateral appendages, bases and/or adjacent areas of stems sometimes with a few small colleters. Inflorescences pedunculate or sessile, few- to many-flowered, cymose but appearing umbellate and/or racemose, bracteate; pedicels longer than calyces, ebracteolate. Calyx about ' as long as corolla or less, free nearly to base, usually with nonglandular pu- bescence abaxially, glabrous adaxially; lobes acute or sometimes subotuse, bases variously imbricate, with minute squamellae. Corolla much shorter than leaves, free nearly to base to united throughout the basal 4 (in C. Blodgettii, C. North- ropiae), campanulate (but lobes mostly free), campanulate-salverform (in C. Northropiae), or campanulate-urceolate (in C. Blodgettii), glabrous or with dense, nonglandular pubescence adaxially (in C. Blodgettii and C. Northropiae), glabrous abaxially; lobes valvate or variously imbricate, erect proximally but ascending, spreading, or reflexed distally. Outer corona absent. Inner corona exceeding or (in C. scoparium) exceeded by anther head, stipitate to substipitate or sessile, comprising distinct or (in C. scoparium) basally connate, erect to ascending segments, each basally free from or (in C. Blodgettii) adnate to corolla [or rarely absent]; segments 5S, laminate or (in C. Northropiae) fleshy and sub- terete (solid, not conduplicate), apically acute, subobtuse, or (sometimes in C. angustifolium) emarginate, or (in C. /aeve) with 2 linear, erect lateral append- 1989] ROSATTI, APOCYNINEAE 481 ages. Anther head cylindrical, immediately subtended by or (in C. Blodgettii) seemingly (by adnation) separated from base of inner corona; anther wing slots directed outward. Anthers each with an apical, hyaline flap erect to arching over style-stigma head, unappendaged abaxially. Style-stigma head short-cy- lindric, with flat, conical, or bilobed apex. Corpuscula exceeded by anther head; translator arms subhorizontal; pollinia pendulous, completely fertile. Fruits follicular, 1 or sometimes 2 from each flower (the pair, when present, from single flower usually divergent less than 90°), narrowly fusiform to fusiform- ovoid, acute to attenuate-caudate apically, dehiscent along adaxial suture, erect on erect pedicels or pendulous on pendulous ones; surface smooth (or reportedly sometimes in C. /aeve) winged longitudinally, with various degrees of non- glandular pubescence. Seeds many, ovate-lenticular, each with a tuft of tri- chomes at apex; seed-coat surface usually smooth, lateral wings entire or re- motely crenate. (Including, among others, Ampelamus Raf., 1819; Amphistelma Griseb., 1861, at least in part; Epicion Small, 1933; Enslenia Nutt., 1818, at least in part; Enslinia Reichenb., 1828, at least in part; Gonolobus Michx., 1803, in part; Lyonia Ell. non Nutt., of the Ericaceae, 1821; Mellichampia Gray, 1887; Metalepis Griseb., 1866; Metastelma R. Br., [preprinted 1810] 1811, at least in part; Roulinia Dene., 1844; Rouliniella Vail, 1902; and 7y- lodontia Griseb., 1866.) Lectotype species: C. acutum L.; see Meyer, Com- ment. Pl. Afr. Austral. 216. 1837 [1838]. (Name from Greek, kynos, dog, and ancho, to strangle or lace up, perhaps in reference to the similarity of the plants in morphology and in use to those of Apocynum, and possibly to their usually lianous habits, as well.)— VINE MILKWEED, SAND VINE. Considered here in the broad sense, Cynanchum comprises 200 or more species (Spellman, 1975b) in tropical and to lesser extents temperate regions throughout the world. As with some of our other genera of Asclepiadaceae, Cynanchum owes its large size mainly to the taxonomic reorganization of the family by Woodson. Small (1933) had placed our five species, all indigenous in the Southeast, in five separate genera, the names of which comprise only a minority of those now generally considered to be synonymous with Cynan- chum. As originally proposed by Linnaeus, Cynanchum included five species. Only one or, depending on taxonomic interpretation, two of these were retained in the genus by Brown, who nevertheless included a total of 15 species in six unnamed “‘sections” that he thought would eventually be accepted as genera. Indeed, genera were subsequently proposed for many of Brown’s “sections,” as well as for other plants now generally thought to be congeneric. The con- solidation of these genera appears to have begun with Schumann, who in reverting to a system more in line with Brown’s, included under Cynanchum 11 generic synonyms (and over 100 species) that had been in use for plants throughout the world. Woodson listed two of these genera plus 21 other generic synonyms for plants in North and Central America and stated that several more eventually would be added for species in South America. It is difficult to know how Schumann’s concept of Cynanchum, which he assigned to subtribe Cynanchinae of tribe Asclepiadeae, compares to the one employed here. Only one of our species, C. /aeve (Michx.) Pers., was actually 482 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 mentioned in his treatment of the family, and it was included (as Enslenia albida Nutt.) in a different subtribe of the Asclepiadeae, the Asclepiadinae. (For this reason and others, these subtribes are not recognized here; see family discussion.) Some of our other species were included, for example, by Small (1933), in genera also separated from Cynanchum by Schumann: C. Blodgettii (Gray) Shinners was placed in Metastelma R. Br., and C. scoparium Nutt. was included in Amphistelma Griseb.; Schumann included Amphistelma in Meta- stelma, which he placed with Enslenia in the Asclepiadinae. Small (1933) included C. angustifolium Pers. (as C. palustre (Pursh) Heller) in Lyonia Ell. (non Nutt.), which was ignored by Schumann, and C. Northropiae (Schlechter) Alain in the newly proposed Epicion Small. Schumann recognized two sections in Cynanchum: Cynoctonum (E. Meyer) K. Schum. (= sect. Cynanchum), with completely to nearly completely fused corona segments, and Vincetoxicum (Moench) K. Schum., with basally fused to almost entirely free corona segments. Woodson based six subgenera on characters of the inflorescence, corolla, and corona. However, the subgenera are difficult to interpret because of the existence of specimens assignable to the genus that are not accommodated by the key (see below), and because only a few representative species were listed for each. The inflorescences were said to be umbelliform in subg. Metastelma (R. Br.) Woodson (corolla lobes ascending or spreading) and subg. ‘‘Cleistolobus’’° (corolla lobes inflexed and hoodlike) but racemiform to corymbiform in the other four. The corollas were reported to be urceolate in one of these, subg. Tylodontia (Griseb.) Woodson, and campanulate to rotate-subcampanulate in the other three. The segments of the corona were said to be broad, mostly connate, and emarginate in subg. Metalepis (Griseb.) Woodson, but elongate and free or only basally united in the other two. They were indicated as being acuminate and entire or with some very obscure lateral lobes in subg. Melli- champia (Gray) Woodson, but deeply bifid (i.e., with two long, linear lobes apically) in subg. Ampelamus (Raf.) Woodson. As defined by Sundell (p. 11), subg. Me/lichampia comprised 11 New World species and included as three (of four) sections ‘elements from three of Wood- son’s (1941) subgenera,” subg. Ampelamus, subg. Mellichampia, and subg. Metalepis.*' Sundell described the subgenus as having an inner corona of five segments united basally into a short tube. However, it is difficult to evaluate the group because of the inclusion of plants that do not fit this description and the exclusion of others that do (see below). Five species of Cynanchum, some of which appear not to be accommodated by any of the relevant treatments outlined above, occur in the southeastern United States. Cynanchum laeve (Gonolobus laevis Michx., Enslenia albida Nutt., Ampelamus albidus (Nutt.) Britton) is readily (and, among our species, The name of this subgenus was not validly published by Woodson, since it was accompanied in 1941 neither by a Latin description or diagnosis nor by a reference to a previously and effectively published Latin description or diagnosis (i.c., it was not published in accordance with Art. [ICBN, 1988], the effective date of which is January 1935). Therefore, it is enclosed in auoimton marks ne re. “Sundell was evidently the first to unite these three acewiee of Woodson (the names of which were published simultaneously), so the name he chose has priority 1989] ROSATTI, APOCYNINEAE 483 uniquely) placed in Woodson’s subg. Ampelamus by virtue of its racemiform inflorescences and deeply bifid corona segments. It was included in Sundell’s subg. Mellichampia despite its entirely free corona segments, although because of this fact it was isolated in the monotypic sect. Ampelamus (Raf.) Sundell. (None of our other species was mentioned by Sundell.) Cynanchum laeve is unique among our representatives of the genus in having cordate leaves. (Sar- costemma cynanchoides Dene., which enters our area in Arkansas and which is sometimes confused with C. /aeve, has narrower cordate leaves.) It 1s found in alluvial thickets and old fields, as well as along stream banks and roadsides, from Pennsylvania to Kansas and generally southward to the Florida Panhandle and central Texas. Cynanchum scoparium (Amphistelma scoparia (Nutt.) Small) has shallowly lobed coronas and would appear to belong to subg. Metalepis sensu Woodson. However, the lobes are not emarginate and the inflorescences are sometimes umbellate, so it is unclear to which subgenus in Woodson’s system it belongs. On the basis of its basally connate corona segments, the species might also appear to belong to subg. Mellichampia sensu Sundell, and presumably to sect. Metalepis (Griseb.) Sundell. However, C. scoparium lacks the stipitate anther heads (they are sessile) that in part characterize this section, and the species was not treated by Sundell. In addition to its sometimes umbellate inflores- cences and basally connate corona segments, C. scoparium is unique among our species of Cynanchum in several ways: the leaves (which are linear-lan- ceolate to linear but nevertheless petiolate) are deciduous relatively early, so the plants are often leafless or nearly so; the calyx lobes are obtuse (vs. acute); and the lobes of the corona do not exceed the anther head. Cynanchum sco- parium is known from hammocks and pinelands in Beaufort County, South Carolina, and in central and southern Florida. Cynanchum Northropiae (Metastelma Northropiae Schlechter; Epicion Northropiae (Schlechter) Small) is the only one of our species that can be placed in subg. Mellichampia sensu Woodson—an assignment primarily due to its racemose inflorescences. Its corona structure, however, suggests that it is more closely related to species in Texas and Mexico that, on the basis of umbellate inflorescences and ascending or spreading corolla lobes, are assignable in Wood- son’s system to subg. Metastelma (see Henrickson). In addition, the species has ovate to elliptic leaves with mucronate or acuminate (rarely emarginate) apices; corollas united in the basal third, with lobes that are densely pubescent adaxially; and clearly stipitate coronas, with segments that are distinct (and free from the corolla), narrowly lanceolate, radially (1.e., laterally) compressed, and adaxially grooved. The species is known from hammocks, pinelands, and cut-over lands in southern Florida, including the Keys, and in the West Indies. Two of our species are assignable to subg. Metaste/ma, although on the basis of corona structure they appear to be related neither to Cynanchum Northropiae nor to the species in Texas and Mexico discussed above. Cynanchum Blodgettit (Metastelma Blodgettii Gray) has linear-lanceolate to linear leaves with acute and/or sometimes mucronate apices; corollas united in the basal third (or less) and urceolate (a condition partially characterizing subg. 7y/odontia), with lobes densely pubescent adaxially (sometimes only near the apex); and sessile coronas 484 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 with lanceolate segments distinct from one another but basally adnate to the corolla and tangentially (i.e., dorsiventrally) compressed. The plants grow in hammocks and in rocky or sandy soil in pine and pine-palm woods in southern Florida, including the Keys. Although the species was reported from southern Texas by Small (1933), it was not treated by Correll & Johnston, and I have seen no specimens from the state. Cynanchum angustifolium (C. palustre (Pursh) Heller; Lyonia palustris (Pursh) Small) is also assignable to subg. Metaste/ma. The leaves are linear with acute apices, but they are sessile and usually reflexed as well, two conditions not evident in our other species. In addition, C. angustifolium has glabrous, rotate- campanulate corollas divided nearly to the base, and stipitate to substipitate coronas with ovate to elliptic, distinct segments that are obtuse, retuse, or emarginate. The plants occur in coastal marshes with saline or brackish water and in sandy (probably moist) or marly soil of adjacent hammocks and dunes from North Carolina to southern Florida, including the Keys, and from there to Texas. Several species of Cynanchum, including C. /aeve (Bailey et al.), are cultivated in North America. Standley & Williams’s comment that Cynanchum in North America would be a good subject for dissertation research appears still to be true. REFERENCES: Under Sees for the Apocynineae, see BAILEY et al.; BROWN; CORRELL & JOHNSTON; SCHUMANN; SMALL, 1933; and STANDLEY & WILLIAMS. Under references for the Asclepiadaceae, see SPELLMAN, 1975b; and Woopson. HeEnrIckson, J. Notes on lanes (Asclepiadaceae). Sida 12: 91-99. 1987. [Three species in Texas and northern Mexico evidently do not enter the Southeast but appear to be related to C Northropiae; very helpful illustrations.] SUNDELL, E. The New World species of Cynanchum L. subgenus ams (A. Gray ex Wats.) Woods. (Asclepiadaceae). Evol. Monogr. 5: 1-63. 5. Sarcostemma R. Brown, Mem. Wernerian Soc. 1: 50. 1811. Terrestrial (often in xeric habitats) to aquatic (sometimes rooted in standing fresh, brackish, or salt water), scandent and/or twining, perennial herbs (usually woody toward base) or suberect to prostrate subshrubs [or erect succulents] with white latex; axes and leaves glabrous or with various degrees of nonglan- dular pubescence [or sometimes glaucous]. Leaves opposite, sometimes ca- ducous and the plants leafless or nearly so; blades ovate, elliptic, lanceolate, linear, or filiform [or scalelike], apices acute, acuminate, or mucronate, margins often revolute, bases attenuate, truncate, obtuse, or cordate [or hastate]; petioles distinct, without lateral appendages, bases and/or adjacent areas of stems usu- ally with colleters. Inflorescences pedunculate [or sessile], few- to many-flow- ered, cymose but appearing umbellate [or racemose or corymbose], rarely with 1989] ROSATTI, APOCYNINEAE 485 both flowers and pedicel scars present, bracteate; pedicels longer than calyces, ebracteolate. Calyx about 4 as long as corolla or less, free nearly to base, usually with nonglandular pubescence abaxially, glabrous adaxially; lobes acute, bases variously imbricate, with (or, reportedly, sometimes without) minute squa- mellae. Corolla much shorter than leaves, free nearly to base to united through- out the basal 13, subrotate or subcampanulate [or sometimes campanulate or salverform], glabrous or with nonglandular pubescence abaxially, mostly gla- brous adaxially; lobes valvate or variously imbricate, spreading [to nearly erect] basally, ascending, spreading, or reflexed apically. Outer corona [rarely absent] discoid, membranaceous or fleshy, mostly entire, basally adnate to corolla tube but apically free, forming an erect or ascending annulus. Inner corona exceeding to sometimes exceeded by anther head, substipitate or sessile, composed of distinct, mostly erect segments, each basally free from or (at least in S. clausum) adnate to annulus; segments >, fleshy, vesicular, ovoid [or nearly spheroid]. Anther head cylindrical, i subtended by base of inner corona; anther wing slots directed eee Anthers each with an apical, hyaline flap inflexed over style-stigma head, unappendaged abaxially. Style-stigma head short-cy- lindric, with flat, bilobed, or umbonate apex. Corpuscula partially exserted from anther head; translator arms arching downward; pollinia pendulous, com- pletely fertile. Fruits follicular, | or sometimes [in some species, usually] 2 from each flower (the pair, when present, from single flower usually divergent less than 90° [to about 180°), lance-ovoid, narrowly fusiform to fusiform-ovoid [or obclavate], acute to attenuate apically, dehiscent along adaxial suture, erect on erect pedicels to pendulous on pendulous ones; surface shallowly striate longitudinally, with various degrees of nonglandular pubescence. Seeds many, ovate-lenticular, each with a tuft of trichomes at apex; seed-coat surface rugose and/or papillate, lateral wings (at least in S. clausum) coarsely serrate. (In- cluding Ceramanthus (Kunze) Malme, 1905; Funastrum Fourn., 1882; Oxy- stelma R. Br., 1809; Philibertella Vail, 1897; and Philibertia HBK., 1819.) LECTOTYPE SPECIES: S. viminale (L.) R. Br. (Euphorbia viminalis L.); see Holm, Ann. Missouri Bot. Gard. 37: 506. 1950. (Name from Greek, sarkos, flesh, and stemmatos, crown or wreath, probably in reference to the fleshy (and inflated) segments of the corona.)— MILK WITHE. Considered here in the broadest sense, Sarcostemma is a genus of approx- imately 34 species, 22 in the New World (Holm) and about 12 in an area from Africa to Australia (Huber). There has been a great deal of confusion regarding Sarcostemma and several other genera with which it is here combined. Much of this has had to do with imprecise and/or inadequate treatment of the plants involved, shortcomings that are best appreciated by a brief consideration of their taxonomic history According to Holm, the most recent monographer of the genus, Sarcostemma originally included only succulent-stemmed, scale-leaved lianas of Asia and Africa with rotate corollas and both outer and inner coronas, the outer annular and the inner with five separate, inflated segments. In fact, however, in the protologue of Sarcostemma, Brown said nothing about succulence; indicated 486 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 that the stems were without leaves (although he was an referring to plants with scale-leaves) or that the leaves were far apart (‘‘fo . distantibus’’); reported the genus from eastern India, Africa, and ae as well as New Caledonia; and did not address the issue of inflation of the inner corona seg- ments (he and many authors have referred merely to the fleshiness of the segments). Evidently, the summary provided by Holm was based on consid- eration of the species mentioned by Brown as they are now understood and/ or on information that has come to light subsequently; generally, the same approach seems to have been taken in the remainder of his account. In addition to Sarcostemma, Brown also established Oxyste/ma R. Br. for plants of eastern India and possibly Australia with laminate foliage, rotate corollas, and single (i.e., only inner) coronas, without addressing the subject of corona-segment inflation. From populations in the New World, Kunth (in Humboldt er al.) included in Sarcostemma some with lanceolate—albeit membranaceous—leaves and rotate corollas, and he erected Philibertia HBK. for others with cordate leaves and partly described with the phrase “corolla urceolato- (campanulato-?) ro- tata,” without addressing corona-segment inflation for either. Decaisne later described the corollas of Sarcostemma solanoides (HBK.) Dene. (Philibertia solanoides HBK.), the only species included in Philibertia by Kunth, with the word “pelviformi” (i.e., basin or saucer shaped), which would seem to lie somewhere between rotate and campanulate Decaisne adopted a broad view of Sarcostemma, in which he recognized two sections. Section Eusarcostemma (= sect. Sarcostemma) included plants with rotate corollas and was divided into three unnamed groups, one comprising Old World species with scale-leaves (‘‘aphylla” presumably in the sense that the foliage is not laminate), another of New World plants with linear to linear- lanceolate leaves (including S. clausum (Jacq.) Schultes, one of our two species), and a third of New World species with cordate leaves and deeply divided (but nevertheless rotate) corollas (including S. cynanchoides Dene., the other of our two species). Section PHILIBERTIA (HBK.) Dene. included New World species with cordate leaves and urceolate-rotate (possibly campanulate in the sense of others) corollas. Despite employing a broad concept of Sarcostemma, Decaisne continued to recognize Oxystelma R. Br., in which he included Old World species with rotate corollas, as distinct. His description of Oxyste/ma differed significantly from Brown’s account of the genus in at least two respects. Decaisne referred to an annulus on the corolla tube (““fauce annulo. . .”’), that is, an outer, annular corona as now interpreted, and described the segments of the (inner) corona, possibly for the first time, as being at least partially inflated. Even so, Decaisne did not mention this character state for Sarcostemma, even though he included in the latter our two species, each of which has inner corona segments that are clearly inflated. Bentham (in Bentham & Hooker) included in Sarcostemma plants from Africa, Asia, and Australia with succulent stems, scalelike leaves (again, stems were said to be leafless), subrotate corollas, and at least sometimes subsaccate 1989] ROSATTI, APOCY NINEAE 487 (i.e., somewhat inflated) inner-corona segments. For reasons that are not en- tirely clear, he referred all of the New World species in question to Philibertia, which he described as having very widely campanulate to subrotate corollas and, at least occasionally, saccate segments of the inner corona. Significantly, he also noted an affinity between Sarcostemma and Oxystelma, and between Oxystelma and Philibertia; in fact, his descriptions of the last two genera were very similar. However, although no such terms were applied to Oxyste/ma, he indicated that the inner corona segments were at least sometimes saccate in Philibertia. (He also noted other differences in the segments.) Fournier adopted very narrow generic concepts in establishing the New World genus Funastrum Fourn., which was described in part as having fleshy, vesicular corona segments; neither corolla shape nor the presence or absence of an outer corona was specified. Of the two species included, one actually belongs in Cynanchum L. according to Holm, while the other, F. angustissimum (An- dersson) Fourn. (Asclepias angustissima Andersson, = Sarcostemma angustis- simum (Andersson) R. Holm), of the Galapagos Islands, was indicated as having axillary umbels. Nevertheless, Holm (p. 535) reported that it is the only species of Sarcostemma with a “‘truly terminal inflorescence” (i.e., one that is not overtopped by sympodial growth; he actually considered the inflorescence to be terminal throughout the family), and on this basis he assigned it to the monotypic ser. ANGUSTISSIMA R. Holm. Schumann also employed rather narrow generic concepts, maintaining in Sarcostemma only Old World species with prostrate or twining, succulent stems, scale-leaves (his “‘blattlose Straucher’’), and rotate corollas. He divided the genus into a sect. Eusarcostemma (= sect. Sarcostemma), with both outer and inner coronas (the former annular and, surprisingly, five- to ten-lobed; the latter sometimes partially vesicular), and a monotypic sect. Sarcocyphula (Har- vey) K. Schum., with only an outer one. Oxystelma and Philibertia were two other genera recognized by Schumann. The former included species in both Eastern and Western hemispheres, while the latter was used only for plants in the New World. Although the descriptions of the two genera were essentially identical (e.g., each was described as having twining, presumably nonsucculent stems; laminate foliage; rotate or campan- ulate corollas; and both outer and inner coronas, the former annular and the latter with vesicular segments), the key indicated that corollas were campan- ulate in Oxyste/ma and rotate in Philibertia. Schumann accepted Funastrum in exactly the same sense as Fournier had proposed it; that 1s, as including the same two South American species. According to Schumann, these are twining subshrubs with (presumably) nonsucculent stems, linear basal and scalelike upper leaves, rotate corollas, no outer coronas, and inner coronas with vesicular segments. It is surprising and often regarded as one of many unsatisfactory aspects of Schumann’s classification that, within the Asclepiadeae, Sarcostem- ma and Funastrum were placed in two separate subtribes, while Oxystelma and Philibertia were included in a third. Vail proposed Philibertella for the same American plants that had constituted Philibertia in Schumann’s system, evidently because she felt the plants did not 488 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 adequately match Kunth’s (in Humboldt et a/.) description of Philibertia and because the only species Kunth had included in the genus was generally placed by that time in Oxyste/ma. Nevertheless, she indicated that corollas in Phili- bertella were either campanulate or rotate, so the distinction she attempted to make remains unclear Kunze established subg. CERAMANTHUS Kunze for New World members of Sarcostemma with urceolate corollas. Malme used the same name in elevating the group to generic status, even though it is at that rank a later homonym for a genus of Euphorbiaceae (i.e., Ceramanthus Hassk.) and several other generic names were available for the taxon that he envisioned (i.e., he included Fu- nastrum, Philibertella, Philibertia, and others as synonyms). Schlechter concluded that the New World species in question should be separated at the generic level and therefore resurrected Funastrum and Phili- bertia. Among other differences, Funastrum was indicated to have (compared to Philibertia) more rotate (vs. more campanulate) corollas, stipitate (vs. sessile or substipitate) anther heads (i.e., anther heads held above vs. immediately or nearly immediately subtended by inner coronas), and entire (vs. two-cleft) style- stigma head apices. Following Decaisne in taking a broad view of Sarcostemma, Woodson united within it Funastrum, Philibertia, and others; Oxystelma was ignored because he was treating only New World plants. Although Holm formally treated only New World taxa in his revision, he adopted an even broader generic concept in adding Oxyste/ma to the taxa that had been included in Sarcostemma by both Decaisne and Woodson. Unlike Woodson, who did not subdivide the genus, Holm recognized three subgenera, one with three sections and another with five series. In view of the discussion above, his treatment seems most tenable and therefore is followed here. Subgenus Eusarcostemma (Dene.) R. Holm (= subg. SARCOSTEMMA), al- though based on sect. Eusarcostemma Dene. (26 species, in both Eastern and Western hemispheres), was restricted to those species from Africa to Australia with succulent stems and scalelike leaves (i.e., the genus as originally delimited). Although the species included in this subgenus were not listed by Holm, the key indicated that the group was also characterized by rotate-subcampanulate corollas; an annular outer corona adnate to the filaments; and ovoid, vesicular (i.e., inflated) inner-corona segments. The two other subgenera recognized by Holm included plants with nonsuc- culent stems and laminate foliage. Subgenus OxysTELMA (R. Br.) R. Holm was characterized by campanulate corollas; an annular outer corona either adnate to the corolla or more or less obsolete; and variously ovoid or spheroid, ve- sicular inner-corona segments. It included sect. EuoxysTELMA R. Holm (Oxy- stelma sensu Brown), with a single species in Africa and Asia, as well as sect. PeENTACYPHUs (Schlechter) R. Holm (Pentacyphus sensu Schlechter) and sect. PHILIBERTIA (HBK.) Dene. (Philibertia sensu Kunth), with two and six species, respectively, in South America Subgenus CERAMANTHUS Kunze (Ceramanthus (Kunze) Malme), comprising 14 New World species, corresponded to Philibertia as conceived by Bentham (i.e., the plants were said to have rotate-subcampanulate or salverform [rotate, as interpreted earlier] corollas), Funastrum of Fournier, and Philibertella of 1989] ROSATTI, APOCYNINEAE 489 Vail. It was otherwise indicated to have an annular outer corona (absent in one species) free from the corolla, and variously ovoid or spheroid, vesicular inner-corona segments. My observations of Sarcostemma clausum (Jacq.) Schultes and S. cynanchoides (discussed more fully below) indicate that in each the outer corona is adnate to the summit of the corolla tube. Near the apex it is sufficiently free from the corolla to form an annulus, which is fused with the bases of the inner-corona segments in S. clausum but free from them in S. cynanchoides. Subgenus CERAMANTHUS was divided into five series. One of these, ser. FLAVA R. Holm, included only a single, South American species with flowers lacking an outer corona. The other four differed primarily in characters of the inflorescence. Two series of subg. CERAMANTHUS are represented in the southeastern United States, each by a single species. Series CLAUSA R. Holm (peduncles usually as thick as or thicker than subtending internodes, or lacking; when lacking, and sometimes otherwise, leaf bases obtuse to cuneate) included Sarcostemma clausum (leaves often caducous, elliptic to ovate). An extremely variable species (Holm listed nearly 60 synonyms, in six genera), S. clausum is found mostly in savannas or sandy areas in Mexico and northwestern South America but sometimes on wooded slopes, on flood lains, or rooted in standing water in Colombia, and Venezuela, has salverform corollas with constricted tubes (vs. rotate-subcampanulate corollas with unconstricted tubes), and the plants are glaucous (vs. at least somewhat pubescent) throughout. Series CYNANCHOIDES R. Holm (peduncles not as thick as subtending inter- nodes, or lacking; when lacking, and sometimes otherwise, leaf bases cordate) included Sarcostemma cynanchoides, represented in our area, and eight other species distributed from the southwestern United States to central Argentina. Although in material I have seen from the Southeast the peduncle is often thinner than the subtending internode in S. c/ausum (in contrast to the usual condition in ser. CLAUSA), it is relatively even narrower in S. cynanchoides. Among specimens I have examined, the outer corona annulus 1s fused to the bases of the inner-corona segments in S. clausum, while these structures are free in S. cynanchoides. Two subspecies were recognized in Sarcostemma cynanchoides by Holm. Subspecies cynanchoides was said to differ from subsp. Hartwegii (Vail) R. Holm in its broadly (vs. narrowly) lanceolate leaves with cordate or rarely truncate (vs. hastate or rarely truncate or obtuse) bases. Subspecies cynan- choides was indicated by Holm to grow in sandy soils of rivers and ditches and in rocky canyons from southern Arizona to the central parts of Oklahoma and Texas, as well as in adjacent parts of Mexico, and it has since been observed to enter our area in Arkansas. Smith indicated it for Franklin, Izard, and perhaps Pope counties, and I have seen a specimen (Thomas & Reid 20640, GH) from cliffs near the White River in Izard County that is assignable to this subspecies but misidentified as Cynanchum laeve. Subspecies Hartwegii occurs in more xeric habitats (e.g., dry, sandy, or rocky soils of arroyos and plains) and is distributed farther south and west, to central Mexico and southern California. Sarcostemma cynanchoides subsp. cynanchoides begins to flower nearly two 490 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 months later than subsp. Hartwegii (Holm). Nevertheless, putative hybrids have been found in the area of geographic overlap between the two, mostly along the border between the United States and Mexico, from Arizona to western Texas. The low number of such hybrids might be related to the fact that the species is poorly collected in that area. As defined here, Sarcostemma is most closely related to Cynanchum and Blepharodon Dene. The latter includes about nine species and, according to Woodson, is perhaps divisible into a North American group of twining plants and a South American assemblage of erect, herbaceous ones. Blepharodon was considered to be in part intermediate 1n corona structure between Cynanchum and Sarcostemma by Holm (p. 504), who stated, without mentioning the groups proposed by Woodson, that the genus 1s divisible into one group with corona segments consisting of “two radially arranged, laminate lobes partially fused by their margins” and another in which such segments are “‘semi-vesicular.” The flowers of Blepharodon and Cynanchum lack the outer (annular) corona characterizing Sarcostemma, although according to Holm they sometimes have structures that might be mistaken for one. Despite the foregoing, the annular outer corona and the five inflated segments of the inner corona of the African and Asian genus Pergularia L. were consid- ered by Holm to represent only superficial resemblances to Sarcostemma. Among other differences, Pergularia supposedly has a five-lobed, denticulate (vs. entire; however, see Schumann) outer corona and spurred (vs. not spurred) inner-corona segments. The strong, flexible stems of Sarcostemma clausum (subg. CERAMANTHUS) are used in Guatemala for stringing fish and as a substitute for twine (Standley & Williams). Holm remarked that species of sect. PHILIBERTIA sensu Holm were often grown in English greenhouses during the nineteenth century and that those belonging to subg. SARCOSTEMMA sensu Holm are still commonly cultivated; in the latter group is S. viminale, which is sometimes sold as “Eu- phorbia pendula” (Bailey et al.). REFERENCES: Under references for the Apocynineae, see BAILEY et a/l.; BENTHAM & HOOKER; BROWN; DECAISNE; HUBER; HUMBOLDT et al.; SCHUMANN; SMITH; and STANDLEY & WILLIAMS. Under references for the Asclepiadaceae, see FoURNIER and Woopson. Hoi_m, R. W. The American species of Sarcostemma R. Br. (Asclepiadaceae). Ann. Missouri Bot. Gard. 37: 477-560. 1950. [Many aspects of morphology and polli- nation biology discussed in detail.] Kunze, G. Pugillus tertius plantarum. Linnaea 20: 1-64. 1847. [Asclepias, and other Asclepiadaceae, 26-28. MaA.ME, G. O. A. sera de nonnullis Asclepiadaceis austroamericanis. Ark. Bot. 4: 1-19. pls. 1, 2. 1905. — R. paneer H. B. et Kth. und Fuvastrum Fourn. Repert. Sp. Nov. 13: 279-287. 1914 VaIL, A. “a St udies in the Asclepiadaceae.—I. Notes on the genus Philibertella in the United States. Bull. Torrey Bot. Club 24: 305-310. 1897. 1989] ROSATTI, APOCYNINEAE 491 6. Asclepias Linnaeus, Sp. Pl. 1: 214. 1753; Gen. Pl. ed. 5. 102. 1754. Terrestrial to semiaquatic (sometimes in brackish water), erect, perennial (or the introduced 4. curassavica®’ [and perhaps 2 other species] annual) herbs [or small shrubs] with white latex (only A. /ongifolia and A. tuberosa without latex); axes and leaves glabrous or with various degrees of nonglandular pubescence. Leaves usually decussate, _ often irregularly alternate and/or subwhorled whorled, usually in 4’s, at node subtending inflorescence 1 in A. quadrifolia), times very short and/or indistinct (especially when leaves linear or filiform), without lateral appendages, bases and/or adjacent areas of stems sometimes with a few small colleters. Inflorescences pedunculate, few- to many- [rarely single-]flowered, cymose but usually appearing umbellate, bracteate or rarely ebracteate; pedicels longer than calyces, ebracteolate. Calyx about 2 as long as corolla but usually much less, free nearly to base, glabrous or with various degrees of nonglandular pubescence abaxially, glabrous adaxially; lobes usually acute, bases sometimes slightly imbricate, usually with minute squamellae. Corolla much shorter than leaves, free nearly to base, essentially rotate (but see below), mostly glabrous abaxially and ay. lobes variously imbricate or sometimes valvate, reflexed or spreading proximally but reflexed or some- times spreading or ascending (erect only in A. pedicellata) distally (when re- flexed or spreading distally, often ascending apically). Outer corona absent. Inner corona exceeding to exceeded by anther head, stipitate to substipitate or sometimes sessile, composed of distinct, essentially erect or ascending (then sometimes basally deflexed) hoods (i.e., segments), each basally free from (but sometimes in contact with) corolla; lobules (i.e., small, erect, bifid or less often entire appendages) alternating with hood bases; hoods 5, fleshy, conduplicate (and therefore vessel- or tubelike), each open apically (openings not appressed against anther head) and exceeding to usually exceeded by horns (1.e., exserted, elongate, terete or laterally flattened, adaxially adnate appendages originating within hoods), or sometimes (in subg. AscCLEPIAS, subg. ASCLEPIODELLA) €X- ceeding crests (i.e., nonexserted, ridgelike, rounded, adaxially adnate append- su ah unappendaged within, or (in 4. pedicellata) each apically opened (openings not This and epithets for many other species in the genus customarily have been given the feminine form, even though Asclepias would appear to be masculine and, in fact, was indicated as such by Stearn (Bot. Latin, 306. 1966). Nevertheless, as indicated by nearly all of the apne ae 7 Species peeke (1753), Linnaeus treated it as feminine. According to Art. 76.1 (IC 1988), * generic name retains the gender assigned by its author, unless this is contrary to nen ae which, according to Note 1, ‘usually maintains the classical gender of a Greek or Latin word, when this was well established.” The first example under this recommendation included names oie: than Asclepias) * ‘for which botanical usage has reestablished the classical gender despite another choice by Linnaeus.” Therefore, it appears that if Asc/epias is masculine according to “botanical tradition,” Tawaeue: should probably be overruled and the name should be treated as masculine — Asclepias. a-f, A. syriaca: a, power a oie reflexed; corona stipitate ds open at apex [openings not appressed an anther head] and here slightly Bara ee hom and anther head, abaxial surfaces exceeding those adjacent to anther head ot recurved apically, margins auriculate; anther head [excluding wings and eal re shorter than broad, im- mediately subtended by hoods; corpusculum at summit of outwardly projecting slot fl ove ith h e | hyaline flap terminating each of 2 anthers), x 12; d, aDiner, adaxial side, with 2 ee (partly concealed) each connected by a translator arm, a corp 1989] ROSATTI, APOCYNINEAE 493 appressed against anther head) but inwardly geniculat te (and the openings there- fore concealed) and apparently unappendaged within, or (in subg. ACERATES) each apically opened (openings appressed against anther head) and exceeding crests or apparently unappendaged within, or (in subg. ANANTHERIX, subg. ASCLEPIODORA) each apically enclosed (by conduplication) and apparently un- appendaged within. Anther head cylindrical, immediately subtended by or (in A, pedicellata) held well above and separated by a long stipe from base of inner corona; anther wing slots directed outward. Anthers each with an apical, hyaline flap inflexed over style-stigma head, unappendaged abaxially. Style-stigma head cylindrical, with flat or depressed apex. Corpuscula partially exserted from anther head; translator arms arching downward; pollinia pendulous, completely fertile. Fruits follicular, 1 or very rarely 2 from each flower (the pair, when present, from single flower usually divergent less than 90°), narrowly fusiform to infrequently fusiform-ovoid or broadly ovoid, acute to attenuate-caudate apically, dehiscent along adaxial suture, erect on deflexed pedicels (the latter therefore S-shaped) or less frequently erect on erect ones (pendulous on pen- dulous pedicels in A. perennis); surface smooth or with elongate tubercles [or spines], nearly glabrous to rarely with dense nonglandular pubescence. Seeds many, ovate-lenticular, each with (or without, in 4. perennis) a tuft of trichomes at apex; seed-coat surface usually smooth, lateral wings usually entire and more or less flat, sometimes crenate and/or undulate. (Including Acerates Ell., 1817; Anantherix Nutt., 1818; Asclepiodella Small, 1933; Asclepiodora Gray, 1877; Biventraria Small, 1933; Oxypterix Greene, 1897; Podostemma Greene, 1897; and Podostigma Ell., 1817. Excluding Gomphocarpus R. Br., [preprinted 1810] 1811; Schizoglossum E. Meyer, 1838; and Xysmalobium R. Br., [preprinted 1810] 1811.) Lecroryee species: 4. syriaca L.; see Hitchcock & Greene, Prop. Brit. Bot. 136. 1929. (According to Linnaeus [Critica Botanica, p. 76, 1737; p. 60 in A. Hort, English translation, 1938], named for the physician Aesculapius; Asclepias listed by Linnaeus [Critica Botanica, p. 102, 1737; p. 83 in A. Hort, English translation, 1938] as an ancient Greek name, but later [Philosophia Botanica, p. 141, 1751] asa name provided by Tournefort; according to Morse, 1985b, named for Asclepios, the Greek god of healing.)— MILKWEED. Considered here in neither the broadest nor the narrowest of senses, Asclepias includes more than 100 North and Central American as well as “‘less than a dozen poorly differentiated” southeastern South American species (Woodson, 1954, p. 28) and another translator arm to | pollinium (partly drawn) from adjacent anther, x 12; e, pollinarium, abaxial side (1 pollinium from each of 2 adjacent anthers, 2 translator arms, and 1 corpusculum with surface grooved for attachment to insects; pollinia and translator arms oriented as in intact anther, x 16; f, gynoecium, with androecium re- moved (the 2 carpels mostly free in ovule-bearing and stylar regions, nied apically into style-stigma head; | vertical furrow, of 5, of style-stigma is Mos nter], with stigmatic surface [below] and position of corpusculum at summit o cial walls eu chamber), x 12. g, A. connivens: pollinarium (cf. “‘e’’), Mes characteristic long tran lator arms, x 12. h, A. pedicellata: side view of flower wit olla lobes remov oe (sepals spreading; c corolla lobes erect; corona sessile, es ad open but geniculate apically; anther head held well above hoods), ~ 5. 494 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 GURE 9. Asclepias. a, b, . ‘taca: a, flower with very young fruit, cut longitudi- nally, after fall of corolla, — ecium, and s tyle-stigma head but before disappearance of calyx and probable abortion of | a sowing 3 calyx lobes, 2 with squamellae at base, 2 very young follicles, mostly free, w seeds, x 8; b, nearly mature follicle, pie on “dete xed pedicel, with aborted carpel at lower right, x 1/2. c-f, 4. incarnata subsp. pulchra: c, distal part of stem with some inflo- rescences and infructescences paired at nodes, young follicles erect on erect pedicels, x Y2; d, mature, dehisced follicle, releasing seeds Won apical tufts of trichomes Wy. @, seed, apical tuft of trichomes removed, seed co Wee laterally, x 3; f, embryo, somewhat expanded from soaking in water ee as in ‘“‘e”’), x 4, Here excluded from Asclepias is Gomphocarpus R. Br., which was tentatively included by Woodson (1954). According to Pobedimova, the latter includes about 100 species in central and northern Africa, about ten in Central and South America, and one in Arabia. Although in both genera the segments of the inner corona are conduplicate (hood shaped), those in Gomphocarpus lack the adaxial appendages (horns or crests) characteristic of but not universally present in Asclepias (some of our species at least appear to lack them: see below). Asclepias may also be related closely to two genera of tropical and southern Africa with laminate (i.e., flat, not conduplicate) inner-corona seg- ments. These structures have and do not have adaxial appendages in Schizo- glossum E. Meyer (about 25 species) and Yysmalobium R. Br. (about ten species), respectively (see Schumann). Good informally recognized four assem- blages of Asclepiadaceae with representatives in both the Old and New worlds: one of them was composed entirely of these four genera. 1989] ROSATTI, APOCYNINEAE 495 Here included in Asclepias are a number of segregate genera that were rec- ognized, for example by Small, as being monotypic and confined to the South- east (Anantherix Nutt., Asclepiodella Small, Oxypterix Greene, and Podostigma Ell.), monotypic and geographically centered in but not limited to the Southeast (Biventraria Small), or composed of several species and relatively widespread in North and Central America (Acerates Ell. and Asclepiodora Gray). According to Woodson (1954), the natural distributions of 27 species of Asclepias include at least part of the southeastern United States, while ne of five others (4. asperula (Dcne.) Woodson, A. Engelmanniana Woodson, A. linearis Scheele, A. oenotherioides Cham. & Schlecht., and A. Sullivantit En- gelm.) were shown to approach this area closely from the west; A. Sullivantii has since been recorded from Arkansas (Smith). The showy-flowered A. curas- savica L., native to Central America and/or the Antilles and/or South America, has been widely introduced in the tropics and subtropics of both the Old and New worlds, including, in North America, parts of the Coastal Plain, from Florida to Texas, and southern California.*3 In most cases, these 33 species can be — distinguished using relatively few vegetative and floral characters (see below). a een (1954) recognized nine subgenera (one of which included nine series), 108 species, and 14 subspecies in his treatment of North and Central American plants. Of these, seven or possibly eight subgenera and seven series are represented in the southeastern United States. Although Woodson (1954) considered these species groups to be mostly unnatural, t they are nevertheless presented in a different order for logistical purposes, and in some cases the characteristics defining them have been modified to reflect my studies of the 33 species that occur or possibly occur without cultivation in the Southeast. Lynch & Martin have suggested that cardenolide profiles obtained from thin- layer eee would be of value to the infrageneric classification of Asclepia este all of the herbarium specimens I have seen are readily identified using criteria presented below. A much longer discussion, and in some cases liquid- preserved or fresh material as well, would be needed to identify the remainder. Vegetative and floral characters have been used instead of ones involving the fruits, at least in part because the latter are not nearly as variable within the genus, are unknown for some species, and are not present on most specimens. Because the following is limited to features discernible on herbarium specimens, 38Whether or not Asclepias curassavica is truly naturalized anywhere in North America is not entirely clear. Long & Lakela stated that it is a pantropical weed that has become natur alized on the Coastal Plain from Florida to Texas and in California, end wvondealin included it amon plants ‘native or naturalized” in central Florida. However, it species of the Florida Panhandle (lel or eas (Jones & Coile), nor r among “some exotic ornamental species that persist after the abandonment of land” (Clewell, p. 1) included with the former. MacRoberts listed the species among those growing without cultivation in Louisiana but did not indicate whether or not the plants were reproducing. Saree & Johnston (p. 1233) referred to it in their treatment of plants in Texas simply as an “alm ubiquitous waif” of the tropics and subtropics in the New World, and Munz (p. 457) eee that in Selorala it is occasionally cultivated and escaped, ‘tas at Mandeville Canyon, Santa Monica Mts 496 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 several characters often used in identification (e.g., color of flowers; shape of abaxial surfaces of hoods) have been avoided. Asclepias pedicellata Walter (Podostigma pedicellata (Walter) Vail), of the Coastal Plain from Florida to North Carolina, is our only representative of subg. PoposTIGMA (EIl.) Woodson. The species appears to be unique within the genus in its erect or nearly erect corolla lobes (vs. reflexed or less often spreading and/or ascending). Furthermore, whereas the anther head is held above and separated from the hoods by a long section of the column in A. pedicellata (see FiGure 8h), it is immediately subtended by them in all other species in the genus, including the other three included in subg. PopostiGMA by Woodson (1954). In fact, although Woodson (1954, p. 48) stated that the anther head is “borne ona stipe. . . high above the hoods” in subg. Popostic- MA, it is only slightly elevated in these other three species, which in this respect grade into subg. ASCLEPIODORA (Gray) Woodson. The hoods are closed at the apex in species of two of our subgenera. In the monotypic subg. ANANTHERIX (Nutt.) Woodson they are spreading or erect at the base and converge above the anther head, and the coronas are stipitate to substipitate (i.e., the column is visible between the bases of the hoods and the corolla). The single representative of subg. ANANTHERIX, Asclepias connivens Baldwin (Anantherix connivens (Baldwin) Feay), occurs on the Coastal Plain in the northern half of Florida and in the southern parts of Georgia, Alabama, and Mississippi. (For the pollinarium of A. connivens, see FIGURE 8g.) In subg. ASCLEPIODORA (Gray) Woodson (seven species, two in the Southeast) the hoods are deflexed at the base but otherwise ascending, although not con- vergent above the anther head, and the coronas are sessile (i.e., the column is not visible between the bases of the hoods and the corolla). Asclepias viridis Walter (Asclepiodora viridis (Walter) Gray), with hoods that are about one-half or less as long as the spreading or ascending corolla lobes, is known from all of the Southeastern States except the Carolinas, although it appears to be most common from Nebraska and Missouri south to eastern Texas and Louisiana and is sporadic east of the Mississippi River. The distribution of 4. asperula, with hoods that are nearly as long as the spreading or ascending corolla lobes, is centered even farther west, in the southwestern United States, and very closely approaches but evidently does not include Arkansas. The hoods are open at the apex (see FiGure 8a, b) in species of the other five subgenera under discussion, although the openings are appressed against the anther head and/or column in subg. AcERATES (Ell.) Woodson (five species, three in the Southeast). The coronas are sessile in Asclepias viridiflora Raf. but clearly stipitate in 4. hirtella Woodson and A. longifolia Michx. (Acerates longifolia (Michx.) Ell., Acerates floridana (Lam.) Hitche., Acerates delticola Small) Asclepias viridiflora is widespread in the eastern two-thirds of the United States as well as in some adjacent parts of Canada, and it occurs throughout the Southeast; however, its distribution appears to be centered in the lower Midwest and adjacent parts of the Central Plains, and it has only recently been reported from Florida (Anderson). Asclepias hirtella (inflorescences lateral, hemispheric to nearly spherical; hoods with small crest at base within) is geo- graphically centered in the lower Midwest but is known from all of our states 1989] ROSATTI, APOCYNINEAE 497 except Florida and the Carolinas, whereas A. /ongifolia (inflorescences mostly terminal, less than hemispheric; hoods unappendaged within) is distributed along the Coastal Plain from Louisiana to Maryland. The hoods are open at the apex and the opening is not appressed against the anther head in the remainder of the 33 species of the Southeast. The hoods at least appear to lack internal appendages in subg. PoLyotus (Nutt.) Woodson (three species, one or possibly two in the Southeast). The hoods are apically trilobed in Asclepias stenophylla Gray, with the shorter middle lobe evidently representing an otherwise completely adnate horn, whereas they are apically unlobed and truly unappendaged in 4. Engelmanniana. Asclepias stenophylla is geographically centered in eastern Nebraska, western Missouri, and Okla- homa east of the panhandle and is probably fairly common in northwestern Arkansas. Asclepias E is distributed more to the west, from which it approaches but apparently does not enter the Southeast. Each hood has an internal appendage in the remainder of the 33 Southeastern species. In most the appendage is an exserted horn (exceeded by to exceeding the hoods), while in some it is a nonexserted crest that probably represents a reduced horn. The coronas are sessile or very nearly so in two of the subgenera involved, subg. ASCLEPIODELLA (Small) Woodson and subg. PoDOSTEMMA (Green) Woodson. The hoods do not exceed the anther head in subg. AscLE- PIODELLA, a group of seven species, two of which are represented in the South- east by plants with linear to filiform leaves. The hoods of Asclepias cinerea Walter (southern South Carolina to northwestern and central Florida) have horns, while those of A. Feayi Chapman (Asclepiodella Feayi (Chapman) Small) (restricted to peninsular Florida) have crests. The hoods exceed the anther head by between about one- and two-thirds their lengths in the six species of subg. PoposTEMMA, one of which, Asclepias oenotherioides Cham. & Schlecht., with broader than linear leaves, is distrib- uted from eastern Texas (near to but evidently not in Louisiana) to northwestern Mexico, southward to Costa Rica. The coronas are not always completely sessile in Asclepias cinerea (subg. ASCLEPIODELLA) and 4. oenotherioides (subg. PODOSTEMMA). The plants in- volved therefore might be confused with members of the last subgenus to be considered, subg. AsCLEPIAS, except that the hoods in A. cinerea have very prominent, erect, marginal auricles that clearly exceed the rest of the hood and those of A. oenotherioides are much narrowed toward the base. The coronas are stipitate or sometimes substipitate in subg. ASCLEPIAS, al- though in some cases the column is hidden by deeply saccate and/or deflexed hoods. Woodson (1954) recognized nine series™ in the subgenus, seven of which are represented in the southeastern United States. In three of these (sers. “‘In- carnatae,” ‘““Tuberosae,” and “‘Exaltatae”) the anther heads, excluding wings and terminal flaps of the anthers, are usually about as long as to longer than 34The names of these series were not validly published by Woodson (1954), since eee were ac- companied neither by a Latin description or diagnosis nor by a reference to a previously and eff published Latin poceca Dood or claenO nt (i.e., they were not st published | in Pe eordanee ea Ad 36. \ (ICBN, 1988], t which is January 1935). Therefore, they are enclosed in quotation marks here. 498 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 broad, while in four (sers. “‘Roseae,” ““Syriacae,” ““Macrotides,” and “‘Purpuras- centes”’) they are usually shorter than broad. Subgenus AscLepiAs ser. “‘Incarnatae” includes about 16 species, four or perhaps five of which occur in the Southeast. The hood margins are without auricles and the leaves are opposite in all of these except Asclepias verticillata L. (throughout our area and the eastern United States in general, also in adjacent parts of Canada, and sporadically to as far west as Arizona), in which the hood margins are usually auriculate and the leaves on individual plants are whorled (in 3’s, 4’s, or sometimes 5’s) or, infrequently, usually whorled and otherwise opposite. The hoods clearly exceed the anther head in A. curassavica (see above) but range from slightly exceeding to clearly being exceeded by it in three other species. The inflorescences are usually paired at the upper nodes and solitary below in 4. incarnata L., while in A. perennis Walter (leaves broader than linear) and A. linearis Scheele (leaves linear to filiform) they are nearly always solitary. Asclepias incarnata is widespread in the eastern three-fourths of the United States and adjacent parts of Canada but is most concentrated in the area north of ours. It seems to be fairly common in Florida, eastern Tennessee, and the Carolinas but is poorly known if not absent in other parts of the Southeast. Asclepias perennis has a curious and possibly artifactual distribution, involving the Coastal Plain from Texas to South Carolina and the lower Ohio River valley. Asclepias linearis is endemic to southeastern Texas, according to Woodson (1954) and Correll & Johnston. Although Small indicated that it extends into our area to as far east as Alabama, the species was listed for neither Louisiana (MacRoberts) nor Arkansas (Smith). In subg. Ascepias sers. “Tuberosae” and “Exaltatae” the hood margins have auricles and the leaves are opposite or irregularly alternate. The auricles are Closer to the hood bases and the hoods clearly exceed the anther head in ser. ““Tuberosae.” This series is represented in our area by three species, one, A. tuberosa L. (with irregularly alternate leaves), widely distributed (to as far west as Arizona) but primarily eastern North American (including southern parts of Canada), and two (with opposite leaves) of the Coastal Plain (from eastern Texas to southern New York). 4sclepias rubra L. (absent in peninsular Florida) has acute hood apices and broadly ovate to lanceolate leaves, while A. lanceolata Walter (common throughout Florida) has rounded hood apices and linear-lanceolate leaves. The auricles are closer to the hood apex (but usually do not exceed the rest of the hood by as much as they do in Asc/epias cinerea of subg. ASCLEPIODELLA), and the hoods never exceed the anther head by much (they range from slightly exceeding it to being slightly exceeded by it) in ser. “Exaltatae.” Two of the nine species in this group are represented in the southeastern United States. Asclepias exaltata L. (leaf-blade bases not clasping stem, peduncles usually lateral and much shorter than subtending leaves), of the northeastern United States and adjacent Canada, enters our range along the Appalachian Mountains to about as far south as northern parts of Georgia and perhaps Alabama. Asclepias amplexicaulis Small (leaf-blade bases clasping stem, peduncles usu- ally terminal and much longer than the subtending leaves) is widespread in the 1989] ROSATTI, APOCYNINEAE 499 eastern two-thirds of the United States (and probably occurs in southern Can- ada), including our area, although it is not known to occur in peninsular Florida. The basis for Woodson’s (1954) inclusion of two of our species and 11 others in ser. “Roseae” is not altogether clear. Among our representatives of this series, the abaxial surface of each hood is exceeded by the surfaces adjacent to the anther head and is not recurved apically in Asclepias tomentosa Ell. (hoods slightly to clearly exceeded by anther head) but clearly exceeds the surfaces adjacent to the anther head and is recurved apically in A. obovata Ell. (hoods clearly exceeding anther head). Both species are known from scattered localities along the Coastal Plain from eastern Texas to the Carolinas, except that 4. obovata (Acerates obovata (Ell.) Eaton) has been reported from neither North Carolina nor peninsular Florida; Clewell (p. 251) indicated that it is a ruderal in the Florida Panhandle in ‘Jackson and Gadsden to Bay Cos.” The abaxial surface of each hood exceeds the surfaces adjacent to the anther head but is not recurved apically in sers. ““Syriacae,” ““Macrotides,” and **Pur- purascentes.”’ The marginal auricles of the hood generally are sharply incised and acute at the apex in ser. ‘““Syriacae”’ but shallow and broad in sers. “Mac- rotides”’ and ‘‘Purpurascentes.” Four of the 11 species belonging to ser. ““Syriacae”’ occur in the southeastern United States. Asclepias viridula Chapman has linear to filiform, opposite leaves and appears to be endemic to northern Florida and perhaps adjacent parts of Georgia, while our other three species have wider, nearly always opposite leaves and are more widespread. Asclepias humistrata Walter (leaf-blade bases clasp- ing stem, leaves opposite) occurs on the Coastal Plain from Louisiana to North Carolina, while 4. quadrifolia Jacq. (leaf-blade bases not clasping stem, leaves whorled or subwhorled at or near first node beneath inflorescence but otherwise opposite) and A. syriaca L. (leaf-blade bases not clasping stem, leaves opposite) mostly grow in the northeastern quarter of the United States and adjacent parts of Canada, including in our area northern Arkansas and the Appalachian Moun- tains to about as far south as Georgia and Alabama. Asclepias syriaca is wide- spread in the northeastern quarter of the United States and adjacent parts of Canada; it may be native to the northern parts of our range but naturalized farther south (see Woodson, 1954; Broyles & Wyatt). One of the eight species included in ser. ‘“‘Macrotides,”’ Asclepias Michauxii Dene. (A. angustifolia Ell.), occurs in the Southeast. It 1s endemic to the Coastal Plain from Louisiana to South Carolina, including peninsular Florida, and can be distinguished from our species of ser. ““Purpurascentes” by its irregularly alternate, filiform to linear (vs. opposite, broader than linear) leaves. Three or four of a total of eight species comprising ser. ““Purpurascentes”’ occur in the Southeast. The hoods are acute to narrowly rounded apically and exceed the anther head by more than half their length in Asclepias purpurascens L. and 4. Curtissii Gray (Oxypterix Curtissii (Gray) Small). Asclepias purpu- rascens (plants erect) occurs primarily in the northeastern quarter of the United States and in adjacent parts of Canada but enters our range in parts of Louisiana, Arkansas, Tennessee, and North Carolina, whereas 4. Curtissii (plants decum- bent) is endemic to peninsular Florida. The hoods are broadly rounded to 500 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 truncate apically and exceed the anther head by less than half their length in A. variegata L. (Biventraria variegata (L.) Small) and A. Sullivantii. Asclepias variegata (leaf-blade bases not clasping stem) is widespread in the area southeast of a curve from about southern New England to Arkansas and eastern Texas but has not been reported from peninsular Florida. Asclepias Sullivantii (leaf- blade bases clasping stem) is mainly distributed in the southern half of the Midwest (reported also from southern Canada) and enters our area only in Jackson County, Arkansas (Smith). Woodson (1954) recognized infraspecific taxa in only three of the 33 species mentioned above. On the basis of morphological, ecological, and geographic evidence, two subspecies, both of which occur in the Southeast, were recognized in Asclepias incarnata. Subspecies pulchra (Ehrh.) Woodson (plants conspic- uously pubescent instead of essentially glabrous, leaves broader than in the typical element, and other minor differences) occurs along the Atlantic Coast (as well as in adjacent, inland areas) in Florida and from Maine and Nova Scotia to North Carolina, where it frequently grows in brackish tidal marshes, while subsp. incarnata has a more inland distribution (essentially throughout the Midwest, but with widely scattered stations elsewhere) and is only associated with fresh water. The possibility that the morphological characteristics of subsp. pulchra are induced by saline environments apparently has not been investi- gated. Woodson (1954) recognized four subspecies in Asclepias tuberosa based on leaf shape and to some extent geography; three of them are known from the Southeast. However, he had earlier demonstrated “extremely active” introgres- sion among them (Woodson, 1953), and I have observed them to be morpho- logically indistinct. Two subspecies were treated in A. asperula, the more eastern of which, subsp. capricornu (Woodson) Woodson (sessile to subsessile instead of clearly pedunculate inflorescences, in addition to other differences) was re- ported from a locality in McCurtain County, Oklahoma, within about 40 miles of Arkansas and was said sometimes to spread to pastures from more natural habitats. According to Woodson (1954), few North American species of Asclepias occur at elevations above 2000 meters, and a minority (perhaps as many as 12 of ours) are distributed as far north as Canada. Most of the species discussed above appear to be optimally adapted to relatively open, dry habitats such as pine barrens, flatwoods, prairies, glades, thickets, and one dunes. In pane to A. incarnata (see above), other species at least someti urring in moister habitats than these include 4. asperula (desert swales), A. connivens Ga low areas in woods), A. curassavica (moist places), A. Engelmanniana (washes, bottoms, swales), 4. exa/tata (moist woods and meadows), A. Fie (swamps, low pine lands), 4. perennis (low swampy ground), 4. rubra (bogs, marshes, wet meadows, low pine barrens), and A. syriaca (alluvial bottoms); A. lanceolata is found in wet pine barrens and low glades, as well as in both fresh water and brackish marshes. Although Woodson (1954) stated that, among the 33 species in our area, only 4. curassavica, A. syriaca, and possibly 4. incarnata are to any extent weedy, he indicated that 4. amplexicaulis, A. Feayi, A. hirtella, A. 1989] ROSATTI, APOCYNINEAE 501 obovata, A. oenotherioides, A. purpurascens, A. Sullivantii, and A. viridiflora often spread to roadsides, old fields, and/or railways from more natural habitats. Natural hybrids between species of Asc/epias appear to be very rare, pre- sumably because of mechanical isolation effected by sets of rather exact pol- lination requirements. Natural and fertile artificial hybrids have been reported in subg. ASCLEPIAS between A. exaltata, of ser. “Exaltatae,” and A. syriaca, of ser. ‘“Syriacae”’ (see Kephart & Heiser; Kephart et a/.). Included by Woodson (1954) in a list of nine apparent but untested hybrids are some specimens combining the characteristics of A. amplexicaulis (ser. ““Exaltatae’’) and A. humistrata (ser. ““Syriacae”’), from Florida; A. amplexicaulis and A. syriaca, from New York; and 4. syriaca and A. viridiflora (subg. ACERATES), from Indiana. Other specimens intermediate between one of the 33 species in the Southeast and another species were also reported. These involve A. verticillata and the extraregional A. pumila (Gray) Vail (both of ser. “Incarnatae”’); A. syriaca and the extraregional A. speciosa Torrey (ser. ““Purpurascentes’’), be- tween which hybrids were obtained experimentally (see also Thomson & Wag- ner); and A. oenotherioides and the extraregional A. Emoryi (Greene) Vail (both of subg. PoposTemMMa), between which intermediates are frequent in nature. Plants grown by Moore (1946b) from seeds probably resulting from natural pollinations and produced by plants of 4. speciosa growing in experimental plots with A. syriaca were intermediate in flower, stem, and leaf character states at least in part defining these two species. Moore (1946b) attempted to produce artificial hybrids involving several of the species discussed here, for the most part unsuccessfully. Although ovary enlargement, presumably resulting from successful fertilization, was observed for crosses of Asclepias tuberosa (ser. ‘“‘Tuberosae’’) with A. Sullivantii (ser. ‘“‘Purpurascentes’’), A. curassavica (ser. ‘“Incarnatae’’) with A. syriaca (ser. “Syr- jacae’’), and A. curassavica with A. speciosa (ser. ‘‘Purpurascentes’’), the fruits and seeds did not mature. Moore’s list of even-less-successful artificial crosses (i.e., those after which no ovary enlargement was observed) included various other combinations involving these species and A. incarnata, A. Sullivantii, and A. tuberosa, to which Woodson (1954) added A. curassavica crossed with A. quadrifolia (ser. ““Syriacae’’) and A. purpurascens (ser. ““Purpurascentes’’), A. verticillata crossed with A. incarnata (both of ser. “Incarnatae’’), and others involving these and two extraregional species. Cytological evidence has not been useful in the evaluation of putative hy- brids, primarily because all of the species involved are of the 2 = 22 cytotype. Otherwise it has been of interest in the sense that the chromosomes of Asclepias curassavica are significantly smaller than those of the other species investigated by Moore (1946b), all native to North America, consistent with the idea that this species is a more distantly related introduction to the continent. Even though fruits are not especially variable within the genus, the spatial relationships between the follicle and pedicel appear to be of phylogenetic importance (Woodson, 1954). During fruit maturation in most species, the pedicel becomes deflexed while the follicle remains erect. Despite being com- mon within the genus and (supposedly) in the rest of the Asclepiadeae (but see SUZ JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 generic descriptions here), this condition was thought to be relatively advanced because it is unknown in the rest of the Asclepiadaceae. The production of erect follicles on erect pedicels or pendulous follicles on pendulous ones is more characteristic of the family as a whole, and although less common in Asclepias, was therefore considered to be more primitive than erect follicles on deflexed pedicels. Erect or pendulous pedicels are most frequent in subg. ASCLEPIAS, while erect follicles on deflexed pedicels are unknown in ser. “Incarnatae”’ (cf. FIGURE 9c); these taxa were considered on the basis of floral morphology to be the most primitive within their genus and their subgenus, respectively. Despite the foregoing, erect follicles on erect pedicels or pendulous follicles on pendulous ones occur as well in other subgenera of Asclepias and in other series of subg. AscLepiAs. Among the Southeastern species, erect follicles and pedicels are known in A. incarnata, A. linearis, A. curassavica, and A. verticillata of ser. “Incarnatae,” as well as in A. quadrifolia and A. viridula of ser. “Syr- iacae,” A. Michauxti of ser. ‘‘Macrotides,” A. cinerea and A. Feayi of subg. ASCLEPIODELLA, and A. stenophylla of subg. PoLyotus. Among this same group, pendulous follicles and pedicels (and, incidentally, noncomose seeds) occur only in A. perennis of ser. “‘Incarnatae’’, the fruits of A. Curtissii and A. pedi- cellata are evidently unknown. Pendulous follicles and pedicels also occur in extraregional species of ser. ‘“‘Roseae”’ and subg. PoposTEMMA. Woodson (1954) considered most of these species to be primitive within their respective series and subgenera on the basis of floral morphology. The precise location of nectar production in flowers of Asclepias has been controversial. The hoods were generally accepted as the site, probably because of the conspicuous accumulation of nectar there, until, according to Galil & Zeroni (1965), Stadler concluded in 1886 that it was produced in the stigmatic chambers as well. More recent workers reverted to the original idea, and some even regarded the horn as its ultimate source. Working with A. curassavica in Israel, Galil & Zeroni (1965) showed that nectar accumulates both in the hoods and in the stigmatic chambers, although histochemical and other tests dem- onstrated that it is produced only in the latter (see also Schnepf & Christ). Using very small pipettes and a solution of ten percent sucrose with a few drops of india ink, they also established that liquid passes from the base of each stigmatic chamber into the two nearest hoods through passageways largely defined by the lobules of the corona and the bases of the two adjacent anthers; that is, each lobule causes the nectar to flow from the base of the stigmatic chamber laterally into the bases of the two nearest hoods, instead of down the on pelween them. The dently allow for lation of the massive amounts of nectar required by insect visitors large enough both to fly the great distances often separating individual plants and effectively to engage the pollination mecha- nism involved, which in various ways favors outcrossing. In many cases during nectar-feeding visits, one (or more) of an insect’s appendages (usually a leg or a head part) becomes wedged in the abaxial groove of the corpusculum, to which it is sometimes guided from below by the outwardly projecting, apically narrowing slot (and associated, upwardly oriented bristles) between the anther wings. Visitors of sufficient strength depart for another flower intact, carrying 1989] ROSATTI, APOCYNINEAE 503 away the corpusculum with attached translator arms and pollinia, while others remain trapped or leave behind the wedged part because the pollinia are too heavy and/or too tightly held within the anthers. According to most accounts, the translators begin to rotate within a few minutes after removal from the flower, continuing until the wide axes of the pollinia are perpendicular to their original orientation and their more sharply convex sides are adaxial. This action orients pollinia such that the same pol- linator behavior that effected their removal from the anthers of one flower can result in their insertion, one at a time, into the outwardly projecting anther wing slots of another, with their more sharply convex sides inward, adjacent to the receptive surfaces within the stigmatic chambers. Because such effective in- sertion of pollinia into the slots is substantially more likely after translator rotation, delay of the latter until after the pollinator is likely to have departed for another individual is generally thought to promote outcrossing. However, my observations of Asclepias syriaca indicate that the translator arms are under tension while the pollinarium is in place, and that they begin to rotate im- mediately upon removal. Pollinator activity at a flower visited after translator rotation in the polli- narium being transported sometimes results in the insertion of one pollinium of the pair into a stigmatic chamber, usually from somewhat below. As the insect moves its trapped appendage toward the apex of the flower in attempting to free itself, the inserted pollinium may become detached by pressure exerted against the anther wings forming the anther slot, so that the inserted pollinium remains in the stigmatic chamber. At other times the translator arm or the insect’s appendage to which it is attached becomes wedged in the abaxial groove of the resident corpusculum, forcing detachment of the inserted pollinium, perhaps along with the translator arm and sometimes the insect’s appendage as well. Continued pollinator movement may also result in removal of the resident pollinarium in the manner described above. If an inserted pollinium remains attached to its translator arm and to the insect’s appendage and if either of the latter is wedged in the resident corpusculum upon its removal, the inserted pollinium will be removed along with the resident pollinarium. Successive repetitions of such occurrences often result in chains of pollinaria. Wyatt (1976) advanced the idea that inserted pollinia are more likely to remain so when the associated resident pollinarium (especially the corpuscu- lum) is still intact and in place. (For extraregional species, see also Lynch.) That removal of resident pollinaria is possible without upsetting a previous insertion was offered to explain unexpected observations that the numbers of insertions into chambers associated and not associated with intact resident pollinaria are frequently about equal. However, Wyatt (1978) later provided experimental evidence from manipulated flowers in natural populations of Asclepias tuberosa in North Carolina that insertions were in fact more numerous in chambers associated with intact resident pollinaria. He believed that these findings confirmed his original hypothesis that the corpusculum helps to ensure that the translator arm will be broken and the inserted pollinium will remain behind. Such a mechanism presumably favors outcrossing, and it perhaps simultaneously decreases the wasteful placement of pollinia, to the extent that 504 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 insertion of pollinia would be less likely into the stigmatic chambers of flowers that have already been pollinated (because one or more of the five corpuscula would have been removed). Most evidence suggests that Hymenoptera generally are more frequent pol- linators of Asclepias flowers than are Lepidoptera, and that the latter in turn are more frequent than Coleoptera, Diptera, and Hemiptera. Such was indicated in Robertson’s (1928) compilation in Illinois, in which visitors with pollinaria, among others, were listed for each of eight species represented in the Southeast (A. incarnata, A. purpurascens, A. Sullivantii, A. tuberosa, A. longifolia, A. syriaca, A. verticillata, and A. viridiflora), ruby-throated hummingbirds (4r- chilochus colubris) were reported to visit the flowers of the first four of these, although none was observed with pollinaria. More detailed and recent studies also have indicated that, in general, Hy- menoptera are more important than Lepidoptera in the pollination of Asclepias flowers (for exceptions, see Bierzychudek; Wyatt, 1980a). Wyatt (1976) found that, in comparison with localities farther east, many pollinaria were removed but few inserted among 15 populations of 4. tuberosa in eastern Kansas and western Missouri that were visited by relatively small numbers of Hymenoptera (mostly bumblebees, honeybees, and wasps) and large numbers of Lepidoptera (primarily monarch, swallowtail, and sulfur butterflies). Although butterflies (especially monarchs) were frequent visitors to flowers of A. syriaca in a coastal area in Maine, bumblebees were usually ten to 20 times more numerous and carried pollinaria four times as often (Morse, 1985b). Morse (1985b) also noted that, although at night flowers of Asclepias syriaca were quite fragrant, produced considerably more nectar than during the day, and were visited by large numbers of moths, Lepidoptera were responsible for only five to 25 percent of the total number of pollinial insertions (day and night) (but see Willson & Bertin). Nevertheless, Bertin & Willson demonstrated with 4. syriaca and A. verticillata that, although insect visitors and pollinial insertions were less frequent at night, seed-set and “pod-set” were equal or greater, suggesting that nocturnal pollinators may be superior to diurnal ones. Bertin & Willson also suggested that the predominance of diurnal visitors may be due largely to the introduction of the honeybee (Apis mellifera). Data provided by Willson, Bertin, & Price indicated that for Asclepias ver- ticillata Hymenoptera were more frequent carriers of pollinaria than were Lepidoptera at three of four study sites in Illinois, and that whereas the former were more important than the latter at the least disturbed site (70 percent vs. 0 percent ofall vectors, respectively), the opposite was true at the most disturbed site (15 percent vs. 51 percent). Pollinaria of three species of Asclepias represented in the Southeast were found by Kephart (1983) on a wide variety of insects collected from experi- mental plantings in Indiana. Although close associations between individual plant and pollinator species were not evident, 4. syriaca was associated with a narrower range (Hymenoptera, including bees in Apis and Bombus and wasps in Polistes and Sphex; Lepidoptera in Cisseps, Danaus, and Speyeria,; Coleop- tera in Chauliognathus; and Diptera and Hemiptera in unspecified genera) than either A. incarnata (bees in Apis, Bombus, and Xylocopa, wasps in Cerceris, 1989] ROSATTI, APOCYNINEAE 505 Myzinum, Polistes, and Sphex; Lepidoptera in Battus, Cisseps, Danaus, and Speyeria; Coleoptera in Chauliognathus;, and Diptera and Hemiptera in un- specified genera) or A. verticillata (all taxa mentioned above for A. syriaca and A, incarnata except Speyeria; also wasps in Entypus). Beare & Perkins’s field observations of two populations of Asclepias tuberosa in south-central Oklahoma revealed the existence of fruiting and nonfruiting individuals. The slots over the stigmatic chambers were significantly wider than the pollinia were thick in plants that developed fruits, while the reverse was true of those that did not (pollinial thickness was constant). Fruiting in- dividuals clearly had greater numbers of pollinial insertions and greater per- centages of intact pollinaria as well, supporting Wyatt’s (1976) hypothesis that the likelihood of pollinial insertion is enhanced by the presence of intact pol- linaria. Dimorphism in slot width would favor outcrossing to the extent that flowers with narrow slots are much less likely to be pollinated in general (3.45 percent vs. 96.55 percent insertion) and therefore to be self pollinated in par- ticular. Evidently, nonfruiting plants act primarily as pollinial donors. A variety of explanations, including infrequent, ineffective, or disruptive insect visitation, resource limitations for developing ovules and fruits, and/or various levels of physiological incompatibility, have been offered to account for the low percentage (ca. | percent according to Woodson, 1954) of flowers that develop into mature fruits in species of Asclepias. That pollination does not limit fruit-set has been suggested by various studies of Asclepias syriaca in North America in which the percentage of flowers with inserted pollinia has ranged from about 65 among natural populations in south- eastern lower Michigan (Sparrow & Pearson) to about 95 among experimental plants in southeastern Ontario (R. J. Moore, 1947). However, much lower (but unspecified) values were obtained for 4. tuberosa in natural populations throughout much of the eastern half of the United States, perhaps at least in part because pollinia of this species are thought to be removed but not inserted by some species of butterflies (Wyatt, 1976). Low insertion rates in Asclepias tuberosa might also be due to nectar-robbing ants, on which pollinaria of this species have been observed in large numbers (Wyatt, 1976). Indeed, insertions were more frequent among inflorescences of A. curassavica in Costa Rica from which ants were excluded (Wyatt, 1980a). Thus, nectar-robbing can reduce both the number of pollinaria and the amount of nectar available to legitimate pollinators, some of which, especially in A. curassavica (Galil & Zeroni, 1965), have been observed to remove nectar di- rectly from the stigmatic chambers of flowers in which the hoods were emptied, resulting in the relatively ineffective placement of pollinaria on head parts (see also Fritz & Morse). Wyatt (1980a) suggested that the opposite, sessile, am- plexicaulous leaves of several species (e.g., 4. amplexicaulis, A. humistrata, A. Sullivantii) might reduce the number of ants reaching inflorescences and thereby enhance sexual reproduction. Fruit-set among species of Asc/epias may also be limited by physiological incompatibility systems and resource limitation. Sparrow & Pearson found that only flowers pollinated with pollinia from distant plants or from clones producing fruits of a different type (see Sparrow) developed mature fruits. All 506 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 other pollen was either completely ineffective or resulted in fruit abortion at various stages early in development. Willson & Price (1980) demonstrated that various aspects of seed and fruit production by uncultivated plants of A. in- carnata, A. syriaca, and A. verticillata in Illinois were generally increased by the addition of nitrogen, phosphorus, and potassium and were decreased by defoliation and perhaps by shading. Not only is fruit-set in Asclepias infrequent overall, but almost invariably only one of the two carpels per flower develops into a mature fruit. Sparrow & Pearson reported that in artificially pollinated flowers of A. syriaca, pollen tubes from a single stigmatic cavity enter only a single style, suggesting that in general the ovules of one carpel may be fertilized only by pollen in two adjacent chambers, and that those in the other might receive pollen tubes only from the other three. If this is indeed true, fertilization of both carpels would require pollinial insertion in both groups of stigmatic chambers. The rarity of natural hybrids between species of Asclepias, despite the close geographic and/or ecological proximity of many, has usually been attributed to mechanical isolation effected by a specific fit between stigmatic chambers and pollinia (1.e., the “lock and key” of many authors, the “tongue-in-groove”’ of Beare & Perkins) of individual species (see especially Holm). Vector spec- ificity has not been invoked because pollinaria of several diverse species have often been observed on individual insects. Kephart & Heiser studied pollination in natural and experimental popula- tions of three sympatric, mostly self-incompatible species of Asclepias in In- diana, 4. incarnata, A. syriaca, and A. verticillata. Despite statistically signif- icant differences among the species in the sizes of both stigmatic chambers and pollinia, effective pollinations were observed for all combinations except those involving the relatively large pollinia of 4. syriaca and their insertion into the comparatively small stigmatic chambers of 4. incarnata and A. verticillata: pollinaria of A. tuberosa (and other species) were also occasionally found in the stigmatic chambers of all three species. Thus, Kephart & Heiser concluded that, while such mechanical barriers to hybridization are neither as effective nor as significant as was previously thought, they and other pre-pollination barriers may nevertheless sometimes contribute to decreased interspecific pol- lination. Indeed, Macior’s field observations of Asclepias incarnata, A. syriaca, and A. verticillata in southeastern Wisconsin suggested that other floral character- istics affect the positioning of pollinaria on pollinators. Poor footing on flowers and wide spaces between hoods (as in A. syriaca) were positively correlated with pollinarial placement on insect pulvilli, while good footing and narrow spaces (as in A. incarnata and A. verticillata) seem to favor placement on the tarsal claws subtending them, at least in part because the tarsus in such cases is less likely to be drawn through the stigmatic chambers directly from below. Macior also implied that behavioral diversity among insects might in some cases decrease the likelihood of cross-pollination. That the foregoing may be, nevertheless, ineffective in preventing the misplacement of pollinia is indicated by his observations of chains composed of pollinaria belonging to two or more of these and other species. 1989] ROSATTI, APOCYNINEAE S07 The failure of all of 279 attempted artificial crosses involving Asc/lepias incarnata, A. syriaca, and A. verticillata as reported by Kephart (1981) indicates that post-pollination barriers are considerably more important than those dis- cussed above. Pollen-tube penetration into the styles and, in some crosses between A. incarnata and A. verticillata, into the ovules as well suggests late pre- or early post-fertilization, physiological incompatibility. The evidently weaker barrier between 4. incarnata and A. verticillata may imply greater genetic similarity between them, consistent with their separation from 4. syr- iaca at the level of series by Woodson (1954). M insects are rendered unpalatable by ingestion of plants in several species of Asclepias (as well as other Apocynineae), due primarily to the presence of cardiac glycosides in the latex (e.g., see Brower ef al.; Duffey & Scudder; Erickson; Euw et al.; Isman et a/.; Rothschild; Seiber et a/.). The resemblance of the palatable viceroy butterfly to the unpalatable monarch is perhaps the best-known and most frequently taught example of Batesian mimicry. Bailey and colleagues listed about 20 of our species of Asclepias as being cultivated in North America. Some are grown as ornamentals, while others have been evaluated as potential crop plants for a variety of purposes (see family discussion). Several species in the southeastern United States were used by ‘‘Indians” in various ways: latex for chewing gum; bark fiber to make rope and string; buds, young shoots, leaves, and fruits for food; and dried roots for medicine (Uphof). Asclepias curassavica 1s one of the most common weeds in the lowlands of Central America, where the latex is placed in the nostrils to treat head colds and in cavities to relieve toothache; it 1s also sometimes used internally as an emetic and purgative and is believed to be useful in the treat- ment of poisonous bites (Standley & Williams). The many uses of A. incarnata, A. syriaca, and A. tuberosa have been discussed by M. I. Moore. In addition to the literature discussed above, see annotations of the references below for information about overall biology (Bhowmik & Bandeen), general ecology (Blakley & Dingle; Price & Willson, 1976, 1979; Wilbur, 1976), pol- lination biology and breeding systems (Boyden; Chaplin & Walker; Corry, 1883a, 1883b; Dailey et a/.; Frost; Galil & Zeroni, 1969; Kephart, 1979; Morse, 1981, 1982; Morse & Fritz, 1983, 1985; Pleasants & Chapin; A. K. Southwick & Southwick; E. E. Southwick; Willson & Price, 1977), fruit and seed produc- tion (Bell; Bookman, 1983a, 1983b, 1984; R. J. Moore, 1946a; Morse, 1987; Queller, 1983, 1985; Wilbur, 1977), seed germination (Baskin & Baskin; Farmer et al.: Jeffery & Robison; Kis; Morse & Schmitt; Oegema & Fletcher), and the taxonomic significance of latex (Moyer). REFERENCES: Under references for the Apocynineae, see BAILEY ef a/.; CLEWELL; CORRELL & JOHNSTON; Jones & Corte; LonG & LAKELA; MACROBERTS; MUNZ; POBEDIMOVA; ROBERTSON (1928); SCHUMANN; SMALL (1933); SmitH; STANDLEY & WILLIAMS; UPHOF; and WUNDERLIN. Under references for the Asclepiadaceae, see Goop and SPELLMAN (1975b). Under references for Sarcostemma, see HOLM. ANDERSON, L. C. Noteworthy plants from north Florida. Sida 10: 295-297. 1984. [4. viridiflora new to state (Gadsden Co.). 508 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 BaskIN, J. M., & C. C. BAskin. Germination of common milkweed (Asclepias syriaca L.) seeds. Bull. Torrey Bot. Club 104: 167-170. 1977. [Seeds obtained and experi- ments conducted in Kentucky; initially dormant at maturity in October; cold-strat- ified seeds germinated about equally in light and dark; some evidence that longer periods of stratification (nine weeks vs. one to three) allowed for higher germination rates at lower ols a aaag ] Beare, M. H., & W.E. Perkins. Effects of variation in floral morphology on pollination sae ae in sce tuberosa L., butterflyweed (Asclepiadaceae). Am. Jour. 9: 579-584. BELL, c R. Pollen on in relation to ovule number in two species of Asclepias (Asclepiadaceae). (Abstract.) ASB Bull. 21: 38. 1974. [4. syriaca weedier and with more seeds per follicle than 4. exaltata (250 vs. 75).] & J. M. FARMER. Dynamics of pollen flow in Asclepias syriaca (Asclepiadaceae). P. 145 in Second International Congress of Systematic and Evolutionary Biology, Vancouver, B. C., Canada, ay 17-24, 1980. The University of British Columbia, Vancouver, B. C., Canada. BerTIN, R. I., & M. F. WiLttson. Effectiveness of diurnal and nocturnal pollination of two nillkwceds, Canad. Jour. Bot. 58: 1744-1746. 1980. BuHowmik, P. C., & J. D. BANDEEN. The biology of Canadian weeds 19. Asclepias syriaca L. Canad. Jour. Pl. Sci. 56: 579-589. 1976. [Morphological variation, economic importance, geography, ecology, growth and development, reproduction, hybrids, population in cs, responses to herbicides and other human manipulations, ef- fects of parasi BIERZYCHUDEK, 2 deat Lantana, and Epidendrum: a floral mimicry complex? as 13(Suppl., Saat Botany): 54-58. 1981. [Butterfly species in five milies by far the most common visitors to A. curassavica in Costa Rica, and the only ones to carry pollinia: bumblebees observed to visit flowers but not to carry BLAKLEY, N. R., & H. DINGLE. Competition: butterflies eliminate milkweed bugs from a Caribbean island. Oecologia 37: 133-136. 1978. [By eliminating the food plant, A. curassavica. BooKMAN, S. S. Evolutionary and ecological bases for selective fruit maturation in Asclepias. 118 pp. Unpubl. Ph.D. dissertation, Washington State Univ., Pullman. 1982.* (Diss. Abstr. B. 43(2): 326, 327. 1982.) ——. Costs and benefits of flower abscission and fruit abortion in Asclepias speciosa. Ecology 64: 264-273. 1983a. [Ninety-seven percent of ovaries fail to develop into mature fruits; relatively low cost of producing surplus ovaries at least partially offset by benefits derived from competition between them during fruit development (i.e only the most fit mature), but see BOOKMAN, 1983b. . Effects of pollination timing on fruiting in Asclepias speciosa Torr. (Asclepia- daceae). Am. Jour. Bot. 70: 897-905. 1983b. [Within an umbel, earlier pollination associated with greater likelihood of fruit maturation. Evidence for selective fruit production in Asclepias. Evolution 38: 72-86. 1984. jo enter each ovary; 70 percent failure rate attributed mostly to fruit abortion caused by post-fertilization self-incompatibility and competition between fruits of different paternal parentage. ] Boypen, T. C. Floral mimicry by Clete ibaguense (Orchidaceae) in Panama. Evolution 34: 135, 136. 1980. [Non-nectar-producing species mimicking two that do, Lantana Camara L. ence aa A. curassavica;, each of ten captured monarch butterflies with pollen or pollinia of all three species. Brower, L. P., P. B. McEvoy, K. L. WiLtiAMson, & M. A. FLANNERY. Variation in cardiac glycoside content of monarch butterflies from natural populations in eastern North America. Science 177: 426-429. 1972. [Concentration mean and variance decrease southward 1989] ROSATTI, APOCYNINEAE 509 Broy es, S. B., & R. Wyatt. Asclepias syriaca in Georgia. Bull. Torrey Bot. Club 115: 297, 298. 1988. [Well-established, evidently naturalized population in Barrow Co.] CHAPLIN, S. J.. & J. L. WALKER. Energetic constraints and adaptive significance of the floral display of a forest milkweed. Ecology 63: 1857-1870. 1982. [4. quadrifolia.] Corry, T. H. On ae mode of development of the pollinium in Asclepias Cornuti, Decaisne. Trans. Linn. Soc. London, II. Bot. 2: 75-84. p/. 16. 1883a. [A. syriaca.] n the cae a development of the gynostegium, and the mode of fertil- ization in Asclepias Cornuti, Decaisne (A. syriaca, L.). [bid. 173-207. pls. 24-26. 1883b. [Extensive observations of pollination; helpful illustrations. ] DalLey, P. J., R. C. Graves, & J. M. KinGso.ver. Survey of Coleoptera collected on the common milkweed, 4sclepias syriaca, at one site in Ohio. Coleopterists Bull. 32: 223-229. 1978. [Species, total number of individuals, and dates collected.] a S. S., & G. G. E. Scupper. Cardiac glycosides in North American Asclepia- aceae, a basis for unpalatability in brightly coloured Hemiptera and Coleoptera. ae Insect Physiol. 18: 63-78. 1972. [Nineteen species of Asclepias with different numbers and types of cardenolides: content also varies from tissue to tissue within Erickson, J. M. The utilization of various Asclepias species by larvae of the monarch butterfly, Danaus plexippus. Psyche 80: 230-244. 1973. [Larvae reared on A. curassa- vica seem no less viable than those reared on less toxic species, including A. incarnata, Euw, J. V., L. FisHetson, J. A. PARSONS, T. REICHSTEIN, & M. ROTHSCHILD. Cardenolides (heart poisons) in a grasshopper feeding on milkweeds. Nature 214: 35-39. 1967. [Incorporated in tissues and in fluid sie from defensive glands. ] FARMER, J. M., S. C. Price, & C. R. Bett. Population, temperature, and substrate influences on common milkweed (4 eee syriaca) seed germination. Weed Sci. 34: 525-528. 1986. [Highest percent germination obtained (all substrates) ee an alternating temperature regime of 16 hours at 20°C and eight hours at 30°C; c relation strong with seed source, weak with seed size.] Finn, W. W. Male cells in angiosperms I. ae and fertilization in Asclepias Cornuti. Bot. Gaz. 80: 1-25. 1925. pls. 1-3. [A. syriaca. Fritz, R. S., & D. H. Morse. Nectar parasitism of . ner syriaca by ants: effect on nectar levels, pollinia insertion, pollinaria removal and pod production. Oecologia 50: 316-319. 1981. [Higher fruit production in ean plants despite no dif- ferences in rates of pollinia insertion. Frost, S. W. Insects and pollinia. Ecology 46: 556-558. 1965. [Pollinia from ae species of Asclepias on a single pollinator; of Lepidoptera taken at night near Un versity Park, Pennsylvania, pollinia “in noticeable numbers” only on Halisidota tessellaris, a moth; insects taken with pollinia on flowers of A. (ncarnata, A. syriaca, and A. tuberosa included mostly Hymenoptera. Gacer, C. S. The development of the pollintum and ee -cells in Asclepias Cornuti, Decaisne. Ann. Bot. 16: 123-148. p/. 7. 1902. [A. syri GautL, J., & M. ZERONI. Nectar system of Asclepias curassavica. Bot. Gaz. 126: 144- 148. 1965 & : the organization of the pollinium in Asclepias curassavica. Ibid. 130: 1-4. 1969. [Pollen tubes grew toward convex side of pollinium (side normally adjacent to stigmatic surface after insertion), toward area of tiny notches, even after cuticle removed from other side, suggesting a complex control system. Gitmartin, A. J. Numerical phenetic determination of the taxonomic status of Asclepias Giffordii. Bull. Torrey Bot. Club 104: 496-505. 1980. [Previously thought to be a Sia peace A. es Bentham and A. spec iosa ee Wer Of western North Ame synonymy under t to be as distinct as eae species in family. IsMAN, M. B., S. 8S. Durrey, & G. G. E. Scupper. Cardenolide content of some leaf- and stem-feeding insects in temperate North American milkweeds (Asclepias spp.). 510 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Canad. Jour. Zool. 55: 1024-1028. 1977. [Suggested that relatively ai insects feed- ing primarily on milkweeds used the obtained cardenolides in defe JEFFERY, L. S., & L. R. Ropison. Growth characteristics of common Seer Weed Sci. 19: 193-196. [Effects of time, temperature, and moisture on seed ger- mination in 4. syria KEPHART, S. R. The floral ecology and reproductive isolation of three sympatric species of Asclepias. 161 pp. Unpubl. Ph.D. dissertation, Indiana Univ., Bloomington. 1979. (Diss. Abstr. B. 40(7): 2976. 1979 [1980].) [Post-pollination incompatibility more important than pre-pollination barriers between species, except that pollinia of A. syriaca are too large to be inserted in flowers of A. incarnata and A. verticillata. Breeding systems in Asclepias incarnata L., A. syriaca L., and A. verticillata L. Am. Jour. Bot. 68: 226-232. 1981. [Some wel ce yielded mature follicles in A. incarnata (29 percent) and A. syriaca (45 percent), but not in A. verticillata.] The partitioning of pollinators among three species of Asclepias. Ecology 64 120-133. SER, JR. Syren ae in Asclepias: lock and key hypothesis ee ee | eee 34: 738-746. 1980. , R. Wyatt, & D. PARRELLA. ate idization in North American Asclepias. 1. Morphological evidence. Syst. Bot. 13: 456-473. 1988. [Biometric study of artificial hybrids and natural populations in Michigan, Virginia, and West Virginia revealed hybridization, backcrossing, and introgression between A. exaltata and A. syriaca; principal components from analysis of 17 vegetative and 13 reproductive characters suggested intermediate placement of putative F, hybrids between distinct parental species; seeds from artificial crosses yielded fertile F, hybrids. Kis, B. Effects of after-ripening upon germination and growth of Asclepias tuberosa. Bull. Bot. Club Wisconsin 16(1): 3-7. 1984. [Include inhibition of damping-off and greater initial growth rates.] Kiemow, K. M., & D. J. RAyYNAL. Demography of the polycarpic herbaceous perennial . Asclepias tuberosa (butterflyweed) in a chronically unproductive site. Am. Midl. Nat. 115: 374-385. 1986. [Decline in population size during four relatively dry years in an abandoned limestone quarry. ] Lucansky, T. W., & K. T. CLlouGH. Comparative anatomy and morphology of Asclepias perennis and Asclepias tuberosa subspecies Rolfsii. Bot. Gaz. 147: 290-298. 1986. [Leaves of both have greater length-to-width ratios in dry, sunny sites than in wet, shaded ones; secondary growth in stems of both taxa involves cambial activity; A. perennis with slender, adventitious roots, 4. tuberosa subsp. Rolfsii with a thick taproot; many other details.] Lyncu, S. P. The floral ecology of Asclepias ee Woods. Madrono 24: 159-177. 1977. [Corpusculum thought to serve as point of translator-arm wedging and break- age in this northern-Californian endemic; ie observations also relevant to genus as a whole.] & . Martin. Cardenolide content and thin-layer chromatography profiles of monarch es eles st ela L., and their larval host-plant milkweed, Asclepias viridis Walt., northweste ny Louidana. Jour. Chem. Ecol. 13: 47-70. 1987. eee seat for plant eu 21 spots resolved for A. viridis; value to infrageneric taxonomy of Asclepias suggested. ] Macior, L. W. Insect adaptation and behavior i m Ee tEDIaS aay Bull. Torrey Bot. Club 92: 114-126. 1965. [Frequency and of pollinia of A. incarnata, A. syriaca, and 4. verticillata on species ee wasps and social cee in southeastern Wisconsin: detailed descriptions of insect behavior based on cinematographic rec- ords; insect morphology and behavior related to floral morpholog Moore, dies I. The common milkweed. Ontario Nat. 12: as 1972. [A popular summary of man es of A. incarnata, A. syriaca, and A. tuber. Moore, R. T. ier ene on rubber-bearing plants I. pee of normal and 1989] ROSATTI, APOCYNINEAE oe ee aborting seeds in Asclepias syriaca L. Canad. Jour. Res. C. 24: 55-65. 1946a. [His- tological abnormalities in ovules accompanying fruit abortion.] IV. Cytogenetic studies in Asclepias (Tourn.) L. Ibid. 66-73. 1946b. [See Wyatt, 1976.] Investigations on rubber-bearing plants V. Notes on the flower biology and pod yield of Asclepias syriaca L. Canad. Field-Nat. 61: 40-46. 1947. [See Wyatt, 1976.] Morse, D. H. Modification of bumblebee foraging: the effect of milkweed pollinia. Ecology 62: 89-97. 1981. [Slowed foraging of two bee species by about 25 percent, bees with pollinia moved to other plants more frequently than those without, bees foraging on milkweed “regularly” lost claws or segments of legs.] The turnover of milkweed pollinia on bumble bees, and implications for out- crossing. Oecologia 53: 187-196. 1982. [Average retention time slightly over one day on mouth parts, six hours on feet. . Costs ina milkweed and bumblebee mutualism. Am. Nat. 125: 903-905. 1985a. [A. syriaca.] —. Milkweeds and their visitors. Sci. Am. 253: 112-119. 1985b. . Role of pollen and ovary age in follicle production of the common milkweed Asclepias syriaca. Am. Jour. Bot. 74: 851-856. 1987. [Over five-day life of flower, viability of pollinia was constant, receptivity of stigmas “decreased threefold.”] & R.S. Fritz. Contributions of diurnal and nocturnal insects to the pollination of common milkweed (Asclepias syriaca L.) in a pollen-limited system. Oecologia 60: 190-197. 1983. [Eight times as many fruits matured from flowers exposed only to diurnal pollinators (mostly bumblebees), as compared to those exposed only to nocturnal ones (moths), even though flowers produced four times as much nectar at night. Variation in the pollinaria, anthers, and alar fissures of common milkweed (Asclepias syriaca L.). Am. Jour. Bot. 72: 1032-1038. 1985. [Variation mostly in shape rather than size; a clear separation into donor and recipient clones was not exhibited. & J. Scumitt. Propagule size, dispersal ability, and seedling performance in Asclepias syriaca. Oecologia 67: 372-379. 1985. [Heavier seeds might not have been dispersed as far but had superior germination and survivorship.] Moyer, L. S. Chromosome numbers and electrophoresis of latex in Asclepias. Bot. Gaz. 97: 860-865. 1936. [Suggested that mobility curves and isoelectric ce of latex differ among species, including 4. curassavica, A. incarnata, and A. sy Orcema, T., & R. A. FLetcHer. Factors that influence dormancy in mnilicaced seeds. Canad. Jour. Bot. 50: 713-718. 1972. [Kinetin and gibberellic acid most effective Pratak, D. M., J. PATEL, C. E. Torren, R. P. Swenson, P. BRown, & G. R. PETTIT. Cell growth inhibitory glycosides ae Asclepias amplexicaulis. four. Nat. ee 48: 470, 471. ey [Activity against cell line from a human nasopharnyx carcinoma.] = J. M., & S. J. Cuaptin. Nectar production rates of Asclepias quadrifolia: uses aan consequences of individual variation. Oecologia 59: 232-238. 1983. ee production correlated with removal but not insertion of pollinia.] Price, P. W., & M. F. Wittson. Some consequences for a parasitic herbivore, the milkweed longhorn beetle, Tetraopes tetrophthalmus, of a host-plant shift from Asdepias [sic] syriaca to A. verticillata. Oecologia 25: 331-340. 1976. [Larvae para- sitic on rhizomes; A. syriaca with larger stems and rhizomes, growth in cooler sites. ] bundance of herbivores on six milkweed species in Illinois. Am. Midl. Nat. 101: 76-86. 1979. [A. incarnata, A. Sullivantii, A. syriaca, A. verticillata, A. amplexicaulis, and A. tuberosa occur in this order on a gradient ee wet to d 1 I i ressure. ] QUELLER, D. C. Sexual selection in a hermaphroditic plant. Nature 305: 706. 707. 1983. [Evolution of flower number, umbel longevity, and seasonal distribution of flowers in A. exaltata through the male-competition component of sexual selection.] Sh2 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 ———.. Proximate and ultimate causes of low fruit production in Asclepias exaltata. Oikos 44: 373-381. 1985. [Primarily low availability of nutrients: incompatibility systems, fruit herbivory, and insufficient pollination excluded as major causes. | Reznicek, A. A. John Goldie’s 1819 collecting site near Lake Simcoe, Ontario. Canad. Field-Nat. 94: 439-442. 1980. [4. tuberosa reported in 1819, not found in 1976.] Ropertson, C. Notes on the mode of pollination of Asclepias. Bot. Gaz. 11: 262-269. pl. 8. 1886. [Detailed accounts of several species; illustrations of J. Sullivantii.] ROTHSCHILD, M. Secondary plant substances and warning coloration in insects. Symp. Roy. Entomol. Soc. London 6: 59-83. 1973. [Table / isa list of insects, food plants, and toxins stored by insects; cardenolides from Apocynineae found in Orthoptera, Hemiptera, Homoptera, Coleoptera, Neuroptera, Lepidoptera, and Diptera.] SAFWaAT, F. The initiation of vascular cambium and production of secondary xylem in flower bud pedicels of Asclepias curassavica L. in culture. Ann. Missouri Bot. Gard. 56: 251-260. 1969. [Light micrographs s.] ScHnepF, E., & P. Curist. Unusual transfer cells in the epithelium of the nectaries of Asclepias syriaca L. Protoplasma 105: 135-140, 1980. [Confirmation of nectaries ae in stigmatic a SerBer, J. N., L. P. BRower, S. M. Lee, M. M. McCuesney, H. T. A. CHEUNG, C. J. NELSO T. R. WATSON. seen connection between overwintering monarch eae from Mexico and their larval food plant, Asclepias syriaca. Jour. Chem. Ecol. 12: 1157-1170. 1986. [Contains “virtually the same cardenolides as 4. spe- ciosa,” which are of lower emetic potency than those of 4. eriocarpa, a California species; increase in abundance and distribution of 4. syriaca has rendered overwin- tering monarchs in Mexico more vulnerable than those in California. ] SHANNON, T. R., & R. Wyatt. Pollen peace - Asclepias exaltata: effects of flower age, drying time, and pollen source. Syst. Bot. 11: 322-325. 1986a. [Varied greatly among plants, but little during first four days reduced by more than 50 percent 24 hours after removal of pollinia (probably due to drying), suggesting limitations to eg dispersal distances & > reproductive biology of Asclepias exaltata. Am. Jour. Bot. 73: 11-20. 1986b. Flowering plants smaller than nonflowering ones; overall fruit-set low (about 2 percent of flowers), but za fruits had high probability of maturing. ] SoutHwick, A. K., & E. E. Soutuwick. Aging effect on nectar production in two clones of Asclepias syriaca. Oecologia 56: 121-125. 1983. [In northern Michigan, daily peak at 0800, overall peak 50 hours after flower opening, cessation after 120 hours. SouTHWICcK, E. E. Nectar biology and nectar feeders of common milkweed, Asclepias syrtaca L. Bull. Torrey Bot. Club 110: 324-334. 1983. [Wasps and bumblebees most abundant in northern Michigan, honeybees in New York. Sparrow, F. K. Types of pods of Asclepias syriaca found in Michigan. Jour. Agr. 73: 65-80. 1946. [Forty-nine eer thought to be genetically based, described and illustrated; reference made to ore. & . PEARSON. Pollen pee t in Asclepias syriaca. Jour. Agr. Res. 77: 187- 199, 194 ae SREEDEVI, P., & A. N. NAMBoopiri. Cytology of normal and seasonally aberrant mi- rane in Asclepias and Dregea (Asclepiadaceae). Cytologia 44: 377-384. 1979. [Also Calotropis and Telosma: transient sterility in natural populations due to microsporogenic aberrations (including formation of univalents) brought on by eat.] summer he THOMSON, J. W., & W. H. WAGNER, JR. Hybrid swarms of the prairie Asclepias speciosa with the common Asclepias syriaca. (Abstract.) P. 264 in R. L. Stuckey & K. J. Reese, eds., The prairie peninsula—in the “shadow” of Transeau. Ohio Biol. Surv. Notes No. 15. 1981. [Hybridization and backcrossing in western Minnesota J WuitrForp, P. C., & P. B. WHitrorp. Effects of trees on ground cover in old-field succession. Am. Midl. Nat. 99: 435-443. 1978. [Comparison Ga and cover 1989] ROSATTI, APOCYNINEAE 513 under canopy of Quercus ellipsoidalis and in open field; A. amplexicaulis not present but 4. verticillata with greater values in latter. Witpur, H. M. Life history evolution in seven milkweeds of the genus Asclepias. Jour. Ecol. 64: 223-240. 1976. [A. exaltata, A. incarnata, A. eae A. syriaca, A. see ‘ verticillata, and A. viridiflora in southeastern Michiga gule size, number, and dispersion pattern in pe ee and ie Am. mae Tl: 43-68. 1977. [A. exaltata, A. incarnata, A. eae A. syria A. tuberosa, A. verticillata, and A. viridiflora in southeastern Michiga WILLson, M. F., & R. I. Bertin. Flower-visitors, nectar production, and Nclicreacence size of Asclepias syriaca. Canad. Jour. Bot. 57: 1380-1388. 1979. [Maximum pro- duction of nectar sugar in afternoon and evening: major carriers of pollinaria honey- bees in afternoon, noctuid moths at night; larger inflorescences attracted greater numbers of visitors (both carriers and thieves).] Price. Nectar production and flower visitors of Asclepias vertcilaa, mh. Midl. Nat. 102: 23-35. 1979. [Wasps most frequent native visitors natural habitats but infrequent at disturbed site; few visitors observed to carry aliaatik several appeared to rob nectar.] & RICE. The evolution orf cate size in Asclepias (Asclepiadaceae). Evolution 31: 495-511. 1977. [A. incarnata, A. syriaca, A. verticillata; less extensive observations on A. curassavica, A. Sullivantii.] & ———. Resource limitation of fruit ae seed production in some Asclepias species. Canad, Jour. Bot. 58; 2229-2233. & B. J. Ratucke. Adaptive design of - floral display in Asclepias syriaca L. Am. Midl. Nat. 92: 47-57. 1974. [Apparent balance between a ae to limit inflorescence size and to increase number of Bs considered adapti Witson, K. J., B. H. PETERSEN, & D. D. Biessoer. Immunocytochemical sdentileanibn of laticifers. Protoplasma 122: Aa 90. 1984. [A. syriaca, A. tuberosa.] Woopson, R. E., Jk. Some dynamics of leaf variation in Gis tuberosa. Ann. Missouri Bot. Gard. 34: 353-432. pl. 45, 1 foldout diagram. 1947. [Details about morphology, reproductive biology, cytogenetics, ecology, and hybridization. ] . Biometric evidence of natural selection in Asclepias tuberosa. Proc. Natl. Acad. Sci. 39: 74-79. 1953 . The North American species of Asclepias L. Ann. Missouri Bot. Gard. 41: |- 211. 1954. Butterflyweed revisited. Evolution 16: 168-185. 1962. [Reevaluation of pop- ulations studied 14 years earlier (Woopson, 1947) indicated that subspecies interior and tuberosa underwent surprisingly rapid genetic amalgamation; but see WyaTT & ANTONOVICS. The geography of flower ae in es Evolution 18: 143-163. pl. J. 1964. [Several paragraphs by S. GHT; red-orange predominant in eastern. and northern Missouri and adjacent areas, ee which yellow increases in frequency in all directions.] Wyatt, R. Pollination and fruit-set ; cla a reappraisal. Am. Jour. Bot. 63: 845- 851. 1976. Hae reports (R ORE, 1946b, 1947) that insertion of pollinia in adjacent stigma iS pee for fruit set in A. curassavica, A. incarnata, A, Siliania. A. syriaca, A. tuberosa, and others, a single insertion is adequate in A. tuberosa Population dynamics and reproductive biology of butterflyweed, Asclepias tu- berosa L. 191 pp. Unpubl. Ph.D. dissertation, Duke Univ., Durham, North Carolina. 1977.* (Diss. Abstr. B. 38(8): 3531. 1978.) [First documented report of (low levels of) self- pene in Asclepias.] Experimental evidence ae the role of the corpusculum in Asclepias salar con Syst. Bot. 3: 313-32 78. The impact of nectar- Seta aa on the pollination system of Asclepias cu- 514 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 rassavica. Bull. Torrey Bot. Club 107: 24-28. 1980a. [Primary pollinators in Costa Rica and Jamaica evidently monarch, queen, and other butterflies of the genus Danaus. | ———.. The reproductive biology of Asclepias tuberosa 1. Flower number, arrangement, and fruit-set. New Phytol. 85: 119-131. 1980b. [Populations in North Carolina and Virginia studied.] II. Factors determining fruit-set. bid. 88: 375-385. 1981. [Four- to five-week overlap in flowering time within and among inflorescences on a single plant in this self-incompatible species; fruit-set predictions from model very close to the 0.33 percent observed in n & J. ANTonovics. Butterflyweed re-revisited: spatial and temporal patterns of leaf shape variation in Asclepias tuberosa. Evolution 35: 529-542. 1981. [Trends observed by Woopson (1962) from 1946-1960 due to phenotypic aces induced by weather patterns specific for sampling years, not to natural selection. & T. R. SHANNON. Nectar production and pollination of Asc eins exaltata. Syst. Bot. 11: 326-334. 1986. [Flowers available to pollinators for about six days; most nectar produced at night; sucrose concentration increased from morning to afternoon (due to evaporation?); pollination rates, fruit production positively correlated with J sucrose concentration 1989] AL-SHEHBAZ, DACTYLOCARDAMUM ne DACTYLOCARDAMUM (BRASSICACEAE), A REMARKABLE NEW GENUS FROM PERU IHSAN A. AL-SHEHBAZ! Dactylocardamum imbricatifolium, a new genus and species, is described from the Sierra Blanca of Peru. Its relationships to Eudema and Xerodraba are discussed. Scanning electron micrographs of leaves, fruits, and seeds of Dactylocardamum and leaves of Xerodraba are presented. During the past several years I have been studying the systematics and evolution of various South American genera of the Brassicaceae (Cruciferae). Almost every genus that I have examined critically (see, for example, Al- Shehbaz, 1986, 1989a, 1989b, 1989c, 1990) included one or more novelties. However, the discovery of a new genus with unique morphology and evident relationships to well- denned: pmonophylene genera was rather surprising. t fter called Dactylocardamum, which literally means a digitate nicer reference to the fingerlike branches) was collected from Cordillera Blanca, Peru, and was misidentified as Xerodraba Skottsb. A close examination of the plant immediately revealed that it belongs to an undescribed genus, not only because it has several unique features that separate it readily from its nearest relatives, Eudema Humb. & Bonpl. and Xerodraba, as well as from all other genera of the Brassicaceae, but also because its type locality is separated from the nearest known range of Xerodraba by at least 3200 air kilometers. The overall growth habit and certain aspects of the leaves of Dactylocar- damum are fairly similar to those of Yerodraba. However, the two genera are readily distinguished from each other by the nature of their fruits, the length of the style and the fruiting pedicels, and the texture of the leaves (see FIGURE 1 and TABLE). On the other hand, Dactylocardamum resembles Eudema in the texture of the leaves and in a few characters of the fruits. However, it differs markedly in habit, leaf persistence, number of flowers per branch, pedicel length, and fruit position (see TABLE). The flowers in Eudema are few to several on each branch, and they mature in a racemose fashion (Al-Shehbaz, 1990), while those of Dactylocardamum are solitary. The three genera are easily sep- arated from each other by the following key: A. Pee appressed, densely imbricate, persistent as a whole; branches fragile, fingerlike; flowers | at tip of each branch. B. tates thin, papery; ee subterete, indehiscent, completely included and sand- wiched between leaf bases; septum absent; style 2.5-3 mm long; valves mem- branaceous; fruiting pedicels to 0.4 mm long. ............ Dactylocardamum. B. Leaves thick, fleshy; fruits flattened parallel to septum, dehiscent, exserted above 'Amold Arboretum, Harvard University, 22 Divinity Avenue, Cambridge, Massachusetts 02138. © President and Fellows of Harvard College, 1989. Journal of the Arnold Arboretum 70: 515-521. October, 1989. JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 Phenetic comparison of Dactylocardamum, Eudema, and Xerodraba. GENUS CHARACTER Dactylocardamum Eudema Xerodraba Fruits Flattening Not flattened, Latiseptate or Latiseptate subterete subterete Dehiscence Indehiscent Dehiscent (rarely Dehiscent indehiscent) Valve texture Membranaceous Somewhat leath- Somewhat leathery ery Septum Absent Complete or per- Complete or rarely perforated Position Sandwiched be- Style length (in mm) Fruiting-pedice length (in mm) Flowers per branch eaves Length (in mm) Arrangement Blade orientation Texture Persistent portion Branch shape tween leaf bas- es 2.5-3 =< 0.4 l Densely imbri- cate Appressed Thin, papery Entire leaf Fingerlike forate Exserted above leaves 0.4-3.5 2-20 Few to several 2 Loose, not imbri- cate Divaricate to as- cendin Thin, not papery Petiole Various, but nev- er fingerlike Exserted above leaves 1.6-9 Densely imbricate Appressed Thick, ar fleshy Entire st Fingerlike leaf rosette; septum complete or faa aa style 0.4-1 mm long; valves Xer somewhat leathery; reales aces 7 A. Leaves neither appressed n 1.6-6 mm odraba. ero almost ee with only petioles ee branches neither fragile nor eer fen few to several at tip of each bra The evolutionary relationships among the three genera is rather interesting. Although Dactylocardamum resembles Xerodraba in several aspects of habit and foliage, the bizarre vegetative morphology probably evolved independently in the two genera. On the other hand, Dactylocardamum does show certain affinities to Eudema, particularly in some aspects of the fruits and in leaf pubescence, and they may be more closely related to each other than either is to Xerodraba. However, Dactylocardamum is unique in the Brassicaceae in having axillary fruits that are completely concealed and sandwiched between the appressed, densely imbricate leaf bases (see FiGURE 2a-c). These hidden fruits can be seen only after the removal of the subtending leaves. Although the solitary flower of D. imbricatifolium is terminal in the bud stage and during anthesis, the developing fruit is gradually pushed to an axillary position due to the production of new leaves by the terminal leaf bud. The fruits of a given 1989] AL-SHEHBAZ, DACTYLOCARDAMUM 517 Fic anning electron micrographs of Dactslocardamun and ann a, im aie (holotype): a, leaf, b, portion of branched stem; c, top view of branch p. d, X. pectinata (Pisano & Cardenas 4749, HIP), ee e, X. lepana esi 19, > leaves; f, XY. patagonica (Donat 181, GH), leaves. Scale bars mm. ee can therefore be separated by several to many obsolete to very short no Dactylocardaman is easily distinguished from all other genera of the Bras- sicaceae by a unique combination of characters: the septum 1s completely lacking, the inflorescence is reduced to a terminal subsessile flower, the axillary fruits are sandwiched and completely hidden between the densely imbricate, wholly persistent leaves, the fruiting pedicel is reduced to only 0.4 mm long, the internodes are obsolete or reduced to only 0.05—0.2 mm long, and the branches are subdichotomous and fingerlike (see FIGURES la-, Two other South American genera of the Brassicaceae, Delpinophytum Speg. and Lithodraba Boelcke (Boelcke, 1952; Boelcke & Romanczuk, 1984), produce dense cushions and have growth habits fairly similar to those of Dactylocar- as ay Os hk (arn eer Sr ‘as CUD wi me <« FIGURE2. D / l imbricatifoli l itudinal ; git of branch show- ing fruit with | seed and densely imbricate leaves; b, fruit surrounded by several leaves: c, portion of branch showing fruit on short peduncle and scars of removed leaves: d, seed. Scale bars = | mm. 1989] AL-SHEHBAZ, DACTYLOCARDAMUM 519 damum and Xerodraba. Lithodraba is somewhat related and is perhaps a sister group of Englerocharis Muschler. It is easily distinguished from the new genus in having racemose inflorescences, four stamens, long nonimbricate leaves, septate fruits, and ringlike nectar glands. De/pinophytum is indistinguishable from Xerodraba in almost every aspect except the type of fruit flattening: the former has angustiseptate fruits, while the latter has latiseptate ones. Although Schulz (1936) placed Delpinophytum in the tribe Lepidieae DC. and Xerodraba in the Sisymbrieae DC., I believe that they are sister genera and may even be congeneric. Dactylocardamum imbricatifolium Al-Shehbaz, gen. et sp. nov. + re “1 A 4 b v imbricati-foliosis, usque ad 3 cm longis, 3—4.5 mm latis. Folia minuta oblongo- obovata, imbricata, persistens, subappressa, papyracea, superne pilosi-ciliata, 2.5-3.5 mm longa, 1.1—-1.6 mm lata. Flores solitares apicales subsessiles. Sepala erecta oblonga, membranacea, 2.5-3 mm longa. Petala lineari-spathulata, 4— 4.5 mm longa, ca. | mm lata. Pedicelli fructiferi crassi, axillari, usque ad 0.4 mm longi. Siliqua ovoidea, membranacea, aseptata, glabra, brevissime pedun- culata; stylo filiforme 2.5-3 mm longo. Semina | vel 2, oblonga vel ovata, pendula; cotyledones incumbentes. Perennial herb forming dense cushions, woody and much branched at base, the ultimate branches fingerlike, thick, 1-3 cm long, 3-4.5 mm in diameter. Leaves persistent, papery, densely imbricate, separated by internodes 0.05-0.2 mm long, subappressed, oblong-obovate, 2.5-3.5 mm long, 1.1-1.6 mm wide, obtuse at apex, ciliate and sparsely pubescent on the upper surface of the green, distal, bladelike half, the petiolelike lower half conspicuously flattened, straw colored, 1.4-1.9 mm long, 0.8-1 mm wide; trichomes unbranched, 0.5-1.2 mm long. Flowers solitary, subsessile, each terminating a branch, soon becom- ing lateral as young fruit develops. Sepals erect, membranaceous, oblong, 2.5- 3mm long, 1-1.2 mm wide, sparsely pubescent below rounded apex, nonsaccate at base. Petals spatulate-linear, 4-4.5 mm long, ca. 1 mm wide, attenuate at base, obtuse at apex, somewhat crisped distally. Stamens 6, somewhat tetra- dynamous; filaments erect, filiform, 2.5-3 mm long; anthers oblong, 0.7-0.8 mm long. Nectar glands 4, | on each side of each lateral stamen. Fruiting peduncles thick, axillary, to 0.4 mm long. Fruits ovoid, sessile, sandwiched and completely hidden between leaves, indehiscent except when subjected to force, somewhat inflated, 1.2-1.8 mm long, 1-1.4 mm wide; valves mem- branaceous, glabrous; septum lacking; style filiform, 2.5-3 mm long in flowering or fruiting material; stigma entire. Seeds 1 or 2 per fruit, oblong to ovate, plump, 1-1.2 mm long, 0.8-0.9 mm wide, dark brown, reticulate, wingless, pendulous from short, subapical funicles; cotyledons incumbent or obliquely so. Type. Peru, Cordillera Blanca, Llanganuco Valley, Pisco Creek, 15,500 ft [ca. 5085 m] alt., August 1959, J. C. Tothill 144 (holotype, uc; fragment, GH). JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 520 bo a) - aa ~ oO Oo a geo Be ae: o — ’ b f, pistil; g, fruit with part of style; | mm. Drawn from the holotype by the author. rege ak vas Je aS &G 5 a sae os se SO Re eC “2 Se “3 Ras: ¢ QOS ae SRC : aces er po etee % Cy: EAN Tae 3 ee ee Ko Regt ah eee UE eR ORR : SOC MS ete. ECS ne anette te ee Ma ON # ere ss ae cs Rene ee heres a SS Se ea C2 LONG NG Get tos eo , petal; e > sepal,; Cx se Ser eet Kee SEES b-h tJ Dactylocardamum imbricatifolium: a, portion of plant; one stamen removed), d fruit. Scale bars: a, | cm; mene oes ce Pipers § ZX CA = SAN < ean < ESE SOSteS S . OTS Cara aoa : | ( 1989] AL-SHEHBAZ, DACTYLOCARDAMUM S21 Additional field notes on the type read: ‘““Mat plant forming rather tight cushions over rocks of lower talus slope. South exposure, associated with other mat plants.” ACKNOWLEDGMENTS Iam most grateful to Reed C. Rollins for his critical review of the manuscript, and to Donald H. Pfister for obtaining funds from the Harvard University Herbaria that supported the SEM portion of the research. I am thankful to Elizabeth A. Shaw for correcting the Latin description, to Thomas Duncan and Barbara Ertter for permitting the deposition at Gu of fragments of the holotype, to Trisha Rice for the SEM work, to Barbara Nimblett for typing the manu- script, to Michael A. Canoso for obtaining the loan, to Elizabeth B. Schmidt and Stephen A. Spongberg for their editorial advice, and last, but certainly not least, to my wife, Mona, for her continuous support. LITERATURE CITED AL-SHEHBAZ, I. A. 1986. Lepidium solomonii (Cruciferae), a new species from Bolivia. Ann. Missouri Bot. Gard. 73: 830, 831. a. The South American genera Brayopsis and Englerocharis (Brassi ) Nordic J Bot. 8: 619-625. 89b. Systematics and evolution of Schizopetalon (Brassicaceae). Harvard Pap. oe 1: 10-46. . 1989c. New and noteworthy Draba (Brassicaceae) from South America. J. Amold Arbor. 70: 427-437. ——. 1990. sc limits and taxonomy of Brayopsis and Eudema (Brassicaceae). Ibid. 71: in p ee O. 1952. Nous nuevo género de cruciferas de la Argentina. Darwiniana ai — &M.C. RomanczuK. 1984. Cruciferae. Fl. Patagonica 4A: 373-544. ScHuuiz, O. E. 1936. Cruciferae. Jn: H. HArMs, ed., Nat. Pflanzenfam. ed. 2. 17B: 227- 658 BOOK REVIEW 573 BOOK REVIEW Grass Systematics and Evolution, edited by Thomas R. Soderstrom, Khidir W. Hilu, Christopher S. Campbell, and Mary E. Barkworth. Smithsonian Institution Press, Washington, D. C., 1988. xiv + 474 pp. illustrations. ISBN 0-87474-300-1. $45 hardcover. In July, 1986, some 150 students of grass systematics from North and South America, Europe, Africa, Asia, Australia, and New Zealand gathered at the Smithsonian Institution’s Natural History Museum to discuss the systematics and evolutionary biology of grasses. The genuinely international character of the International Symposium on Grass Systematics and Evolution should not have been a surprise, for the principal organizer was Thomas Soderstrom. Soderstrom was known for his enthusiastic encouragement of—and collabo- ration with—numerous members of the worldwide community of grass sys- tematists. His death in 1987 prevented him, however, from witnessing the publication of this attractive, cleanly edited volume of papers from the sym- posium. The book’s 33 papers are organized under seven headings (Structural Di- versity; Biochemical Diversity, Reproductive Biology; Evolution; Systematics of Major Groups; Taxonomic Data and Analysis; and Grasses: Past, Present, and Future). The majority of the contributions are review papers, and with few exceptions the coverage under each heading is rather complete (e.g., under Biochemical Diversity are treatments of photosynthetic pathways, chloroplast DNA, nuclear DNA, amino-acid profiles, and isozymes). Following the papers are a composite 71-page bibliography and an index to taxonomic names. The organization of this volume directly reflects that of the symposium, with each major section more or less corresponding to the proceedings of a given morning or afternoon of the conference. The conveners apparently called for a survey of grass systematics, and the contributors provided it. While symposium volumes often comprise odd com- binations of articles that take off in their own directions, this one represents a theme developed. Unfortunately, comprehensiveness and consistency seem to have been provided at the cost of original analysis and synthesis. Even as reviews, several of the articles discuss little that 1s new; they repeat the contents of earlier reviews, cite a few of the most recent papers, and sign off. Others do present fresh overviews of segments of the literature that were in need of such treatment. Although the majority of the offerings are reviews, the volume also includes some significant original contributions (e.g., Barkworth and Everett on generic delimitation and alignments in the Stipeae, Campbell and Kellogg on cladistic structure of the Poales, Kellogg and Campbell on cladistic structure of the Poaceae, Soderstrom and Ellis on tribal delimitation in the Bambusoi- deae, and Zuloaga on infrageneric classification of Panicum). © President and Fellows of Harvard College, 1989. Journal of the Arnold Arboretum 70: 523, 524. October, 1989. 524 JOURNAL OF THE ARNOLD ARBORETUM This volume also gives the reader a sense of the level of satisfaction in the agrostological community with the now-conventional five-subfamily system; it seems to me that there is more complacency than there should be. The Festucoid/Panicoid dichotomy, which served agrostologists well in its time, has finally passed away. Indeed, the 1959 Montreal grass symposium, to which the Washington symposium might be compared, is remembered today in large part for the boost it gave to the five-subfamily system, which was then emerging. One could not have expected a new synthesis to appear on cue in 1986, but a little more questioning of the current system might have gone a long way. The subfamily Arundinoideae, for instance, is widely acknowledged to be an un- natural assemblage. Most of the authors surveying particular sorts of characters mention this in passing but seem content to catalog the odd character com- binations that occur in the group, note the need for further analysis, and then move on to the other, more clearly defined subfamilies. Specific attention to the realignment of arundinoid elements, and to other outstanding problems with the current system, would have evidenced a greater interest in progress. Despite these problems, the volume does include many useful reviews, as well as a handful of original contributions. Even the weaker reviews provide citations of the key literature. Consequently, the book’s greatest success is as a wide-ranging survey of grass systematics, and as the best available point of entry into the enormous literature of this discipline. Properly supplemented, it might even serve as the principal text for a course on the subject. Agrostol- ogists would be well advised to have a copy close at hand.—JERROLD I. Davis, L. H. Bailey Hortorium, Cornell University, Ithaca, New York 14853. 1989] INDEX 525 INDEX Abies, 249, 251-254, 262, 298-305 — subg. Abies, 299 — — sect. Balsameae, 299, 300, 302 — subg. Pseudotorreya 3 | wn ra ° ed Q 4 pel) a | fa) 3 50) 2. Zz we oO No — araragi, 292 — balsamea, 254, 299, 300, 302 — — var. fraseri, 299 _ r. phanerolepis, 300 — eA 298, 299 — concolor, 300-302 fraseri, 254, 299-302 — grandis, 300 — lasiocarpa, 299, 300, 302 | 0.3 of ag 3 tO \O [oe] — sibirica, 298 — spectabilis, 253 — tsuga, 292 Abietaceae, 248 Abietineae, 248 99 Fees 341, 348 Acrostichum aureum, Adenanthera peregrina var. peregrina, 145 Adenium, 315 Adoxaceae, 309 Agave, 34 Allamanda, 312, 314-316, 329, 330, 332- 337, 340, 341, 344, 345, 354, 377-379 = cathartica, 377-379, 3 44 AL-SHEHBAZ, IHSAN A. Dactylocardamum (Brassicaceae), a Remarkable New Ge- nus from Peru, 515-521 AL-SHEHBAZ, IHSAN A. New or Noteworthy Draba (Brassicaceae) from South Amer- ica, 427 AL-SHEHBAZ, IHSAN A., and BERNICE G. SCHUBERT. The Dioscoreaceae in the Southeastern United States, 57-95 Alstonia, 314, 329, 330, 332, 339-341, 343, 344, 346 — macrophylla, 346, 354 Amaranthaceae, | Amaryllidaceae, 150 Amischotolype, 102 Ampelamus, 481 — albidus, 482 Amphistelma, 481, 482 — scoparia, 48 Amsonia, 312, 329, 330, 332, 339, 340, 344, 345, 355, 358-363, 365 — subg. Amsonia, 361, 362 — subg. Articularia, 359, 361 — subg. Longiflorae, 359 — subg. Sphinctosiphon 359, 361, 362 ria, 361, 362 — sect. Euamsonia, 35 — sect. Sphinctosiphon, 359 _ angustifolia, , 359, 361- 363 — — var. ciliata, 361 — — var. filifolia, 361 — illustris, 362 1, 362 — tabernaemontana, 358-363 Anantherix, 493, 495 — connivens, 496 Anaphalis, 142 Anaxagorea javanica var. tripetala, 241 — luzonensis, — petiolata, 241 — phaeocarpa, 241 1 Androsyne, 68 526 Aneilema, 99, 101, 102, 123 i 23 Angadenia, 330, 332, 333, 339, 345, 355, 380, 382-384, 389 amosa, 241 reer The Biosystematics and Evo- lution of the Polyalthia hypoleuca Com- plex of Malesia, I. Systematic Treat- ment, 153-24 Sete 153, 154, 166, 170, 173, 175, 196, ae 202, 204, 212, 219, 225. on 229, — tribe Tee 171, 173, 176 — tribe Unoneae, 166, 199 — — subtribe Xylopineae, 166 — tribe Uvarieae, 166, 171 Apocynaceae and Asclepiadaceae: The Genera of Suborder Apocynineae in the Southeastern United States, 307-401, 443-514 Apocynaceae, 79, 307, 309-319, 328-354, 359, 378, 381, 383, 385, 387, 389, 393, 446, 447, 451, 454 — subfam. Apocynoideae, 309-312, 314, 315, 317, 329, 331-334, 336, 337, 339, 343-345, 354, 365, 378, 379, 390, 392, 395, 452-454 tribe Apocyneae, 311, 331-333, 339, ae 354, 390, 392 — — tribe Echiteae, 331, 347, 379 — — tribe Parsonsieae, 312, 333, 339, 392 — subfam. Cerberoideae, 317, 334, 453 — subfam. Echitoideae, 331, 332, 379, 392 — — tribe Echiteae, 332 — — tribe Parsonsieae, 332 333, 339, 345, — subfam. Plumerioideae, 309-311, 314, 317, 329, 331-333, 336, 339, 342-345, 354, 356, 359, 365, 378, 452, 453 — — tribe Allamandeae, 313, 334, 341, 377 a as Landolphiinae, 313, 332, 336, 339 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 Apocynaceae subfam. Plumerioideae tribe Arduineae subtribe Melodininae, 313, 332, 334, 336, 339, 346 — — tribe Carisseae, 331, 334, 335, 342 — tribe Pleiocarpeae, 312, 313, 331- ae 335, 336, 338, 340, 342, 343, 345 ribe Plumerieae, 331-333, 339, 342, "ae 345, 356, 357, 359, 365 subtribe Alstoniinae, 332, 333, ~ 330. 345, 346, 358, 359, 365 — — — subtribe Cerberinae, 332, 334, 347, 57 ees subtribe Plumeriinae, 332, 341 — — — subtribe Rauvolfieae, 333, 334, 342, 343 Les) subtribe Rauvolfiinae, 332, 333, 339, 345, 347, 356, 357 subtribe Tabernaemontaninae, 332, 333, 339, 340, 346, 359 tribe Tabernaemontaneae, 342, 343 Apocynineae (Apocynaceae and Asclepia- daceae) in the Southeastern United States, The Genera of Suborder, 307- 401. 443-514 Apocynineae, 307-401, 443-514 Apocynum, 311-313, 317, 318, 329, 330, 332, 333, 337-340, 344, 345, 355, 376, 390-401, 452, 481 — androsaemifolium, 315, 337, 338, 391- 394, 396-400 — — var. androsaemifolium, 399 — — var. glabrum, 399 — — var. incanum, 397 — var. intermedium, 397 = ee cr enn 337, 392, 393, 397-400 var. pubescens, 399 — hypericifolium, 397, 399 — medium, 393, 397-399 — — var. floribundum, 397 a. Cane. 332 1989] Aristolochiaceae, 60 Armillaria matsutake, 252 Artabotrys harmandii, 241 — siamensis, 241 — zeylanicus, 241 Asclepiadaceae and Apocynaceae: The Genera of Suborder Apocynineae in the Southeastern United States, 307-401, —-514 Asclepiadaceae, 147, 307, 309-319, 328- 330, 332, 333, 335-337, 341, 342, 344, 378, 391, 443-4 — subfam. Asclepiadoideae, 310-312, 450- 454, 461 —— ‘tribe Asclepiadeae, 317, 450, 451, 454, 477, 481, 482 — — — subtribe Asclepiadinae, 482 — — — subtribe Cynanchinae, 481 — — tribe Gonolobeae, 450, 451, 453, 454, 461, 463 oo tribe Pergularieae, 450, 453, 454 — subfam. Periplocoideae, 310-312, 317, 337, 444, 445, 450-454 — subfam. Secamonoideae, 310-312, 445, 450, 452-454 Asclepiadeae, 443 Asclepias, 312, 315, 392, 400, 444-447, 451, 452, 454, 455, 460, 464, 469, 470, 491-514 — subg. Acerates, 493, 496, 501 — subg. Anantherix, 493, 496 — subg. Asclepias, 491, 497, 498, S01, 502 — subg. Asclepiodella, 491, 497, 498 — subg. Asclepiodora, 447, 493, 496 — subg. Podostemma, 497, 501, 502 — amplexicaulis, 498, 500, 501, 505 — angustifolia, 499 — angustissima, 487 — curassavica, 491, 495, 498, 500-502, 5, 507 — curtissii, 499, 502 — emoryi, 5 — engelmanniana, 495, 497, 500 — exaltata, 498, 500, 501 — feayi, 497, 500, 502 — hirtella, 496, 500 INDEX S27 Asclepias ae: 499, 501, 505 — incarnata, 498, 500-502, 504, 506, 507 Asclepias incarnata subsp. incarnata, 500 — — subsp. pulchra, 494, 500 — lanceolata, 498, 500 — linearis, 495, 498, 502 — longifolia, 491, A 497, 500, 504 — michauxii, 499, 502 — obovata, 499, 501 — oenotherioides, 495, 497, 501 — pedicellata, 444, 491, 493, 496, 502 — perennis, 493, 498, 500, 502 — pumila, 501 _ purpurascens, 499, 501, 504 — quadrifolia, 491, 499, 501, 502 — rubra, 498, 500 — speciosa, 501 — stenophylla, 497, 502 — sullivantii, 495, 500, 501, 504, 505 — syriaca, 455, 492-494, 499-501, 503- 507 — tomentosa, 499 — tuberosa, 491, 498, 500, 501, 503-507 — variegata, 5 — verticillata, 498, 501,502, 504, 506, 507 — viridiflora, 496, 501, 504 — viridis, 496 — viridula, 499, 502 Asclepiodella: 493, 495 — feayi, 497 Asclepiodora, 493, 495 — viridis, Asimina triloba, 241 Asystasia gangetica, 148 Athyrocarpus, 102 Avetra, 58, 59, 61 Basket-plant, 117 Beaumontia, 348 Begonia heracleifolia, 146 Begoniaceae, 146 Biosystematics and Evolution of the Poly- althia hypoleuca Complex (Annonaceae) of Malesia, The, I. Systematic Treat- ment, 153-246 Biventraria, 493, 495 — variegata Blepharodon, 490 Bluestar, 3 Bocageopsis, 170, 173 — canescens, 241 — mattogrossensis, 241 4 Boletus edulis, 252 528 Boraginaceae, 345 Borderea, 58, 59, 61, 69 BorRNSTEIN, ALLAN J. Taxonomic Studies in the Piperaceae—I. The Pedicellate Pipers of Mexico and Central America (Piper Subg. Arctottonia), 1-55 Botryosicyos, 68 Brassicaceae: Dactylocardamum, a Re- oo New Genus from Peru, 515- a ee New or eb Draba from South America, 427-4 5 — alicastrum, 6, 39, 47 Buddlejaceae, 309, 341 Bursera, 6, 18, 34, 44, 47 Button-rods, 137 aed 160 Calliam Callisia, OR. 99, 101, 107, — sect. Brachyphylla, 119 — sect. Callisia, 120 — sect. Cuthbertia, 98, a | 108, 117-120 LL, 116 19 120 — sect. Leptocalisia —~ 119 — cordifolia, 118- = macdougalli 120 antha, 119 — ciniiflors 118, 119 — navicularis ‘oO * - soconuscensis, 120 warszawicziana, 119 Calophyiluin sinncel 160 — canum Calopogonium mucunoides, 145 Calotropis, 45 — procera, 447, 451 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 Campelia, 100, 112 — standleyi, 112 Caprifoliaceae, 309 Carandas, 331 Cardiopteridaceae, 68 Cardiopteris, 68 Carissa, 312, 314, 316, 329-332, 339, 340, 344, 346 — arduina, 346 — bispinosa, 346 — carandas, 346 — grandifolia, 315, 346, 354 — macrocarpa, Cartonema, 98, 99, 101 Cartonemataceae, 99 Carya, 468 Cassia javanica, 145 Catharanthus, 312, 314, 315, 319, 329, 330, 332, 337-345, 355, 363-369, 371, 374 368 — roseus, 317, 319, 336, 337, 345, 364- 368 — scitulus, 368 Cathaya, 248-253, 266, 286 Cedrus, 248-254, 300, 302 — de 302 Celastraceae, 144 elastrus grenadensis, 144 — racemosus, 144 Celosia argentea, 149 Central America, The Pedicellate Pipers of Mexico and (Piper Subg. Arctottonia). aa Studies in the Piperaceae— -5 Pa ere 250 Chomolaena integrifolia, 147 Chrysobalanaceae, 160 Chrysothemis pulchella, 148 Cladophyllaceae, 58 Cleistochlamys kirkii, 241 Cleistopholis glauca, 241 — ae — staudti Climbing ae 387 Climbing milkweed, 463 Cochliostema, 101 a, 155 1989] Coleotrype, 99, 100 Commelina, 98, 100-102, 107, 124-130 — africana, 101, 128 - appendiculata, 125 8 — cabolinianas | — dianthifolia, 125 — diffusa, 102, 126, 128 — — var. aquatica, 126 — erecta, 125, 128 — forskalaei, 128 — gigas, 125 — graminifolia, 102 — hirsuta, 128 — scapiflora, 1 — undulata, 128 — virginica, 102, 103, 125 — zambesiaca, |! 2 Commelinaceae in the Southeastern United States, The Genera of, 97-130 Commelinaceae, 97-130 — subfam. Lia teks 98-101 — subfam. Ssaneeye aaa ae, — — tribe Commeline ss 101, 122 — — tribe edeemiicae 98, 99,101, 108 ubfam. Tradescantioideae, 98 98 Compositae, 147 Convolvulaceae, 77, 148, 318 Cordia, 32 Crematosperma, 173 oma rot Crematospermum macrocarpum, 241 CRINS, WILLIAM he Tamaricaceae in the Southeastern United States, 403-425 Cruciferae, 515 Cryptostegia, 455 — grandiflora, 447, 451, 455, 460 Cuphea hyssopifolia, 146 INDEX 529 Cupressaceae, 248, 249 Cuthbertia, 98, 99, 108, 117, 118 e, 464 Cymbispatha, 108, 111 Cynanchum, 314, 444, 447, 451, 452, 455, 460, 480-484, 487, 490 — sect. oe 482 — acutum, 48 _ Serr rem 480, 482, 484 — blodgettii, 460, 480-483 — clausum — laeve, 473, 480-484, 489 — northropiae, 444, 460, 480, 482, 483 8 — parviflorum, 147 — scoparium, 444, 460, 480, 482, 483 Cyperaceae, 150 Cyperus iria, 150 Dactylocardamum (Brassicaceae), a Re- able New Genus from Peru, 515-— Dactylocardamum, 515-521 — imbricatifolium, 515-521 Dasymaschalon, 169 Dayflower, 125 Delpinophytum, 517, 519 Desmos, 17 — cochinchinensis, 241 — dasymaschalus, 241 Dichorisandra, 100-102 oe 2 62, 67-95 sec — sect. Dees. 71, 73 — sect. Enantiophyllum, 73, 76 — sect. Lasiophyton, 75 — sect. M — sect. Macropoda, 69 530 Dioscorea sect. Opsophyton, 71, 73 — — subsect. Euopsophyton, 7 | | nn Oo oO Q 0 idee tas C4] Sy ae 8 ® 25 Q ie} Qa p ju) sect. Triangulares, 74 — abyssinica, 78 — alata, 73-75, 77-79 — batatas, 73 — brachybot 75 — bulbifera, 6, 68, 71, 73, 74, 78, 79, 150 — cayenensis, 74, — cirrhosa, 75, io — composita, 74, 75, 78 — — x Dioscorea floribunda, 75 — cretacea, 61 — cyphocarpa, 59 — deltoidea, 75, 78, 79 — hirticaulis, 70 — hispida, 75, 78, 79 — hondurensis, 74 _ ae 73 _ ignatenns ae 73, 74, 78, 79 _ sina 69 ar. glabrifolia, 69 _ eel ain 74, 78, 79 — piscatorum, 76, 78, 79 — poilanei, 78 — polygonoides, 61, 79 — prachensilis, 78 — prazeri, 78, — quaternata, He dd. TS — — var. glauca, 69 — reticulata, 7 — rhipogonoides, 75 rotundata, 74, 76, 78 rupicola, 78 a sansibarensis, 78, 79 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Dioscorea sativa, 68 — eats. 78 , 78 — villosa, 68-71, 73, 79 — — subsp. rena 71 — — subsp. — — subsp. hirticalis, 70 — — subsp. paniculata, 69 — — subsp. quaternata, 69 Dioscoreaceae in the Southeastern United States, The, 57-95 Dioscoreaceae, 57-95, 150 m. Stenomeridoideae, 59 — tribe Dea. 9 — tribe Stenomerideae, 59 7 9 Dogbane Family, 328 ena Recent Plant Collections from, 43-152 bra eee ame from South America, or Noteworthy, 427-437 Draba, 427-437 — sect. Chamaegongyle, 433 — sect. Dolichostylis, 431 — affinis, 427, 431 | a 3 OS. oa a n a ve — hammenii, 435 1989] Draba ee 429 431 — peudodnsranhodes 427, 431-435 — ritacuvana, 427, 435, 436 — schultzei, 431 Drimys piperita, 160 Ducampopinus, 262 Dunsterville, G. C. K., an Appreciation, —44] Dunsterville, G. C. K., 439-441 Ebenaceae, 155, 160 Echites, 315, 329, 330, 332, 333, 339, 340, 344, 345, 355, 379-381, 385, 389 — subg. Pseudoechites, 381 — berteroi, 382 — rhynchosperma, 387 — var. umbellatus, 381 acoe a: 463 — pubiflora, 464 Ehretiaceae, 318 Elaphoglossum smithii, 151 Eleocharis cellulosa, 150 Elephantodon, 68 Elleanthus dussii, 149 — ae 473, 482 Enslini 81 Ephedranthus, 170 Epipetrum, a 59, 62 Ericaceae, Be eae in the Southeastern United States, The Genera of, 131-142 INDEX sano 131-142 m. Eriocauloideae, 132, 136 — oe Paepalanthoideae, 132, 138 Eriocaulon, 132, 136-139, 142 — aquaticum, 137 — cinereum, 135-137 — decangulare, 137 _ microcephalum, 136, 137 13 Espeletia, 429, 436 Eudema, 515, 516 Eugenia, 155 Euphorbia pendula, 490 — viminalis, 48 Euphorbiaceae, 18, 149, 318, 391, 488 Eustegia, 44 Evolution of the Polyalthia hypoleuca Complex (Annonaceae) of Malesia, The Biosystematics and, I. Systematic Treat- ment, 153-246 Fagraea, 319 60 ee ittoralis, 150 r, 298 Fischer, 454, 463, 475 10 — angustissimum, 48 Gaultheria domingensis, 147 nera of Commelinaceae in the South- eastern United States, The, 97-130 Genera of Eriocaulaceae in the Southeast- Genera of Suborder Apocynineae (Apo- cynaceae and Asclepiadaceae) in the Southeastern United States, The, 307- 401, 443-514 Gentianaceae, 309, 319, 335, 341 532 JOURNAL OF THE ARNOLD ARBORETUM Gesneriaceae, 148 Gibasis, 98, 101, 107, 120-122 1 — consobrina, 121 — geniculata, 121, 122 — karwinskyana, 121 — linearis, 121 , 494 Goniothalamus eiganteus, 242 — tortilipetalus, Gonolobus, 314, eh 444, 447, 449, 451, 452, 454, 460, 463-465, 467, 471-477. 481 — gonocarpus, 464-466, 469, 472-477 — hirsutus, 472, 474 Gouania lupuloides, 144 Gramineae, 15 Grass Systematics and Evolution [review], 23, 524 Guatteria glauca, 209 — hypoglauca, 2 _ oi. 209, 213, 227 sumatrana, 220 Guttiferae. 160 — ww Hadrodemas, 99 Hairy pipewort, 139 Haloragaceae, 146 Hamatris, 68 Hapalanthus, 120 Haplophyton, 359 — cimicidum, 359 Hat-pins, 137 Hedyots callitrichoides, 146 — commutata, 146 Helichrysum, 142 Heliconia psittacorum, 149 SS Helm _ bulbifera, 7) [voL. 70 Hemlock, 292 Higinbothamia, 59, 68 Himatanthus, 313, 329, 341 Holarrhena, 378 Hololachna ne 405 Hoya, aa Hura — ae 6 Hydrilla verticillata, 150 Hydrocharitaceae, 150 Hydrophyllaceae, 318 Hyperocarpa, 68 Ichnocarpus, 392 Indian hem Indigofera hartwegii, 145 Inga vera subsp. vera, 145 Ipomoea batatas, 77 — obscura, 148 Juglans, 34 Justicia betonica, 148 Keteleeria, 248, ‘oe 251-253, 293, 300 Kopsia, 314, KRAL, ROBERT. as Genera of Eriocaula- ceae in the Southeastern United States, 131-142 Kyllinga nemoralis, 150 Labiatae, 148 Lachnocaulon, 132, 136, 138, 139, 141 142 — anceps, 139 — beyrichianum, 139 — cub 9 — michauxii, 139 Lacmellia edulis, 334 Lagerstroemia speciosa, 146 Landolphia, 345 Lantana camara, 479 Larix, 247-254, 266 Leguminosae, 18, 145, 345 Leiandra, | 1 19 Trees $17, 519 1989] apnea 363-365 a, 364 ee 309, 318, 319, 341-343 — subfam. Plocospermatoideae, 318 2 Lythraceae, 146 Macropiper, 4 Magnolia, 11 Majanthemophyllum petiolatum, 60 Malesia, The Biosystematics and Evolu- tion of the Polyalthia aes Com- lex (Annonaceae) of, I. Systematic Treatment, 153-246 Malmea depressa, 242 Malpighia, | Mandevilla, 348, 380 Mandonia, | Marsdenia, 45 455, 460- — subg. Chthamalia, 464 0 Matelea, 444, 446, 447, 449, 451, 452, 454, 6 ea, 464 — alabamensis, 460, 461, 464-466, 469, 476 — baldwyniana, 464-466, 468, 469 — carolinensis, 462, 464, 465, 467-470, , 474 — cynanchoides, 447, 461, 466, 467, 469 — decipiens, 464, 465, 468-470 — flavidula, 464, 468, 469, 475 — floridana, 464, oe 476 — gonocarpa, 464, 47 — obliqua, 464, 465, ee 469 — palustris, ya — producta _ pubiflora, an 461, 464-467, 469, 476 Mayacaceae, 98-101 Meiocarpidium lepidotum, 242 Meliaceae, 144, 160 Meliola east 206 Mellichamp1 —_— Menispermaceae, 68 Menyanthaceae, 309, 319, 341 recennen 380 INDEX 323 Metalepis, 481 Metastelma, 481, 482 — blodgettii, 483 — northropiae, 483 Mexico and Central America, The Pedi- cellate Pipers of (Piper Subg. Arctotton- ia). Taxonomic Studies in the Pipera- er 1-55 Milk withe, 485 Milkweed, 493 Milkweed Family, 443 Milkweed vine, 478 Mimosa arenosa var. arenosa, 145 Mucuna pruriens f. cochinchinensis, 145 Murdannia, 98, 99, 101, 107, 122-124 123 Moicaina, 404, 405 — germanica, 404, 41 Myriophyllum aquaticum, 146 Myrsinaceae, 147 Myrtama, 404, 405 Nanarepenta, 68 Nerium, 312, 316, 328, 330, 332, 333, 337- 340, 344, 345, 347 — coronarium, ve — divaricatum — oleander, 347, 379 New or Noteworthy Draba (Brassicaceae) from South America, 427-437 Nymphaea rubra, 144 Nymphaeaceae, 144 Ochrosia, 314, 329, fa 332, 334, 339, Oleaceae, 309, 319 534 Oncodostigma, 242 , 242 Orbignya cohune, 6, 24, 44, 47 maa ig Ottonia, 1, = ae : 45 — thiemeanum, 48, 51 Oxandra, 173, — espintana, 242 — laurifolia, 24? — leucodermis, 171, 1, 499 Gone. 450, 485-488 Pachypodium, 315, aa us, 132, 14 Palmae, Pectis elongata var. floribunda, 147 Pennisetum i aan 151 Pentacyphus, 4 Pentalinon, 312, 329, 330, 332, 333, 339, 340, 345, 355, 380, 383-386 — andrieuxil, 385, 386 86 Sei aera 309, S17. 318 Periwinkle, 3 Peru, pee aa (Brassicaceae), a Remarkable New Genus from, 515-521 Pervinca, 370 Petermannia, 58, 59, 61, 62 Petermanniaceae, Phaeanthus, 242 Philibertella, 485, 487, 488 Philibertia, 450, 485-488 — solanoides, 486 eae 99, 117-119 0 — lau JOURNAL OF THE ARNOLD ARBORETUM [VoL. 70 Phyodina micrantha, 119 — navicularis, 119 Phytophthora palmivera, 479 Picea, 249-254, 262, 266, 284-291, 293 — Sect. — sect. — sect. _— sect. — obovata, 286 _ soreness 286, 287, 293, 296 Pinaceae in the oe United States, e Genera of, 247-305 Pinaceae. 247- a — subfam. Abietoideae, 249, 250 — subfam. Laricoideae, 249, 250 — subfam. Pinoideae, 249, 250 Pine, 263 Pine Family, 247 Pinus, 155, 248-254, 262-284, 286. 296 — subg. Diploxylon, 267 . Ducampopinus, 264 — subg. Eupitys, 267 — subg. Haploxylon, 266 — subg. Pinus, 262-264, 267, 273-275 — — sect. Pinus, 267 — — — subsect. Australes, 267, 269, 271- 14° —-— Sup See Contortae, 272-275 . Odcarpae, 273-275 . Pinus, 273 . Sabinianae, 274 273-275 Balioudanne: —— subs — subg. Strobus, 262-264, 2 — sect. Parrya subsect. 4 bsect. Cembroides, 27 — — sect. Strobus subsect. Strobus, 264, 266 — — Su — sect. Diploxylon, 263 — sect. Haploxylon, 263 — sect. Parrya, 266, 274 1989] Pinus sect. Pinea, 266 — sect. Strobus, 266, 273, 274 — subsect. Cembrae, 266, 274 australis, 267, 469 ayacahuite, 253 — banksiana, 272-274 275 brutia, canariensis, 263 coulteri, 273 echinata, 269, 270, 275 edqduus, 2/5 elliottu, 271, zee — var. densa, 27 — var. aaah. ee Daal halepensis, 275 Us DUS palustris, 267, 269, IN lees, ZS — x Pinus echinata, 273 picea, 298 pinea, 275 pungens, 269, 271, 274 5 , 267 eee 264, 266, DOS ar. chiapensis, 266 Beak 262, 263 taeda, 267, 269, 275 72 Piper etre Arctottonia: ees Pedicellate Pipers of Mexico and Central America. Taxonomic Studies in the ee I, 1-55 Piper, 1-5, 11, 13-15 — subg. Arctottonia, 1-55 INDEX 535 Piper subg. Enckea, 14 — subg. Ottonia, 3, 5, 11, 14 — sect. Enckea, 3, 4 — sect ete 44 sect. Steffensia, 3 abalienatum, "1 7,9, 10, 13-15, 17-20, 47 aguilanum, 45, 47 albicaule, 45, 47 amalgo begoniaefolium, 1 brachypus, 6, 11, a 17, 19, 21-23, 25, — brevipedicellatum, 6, 11, 13, 17, 20, 23- 25, 42 chichankanabanum, 48 khuntlananetol £102 .282922 1 — colotlipanense, 6, 10, 13-15, 17, 19, 20, 27-29 — diguetianum, 42, 44 — el-muletonum, 17 — guazacapanense, 4, 6-8, 15, 17, 29-31 mas, 33, 35, — mcvaughii, 35, 47 , 13, 15, 17, 20, 30-33, — michelianum, 17, 33-35, — muelleri, 3, 14-16, 31, a io 40, 51 — neesianum, 3, 6-8, 13, 15, 16, 31, 37- Sl , 50, — DOE ers 6, 9, 17, 22, 40-42, 48 _— palmicol — rosel, 6, a 35, 42-45, 47 — sempervirens, 37 — stipulaceum, 3, 6, 9, 10, 12, 13, 15, 17, 19, 20, 35, 45-48 — tatei, 48 — tejupicoanum, 42 eae 48, 51 — tuxpenyan 48 — Scene eS 6-8, 15, 16, 37, 40, 48- 51 Piperaceae, Taxonomic Studies in the—I. The Pedicellate Pipers of Mexico and Sues America (Piper Subg. Arctotton- , 1-55 neers 2, 3, 14 — tribe Pipereae, 4 Piscidia carthagenensis, 146 Platymitra siamensis, 242 536 Plectranthus verticillatus, 148 Pleiocarpa, 334, 335, 338, 342 utica, 338 Pleurothallis es 149 Plocosperma, 318, 2 Plocospermataceae, ne Plumeria, 313, 329, 33 |.. pe 340, 341, Plumieria, 331 Poacynum, 311, 312, 332, 333, 338, 391- 393 Podocarpaceae, 248, 250-252 ma, 493 — pedicellata, 496 Pollia, 102 Polyalthia hypoleuca Complex (Annona- ceae) of Malesia, The Biosystematics and Evolution of the, I. Systematic Treat- ment, 153-2 Polyalthia, 157, 166-171, 173, 175-177, 197, 199, 200, 205, 224, 243, 244 — sect. Afropolyalthia, 199 — sect. Eu-Polyalthia, 167, 199 — sect. Monoon, 167, 199 — sect. oe 167, 199 — aberrans — affini _ amygdalina, 244 — capuronii, 173, 175, 176, 242 — cauliflora, 168, 170, 242, 244 — celebica, 244 _ cinnamomea, 170, 205, 242, 244 — corticosa, 244 _ diospyrifolia, 177 — discolor, 166, 167, 169, 172, 174, 175, 185-188, 190-194, 196, 198, 202-206, 208, 213, 215-219, 233, 242 — emarginata, 173, 242 — evecta, 244 — flagellaris, 244 — gigantifolia, 242 — glauca, 166-169, 171-173, 175, 177, 179, 185-188, 190-194, 196, 198, 201- 214, 216, 219, 227, 230, 242 — gracilis, 242 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 70 Polyalthia heteropetala, 173, 242 — hypoleuca, 167, 168,171, 172,174,175, 185-188, 190, 192-196, 198, 199, 203, 204, 206, 208, 213,214, 216, 227-233, hepolelica complex, 153-246 — insignis, | — jenkinsii, 170, 175, 200, 242, 244 — lateriflora, 171, 200, 205, 227, 242, 244 — hittoralis, 244 — longifolia, 167, 170, 175, 177, 200, 205, 242, 244 — macropoda, 242 — microtus, 244 — modesta, 244 — motleyana, 244 — multinervis, 168, 169, 175, 185-188, 190, 192-196, 198, 201, 203, 204, 206, 208, 215, aye oo 242 — nemoralis, — obliqua, me po. 244 — oblongifolia, 243 — oligosperma, 173, 175, 176, 243, 244 — oliveri, 175, 243, 244 — ovalifolia, 153, 168, 173, 179, 185-188, 190-194, 196, 202, 203, 207, 208, 213- 216, 218, 227, 230, 233, — parkinsonii, 167, 187, 198, 209, 213 — parviflora, 244 — pedicellata, 243 — plagioneura, 244 — rumphu, 170, 200, 243, 244 — sclerophylla, 171, 200, 205, 243 243 188, 190-194, — socia, 171 — suberosa, 175, 243, 244 — sumatrana, 167-169, 175, 185-188, 190, 192-194, 196, 198, 203-208, 213, 220- 227, 230, 233, 243 — var. macrocarpa, 226, 233 244 — venosa, 244 Polyceratocarpus, 173 — gossweileri, 243 1989] Polyceratocarpus microtrichus, 243 — parviflorus, 243 Polynome, 68 Polypodiaceae, 151 Polypodium aureum var. areolatum, 151 A. The Genera of Pinaceae in the Southeastern United States, 247- 305 Proteaceae, 345 Pseudephedranthus fragrans, 243 Pseudolarix, 248-252, 300 Pseudotsuga, weer 266 — menziesii, 250 Pse udoxandta 173, 176 e Ss . -2 Q psa: 15] Pteris srandifolia, 151 nea mr 155, 468 — lae 469 Rajania, 58, 59, 61, 62 — cordata, Rauvolfia, 312, 315, 329, 330, 332, 333, 340, 341, 344, 345, 347, 356 — glabra, 356 — littoralis, 347 — tetraphylla, sy 339, 345, 347, 355, 357 404, 405 Recent Plant Collections from Dominica, 143-152 Rectanthera, 120 Rhabdadenia, 330, 332, 333, 345, 355, 380, 383, 388-390 — sect. Eglandulosae, 389 — sect. Glandulosae, 332, 383, 389 — sect. Rhabdadenia, — biflora, 379, 389, 390 — pohlii, 389 INDEX 537) Rhamnaceae, 144, 345 Rhazya, 341, 342, 359 oa o8 101 108, 111 — spathacea, 111 Rhynchosia phaseoloides, 146 7 H. The Biosystematics and Evolution of the Polyalthia hypo- leuca Complex (Annonaceae) of Male- sia, I. Systematic Treatment, 153-246 Rollinia, 170 RomERO, Gustavo A. G. C. K. Dunster- ville, 1905-1988, an Appreciation, 439- Rosatti1, THoMaAS J. The Genera of Sub- Rubiaceae, 146, 309, 310, 341-343, 345 Ruizodendron, 170, 243 Saltcedar, 410 Soa wn 446, 447, 450-452, 455, 460, 464, 484-490 — subg. Sooners: 488-490 — — ser. Clausa, 488 — — ser. Cynanchoide, 488 — — ser. Flav — subg. ene 488 4 | wn co oz ge 7*© v7 ce nn oO =| lore) — — sect. Philibertia, 488, 490 — subg. Sarcostemma, 488, 490 — sect. Ceramanthus ser. Angustissima, 487 — sect. Eusarcostemma, 486-488 — sect. Philibertia, 486 — sect. Sarcostemma, 486, 487 — angustissima, 48 — clausum, 485, 486, 489, 490 — cynanchoides, 447, 483, 486, 489 538 Sarcostemma cynanchoides subsp. cynan- choides, 489 — — subsp. hartwegii, 489, 490 — glaucum, 489 SCHUBERT, BERNICE G., and IHSAN A. Dioscoreaceae in the Southeastern United States, 57-95 Secamone, 447, 448, 451-454 Secondatia, 387 Setcreasea, 98, 108, 111 — pallida, Sideroxylon, 32 SODERSTROM, THOMAS R., ET AL., Grass Systematics and Evolution [review], 523, 524 Solanaceae, 148 fe) South Ainenica. New or Noteworthy Draba (Brassicaceae) from, 427-437 Southeastern United States, The Diosco- reaceae in the, 57-95 gscuenre sax United States, The Tamari- aceae in the, 403-425 een United States, The Genera of Commelinaceae in the, 97-13 Southeastern United States, yy Genera of Eriocaulaceae in the, 131-142 Southeastern United Cay The Genera of Pinaceae in the, 24 5 Southeastern United tens The Genera of Suborder Apocynineae (Apocynaceae and Asclepiadaceae) in the, 307-401, 443-514 Spermatophyta, 144 9 8 Spiderwort Family, 97 Spiny-pod, 4 Spiranthes adnata, 149 Spironema, 120 Spruce, 284 JOURNAL OF THE ARNOLD ARBORETUM [voL. 70 Stanfieldiella, 100 Stapelia, 446, 450, 451, 455 Stelechocarpus cauliflorus, 243 Stemodia verticillata, 148 rangler vine, 478 Strobus, 262 — strobus, 266 Strophanthus, 79 Strophis, 68 Syngonanthus, 132, 136, 139-142 — flavidulus, 140, 142 _ umbellatus, 142 Tabernaemontana, 314, 316, 329, 330, 332, 339, 340, 344, 346, 359 — tribe Tamariceae, 404-406 Tamaricoxylon africanum, 406 Tamarisk, 410 Tamarisk Family, 403 Tamarix, 403-406, 410-425 — sect. Oligadenia, 411 — — ser, Gallicae, 411 — — ser. Laxae, 413 — — ser. oo , 411,413 — — ser. Tamari 1, 414 — — ser. Vagianis 413 — africa 410, 4 _ aphylla, 413, 6, ial — canariensis, 410, 411, 413, 414 — chinensis, 413-416 1989] Tamarix ericoides, — gallica, 410, 411, ve 415, 416 — juniperina, 414 — mannifera, 418 — mascatensis, 414 — odessana, 411 — parviflora, 410, 413, 415 — pentandra, 411, 414-417 — ramosissima, 410-412, 414, 415 — rosea, 410 — tetragyna, 410, 413, 416 — tetrandra, 404, 405, 413, 415 62 e, 250 Taxcdiaccan, 248, 249 8 — distichum, 388 Taxonomic Studies in the Piperaceae—I. The Pedicellate Pipers of Mexico and Central America (Piper Subg. Arctotton- ia), 1-55 Taxus, 250 Tear-shrub, 356 Testudinaria, 68 Thelypteris reptans, 151 Thevetia, 312, 314-316, 329, 330, 332, 334, 337-341, 344, 347, 357 — nerufolia, 451 — peruviana, 316, 347, 354, 357, 451 Thyrsanthemum, 10 Tilletia commelinae, 126 Tinantia, 100, 111 Eger, 132 5, 330, 332, 333, 339 340, 344, 345, i 380, 386-388 — sect. Axillanthus, 387 — sect. Eutrachelospermum, 387 — sect. Trachelospermum, 387 — difforme, 339, 387, 388 — jasminum, 346, 387, 388 Trachomitum, 311, 312, 332, 333, 338, 93 — venetum, 391 Tradescantella, 98, 108, 117, 118 — floridana, 11 Tradescantia, 98-102, 107-118, 121 — sect. Austrotradescantia, 112 — sect. Campelia — sect. Cokolornia: 112 — sect. Corinna, 112 INDEX Tradescantia sect. Cymbispatha, 111 _ a ia, 11] . Tradescantia, 108, 109 . Tuberosae, 110 — ser. Virginianae, 108 sect. Zebrina, 111 albiflora, 102 xandersoniana, 102, 109 a 109, 110 hirta, 111 a 111 orchidophylla, 110 — ozarkana, 110 pallida, 111 109, 112, 113 pexata, 110 pilosa, 109 pulchella, 121 reverchonil, 110 rosea, 11 roseolens, 110 rozynskii, 110 schippii, 111 540 Tradescantia sillamontana, 110 sO 2 — spathacea, 111 — subaspera, 102, 109 se Nal oe 109 — tharpu aes "100. 102, 108-110, 112 Trichilia, 32 Trichopodaceae, 58-60 6 Tripogandra, 99-101, 117 Tsuga, 249-252, 254, 262, 291-298, 300 — sect. Hesperopeuce, 252, 292, 295 — sect. Micropeuce, — sect. Tsuga, 247, 251, 296 — americana, 292 — araragi, _ minder 292, 293, 295, 296 — caroliniana, 293, 295, 296 — dumosa, 296 —_ heteropla 293, 295, 296 — x/Jeffrey — longibracteata, 292, 293 a, 252, 292, 293, 295, 296 TUCKER, Gorpon C. The Genera of Com- melinaceae in the Southeastern United States, 97-130 Tylodontia, 481 Tylophora, 450 Umbelliferae, 34 Unona hypoleuca, 209, 213 tii, 209 Ustilago commelinae, 126 JOURNAL OF THE ARNOLD ARBORETUM 252,292, 295, [VoL. 70 Uvaria boniana, 243 — calamistrata, 243 — confertiflora, 243 — sofa, 243 Valerianaceae, 309 Vallesia, 314, 329, 330, 332, 340, 343- , 356-358 Vangueria madagascariensis, 147 Vinca, 312, 315, 319, 328-330, 332, 337, 338, 341-345, 355, 363, 365, 367, 369- 377 — sect. Erecta, 375, 376 — sect. Herbacea, 375, 376 — sect. Major, 375 — sect. Minor, 374, 375 — sect. Pervinca, 374 — sect. Vinca, 374 — sect. Vincopsis, 374, 375 — balanica, 373 — difformis, 373-3 — erecta, 340, 370, oa 373-376 + — libanotica, 374-3 — lutea, 385 — major, 344, 345, 365, 367, 370, 371, 373-375 — media, 373-375 — minor, 316, 317, 337, 344, 345, 36S, 367, 370-376 — pubescens, 374, 375 — rosea, 363, 364 Vincetoxicum, 314, 315, 472-474 472 — gonocarpus, a 474, 476 — suberosum, Vine milkweed, 481 1989] Weldenia, 101 Wild allamanda, 385 Willughbeiaceae, 329 Wrightia, 330 Xerodraba, 515-517, 519 1a, 5 , 517 Xylopia, ae. 200, 202, 243 INDEX Xylopia elliptica, 243 — ferruginea, 243 Xysmalobium, 493, 494 Yam, 68 Yam Family, 57 Zebrina, 98-101, ad 111 — pendula, 102, Zephyranthes rosea, 1 Zippelia, 4 JOURNAL OF THE ARNOLD ARBORETUM HARVARD UNIVERSITY VOLUME 70 1989 Dates of Issue No. | (pp. 1-152) issued LO January 1989. No. 2 (pp. 153-306) issued 25 April 1989. No. 3 (pp. 307-442) issued 14 July 1989. No. 4 (pp. 443-542) issued 11 October 1989. Contents of Volume 70 1989 Taxonomic Studies in the Piperaceae—I. The Pedicellate Pipers of Mexico and Central America (Piper Subg. Arctottonia). ALLAN J. BORNSTEIN ........0. 000000. c cece eee ceeeeeenees 1-55 The Dioscoreaceae in the Southeastern United States. IHSAN A. AL-SHEHBAZ AND BERNICE G. SCHUBERT ............ 57-95 The Genera of Commelinaceae in the Southeastern United States. GORDON C. TUCKER 2.00000 ccc cece cece eee. 97-130 The Genera of Eriocaulaceae in the Southeastern United States. ROBERT KRAL .. 0... ccc ccc beeen eee, 131-142 Recent Plant Collections from Dominica. C. WHITEFOORD ........00 00.0.0 c ccc cece cece cece eee. 143-152 The Biosystematics and Evolution of the Polyalthia hypoleuca Com- plex (Annonaceae) of Malesia, I. Systematic Treatment. STEVEN HW ROGSTAD : 5 h..us eeuwt oedeean pawee eee deeds fais 153-246 The Genera of Pinaceae in the Southeastern United States. ROBERT A: PRICE . yce240.c¥ee eeu cde ee bales Oe oe cbde nk dunn 247-305 The Genera of Suborder Apocynineae (Apocynaceae and Asclepia- daceae) in the Southeastern United States. THOMAS J. ROSATU iy x 49-44 Sek uy yen ha kaya wae eee ee cei ds 307-401 The Tamaricaceae in the Southeastern United States. MVILEIA MG de CORINS. 55 ashaus yee eee ae dea ee bbs ao bh edd 403-425 New or Noteworthy Draba (Brassicaceae) from South America. THSAN A. AL-SHEHBAZ ...0. 0000000 cc cee ee 427-437 G. C. K. Dunsterville, 1905-1988, an Appreciation. GuSsTAVO A. ROMERO ... 00... teen ee, 439-44] The Genera of Suborder Apocynineae (Apocynaceae and Asclepia- daceae) in the Southeastern United States. THOMAS J: ROSATI 9 .i5.2044052-060b8 64 ibe edna eda tana e cua 443-514 Dactylocardamum (Brassicaceae), a Remarkable New Genus from Peru. IHSAN A. AL-SHEHBAZ ... 0.000000 cece 515-521 BOOK Rev 6w¢ 500-66. cahbe8 Gi ae keen bc oda a oesakua eee nae 5232 024 PV oe ig ie Gite os Ge ek Ee Fe §25-541 The Journal of the Arnold Arboretum is a refereed botanical journal that presents the results of original research in systematic and evolu- tionary botany and related disciplines. Beginning in 1988, critical, in- depth reviews of major books published in the above fields will also be included on a periodic but timely basis. In addition to Arboretum staff and associates, we encourage botanists worldwide to submit manu- scripts to the Journal. We particularly welcome those that record the results of studies utilizing Harvard University Herbaria (A and GH) collections or focusing on geographic areas of traditional interest to Arboretum staff. A guide for potential contributors, periodically pub- lished in the Journal, is available on request. Journal of the Arnold Arboretum October, 1989 CONTENTS OF VOLUME 70, NUMBER 4 The Genera of Suborder Apocynineae (Apocynaceae and Asclepia- daceae) in the Southeastern United States. rei LR 245 ee dha g s Seles wide wae bei aes 443-514 Dactvlocardamum (Brassicaceae), a Remarkable New Genus from Peru. IHSAN A. AL-SHEHBAZ ......... 0000 cc cc ccccecucucucucucueus 515-52] BOOK PROVIEW 045638 eeebk ea beddenk 0 34.)400)4. 65 b20c wags ee xes 523, 524 UO gah go peda sa ce bay oe ake ek eA hao aed t uk ada neadedr 525-541 Volume 70, Number 3, including pages 307-441, was issued 14 July 1989.