JOURNAL OF THE ARNOLD ARBORETUM HARVARD UNIVERSITY C. E. WOOD, JR. EDITOR L. I. NEVLING, JR. LAZELLA SCHWARTEN ASST. EDITOR CIRCULATION VOLUME XLI CAMBRIDGE, MASS. 1960 Reprinted with the permission of the Arnold Arboretum of Harvard University KRAUS REPRINT CORPORATION New York 1968 DATES OF ISSUE No. 1 (pp. 1-110) issued January 15, 1960. No. 2 (pp. 111-230) issued March 28, 1960. No. 3 (pp. 231-840) issued July 5, 1960. No. 4 (pp. 341-457) issued October 15, 1960. ge atid Printed in U.S.A. TABLE OF CONTENTS THE GENERA OF THE EBENALES IN THE SOUTHEASTERN UNITED Staves. By C. E. Wood, Jr. and R. B. Channell oc... A Taxonomic REVISION OF ite XII. Section Micro- carpus. By Netta E. Gra STUDIES IN THE GENUS Beer VIII. NomMencLATURAL CuHances. By Richard A. Howard THe GENERA OF THE ARALES IN THE SOUTHEASTERN UNITED SraTes. By Kenneth A. Wilson SrupiEs IN ARTOCARPUS AND ALLIED GENERA, IV. A REVISION OF ARTOCARPUS SUBGENUS Psruposaca. By Frances M. Jarrett STUDIES IN ARTOCARPUS AND ALLIED GENERA, IV. A REVISION OF ARTOCARPUS SUBGENUS PsgupoJgaca (Concluded). By Frances M. Jarrett SomME USEFUL TECHNIQUES IN THE STUDY AND INTERPRETATION oF PoLLEN MorpuHouocy. By I. W. Bailey Tue GENERA OF SARRACENIACEAE AND DROSERACEAE IN THE SOUTH- EASTERN UNITED States. By Carroll E. Wood, Jr. ......0.0.... A REVISION OF THE GENUS CLETHRA IN CHINA. By Shiu-ying Hu PoLyPLoIpy IN ENKIANTHUS (Ericackak). By Hally J. Sax ........ THE GENERA OF HypDROPHYLLACEAE AND POLEMONIACEAE IN THE SOUTHEASTERN UNITED States. By Kenneth A. Wilson ........ STUDIES IN THE GENUS CoccoLoBa, IX. A CRITIQUE OF THE SOUTH AMERICAN Species. By Richard A. Howar STUDIES IN THE GENUS CoccoLoBA, IX. A CRITIQUE OF THE SOUTH AMERICAN Species (Continued). By Richard A. Howard .... Lear VENATION AND is ee IN THE GENUS Raou.ia (Com- PosITAE). By Otto T. Solbri Tue GENERA OF MyrTACEAE IN THE SOUTHEASTERN UNITED STATES. By Kenneth A. Wilson SUPPLEMENT TO A MONOGRAPHIC STUDY OF THE West INDIAN SPECIES OF PHYLLANTHUS. By Grady L. Webster ................ THe ANATOMY OF PHENAKOSPERMUM (Musackag). By P. B. Tomlinson Tue GENERA OF CONVOLVULACEAE IN THE SOUTHEASTERN UNITED srates. By Kenneth A. Wilson BriBpLioGRAPHIC DATA ON THE HILLCREST GARDENS Books, 1911- 1941. By Richard A. Howard 270 279 STUDIES IN ARTOCARPUS AND ALLIED GENERA, V. A REVISION OF PARARTOCARPUS AND Hu.Luertia. By Frances M. Jarrett ...... COMPARATIVE ANATOMY OF THE LEAF-BEARING CACTACEAE, I. OLIAR VASCULATURE OF PERESKIA, PERESKIOPSIS AND QUIA- BENTIA. By J. W. Barley STUDIES IN THE GENUS Cocco.osa, IX. A CriTIQUE ON THE SOUTH AMERICAN Species (Concluded). By Richard A. Howard .... VEGETATION ON GIBBSITIC Sorts IN Hawarr. By J. C. Moomaw and M. Takahashi NOMENCLATURAL CHANGES IN DaAPHNopsis (THYMELAEACEAE). By Lorin I. Nevling, Jr. SEEDLING LEAVES IN PALMS AND aera MorPHOLOGICAL SIGNIF- ICANCE. By P. B. Tomlins THe Drrector’s Report BIBLIOGRAPHY OF THE PUBLISHED WRITINGS OF THE STAFF AND STUDENTS, JULY 1, 1959-JuNE 30, 1960 STAFF OF THE ARNOLD ARBORETUM, 1959-1960 InpEx To Vou. XLI JOURNAL OF THE ARNOLD ARBORETUM VoL. XLI JANUARY 1960 NUMBER | THE GENERA OF THE EBENALES IN THE SOUTHEASTERN UNITED STATES 4 C. E. Woop, JR. AND R. B. CHANNELL OF THOSE COMPRISING THE EBENALES in the Englerian sequence of angiosperm families, four — Sapotaceae, Ebenaceae, Styracaceae, and Symplocaceae — occur in the southeastern United States, and it is these which are considered here as constituting the order. These four families, together with the Hoplestigmataceae, Diclidantheraceae, and Lissocarpa- ceae, also included by Engler and Gilg, are characterized by sympetalous corollas with stamens generally two or three times as many as the corolla lobes (or, by abortion, equal in number to and opposite them) and by superior to inferior, incompletely to completely loculed ovaries with axile placentation. Wettstein, Engler and Gilg, Rendle, Cronquist, and Benson, among others, have retained the four larger farmilies in a single order (although sometimes with additions), but others, as Hallier and Hutchinson, have split the group in various ways. Copeland (see Styracaceae) suggests that the order is a natural group with a collateral relationship to the Ericales and with an ancestry most nearly represented among living plants by the Theaceae. Altogether, the evidence from floral morphology and anatomy, pollen, wood structure, nodal anatomy, and embryology, insofar as this information is EERIE, is in harmony with this view, and no very con- vincing data have yet been presented to the contrary. Various items in the 1 Prepared for a biologically oriented generic flora of the southeastern United States, a joint project of the Gray Herbarium and the Arnold Arboretum made possible through the support of George R. Cooley and the National Science Foundation. The scheme follows that outlined at the beginning of the series (Jour. Arnold Arb, 39: 296-346. 1958). Other published portions of these studies will be found in Jour. Arnold Arb. 40: 94-112, 161-171, 268-288, 369-384, 391-397, 413-419. 1959, and in the pres- ent issue. We are much indebted to the many people who have given freely of advice, information, or materials in connection with the four families treated here. In addi- n to our immediate colleagues, these include L. J. Brass, G. R. See W. H Duncan, R. J. Eaton, R. K. Godfrey, Mrs. J. N. Henry, J. Kucyniak, J. D. Ray, Jr., rd and are used with his kind permission. As in previous papers in this series, ihe illustrations are the work of Dorothy H. Marsh. 2 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI descriptions and discussions which follow will be found to bear on this problem. At least some of the characteristics which often are cited in connection with the interrelationships of the families of the Ebenales need qualifica- tion and a great deal more investigation. For example, although the Sapotaceae are said to have completely septate ovaries, those of at least some (e.g., species of Bumelia and Manilkara) are at anthesis septate be- low but are no more completely so above than those of Styrax, and in in- ferior ovaries in both Styraceae and Symplocaceae a similarly incomplete condition occurs. Moreover, that inferior ovaries characterize five of the thirteen genera of the Styracaceae usually is glossed over in comparisons with the Symplocaceae. It should be remarked, too, that in the Ebenaceae. a single ovule in each locule seems to occur much more frequently than two, and there are indications that this condition may have come about through the development of additional septa which have separated the paired ovules. Indeed, all four families may well provide examples of increase in numbers of flower parts, rather than reduction. (Cf. the perianth and gynoecium of Sapotaceae and Ebenaceae and the androecium of Styracaceae and Sym- plocaceae, for example.) It is also noteworthy in connection with phy- logeny that the ovules of Sapotaceae seem to have a single integument, those of Ebenaceae two, of Styracaceae either two or one (by fusion of the two), and of Symplocaceae one, although as yet very few representatives of these families have been examined. SAPOTACEAE (Sapote FAMILy) Armed or unarmed trees or shrubs with milky sap and alternate [rarely opposite], simple, exstipulate, usually entire and coriaceous leaves. the nodes with 3 traces from 3 leaf-gaps (except some species of Bumelia, 1 from 1). Inflorescences axillary, basically dichasial, ours simple, cymose or umbellate, or the flowers sometimes singly disposed, the pedicels bracte- olate at the base. Flowers complete, regular. Calyx of 4—-9|—12] imbricate, biseriate, or spirally arranged sepals, connate at the base. Corolla sym- petalous, the lobes imbricate in the bud, usually as many as the sepals, sometimes with paired lateral or dorsal appendages. Stamens [twice as many as or| as many as and opposite the lobes of the corolla, epipetalous, distinct, the anthers 2-locular, longitudinally dehiscent; staminodia (when present) alternate with the fertile stamens and the lobes of the corolla. Gynoecium syncarpous, the style 1, the stigma unlobed or with as many lobes as locules, the ovary superior, the locules 1-14, typically 4 or 5, the placentation axile, a single, anatropous, l-integumented ovule in each locule [except in Diplodn with unilocular ovary and 2 ovules]. Fruit an indehiscent berry, often with a thin, leathery to bony outer layer; seeds large, with fleshy endosperm or none. TypE GENuS: Sapota Mill. = Manilkara Adans. A family of about 40 genera and more than 600 species, all woody, and primarily of the tropics of both hemispheres. Six of the approximately 15 1960] WOOD & CHANNELL, GENERA OF EBENALES 3 genera of the Western Hemisphere are represented in the southeastern United States, but only Bumelia occurs outside of peninsular Florida in this area The Sapotaceae are distinguished from the other families of the Eben- ales by the superior ovary which is usually completely septate (note illus- tration of Bumelia, however) and by the presence of a solitary ascending, 1-integumented ovule in each locule. The combination of laticiferous ele- ments in leaves and stems and of two-armed hairs (one arm of which is sometimes suppressed) characterizes the family anatomically. The delimitation of species and of genera, in particular, is difficult in the family, leading to both “splitting” and “lumping” at all taxonomic levels. After a period of neglect, a number of relatively recent studies have clari- fied the taxonomy and nomenclature of many groups, but it is likely that still further changes of both kinds will be made as the various groups be- come better understood. Taxonomic characters of generic significance have included especially the presence or absence of staminodia and endosperm, the presence or absence and location of appendages on the corolla lobes, and the general type of seed-scar (hilum). Unfortunately, very few observations seem to have been made on the biology of the group, and little is known concerning the functioning of staminodia and petal appendages in connection with pollination. The family is the source of a number of economically important timbers; the seeds of several genera provide edible oils; and a number of groups are valued for the refractive rubbery compounds from the coagulated latex (gutta-percha from species of Palaquium, Payena, and Mimusops, chicle from Manilkara Zapota and related plants; balata from Manilkara biden- tata and related species). The family is also well known for a number of excellent tropical dessert fruits (most of which become quite rubbery and inedible when cooked), including the sapodilla (Manilkara Zapota), the sapote or marmalade plum (Pouteria mammosa (L.) Cronq., Calocarpum Sapota (Jacq.) Merr.), the eggfruit or canistel (Pouteria campechiana), and the star apple (Chrysophyllum Cainito). REFERENCES: ASSEM, J. VAN DEN, A. C. vAN BruccEN, Mrs. M. P. M. HERRMANN-ERLEE, M. JeuKEN, A. J. G. H. Kostermans, H. J. Lam, P. vAN Royen, W. VINK. Revision of the Sapotaceae of the Malaysian area in a wider sense. I-III. Blumea 6: 547-— 595. 1952 (Jsonandra, Burckella, Wen ror IV, V. bid. 7: 364-412. 1953 (Ganua, re IVa. eed: 481-483. 1954 (Ganua) ; IIa, Va, VI-IX, IVb. fdid. 8: 201-513. 1957 ea Manilkara, Xan- tolis, Planchonell, ie ' Pouteria, Ganua) ; oe through lower and upper Ree respectively, X 6; k, fruit, x %4; 1, carp from a very large ruit, xX 1; m,e endocarp, cross section a seed, ae (stippled), embryo, ae cavity, < 3; h-j, m, semidiagrammatic. n, H. parviflora: fruit, X VY. 30 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI smaller corollas and smaller fruit with narrower wings often tapering into the pedicel. The other extreme, #. parviflora Michx., a shrub or tree to 10 m. tall, perhaps restricted to sandy woods and bluffs of the Coastal Plain from Georgia and Florida to Mississippi, but possibly entering the Piedmont, has, in its best development, small, rather flaring corollas about 1.2 cm. long, pubescent calyces, and fruits 2.5-3.5 cm. long with narrow wings 3-5 mm. wide tapering into the pedicels (thus giving the effect of a clavate fruit). Most distinctive in northeastern Florida, this plant appears to intergrade with H. carolina; plants from the central areas of South Carolina, Georgia, and Alabama need special study in this respect. Through- out this complex the shape and size of the stony endocarp seem to be roughly the same; the varying shape and width of the four wings produces the wide range of fruit-shapes. The length of the style seems also to be more or less constant: thus it is equalled or exceeded by the corollas of H. carolina (including H. monticola), but protrudes from the short, rather flaring corollas of H. parviflora. The Chinese species, H. Macgregorii Chun, combines a deeply parted corolla (as in H. diptera), four-winged fruits (as in H. carolina), and eight stamens of two lengths (eight equal stamens in H. diptera). The ovary is said to be 3-locular. Halesia (with Nyssa, Carya, and a few others) is of interest in being a genus of the eastern American—eastern Asiatic dis- tribution in which more representatives occur in America than in Asia. In most genera following this pattern the Asiatic contingent appears both to be more numerous and to include more primitive species than the Amer- ican. Halesia is most closely related to Pterostyrax ~~ . Zits 15 species, Japan, China, Burma), Rehkderodendron H. H. Hu (9 species. China, Indochina), Sino jackia H. H. Hu (3 species, fea ee Mellio- dendron Hand.-Mazz. (2 species, China). REFERENCES: Gipson, H. H. American forest trees—15. Silverbell tree, Mohrodendron carolinum eae Britton; Halesia tetraptera Ellis. Hardwood Rec. 20: 10, 1905.* [See also pp. 600-604, of his American Forest Trees, Chi- cago, ao Aah GiersBacH, J., and L. V. Barton. Germination of seeds of the silverbell, Halesia carolina, Contr. Boyce Thompson Inst. 4: 27-37. 1932. [Warm stratifica- tion followed by cold; see also, Woody plant seed manual. U.S. Dep. Agr For. Serv. Misc. Publ. 654: 193, 194. 1948. Goprrey, R. K. Some identities in Halesia (Styracaceae). Rhodora 60: 86-88. Dl, 1257, 1958; |i. ee var. magniflora; adopts H. tetraptera Ellis for A. parviflora Michx. | at full size and which are perhaps somewhat shrunken in drying. The details of i corolla illustrated are hardly compatible with fully developed corollas of either H. carolina or H. parviflora. Fruiting material shrunken in such a way may be mathe by Fox 5265, Rutherford Co., N. C. (GH); Duncan 3349, Clarke Co., Ga. (A) Faxon, Oct. 1896, cultivated, jamaica Plain, Mass. (cir). An inecrucene series 1s fomned by £. J. Palmer 20739, 20763, 26760 [207602], April 11, 12, 1922, from near Page, LeFlore Co., Oklahoma (a): in these collections corollas of plants in various stages of development are approximately 2, 1.2-1.5, and 0.7 cm. long, respectively. 1960] WOOD & CHANNELL, GENERA OF EBENALES al LitTLe, E. L. Nomina conservanda proposals for ten genera of trees and shrubs. Madrono 7: 240-251. 1944. [| Halesia, including history, 250, 251 OrrsTED, A. S. Zur Beleuchtung der Blumen des brasilianischen Theestrauches (Neea theifera Ord. 1863 Pisonia CDOS eel 1866) und des Schneeglockchenstrauches Gales ie Galea Eeelt of O22 5. 1869. [Halesia, 222, 223; reports bisexual ae ane 12 stamens, and in the same Aisne eane: baminate flowers 1/10-1/15 as large with 10-15 sessile stamens and a rudimentary pistil. Merely immature flowers? | Osporn, B. Halesia diptera Ellis, a new tree record from Oklahoma. Proc. Okla. Acad. Sci. 18: 21. 1938. [Misidentification; H. carolina with wings of young fruit obscured in pressing. SARGENT, C. S. Mohrodendron. Silva N. Am. 6: 19-24. pls. 257-259. 1894. [H. carolina, H. diptera.| Sims, J. Halesia tetraptera. Bot. Mag. 23: pl. 910. 1806. (See also J. Lindley, Halesia parviflora. Bot. Reg. 11: pl. 952. 1826. [H. carolina.]) SYMPLOCACEAE (SWEETLEAF FAMILY) Trees or shrubs with alternate, simple, more or less coriaceous, ex- stipulate leaves, the nodes with one trace from one gap; complete, Poona sympetalous flowers: numerous|—4] stamens with ovate anthers; 2—5- locular inferior or half- inferior ovaries, typically with 2 pendulous, axile, 1-integumented ovules in each locule, and drupaceous or baccate fruits. A small unigeneric family of approximately 300 species (placed in eight sections in four subgenera by Brand), of the warmer parts of America, Asia, and Australia (absent from Africa and Europe), the species very numerous in South America, about eight species ranging northward into Mexico, about 18 in the West Indies, and one in the United States, this primarily of our area, but extending into adjacent regions. Many of the species apparently are of local distribution, endemic to islands or moun- tains. The Symplocaceae presumably are closely related to the Styracaceae, but differ in a number of characteristics: stamens usually in several series, often fasciculate (vs. stamens apparently in a single series in Styracaceae) ; ovate anthers (vs. oblong or linear anthers) ; inferior or half-inferior, com- pletely loculed ovary (vs. superior to inferior, incompletely loculed ovary) ; baccate or drupaceous fruit (vs. dry, capsular to indehiscent fruit); simple hairs (vs. stellate or peltate hairs); usually rubiaceous stomata (vs. ranun- culaceous stomata); solitary vessels with elongated pits between vessels and rays in the secondary wood (vs. solitary or clustered vessels with small, round pits between vessels.and rays); and the occurrence of spiral thickenings in the vessels, commonly, and in the fibers, occasionally (vs. the lack of such thickenings). The family is of limited economic importance. A few species are some- times used for their wood, the leaves and bark of several (including S. tinctoria) yield a yellow dye, the roots of some are used in the preparation of tonics, and the leaves of several species are used as a substitute for maté (primarily from /lex species). Symplocos paniculata (Thunb.) Miq. 4 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI (subg. Hopea, sect. Bopua), with terminal clusters of white flowers and methyl-blue fruits, and a few other Asiatic species are sometimes grown as ornamental shrubs. The biological features of the group seem to have been studied hardly at all. 1. Symplocos Jacquin, Enum. Syst. Pl. Carib. 5. 1760; Select. Stirp. Am. Hist. 166. pl. 175, fig. 68. 1763. Deciduous or tardily deciduous shrubs or trees with more or less cori- aceous, sweet-tasting leaves. Flowers in congested or open, racemose or paniculate inflorescences, in ours about 6-14 in short, sessile, close clus- ters or racemes axillary to leaves or their scars, yellow, fragrant. Calyx adherent to the lower portion of the ovary, persistent, the 5 lobes valvate or imbricate. Corolla sympetalous, 3—11-lobed, deeply 5-lobed in ours, the tube short. Stamens [4—|]many, conspicuous, exserted, in several unequal series, epipetalous at the base of the corolla, [free or monadelphous, the filaments connate in a tube, or] in ours pentadelphous, in 5 groups alter- nate with the corolla-lobes, the innate anthers short, orange, the filaments slender, white, distinct except at the base. Pollen usually 3-colpate. Gy- noecium syncarpous; stigma terminal, slightly dilated or capitate, minutely lobed; style linear, the ovary inferior or half-inferior, in ours surrounded at the top by an orange-colored disc of nectariferous spongy tissue, 3[2—5 |- locular; placentation axile, the anatropous (?)1-integumented ovules pendulous, typically 2 in each locule. Fruit drupaceous [or baccate], in ours cylindric-ellipsoid, about 1 cm. long; usually only a single seed de- veloping; embryo straight [or curved], with short cotyledons, and with copious endosperm. (Hopea L. f., 1767, not Roxb., 1814.) TYPE SPECIES: S. martinicensis Jacq. (The name from Greek, symplokos, connected, twisted, entwined, apparently referring to the union of the stamens with each other and with the petals in S. martinicensis.) Symplocos tinctoria (L.) L’Her., sweetleaf, horse-sugar, wild laurel, or yellow-wood, the only species indigenous to the United States, occurs from Sussex County, Delaware, southward to northern Florida and westward to eastern Texas and southeastern Oklahoma. Varietas tinctoria, with glabrous fruits, glabrous or glabrate current stems and tardily deciduous leaves, is widespread on the Coastal Plain, where it is largely restricted to hammocks, the margins of swamps and to sandy soils in association with Pinus, Nyssa, Persea, Magnolia, etc. Varietas Ashei Harbison, with pubescent fruits, persistently hairy stems, and promptly deciduous leaves, is distributed over a limited region of the southern Appalachians in western North and South Carolina, northern Georgia, and southeasternmost Ten- nessee (a single station in Polk County), between 1800 and 4500 feet in altitude, on dry ridges of chiefly acid, red-clay soils, where it is associ- ated with Pinus rigida Mill., Quercus coccinea Muenchh., Q. Prinus L.. and (formerly) Castanea dentata (Marsh.) Borkh. It appears largely to be lacking between mountains and Coastal Plain, although scattered sta- 1960] WOOD & CHANNELL, GENERA OF EBENALES he) tions may be expected in the Piedmont and are known from at least Aiken County, South Carolina, and Wake County, North Carolina. The geo- graphical isolation of these varieties can be determined only after addi- tional stations are sought, particularly in the Piedmont and in northern Alabama (where the species has been reported as far north as Marion, Cull- man. and Morgan counties). Varietas pyvgmaea Fern., based upon dwarf, sterile specimens with small leaves, from white sands of dry pine barrens in Isle of Wight County, southeastern Virginia, is doubtfully distinct. The typical variety occurs in abundance in this general area of the state. Fic. 5. Symplocos. a—j, S. tinctoria var. tinctoria: a, flowering twig, X 1%; b, flow ay s 2; c. detached corolla with stamens, X 2; d, flower with corolla re- moved, - e, same. the ovary in vertical section (semidiagrammatic), x 3; f, cross ee of ovary at anthesis — stippled ace becomes sclerified in fruit (semidiagrammatic), X 6; g, fruiting twig. X 14; h, portion of twig in oblique section to show eee pith, x ™%; i, mature fait, xX 2; Jj, mature fruit, vertical section, the fertile locule with seed to left and center, abortive locules to right Characteristically a shrub, Symplocos ftinctoria occasionally attains arborescent proportions (nearly 30 cm. in diameter and 12 m. tall) in the fertile river bottoms of Mississippi which usually are inundated for several weeks of the year. Brand placed Symplocos tinctoria near S. japonica A. DC. and S. setch- uensis Brand as the only American species of subg. Hopea (L. f.) C. B. Clarke, sect. PALAEOSyMpPLocos Brand, a group distinguished on the basis of the 3-locular ovary and clearly pentadelphous stamens with filiform filaments. However, because the summit of the ovary is glabrous and the united portion of the stamen-filaments is flattened (instead of round in section), Handel-Mazzetti and Peter-Stilbal excluded S. tinctoria from the subgenus (which was then renamed as subg. EosyMPLocos), suggesting that its relationships should be sought among the New World species. In- consistently, sections LopHra and Bosua (also of subg. HopEA) were de- fined by the latter authors to include species with the apex of the ovary either glabrous or pubescent! It may be significant that at least some speci- 34 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI mens of S, tinctoria var. Ashei have a few hairs on the summit of the ovary at anthesis. The subdivisions of Symplocos have been treated by some authors as separate genera, although most have maintained the group as a single genus with well-marked subgenera or sections. Erdtman, noting a variety of pollen types in Symplocos, suggests that pollen morphology “may be in- strumental in subdividing the genus (or —if considered appropriate — in referring the different species to a number of genera now usually regarded as sections etc. under Symplocos).” REFERENCES: Anonymous. Symplocos tinctoria. Natl. Hort. Mag. 15: 268, 271. 1936; 19: 132.1138, pe [Fruiting and flowering plants, respectively. | BENTHAM, G., and J. D. Hooker, Styracaceae. Gen. Pl. 2: 666-671. 1876. rsymbloces, ‘668,] BRAND, A. src easel Pflanzenr. IV. 242(Heft 6): 1-100. 1901. [281 species; 114 described as n Capor, L. reesei Untersuchung der Mateblatter unter Beriicksichtigung ihres Gehaltes an Thein. Bot. Centralbl. 84: 241-251, 275-283, 309-315, 340-345, 369-374. 1900. [Symplocos, 248, 249, 345, 369-371.] CHIRTOIU, M. Remarques sur le Symplocos Klotzsch et les affinités des Sym- plocacées. Bull. Soc. Bot. Genéve II. 10: 350-361. 1918. DECANDOLLE, A. Styracaceae. Prodr. 8: 244- 272. 1844. [Symplocos, 246-258. ] GAYTONDE, WS Ki: CHATTERJEE, and M. L. KHorana. Pharmacognostic studies of Symplocos racemosa, Roxb., and S. baddomei, C. B. Clarke. Indian Jour. Pharm. 12: 290-300. 1950. [Detailed information on mor- phology. GuUrke, M. Symplocaceae. Pflanzenfam. IV. 1: 165-172. 1890. Hose: Mazzettt, H., and E. Perer-STiLBaL. Eine Revision der chinesischen Arten der Gattung nee Jacq. Beih. Bot. Centralbl. 62: 1-42. 1943. [68 spp.; comments on relationships of S. tinctoria.] Harsison, T. G. Symplocos tinctoria Ashei, a new dyebush from the southern mountains. Jour. Elisha Mitchell Sci. Soc. 46: 218-220. 1931 KEARNEY, T. H. Report on a botanical survey of the Dismal Swamp region. ae Ae S. Natl. Herb. 5: 321-550. 1901. [Leaf anatomy of S. tinctoria, 503, 504.] MIERS, I ie the Symplocaceae. Jour. Linn. Soc. Bot. 17: 283-306. 1880. [Splits Symplocos into several genera. | Moutiscu, H. Vergleichende Anatomie des Holzes der Ebenaceen und ihrer Verwandten. Sitz-ber. Akad. Wien 80(1): 54-83. 1879. Mo.urarp, M. Notes de pathologie végétale. Revue Gén. Bot. 10: 87-101. 1898. [IV. Modifications anatomiques déterminées chez un Symplocos par VExobasidium Symploci Ellis, 96-101. pl. 15 SARGENT, C. S. Symplocos. Silva N. Am. 6: 13- 16. pls. 255, 256. 1894. TRAVERSE, A. Pollen analysis of the Brandon lignite of Vermont. U. 3. Dep. Interior Bur. Mines Rep. Invest. 5151: 1-107. 1955. [Symplocos pollen a Symplocos also noted abundant in German brown coal. TurriLL, W. B. Symplocos paniculata. Bot. Mag. 168: pl. 149. 1951. Ursan, I. Addimenta ad cognitionem florae Indiae occidentalis. Bot. Jahrb. 15: 286-361. pl. 9. 1892. [Symplocos, 328-337.] 1960] WOOD & CHANNELL, GENERA OF EBENALES 35 WEHNERT, A. Anatomisch- aa eri Untersuchung der Blatter der Gattung “Sym Pe cos.” Inaug. Diss. Univ. Miinchen. 58 pp. 1906. [69 spp tinctort WItson, E. TL “Symplocos paniculata. Gard. Chron. III. 74: 262, 263. 1923. 2) THE ARNOLD ARBORETUM AND VANDERBILT UNIVERSITY 36 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XLI A TAXONOMIC REVISION OF PODOCARPUS, XII. SECTION MICROCARPUS NetTta E. GRAY Popocarpus ustus (Vieillard) Brongniart and Gris, the only species in section Microcarpus, is a very curious, small, twiggy shrub found in sev- eral places in the mountain forests of New Caledonia. Vieillard and Deplanche, the first collectors, said that the local residents regarded this plant as sacred and attributed marvelous properties to it. The scale-like leaves, similar to some found in section Dacrycarpus, are arranged spirally. covering the surface of the twig and adnate to the stem, often with only a millimeter of free leaf blade. A figure showing a vegetative branch and seeds is shown by Pilger (4, 5) in both of his treatments of the genus. The species has been collected several times since it was described by Vieillard in 1862, but only the most recent collection, that of De Lauben- fels (2) in 1957, has shown it growing parasitically on Dacrydium taxoides Brongn. & Gris. The color of the plant is variously described as reddish, copperish, bronze, or purple. As Luc Chevalier (1) describes it most recently, ‘‘ses rameaux dressés ressemblent 4 une branch de corail rouge et les cOnes femelles au bout des rameaux sont d’un bleu roi légerement argenté.” Indeed, when I soaked some twigs in water in preparation for examination of the leaves, the water turned very reddish purple, the effect of known anthocyanins. In view of the possible absence of chlorophyll, a test was made by analyzing an alcoholic solution from the leaves in a Beck- man photospectrometer. The absorption curve showed the drop at 650 millimicrons indicative of the presence of chlorophyll @a.* These results indicate that the specimen examined most probably was not completely parasitic. It is interesting to note in this connection that the history of the collections shows, in almost every case, that Dacrydium taxoides, the host of this specimen, was collected at the same time as Podocarpus ustus. Since the knowledge of its semi-parasitic nature became available no further collections have been obtained, but one may reasonably suspect that the plant may prove to be a root-parasite on Dacrydium or other gymno- sperms. Podocarpus ustus (Vieillard) Brongniart & Gris, Bull. Soc. Bot. France 13: 426. 1866; Parlatore in DC. Prodr. 16: 521. 1868; Gordon, Pinetum ed. 2. 358. 1875; Warburg, Monsunia 1: 193. 1900; Pilger. Pflanzenr. IV. 5(Heft 18): 58. 1903, Nat. Pflanzenfam. ed. 2. 13: * The author wishes to express her great appreciation to Dr. Herbert Irvin, of the Crime Laboratory, Department of Public Safety of the State of Georgia, for his chloro- phyll determination on De Laubenfels P1065. 1960 | GRAY, REVISION OF PODOCARPUS, XI1 37 245. 1926; Guillaumin, Ann. Mus. Col. Marseille II. 9: 269. 1911, Bull. Mus. Hist. Nat. Paris 18: 100. 1912, Fl. N. Caledonia 11. 1948; Compton, Jour. Linn. Soc. Bot. 45: 425. 1922; White, Wilson & Guillaumin, Jour. Arnold Arb. 7: 77. 1926; Florin, Svenska Vet.-Akad. Handl. III. 10: 270. 1931; Dallimore & Jackson, Handb. Conif. 58. 1923, 1931, 84. 1948. Dacrydium ustum Vieillard, Ann. Sci. Nat. Bot. IV. 16: 56. 1862; Carriére, Conif. 697. 1867 A shrub, usually less than 1 meter tall, with dense, short, erect spreading branches, with the twigs entirely clothed by the copperish to purple or red, spirally placed, persistent adpressed scale leaves (Fic. 1), and with the terminal meristems protected by the youngest leaves. Leaves triangular, carinate, with broadly decurrent bases, 1-2 mm. long, 1—-1.5 mm. broad; midribs not evident. Leaves differentially amphistomatic, having few stomata on the adaxial surface. Leaves without hypodermis or palisade parenchyma. Male strobili terminal and solitary on axillary leaf-covered peduncles up to 6 mm. long, the strobili short-cylindrical, 5 mm. long. Microsporophylls imbricate, sessile, 1.5-2 mm. long, 2-locular, longitudinally dehiscing; apiculi flat, thin, and broadly triangular. Female strobili ter- minal; peduncles 2—5 mm. long, clothed in decussate scale leaves, the up- permost not decurrent, succeeded on the strobilus by about 5 spirally placed longer bracts increasing to 2.5 mm. with longer internodes and blades free from the axis; fertile bracts usually only one, sometimes two, terminal, mostly free from the ovules; no fleshy receptacle. Seed globose, purplish, to 2.5 mm., not crested but often drying so there is an apparent acute apex. DISTRIBUTION: In forests on mountain slopes, up to 1000 m. altitude, in New Caledonia. New Caledonia: Ignambi, ae 1545 (x *): Pic de Pouébo, Vieillard 1269 (Pp), Deplanche 170 (Pp); Mt. Penari, Balansa 3484, 3485 (Pp): oe, Balansa 184 (+P); Mt. Koghi, eae 456, 601 (P); Dumbea, Vieillard 1262 (Pr); slopes above Riviére Bleue, De ae P165 (+c); Forets bas du Pic, des Sources, LeRat 903 (+P); Montagnes de Poila, Vieillard 1267 (+p-Type; +A, K); Baie de Prony, Jeanneney (Pp). No specific locality: Balansa (BM, K, +NY); Pancher, 1879 (+Br, K); Hennecart (Br, +P, tUCLA). In contrast to the leaves of most other podocarps, which are noted for the variety of cell types in their tissues, there is little differentiation in the mesophyll of the leaves of Podocarpus ustus. The walls of the epidermal cells are mostly simple, and the often abundant stomata are mostly on the * The following Dene ee the location of the specimens cited: Arnold ris (P symbol preceding the abbreviation of an herbarium ee that the details of the mee of this specimen have been examined in transverse section 38 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI abaxial surface. The stomata are not arranged in rows, nor are they always oriented longitudinally (Fic. 2), merely being scattered over the surface. Only the smallest amount of vascular tissue connects with that of the stem and there is little vascular tissue in the free part of the blade of the leaf. Just before the bundle disappears the xylem tracheids shorten exceedingly and the vein broadens, with definite short wings of transfusion tissue on each side (Fic. 2), altogether giving a fan-like appearance to the termina- tion of the bundle. The resin canal may continue beyond the end of the bundle but it rarely reaches the tip of the leaf; three resin canals may be present (Fic. 4). The blade of the leaf is swollen or fleshy and packed with roundish cells without conspicuous air spaces between them (Fic. 3). There is no palisade parenchyma and only rarely are a few isolated sclereids of the transfusion type seen. POR y) OS Bae. !. QI e2eu, 5 ry GURES 1-7. Podocarpus ustus. 1, Part of stem, showing scale-leaves (Vieil- lard 1267), * 5. 2, Cleared leaf, showing vascular tissue and stomata in abaxial epidermis (De Laubenfels P165), X 16; S = stoma, X = xylem, TT = trans- fusion tissue. 3, Camera lucida outline of transverse section of leaf (from The solitary ovules are terminal on the twigs (Fics. 5-7). As in other podocarps, a single vascular bundle enters the ovuliferous scale but almost 1960 | GRAY, REVISION OF PODOCARPUS, XII 39 immediately divides, both branches arching over the back to the top of the ovule. One of the specimens of De Laubenfels P165 has pairs of ovules (Fic. 6) and close examination shows them to be separate on the upper- most bracts. The lack of hypoderm in the leaves and the lack of a fleshy receptacle below the ovule definitely separate this species (and thus section Micro- carpus) from section Dacrycarpus. If the scale leaves of Podocarpus ustus be considered merely evidence of modification toward the parasitic habit, then the species seems to be more closely related to section Stachycarpus where the other characters are typical. This position in the genus would be in better accord with the recent genetic studies in Podocarpus by Hair and Beuzenberg (3) who found that P. ustus has a diploid chromosome num- ber of 36. They found chromosome numbers of 36 and 38 in section Stachycarpus and a diploid number in section Dacrycarpus of only 20. LITERATURE CITED 1. CHEVALIER, L. Les coniféres aoe connus en Nouvelle Caledonie. Etudes iWelanésiennes. II. 1: 105-118. 1957 2. Dre LAUBENFELS, D. J. Parasitic conifer ae in New Caledonia. Science 130: 97. July ae 3) HAIR 2 B= and: E: ‘s. BEUZENBERG. Chromosomal evolution in the Podo- carpaceae. ’ Nature 181: 1584-1586. 4. Pricer, R. Taxaceae. Pflanzenreich IV. 5 (Heft 18): 1-124. 1903. 5 . Podocarpaceae. Nat. Pflanzenfamilien. ed. 2. 13: 211-249. 1926. AGNES Scott COLLEGE DECATUR, GEORGIA 40 JOURNAL OF THE ARNOLD ARBORETUM VO. Sit STUDIES IN THE GENUS COCCOLOBA, VIII. NOMENCLATURAL CHANGES RicHaArD A, HowarpbD SINCE THE PUBLICATION of the preceding papers in this series several matters involving nomenclatural changes have come to my attention. These changes concern Coccoloba novogranatensis, a species of the Lesser Antilles, Trinidad and northern South America, and C, nitida and C. guian- ensis, species of South America. In making the necessary corrections, two Vellozo names must be considered and the species concepts of the previous monographers Meisner and Lindau altered for additional taxa. Coccoloba novogranatensis Dr. Armando Dugand has called my attention to a paper of his which was published in Cal/dasia in 1947. I was not aware of this paper in which Coccoloba coronata Jacquin is properly identified and in which Coccoloba novogranatensis Lindau is placed in the synonymy of the Jacquin species. I do not agree completely with Dugand’s treatment, since the proper cita- tion for this species, with additional synonymy, appears to be the following: Coccoloba coronata Jacquin, Enum. Pl. Carib. 19. 1760, Select. Stirp. Amer, Hist. 114. ¢. 77. 1763; Dugand, Caldasia 4: 427. 1947, not Lindau, Symb. Ant. 1: 228, 229, 1899. Coccoloba virens Lindl. Bot. Reg. 21: ¢. 1816. 1835. Coccoloba novogranatensis Lindau, Bot. Jahrb. 13: 192. 1890; Howard, Jour. Arnold Arb. 40: 85-87, 208-9. 1959. Coccoloba dioica Karsten ex Lindau, Bot. Jahrb. 13: 170. 1890. Coccoloba caribaea Urban, Symb. Ant. 5: 337. 1907. Coccoloba waittii Johnston, Sargentia 8: 122. 1949. The type location for Jacquin’s species was cited as Cartagena. Dugand, who is familiar with this area and its vegetation, could not find constant or significant characters to separate Coccoloba coronata Jacq. from C. novo- granatensis Lindau and concluded that they were identical. I can agree with him on this. However, Dugand accepts the synonymy given by Lindau (i.e., that C. punctata L. and C. coronata Jacq. are the same). but I can not agree with this. Lindau’s treatments of C. excoriata, C. punctata, and C. venosa in his monograph of the genus (Bot. Jahrb. 13: 106-229. 1890) and in his later treatment of the West Indian species for Symbollae Antillanae (1: 229. 1890) are confused. This has been discussed previ- ously by Fawcett and Rendle (Jour. Bot. 51: 123. 1913) and by me (Jour. Arnold Arb. 30: 398. 1949). Coccoloba punctata L. (Sp. Pl. ed. 2. 1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VIII 41 523. 1762) is not the same as C. coronata Jacq., but is a synonym of C. venosa L. The name Coccoloba punctata L. does not belong among the syn- onyms given by Dugand. Coccoloba virens Lindl. was illustrated in the Botanical Register on the basis of a specimen which flowered in a British greenhouse in 1833. The place of origin for this plant is uncertain. Lindau first referred this species to the synonymy of his ““Coccoloba punctata” (Bot. Jahrb. 13: 160. 1890) and later to his ‘““Coccoloba coronata” (Symb. Ant. 1: 228. 1899). Cacco- loba virens Lindl. can not be the true Coccoloba venosa L. to which Lindau’s “C. coronata”’ and ‘“‘C. punctata” belong, but it is the present species Coccoloba coronata Jacq. Lindau described Coccoloba dioica, using a manuscript name appearing on an embossed label of a Karsten collection in the Leningrad herbarium. In spite of the specific epithet and the notes on the collection, Lindau failed to note the unisexual nature of the flowers in the published descrip- tion. The Karsten specimen represents a pistillate plant, as the flowers have rudimentary stamens with abortive anthers and no pollen. In a key to the species in his monograph, Lindau distinguished between his C. novo- granatensis and C. dioica by the fact that the former has a pubescent, and the latter a glabrous, lower leaf surface. The Karsten type, which I have been privileged to study, has leaves with pubescence along the midrib and on the primary veins. There is no doubt that the type specimens are com- parable and that C. dioica can be assigned to the synonymy of C. coronata. Coccoloba caribaea Urban and C. waittii Johnston have been placed in synonymy and discussed in earlier papers (Jour. Arnold Arb. 40: 86, 209. 1959). Coccoloba nitida A second correction involves the use of the names Coccoloba nitida, C. marginata, and C. guianensis by Meisner, Lindau and Howard, the three monographers of the genus. In an earlier study (Jour. Arnold Arb. 40: 83-85. 1959) I accepted the interpretation of C. nitida as published by Lindau and as indicated by his annotations on specimens. I had not seen the Humboldt type specimen of C. nitida, and, in fact, made an additional error in considering still another worker’s photograph and annotation of a Martius specimen as the type of the species. Although Lindau cited a Humboldt collection from Colombia in his treatment of the species, he did not state specifically in which herbarium this specimen was located. The collections of Coccoloba of the Berlin herbarium, including the Willde- now herbarium, which I have now seen, do not contain such material. Recently, through the kindness of the Director of the Muséum National d’Histoire Naturelle, Paris, I have examined a sheet labelled primarily “Coccoloba laurifolia Knth.,’ with a smaller annotation of “‘C. nitida.” This specimen was collected at S. Bartholome on the Rio Magdalena in May, 1805, and is numbered 1627. This must be the type of Coccoloba nitida H. B. K., for it agrees with the original description of the species. 42 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI However, it does not agree with Lindau’s description of Coccoloba nitida, nor with the specimen he has so annotated. Lindau apparently did not see the type of C. nitida. Even though he cited the collection, the only avail- able specimen of it does not bear his annotation. The Paris specimen men- tioned is a poor one, consisting of a single short stem with a terminal inflorescence and one and one-half detached leaves. The leaves are insect infested and are abnormal at the tips. The specimen, however, has distinc- tive characteristics in the spathe-like basal ocrea of the inflorescence, the distinctive ridging of the pubescent rachis, and the closely packed fascicles of pedicels and buds. It is clear now that Coccoloba nitida is properly typified by the Hum- boldt collection. As such, it is a distinctive species of Colombia, and the synonymy and citation of specimens given in my earlier papers, as well as those of Meisner in the Flora Brasiliensis and of Lindau in his monograph, are incorrect. To the synonymy of Coccoloba nitida H. B. K. must be added C. micro- neura Meisner. Through the courtesy of the Director of the Royal Botanic Gardens, Kew, I have been able to study the Purdie specimen which is the type of C. microneura. In the original description, Meisner attributes the specimen he saw to the Arnott herbarium, but no material of this species could be found in the Prodromus or Delessert herbaria in Geneva. Lindau, however, saw the specimen at Kew. The Purdie specimen, without number, was collected in Santa Marta, Colombia, and is a full and ample specimen in flower. The species is not known in fruit. While the leaves are somewhat smaller than the type of Coccoloba nitida in all reliable charac- ters, it is evident that C. microneura and C. nitida are conspecific. The correct nomenclature and the specimens examined follow. Coccoloba nitida H. B. K. Nov. Gen. Sp. 2: 176. 1818. Coccoloba microneura Meisner, DC. Prodr. 14: 163. 1856. Colombia. San Bartholome on Rio Magdalena, Humboldt 1627 (p-lectotype) ; Santa Marta, Purdie s.n. (K-type of C. microneura); Dept. El Magdalena, Chimi- chagua, Haught 2228 (F, M, NY). Recently I described Coccoloba darienensis (Jour. Arnold Arb. 40: 159. 1959) based on Allen 934, a fruiting specimen from Darien, Panama. There are minor differences in the number of veins and the general aspect of the leaf, but the close relationship of C. darienensis to C. nitida is obvious. It is possible that additional collections of either species may prove them to be the same and that the range of C. nitida should be extended to Panama. There is still the question of what is the remainder of the Coccoloba nitida concept of Meisner, Lindau and Howard when the Humboldt type is removed and the name applied to an endemic species of Colombia. Of the suggested synonyms of these authors, a Vellozo name is the oldest and most troublesome. 1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VIII 43 Polygonum arborescens Polygonum arborescens Vellozo was described in Florae Fluminensis 162. 1825, and illustrated in the Icones (4: ¢. 43. 1827). The location given as “Habitat silvis maritimis ad ripas fluvii vulgo dicti Taguahy, atque etiam fruticetis Parochiae Campo-grande” can not be located exactly on modern maps, but is presumed to be near the town of Santa Cruz, in the Federal District, or in the western half of the State of Rio, Brazil. Meisner (FI. Bras. 5(1): 38. 1855) refers the species with a question to the synonymy of Coccoloba nitida H. B. K. Meisner’s concept of C. nitida was the first broad one which ignored the Humboldt type and was one which I now consider to be incorrect. Lindau also maintained this erroneous concept and also considered Polygonum arborescens Vellozo as a possible synonym of C. nitida. Vellozo’s description is brief, but the illustration shows several significant characteristics of aid in identifying this plant. It is obviously a liana, with well-developed short-shoots and persisting, nearly foliaceous ocreae which split into lanceolate “‘stipules.”” The flowers are on short pedicels in the inflorescence but the pedicels elongate in fruiting condition. On these characteristics the plant is surely the same as Coccoloba crescentiifolia Chamisso, and a new combination, Coccoloba arborescens, must be made. There is a second Vellozo species and illustration which has been referred to C. crescentiifolia which must be considered. Polygonum frutescens Polygonum frutescens Vellozo (Flor. Flum. 162. 1825, Icones 4: ¢. 44. 1827) has had an unstable history. Vellozo’s desenpcion of this is short (“P. spica simplici, terminali”), but he reports that it occurs in locations similar to his Polygonum arborescens. I can not refer the illustration of P. frutescens with real conviction to any currently recognized species. I believe that it may represent a terminal scrambling shoot of Vellozo’s Polygonum arborescens. However, without knowledge of the ocreae (which can be interpreted in the drawing as deciduous or very short), the point of attachment of the petioles, and the curvature and pubescence of the blade, this can be only an attempt at an intelligent guess. Casaretto was the first to consider Polygonum frutescens Vellozo. He cited this name in synonymy when he described Coccoloba vellosiana (Novarum Stirpium Brasiliensium Decades 70. 1844.) Casaretto’s species, while honoring Vellozo, must be considered to be based on an unnumbered Riedel collection from the vicinity of Rio de Janeiro. Throughout the description Casaretto credits Riedel for the data supplied. The unnumbered collection to which Casaretto refers must be Riedel 674, for the descriptive data on the label “In sylvaticis maritimis inque collib. ciccis R. Janeiro Jul. 1832” compare favorably, although not verbatim, with those given by Casaretto, “Habitat in arenosis maritimis et collibus siccis circa Rio de Janeiro (Riedel).” 44 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI In 1855, Meisner (FI. Bras. 5(1): 36.) reduced Polygonum frutescens Vellozo to the synonymy of his new species Coccoloba gardneri. He also cited in synonymy C. vellosiana Casaretto, although with a question mark. Meisner cited only Gardner 600 from the Serra dos Orgaos, Prov. Rio de Janeiro. I have examined this collection, but fail to find any similarity with the illustration of Polygonum frutescens Vellozo, and therefore be- lieve that Coccoloba gardneri should be reéstablished as a valid species for the consideration of future collectors in the Rio area. The type of this species is Gardner 600 and the synonymy given by Meisner should be ex- cluded. In 1833, Chamisso described Coccoloba crescentiifolia (Linnaea 8: 134- 136.), basing the species on an unnumbered Sellow collection from “Brasilia.” Four specimens bearing labels inscribed ‘“Coccoloba crescentiae- folia N” are in the herbaria at Berlin, Leningrad and Munich. The speci- mens in the Munich herbarium bear a further note which indicates that the collection was made near “‘Praya grande, prov. Rio de Janeiro.” In his monograph (Bot. Jahrb. 13: 173. 1890) Lindau accepted C. crescentiifolia Cham., but placed in its synonymy C. vellosiana Casar. and C. gardneri Meisner, as well as C. fasciculata Weddell. He also listed in the synonymy, with a question mark, Polygonum frutescens Vellozo, as Meisner had done earlier. Lindau also included in his citation of specimens the collection Gardner 600. Thus, none of the monographers has been cer- tain of the interpretation of Polygonum frutescens Vellozo, but all agree on the possibility of its being the same as Coccoloba petro If fur- ther collections or study prove this to be true, the acceptable name will be Coccoloba arborescens (Vellozo) Howard, since no combination of the Vellozo names has been made previously. If further study should show Polygonum frutescens to be distinct from Coccoloba arborescens (C. cres- centiifolia Cham.), then the Vellozo name, being the oldest in use in the Rio area, must be adopted over some other currently accepted name. The correct nomenclature for the concept involving Coccoloba crescentiifolia Chamisso is the following: Coccoloba arborescens (Vellozo) Howard, comb. nov. Polygonum arborescens Vellozo, Flor. Flum. 162. 1825, Icones 4: t. 43. 1827 Coccoloba crescentiifolia Chamisso, Linnaea 8: 134. 1833, “crescentiaefolia.” Coccoloba vellosiana Casaretto, Nov. Stirp. Bras. Decad. 70. 1844, Coccoloba fasciculata Weddell, Ann. Sci. Nat. III. 13: 258. 1849. Coccoloba crescentiifolia var. obtusata Meisner, Fl. Bras. 5(1): 26. 1855. Coccoloba fasciculata Weddell, based on Blanchet 769, was distinguished from C. crescentiifolia on the basis of flat leaves which were less rigid and had shorter petioles. The specimens which I have seen of this collection are of short lateral branches, and there is no doubt that the collection should be identified as C. arborescens. Meisner’s taxon Coccoloba crescentiifolia var. obtusata was based on Claussen 2013. The characteristics Meisner indicated as distinguishing this variety are not consistent in the several examples of the collection 1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VIII 45 which I have seen. As Lindau has already indicated, the variety is not worthy of recognition. Returning now to Lindau’s concept of Coccoloba nitida, two synonyms are easily removed from further consideration. Coccoloba pendula Salz- mann first was published in synonymy by Meisner (FI. Bras. 5(1): 38 1855), and C. recurva Newman was placed in synonymy by Lindau (Bot. Jahrb. 13: 180. 1890). The specimens labelled “C. pendula Salzm.” at Leningrad and Paris are to be referred to C. marginata Benth., a name to be considered shortly. I have been unable to find material labeled ‘“‘C. re- curva Newman” at Geneva and am unable to associate this name. The reference “Coccoloba foliis cordato-oblongis Plum, Pl. am. 137 tab. 146 f. 1” referred with a question to Coccoloba nitida by Lindau. Urban (Repert. Spec. Nov. Beih. 5: 69-70. 1920) has referred the Plumier illus- tration and reference to the species now known as C. venosa L. I am not convinced that either author is correct. The origin of the material which Burmann illustrated is not certain. As Urban suggested that it may have come from Trinidad, I believe its correct identity is C. marginata The specimens Lindau cited in his monograph as Coccoloba nitida are to be distributed among C. arborescens, C. marginata and C. nitida. Al- though I have not seen the collections Lindau cited from Dutch and British Guiana, I suspect that these will prove to be C. marginata. Fried- richsthal 398, reported from Guatemala, was cited by Meisner in the orig- inal description of C. marti. The specimen was attributed to the Vienna herbarium and was reported to have come from “S. Thomas, Guatemala.” Lindau placed C. marti in the synonymy of his C. nitida and cited the Friedrichsthal collection. All of the material of Coccoloba in the Vienna herbarium was lost during World War II, so the identity of this collection may never be established. I have not seen specimens of C. arborescens, C. marginata, or C. nitida in the collections from Guatemala which I have already studied. Coccoloba guianensis In a review of Coccoloba in the Lesser Antilles, Trinidad and Tobago, I accepted the name C. nitida with six taxa as synonyms. The correct name for this species should be C. marginata Bentham, with the following synonymy: Coccoloba marginata Bentham, London Jour. Bot. Hooker 4: 626. 1845. Coccoloba guianensis Meisner, Linnaea 21: 264. 1848. Coccoloba marti Meisner, Fl. Bras. 5(1): 37. 1855. Coccoloba martii var. major Meisner, Fl. Bras. 5(1): 38. 1855. Coccoloba martt var. minor Meisner, loc. cit. Coccoloba nitida var. rotundata Meisner, loc. cit. Coccoloba nitida var. cordata Meisner, loc. cit. Coccoloba trinitatis Lindau, Bot. Jahrb. 13: 182. 1890. Coccoloba douradensis Glaziou, Bull. Soc. Bot. France. IV. 11( Mem. 3f): 571. 1911 (provisional name with mixed type). 46 JOURNAL OF THE ARNOLD ARBORETUM (VOL, LI Lindau recognized Coccoloba guianensis, C. marginata, and C. trinitatis as distinct species. He considered C. martit a synonym of his C. nitida. I have indicated in an earlier paper (Jour. Arnold Arb. 40: 84. 1959) the variations in habit and leaf shape found on one plant of C. marginata in Trinidad, and I remain convinced that C. marginata, C. guianensis, and C. trinitatis are one and the same species. Coccoloba guianensis is the well known and widely used name for this species. Coccoloba marginata, the oldest name for the complex, was considered distinct only on the basis of the slightly recurved leaf margin. The type of C. marginata is Schom- burgk’s second collection numbered 118 (216) from British Guiana. The holotype is in the herbarium at Kew and an isotype is in Paris. All of the specimens which I cited as C. nitida in an earlier paper must be renamed C. marginata Benth. Coccoloba martiit Meisner is typified by a Martius collection from the vicinity of Salgao, on the river San Francisco, in Minas Geraés, Brazil. The specimen illustrates the nonscrambling aspect of the species. Coccoloba douradensis Glaziou was used in a list of Glaziou collections with a question as to its status as a new species. Because only four words were used to describe the plant, the name should be considered a nomen nudum. The collection cited, Glaziou 21978, may prove to be a mixture. As has happened with other numbers of the Glaziou collection, the speci- mens and associated labels or numbers were mixed in the field or in dis- tribution to herbaria. Sheets labeled Coccoloba douradensis, Glaziou 21978, may be referred to Coccoloba densifrons or to C. marginata. 1960] WILSON, GENERA OF ARALES 47 THE GENERA OF THE ARALES IN THE SOUTHEASTERN UNITED STATES ! KENNETH A. WILSON THE ARALES (Spathiflorae), in the view of most authors, include only the families Araceae and Lemnaceae, the latter representing highly re- duced plants of aroid ancestry. A much broader interpretation of the order is the one of Wettstein, who included in his “Spadiciflorae” not only the Araceae and Lemnaceae, but also the Palmae and Cyclanthaceae. Other authors (e.g., K. Fritsch, Ber. Deutsch. Bot. Ges. 50a: 162-184. 1932) have interpreted this group as including also the families Pandanaceae, Sparganiaceae, and Typhaceae. Such an interpretation of the order is subject to considerable disagreement, and these families may indeed not be very closely related to each other. The Arales are here interpreted in the Englerian sense to include only the families Araceae and Lemnaceae, which are generally agreed to be closely allied. ARACEAE (Arum FamI ty) Tuberous or rhizomatous, terrestrial or aquatic (rarely floating) herbs with watery, milky or acrid sap. Leaves simple or compound, basal and solitary or clustered, or cauline and alternate. Inflorescence a spadix (an axis with small, sessile, tightly grouped flowers), subtended by an her- baceous spathe (or spathe absent). Flowers bisexual, with a perianth (in ours), or unisexual, without a perianth (the plants then monoecious or dioecious) ; perianth, when present, of 2-6 parts; stamens 1-6, free or united into a synandrium; gynoecium completely syncarpous, the ovary 1(—many)-locular, superior or inferior (i.e., embedded in the spadix); ovules l—many in each locule. Fruit usually a berry. Seeds with or without endosperm, the embryo large. (Including Pistiaceae.) A large family, primarily tropical or subtropical in its distribution, of over 100 genera grouped into eight subfamilies by Engler (18 tribes by * Prepared for a biologically oriented generic flora of the aie United States, a joint project of the Arnold Arboretum and the Gray Herbarium made possible through the support of George R. Cooley and the National he Foundation, and under the direction of Reed C. Rollins and Carroll E. Wood, Jr. The scheme follows that outlined at the beginning of the series (Jour. Arnold Arb. 39: 296-346. 1958). Other published portions of these studies will be found in Jour. Arnold Arb. 40: 94— 112, 161-171, 268-288, 369-384, 391-397, 413-419. 1959, and in the present issue. In connection with the treatments of the Araceae and the Lemnaceae, I am particularly indebted to Carroll E. Wood, Jr., for his valuable suggestions, advice, and information. The illustrations are the skillful work of Dorothy H. Marsh, and are based on mate- rial collected by Carroll E. Wood, Jr., and Richard J. Eaton. I wish also to express my indebtedness to R. B. Channell, George R. Cooley, Joseph Ewan, R. K. Godfrey, ock, and Edward G. Voss, who kindly assisted in supplying information, aided in the obtaining of ee and helped in various other ways 48 JOURNAL OF THE ARNOLD ARBORETUM [VODs SL Hutchinson) with about 1500 species. Eight genera, representing five sub- families, are native to the United States, while at least two exotic genera, Colocasia and Pinellia, are becoming established. All but three of these (Calla L., Lysichitum Schott, and Pinellia Tenore) are represented in our ea. The family may be recognized by the inflorescence, a spadix, which usu- ally is subtended or enveloped by a spathe. Great diversity exists with re- spect to the nature, position and distribution of the flowers on the spadix. the habitat, habit, leaf morphology, flower structure and other morpholog- ical characters. The subdivision of the family is based not only on floral but also on anatomical characters, such as the presence or absence of raphides and the occurrence of latex. Although there appears to be no doubt that the family represents a nat- ural group, much speculation exists with respect to its origin and phylo- genetic relationships. Engler considered the Araceae to be derived from the Palmae through the Cyclanthaceae. Hutchinson, on the other hand, thought the group to be monophyletic and derived from the stock of the tribe Aspidistreae of the Liliaceae. Other authors have considered it as haviag evolved from the ancestors of the Liliaceae through various phylo- genetic lines. The Araceae as a group are notable for the occurrence of the needle-like crystals of calcium oxalate (raphides) in ‘‘packets” which occur in small capsules in almost all tissues. These account for the immunity of most Araceae to herbivorous mammals and for the intense irritation or “burn- ing” sensation which is experienced when raw or improperly cooked corms of Colocasia, Xanthosoma, or Arisaema are eaten. In Dieffenbachia, the dumb-cane, widely grown as an ornamental foliage plant, the irritation is so severe as to produce temporary inability to speak. When the cells of the plant are broken in chewing, the capsules absorb water and the needles are discharged with sufficient force to penetrate the tissues of the mouth. The prolonged cooking which is necessary with the “edible” Araceae renders the capsules inactive. (Cf. Colocasia.) Many genera are popular in cultivation as ornamentals, some, such as Philodendron, Syngonium, Pothos, Scindapsus, and Caladium, for their decorative foliage, others, such as Zantedeschia and Anthurium, for their showy spathes. Some of these may possibly escape sparingly in subtropical areas. REFERENCES: Barrav, J. Les aracées a tubercules alimentaires des iles du Pacifique Sud. Jour. Agr. Trop. Bot. Appl. 4: 34-52. 1957. [Includes Colocasia, Xantho- soma, Cyrtosperma, Alocasia, Amorphophallus.] BirpsEy, M. R. The cultivated aroids. 140 pp. Gillick Press, Berkeley. 1951. CAMPBELL, D. H. Studies in the Araceae. Ann. Bot. 14: 1-25. 1900. [Ovule and embryo sac of Dieffenbachia, Aglaonema, Lysichitum, and Anthurium.] (See also, ibid. 17: 665-687. Curys_er, M. A. The development of the central cylinder of Araceae and Lilia- ceae. Bot. Gaz. 38: 161-184. 1904 1960 | WILSON, GENERA OF ARALES 49 Da.itzscH. M. Beitrage zur Kenntniss der eee der Aroideen. Bot. EERE 25: 153-156, 184-186, 217-219, 249-253, 280-285, 312-318. 343- 349. 18 [Includes Anthurium, Spathiphyllum, Rhaphidophora, Monstera, eee Scindapsus, Colocasia, and Acorus. Dav MANN, E. pega) ce in der Bliitenregion einiger Araceen zugleich n Hinweis auf die bargersche Methode. Planta 12: 38-48. 1930 [1931]. aie ae Arisaema consanguinenm, A. amurense, Aglaonema, Arum. Detay, C. Diagramme nucléaire des Aracées. Compt. Rend. Acad. Sci. Paris 222: 1512-1514. 1946 ENGLER, A., and K. Krause. Araceae. Pflanzenr. IV. 23A(Heft 74): 1-71. 1920. [Pars generalis et Index familiae generalis]; 23B(Heft 21): 1-330. 1905 [Pothoideae]; 23B(Heft 37): 1-160. 1908 [Pothoideae, Monsteroid- eae, Calloideae]; 23C(Heft 48): 1-130. 1911 [Lasioideae]; 23Da (Heft 55): 1-134. 1912 ip iilédedroidese “Phindeddeee. Homalomeninae, Schismatoglottidinae]; 23Db(Heft 60): 1-143. 1913 [Philodendroideae: Philodendrinae]; 23Dc(Heft 64): 1-78. 1915. [Philodendroideae: Anubiadeae, Aglaonemateae, Dieffenbachieae. Zantedeschieae. Typhonodor- eae, Peltandreae|]; 23E(Heft 71): 7 139. 1920 | Philodendroideae, Colo- casioideae|; 23F(Heft 73): 1-274. 1920 [| Aroideae, Pistioidea FERNALD. M. L.. and A. C. KINSEY. a. wild plants of eastern North America. 452 pp. Idlewild Press. Cornwall. N.Y. 1943. [Araceae, 111-122. Includes recipes for Arisaema, Peltandra, Calla, Symplocarpus, Orontium, and Acorus. | GaTin, C. L. Premiére contribution a | étude de l!embryon et de la germination des Aracées. Ann. Sci. Nat. Bot. X. 3: 145-180. 1921. [Pothos, Anthurium, Spathiphyllum, Aglaonema, Acorus, Nephthytis, Dieffenbachia, Zante- deschia, Caladium, Arwm. | Gow, J. E. Phylogeny of the Araceae. Proc. Iowa Acad. Sci. 20: 161-168. 1913. {Derived from sepaloid and entomophilous ancestral plants. Arales and Helobiales represent a common line of descent which has diverged in recent times. Studies in Araceae. Bot. Gaz. 46: 35-42. 1908. [Development of the embaon in Nephthytis, oa aaa Aglaonemza. | — . Observations on the hology of the aroids. /bid. 56: 127-142. 191 [Embryo of Aglaonema, cae Pilotenen Arum, Xanthosoma and others. | Harter, L. L. Storage-rots of economic aroids. Jour. Agr. Res. 6: 549+572. 1916.* Jones. G. E. Chromosome numbers and phylogenetic relationships in the Ara- ceae. Diss. Abs. 17: 2394. 1957.* MookerjeEa. A. Cytology of different species of aroids with a view to trace the basis of their evolution. Caryologia 7: 221-291. 1955. [Cytology supports Hutchinson's classification in most cases. | Pritzer, P. Chromosomenzahlen von Araceen. Chromosoma 8: 436-446. 1957. SHARMA. A. K.. and N. kK. Das. Study of karyotypes and their alterations in eae Agron. Lusit. 16: 23-48. 1954. [Colocasia, Alocasia, and Cala- dint W ee, . Mrs. H. J. Dracunculus vulgaris in Tennessee. Castanea 22: 139. Peers ane in an abandoned yard in Rugby. | a0 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI Key TO THE GENERA OF ARACEAE A. Perianth present; flowers (at least the basal) bisexual, frequently proter- ogynous. B. sl isieg lateral on a leaf-like axis; leaves sword-shaped; perianth ents Ge Pat BME. <2o54 ute bed k hes ¥en tens tang eee 1. Acori B. Spadix terminal on peduncle; leaves cordate, ovate, or oblong-elliptic. C. Spathe fleshy, enveloping the subglobose spadix: perianth segments 4: fruit embedded in swollen spadix (multiple ae genase ee ee pi deed oi eek Bad Oe eaMoe ahaa tena eee . Symplocar pus. C. Spathe bract-like, ensheathine base of peduncle F oblong spadix; perianth segments (in lower flowers) 6; fruit a ak not embedded i SAMIR. i ssf ans ge pemaker errr ee eeeees ae 3. Orontium. A. Perianth wanting; flowers unisexual. D. Plants rooted in soil; spadix free from spathe. E. Leaves simple; stamens united into a peltate synangium F. Leaves sagittate or hastate; spadix almost en covered by MOWEES.. cd vo 45-55 ton oe £6 Bee yee ee ea Rea ee 4, Peltandra. FE, Leaves ovate-cordate and peltate; apex of spadix naked or cov- ered with sterile flowers; plants spreading by runners. ......... 5 janis qe dei hp-e sd. Bue ep arya e eeteark ee eat . Colocasia. E. Leaves palmately or pedately divided; stamens 2—4-locular, subsessile. fe Ted Sb bs edoe Launatieaneue sage Sener nes Re ee tapes 6. Arisaema. by, Plants foutine Spadix-adnate to spathe...c..cccraeeiwns ooeees 7. Pistia Subfam. POTHOIDEAE Engl. 1, Acorus Linnaeus, Sp. Pl. 1: 324. 1753; Gen. PL ed. 5, 151. 1754. Plant herbaceous, with a thick, creeping rhizome. Leaves equitant, sword-shaped. Spadix borne laterally on an axis resembling a foliage leaf, spathe. absent (the prolongation of the fertile axis often interpreted as an open spathe). Flowers bisexual, 3-merous, greenish; perianth of 2 whorls of 3 concave tepals; stamens 6, filaments linear; gynoecium with a 2- or 3- loculed superior ovary, each locule with several orthotropous ovules apic- ally attached. Fruit a gelatinous, few-seeded berry. TypE spEcIES: Acorus Calamus L. (Acorus, an ancient Latin name for an aromatic plant.) — SWEET-FLAG. A genus of two or three species, as currently interpreted, from the north- ern temperate regions of the world. It is represented in our area by the widespread Acorus americanus Raf. which grows in low, wet areas, marshes and along the banks of quiet streams. Although generally known as 4. Calamus and considered to have been introduced into the United States from Europe, recent cytological evidence indicates that the American plant is most likely a native one. The American plants are diploid (2n = 24) and fertile, while the European plants, which were introduced from south- ern Asia, are triploid (27 = 36) and sterile. Whether or not the European sterile triploid has been introduced into North America still remains to be shown. The plants of “4A. Calamus” from Siberia and Asia are tetraploid (2n = 48) and fertile. Various characters have been found to be corre- 1960] WILSON, GENERA OF ARALES 51 lated with the degree of polyploidy in A. Calamus. These include the blade width/length ratio, the angle of the spadix insertion, the oil content of the rhizome (greater with polyploidy), water content, and also the amount of calcium oxalate (decreases in polyploids). It has been suggested that each polyploid group be considered a variety of A. Calamus, but perhaps each might better be regarded as separate species. The American diploid (A. americanus Raf., A. Calamus var. americanus (Raf.) Wulff) needs fur- ther study to establish the possible validity of separation of northern and southern plants as was done by Rafinesque. REFERENCES: BUELL, M. F. Acorus Calamus in America. Rhodora 37: 367-369. 1935. ———. Seed and seedling of Acorus Calamus. Bot. Gaz. 96: 758-765. 1935. . Embryogeny of Acorus Calamus. Bot. Gaz. 99: 556-568. 1938. GUENTHER, E. S. Oil of calamus. Chemurgic Digest 2: 138-140. 1943.* Harper, R. M. Is Acorus Calamus native in the United States? Torreya 36: 143-147. 1936. [Apparently not, but was introduced for medicinal pur- poses and other reasons. | Love, A., and D. Love. Drug content and polyploidy in Acorus. Proc. Genetics Soc. Canada 2: 14-17. 1957. [A general review of Wulff’s papers and other evidence. | . Biosystématique du genre Acorus. (Abs.) Ann. ACFAS 23: 100. 1955/ 56 [1957]. Micke, M. Uber den Bau und die ee der Friichte und iiber die Her- kunft von Acorus Calamus L. Bot . 66: 1-23. 1908. [Introduced into Europe in the middle of the 16th a PaRMELEE, J. A., and D. B. O. Savite. Uromyces rusts of Sparganium and Acorus and hige alternate hosts. (Abs.) Proc. Canad. Phytopath. Soc. 2115. 9S Se SCHANTZ, M. von. Uber das atherische Ol beim Kalmus, Acorus Calamus L. Acta Bot. Fenn. 59: 1-138, pls. 1-16. 1958. [Includes notes on anatomy. ] SMALL, J. Calamus. A pocket- lens study. Food 11: 195. 1942.* Wutrr, H. D. Uber die Ursache der Sterilitat des Kalmus (Acorus Calamus L.). Planta 31: 478-491. 1941, . Der Olgehalt verschiedenchromosomiger Rassen vom Kalmus (Acorus ee Zeitschr. Naturf. 1: 600-603. 1946. Olgehalt und Chromosomenzahl des SO gee Kalmus (Acorus Calamus L.). Arch. Pharm. 283: 155-161. . Zur Zytologie, geographischen Verbreitung und eee des Kal- mus. /bid. 287: 529-541. 1954. and B. Hoffmann. Kalziumoxalat-Gehalt und Soar bei Rosa und Acorus. Ber. Deutsch. Bot. Ges. 70: 383-388. 1957 Subfam. CALLOIDEAE Engl. 2. Symplocarpus Salisbury ex Nuttall, Gen. N. Am. Pl. 1: 105, 106. 1818, nom. cons.” Herbs with a stout, erect rhizome bearing numerous large, clustered leaves; entire plant with a strong odor, often compared with that of a * Conservation unnecessary; see Taxon 8: 230, 1959, 52 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XLI skunk. Leaves entire, ovate or cordate, conspicuously veined. Spathe thick, fleshy, purple spotted, ovoid, with a tapering, arched tip, the margin inrolled. Spadix ellipsoid or globose, completely covered by the crowded flowers. Flowers bisexual, proterogynous, maturing basipetally; perianth segments 4, fleshy; stamens 4, opposite the perianth segments, the filaments slender, flattened, the anthers 2-loculed, extrorse; gynoecium with a 1 (-sel- dom 2)-loculed ovary imbedded in the spadix, the solitary ovule pendent, orthotropous; style 4-angled, subulate. Fruits embedded in the spongy tissue of the globose spadix which is roughened by the persistent perianth parts and styles. Seeds naked, without endosperm or seedcoat, subspherical., depressed on the funicular side. (Spathyema Raf., nom. rejic.) Type species: S. foetidus (L.) Nutt. (Name from Greek, symploce, connection, and carpos, fruit, in allusion to the multiple fruit.) — SKUNK-CABBAGE. Symplocarpus foetidus, is known from two widely disjunct areas: eastern Asia (the Japanese plant has been named S. nipponicus Makino), and east- ern North America, where it ranges from Quebec and Nova Scotia, south to ve ae Oo " ee c Fic. 1. Symplocarpus. a-g, S. foetidus: a, habit, showing inflorescence, 3%; b, spadix, X 114; c, section through spadix, showing flowers in vertical sec- tion, X 4; d, fruiting spadix, X 1; e, section through fruiting spadix, showing seeds in vertical section, & 1; f, seed, K 114; g, seedling, X 2. 1960 | WILSON, GENERA OF ARALES 53 New England, North Carolina, Georgia, and Tennessee and west to Ontario, northern Michigan, Minnesota, and Iowa. (It has also been recorded from Florida and Missouri.) It occurs in swamps, wet meadows, and woods where it often forms dense stands. Symplocarpus is one of the earliest plants to flower in the spring. The inflorescence appears barely above the surface of the ground, long before the leaves appear. The flowers are all perfect and proterogynous and mature basipetally on the inflorescence. As the embryo develops it absorbs both endosperm and integuments, so that when mature it possesses no seed coat. Except in the younger stages, dur- ing which it has a monopodial branching system, the plant is sympodial, each branch producing two leaves and a terminal inflorescence, although many of the inflorescences do not mature (Shull). The roots are contractile and function in pulling the plant into the soil. The closest relative of Symplocarpus apparently is Lysichitum, a western American and Asiatic genus of two species, although Hutchinson, primarily on the basis of the more differentiated spathe of Symplocarpus, placed these two genera in different tribes. This view, however, does not seem to be supported by the morphology of the flowers, of the inflorescence, and of the vegetative body (including odor!), which are similar enough to indicate a close relationship between the two. REFERENCES: BONNER, W. D., and C. S. Yocum. Spectroscopic and enzymatic observations on the spadix of skunk cabbage. (Abs.) Pl. Physiol. 31 (Suppl.): xli. 1956.* BReEwsTER, W. Occurrence of the skunk cabbage in an unusual place. Rhodora 11: 63, 64. 1909. [Elevated, dry locality in Massachusetts. | Brown, W. L., Jr. Drosophilid and chloropid flies bred from skunk cabbage. Psyche Boston 63: 13. 1956. Duccar, B. M. Studies in the development of the pollen grain in Symplocarpus foetidus and Peltandra undulata. Bot. Gaz. 29: 81-98. 1900. Gow, J. E. Morphology of Spathyeme Voeuda: Bot. Gaz. 43: 131-136. 1907. Grout, A. J. Leaves of the skunk cabbage. Torreya 3: 6. 1903. [Largest leaf measured 26.5 & 19.5 inches. ] Hackett, D. P. Respiratory mechanism in the aroid spadix. Jour. Exp. Bot. 8: 157-171. 1957.* (See also, Yocum and Hackett. Pl. Physiol. 32: 186— 191. 1957; HacKetTT and Haas. Ibid. 33: 27-32. 58. ROSENDAHL, C. O. Preliminary note on the embryogeny of Symplocarpus foe- tidus Salisb. Science 23: 590. 1906.* . Embryo-sac ae and embryology of Symplocarpus foetidus. Minn. Bot. Stud. 4: 1- 909. SeximoTo, H. Account of i discovery of Symplocarpus nipponicus. (In Japa- nese.) Jour. Jap. Bot. 6: 319-328. 1929. [S. nipponicus Makino. SHULL, J. M. A Methuselah of the plant world. i skunk cabbage. Jour. Hered. 15: 443-450. 1924. [Vegetative morpholo . Spathyema foetida. Bot. Gaz. 79: 45-59. ion [ Vegetative morphol- ogy. ee P. D. An abnormal inflorescence of ee eee Bot. 4: able 329. 1927. [A spadix enveloped by two spat VAN cae AN, W. Cyanide-resistant respiration in skunk AGEN (Abs.) Pl: Physiol. cea. xxix, 1955,.* 54 JOURNAL OF THE ARNOLD ARBORETUM VOL 2e51 Wiuiams, K. A. A botanical study of skunk cabbage, Symplocarpus foetidus. Torreya 19: 21-29. 1919. 3. Orontium Linnaeus, Sp. Pl. 1: 324. 1753; Gen. Pl. ed. 5. 151. 1754. Perennial aquatic herbs with stout, deeply sunken rhizomes. Leaves simple, oblong-elliptic, prominently nerved and long petioled, usually float- ing. Spadix on an elongated scape, oblong, golden yellow, the scape green, with a band of white below the flowers.. Spathe obscure, forming a sheath at the base of the scape, the blade bract-like. Flowers variable: perfect at the base of spadix, with 6 tepals, 6 stamens and a 3-carpellate gynoecium, gradually becoming staminate towards the tip; tepals yellow, 6 in two whorls to 2 in a single whorl; stamens 6-1, 1 or 2 staminodia sometimes present; gynoecium 3- or 2-carpellate; ovary 1-locular with a single basal, anatropous ovule. Fruit a 1-seeded berry with thin pericarp. TYPE AND SOLE SPECIES: Orontium aquaticum L. (Orontium, an ancient name for a plant which is said to have grown in the Syrian river Orontes.) — GOLDEN CLUB. . ~ XSi ee Ia tye aa: RPS oO dP Fic, 2. Orontium. a-g, RX, lee, flowers, X 4; d, spadix, longitudinal section, showing flowers in lateral view and in vertical section, X 4; e, fruiting spadix, X 1%; f, seed with seed coat partially removed to show embryo, X 2; g, seedling, < 1. O. aguaticum: a, habit, X ¥%; b, spadi Orontium aquaticum L. is known only from the eastern United States, primarily on the Coastal Plain, where it occurs growing in the shallow water of pools or streams and on their moist banks or shores, from Florida, north to Massachusetts, and west to Louisiana and Kentucky. The flowers, which are crowded on the spadix, show a remarkable amount of variation, ranging from those with six perianth segments, six stamens and a 3-carpel- late gynoecium, to those with only two perianth segments, one stamen, and one staminodium and no pistil. Numerous intermediate or transitional 1960] WILSON, GENERA OF ARALES 55. conditions may also be found on the spadix. The morphological and bio- logical details of this plant merit further investigation. REFERENCES: CocKERELL, T. D. A. A miocene Orontium (Araceae). Torreya 26: 69. 1926. [Orontum ee spadix from Colorado. RAFFILL, C. P. ntium aquaticum. Gard. Chron. 119: 159. 1946. [Notes on vation. SCHAFFNER, J. H. The flowers of the golden-club. Am. Bot. 43: 99-103. 1937. [ Variation in flower structure. | Subfam. PHILODENDROIDEAE Engl. 4. Peltandra Rafinesque, Jour. Phys. Chem. 89: 103. 1819, nom. cons. Herbs with long, stout-petioled, clustered leaves. Leaf blades sagittate or hastate with 3 prominent palmate nerves. Scape about as long as the petioles. Spathe convolute throughout or dilated and spreading above. Spadix completely covered with flowers, or with the apex naked. Flowers unisexual, a perianth wanting; staminate flowers covering apical portion of spadix, sessile, consisting of 4-8 two-loculed anthers embedded in the mar- gin of a peltate shield-like scale; anthers dehiscing by terminal pores. Car- Fic. 3. Peltandra. a—j, P. virginica: a, habit, X \%; b, inflorescence, * 1%; c, spadix, basal portion, showing staminate and carpellate flowers, spathe re- moved, X 2; d, staminate flower, lateral view, 5; e, carpellate flower, lateral view, m 5; f, carpellate flower, vertical section, showing ae X 10; g, ovule, SOc hk ‘fruitin ng spadix enveloped by base of spathe, de 1, seed, a seed, apical view, seed coat partially removed to show embry 56 JOURNAL OF THE ARNOLD ARBORETUM [vob Sut pellate flowers covering basal portion of spadix, consisting of the 1-loculed gynoecium surrounded by 4 or 5 white, fleshy staminodia; ovules 1- few. basal, amphitropous. Fruits 1-3-seeded berries aggregated in a fleshy head inclosed by the leathery base of the spathe; scape recurved at maturity. (Houttinia Necker, nom. rejic.) Type species: P, undulata Rat. =F, virginica (L.) Schott & Endl. (The name from Greek. pelta, a small shield. and andros, of a man, in reference to the peltiform stamens.) — ARROW- ARUM. Three species, all of which occur in our area, of wet soils or shallow water along stream and pond borders, marshes, swamps, and wet woods in the eastern areas of North America. The best known and most widespread species is Peltandra virginica (2n = c. 44, 88), which is exceedingly vari- able in the shape of the blades, having seven named forms. The northern and southern extremes of this species also differ in the shape of the spathe and in the undulation of the margin of the spathe. Peltandra sagittifolia (Michx.) Morong (P. glauca (Ell.) Feay ex Wood), with an entirely white spathe and red fruits, is found from southeastern North Carolina to Flor- ida, entirely on the Coastal Plain. The third species, P. /uteospadix Fern.. has a white-bordered spathe and yellow spadix and occurs from southeast- ern Virginia, southward to Florida. It has been suggested that P. luteo- spadix may have arisen as a hybrid of P. virginica and P. sagittifolia but no investigations have been made either to confirm or refute this hypothesis. REFERENCES: BarKLey, F. A. Noteworthy plants of Texas. II. A new species of Peltandra. Madrofio 7: 131-133. 1944. [P. Tharpit.| BiakeE, S. F. The forms of Peltandra virginica. Rhodora 14: 102-106. 1912 Dvuccar, B. M. Studies in the development of the pollen grain in Symplocarpus foetidus and Peltandra undulata. Bot. Gaz. 29: 81-98. 1900. [P. virginica. | Epwarps, T. I. The germination and growth of Peltandra virginica in the ab- sence of oxygen. Bull. Torrey Bot. Club 60: 573-581. 1933. . Seed frequencies in Cytisus and Peltandra. Am. Nat. 68: 283-285. 1 : FERNALD. M. L. Are two species passing as Peltandra virginica? Rhodora 42: 430-432. 1940. [Does not answer the question; presents the problem. | A Virginian Peltandra. Rhodora 50: 56-59. 1948. [P. luteospadix. | Gomuens B. Life history of Peltandra virginica. Bot. Gaz. 102: 641-662. 1941. [A morphological study. | Hart, H. T. Delaved germination in seeds of Peltandra virginica and Celastrus scandens, Publ. Puget Sound Biol. Sta. 6: 255-261. 1928. [Seed dormancy. | Trpestrom, I. Notes on Peltandra. Rafinesque. Rhodora 12: 47-50. 1910. [P. virginica vars. heterophylla and angustifolia. | Subfam. COLOCASIOIDEAE Engl. 5. Colocasia Schott, Meletemata Bot. 18. 1832. Perennial herbs with starch-filled corms and tubers. Leaves ovate-cordate and peltate, dark velvety green above. Plants monoecious; inflorescence 1960] WILSON, GENERA OF ARALES ai stoutly peduncled, shorter than the petioles; spathe ovate-lanceolate, con- stricted below the middle; the spadix inclosed by the spathe, with pistillate flowers on the basal portion and staminate flowers above, the two groups separated by a group of sterile flowers, the terminal portion of the spadix naked (or covered by sterile flowers). pean wanting. Staminate flowers of several stamens united into a peltate body, the eee nai laterally adnate or partially free and pendent, dehiscing by an apical slit. Carpellate flowers consisting of the short-styled gynoecium, the ovary 1-locular, with numerous orthotropous ovules on the ovary wall. Fruit a berry. Typr SPECIES: Colocasia antiquorum (L.) Schott, C. esculenta (L.) Schott, or C. acris (R. Br.) Schott, a lectotype apparently not yet chosen. (An old Greek name derived from the Arabic colcas or culcas.) — DASHEEN, TARO. A genus of tropical Asiatic herbs, variously interpreted as consisting of seven or more species. Colocasia esculenta (L.) Schott is widely culti- vated throughout the tropics for its starchy, edible tuber. Numerous culti- vars occur, and several have been introduced into the United States, where the species has been grown in the lowlands of the Coastal Plain from South Carolina to eastern Texas. Colocasia esculenta var. aquatilis Hasskarl has escaped cultivation and is an aggressive weed which forms large clones spreading vegetatively by slender, rapidly growing stolons. This plant has been poorly collected in the southeastern United States, and herbarium ma- terial is rare. It has been reported to be spreading in southern Louisiana, and it is known also from several localities from central to southern Florida growing along streams, marshes, and roadsides. The leaves of naturalized plants observed in southern Florida have a purple centrum and a purple band at the summit of the petiole. No flowering material of this form has been collected in our area. Other clones may well occur, however, and a form lacking purple markings, locally naturalized at Tallahassee, Florida, flowers in September and October (Godfrey) .* Care must be taken to cook the starchy corm thoroughly before it is eaten. Small capsules containing calcium oxalate raphides are found in abundance in all tissues of the plant. If improperly cooked, the capsules discharge the raphides with such force that they penetrate the tissues of the mouth causing great irritation. Prolonged cooking renders the capsules inactive. *A plant grown from a corm collected in Dade County, Florida, in May, 1958, flowered in the greenhouse at the Arnold Arboretum on October 28, 1959. The firm, fleshy spathe was 28 cm. long and light orange in color, except for the green basal portion inclosing the carpellate region of the spadix. The erect, cream-colored staminate region and the sterile appendix of the spadix were exserted from the spathe voluted terminal 18 cm. Flowering lasted two days: the carpellate flowers were re- ceptive the first day and the pollen was shed on the second. The leaves of these plants bore sale markings when collected, but those produced in the greenhouse lacked this colorin 58 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI REFERENCES: Arcengaux, G. A new weed on the spread in Louisiana. Sugar Bull. 24) 3a, 36. 1945. | Colocasia esculenta var. aquatilis. | Hitz, A. F. The nomenclature of the taro and its varieties. Bot. Mus. Leafl. 7: 113-118. 1939. [Colocasia esculenta (L.) Schott, not C. antiquorum Schott. | Kumazawa, S., K. Nrucuti, and F. Honpa. Classification of the taro varieties in Japan. (In Japanese.) Jour. Hort. Assoc. Japan 25: 1-10. [O56 Mourst, M. A. A comparative physiological study of dasheen (Colocasia anti- quorum (Schott). Indian Jour. Hort. 13: 15-19. 1956.* Rao, N.S. A note on the chromosome number in Colocasia antiquorum Schott. Curr. Sci. Bangalore 16: 229. 1947.* [2n = 36.] Ratrensury, J. A. Chromosome numbers in New Zealand angiosperms. Trans. Roy. Soc. N. Zealand 84: 936-938. 1957. [Includes several varieties of Colocasia antiguorum from the Pacific area, 2n = 28, 42.] Sarrorp, W. E. The useful plants of Guam. Contr. U.S. Natl. Herb. 9: 1-416. 1905. [29-71, raphides of Colocasia and their method of discharge. | Sepcwick, T. F. The root rot of taro. Hawaii Agr. Exp. Sta. Bull. 2: 1-21. 1902* YarproucH, J. A. Stomatal count of a single leaf of Colocasia. Proc. lowa Acad. Sci. 41: 71-73. 1934 Younc, R. A. The dasheen; a Southern root crop for home use and market. U.S. Dep. Agr. Farmers’ Bull. 1396. 36 pp. 1924. (Revised 1946.) Subfam. AROIDEAE Engl. 6. Arisaema Martius, Flora 14: 459. 1831. Perennial herbs with globose or subglobose corms each bearing a simple scape sheathed by the leaf petioles. Leaves palmately or pedately divided. Spathe convolute below, spreading and arched above. Spadix obovoid and fertile at the base, above sterile (appendix) and cylindric, clavate or long attenuate. Plants dioecious or at times monoecious with both staminate and carpellate flowers in an inflorescence. Flowers without a perianth. Staminate flowers of numerous 2—4-locular, subsessile anthers opening by a slit or a pore. Carpellate flowers consisting of a gynoecium with a unl- locular ovary containing 5 or 6 erect, orthotropous ovules. Fruits 1-few- seeded red berries, few to many borne on the enlarged, ovoid base of the spadix. (Including Muricauda Small). TYPE SPECIES: A. nepenthoides Mart. (Name from Greek, avis, an ancient plant name, and Aaima, blood, in reference to the red-spotted leaves of some species.) — JACK-IN-THE- PULPIT, INDIAN TURNIP. A genus of over 100 species in 15 sections, of eastern Asia, Africa (Ethi- opia), North and Central America; represented in our area by three to five species in two sections. Section Torruosa Engl., with seven eastern Asiatic and one eastern American and Mexican species, is represented in our flora by A. Dracon- tium (L.) Schott (2n = 28, 56), green dragon or dragon root, widespread from Florida to Texas, north to southern Quebec and Ontario, and also 1960] WILSON, GENERA OF ARALES 59 with disjunct stations in eastern Mexico. The pedately divided leaves and the long-exserted, slender, tapering appendix are characteristic. Our other species belong to sect. Pistittata Engl., which includes about eight species of eastern Asia and three or four of eastern North America. The species of this section have a distinctly stipitate appendix which is more or less cylindrical and shorter than the spathe. Arisaema quinatum (Nutt.) Schott (2m = 28), of rich woods from Florida north to North Carolina and west to Louisiana, is a well-marked plant with a curved spadix and leaves generally with five leaflets. Avisaema Stewardsonii Britton (2m = 28), with the back of the spathe fluted, occurs primarily to the north of our area (Nova Scotia and southern Quebec, southward along the Appalachians at higher elevations in cold, wet woods) reaching its Fic. 4. Arisaema. a—j, A. atrorubens: a, habit, * ‘s; b, spathe, apical view, x a c, staminate inflorescence, spathe removed, <1; d, portion of staminate spadix, x 6; e, staminate flower, lateral view, < 8; f, carpellate inflorescence, spathe removed, < 1; g, carpellate flower, vertical section, See vules, X Sr Ny Ove Ges Oat: ate spadix, unusually large, X WA fruit, eee SHY, showing seed, & 2 k-m, A. Stewardsoniu: k, es lateral v ey, inflorescence, back view, X %; m, spa the, apical view, X %. eae habit ly, 60 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI southern limit in western North Carolina. Most of the taxonomic difficul- ties in the genus in our area center around the widespread A. atrorubens (Ait.) Blume (A. triphyllum sensu Huttleston) (22 = 28, 56), a plant primarily of woodlands in the northeastern United States, southward to South Carolina and Tennessee, and its close relatives. The status of this familiar and abundant plant and that of the smaller, later-flowering A. triphyllum (L.) Schott (A. triphyllum subsp. pusillum (Peck) Huttleston) (2n = 28) are still unsettled. Although the two are quite distinctive in some areas, in others forms of intermediate morphology occur; A. acumti- natum Small, of the southeastern Coastal Plain, is perhaps one of these latter. All of these species have been distinguished on morphological grounds and, for the most part, appear to be distinct ecologically. In instances in aries Arisaema atrorubens and A. Stewardsonii come in contact (to the north of our range), hybrid colonies may occur. Such local hybridization and intergradation has led to the interpretation of the latter as a subspecies of the former. However, with local exceptions, these plants appear to rep- resent distinct groups which maintain themselves as natural units. Further careful population studies are essential. Collectors should note conformation (especially fluting of 4. Stewardsonii), coloration and mark- ings of the spathe, width of flange at the upper margin of the spathe, and diameter of fruits. Careful notes on habitat, flowering time, and variability of colonies should be taken. Apart from the differing taxonomic interpretation, the nomenclatural application of the name A. ériphyllum L. has been a matter of controversy. It is used here in the sense of Fernald (1940) for the plant formerly known as A. pusillum Peck; almost all literature references to A. triphyllum prior to 1940 apply to the plant here called A. atrorubens, Numerous morphological studies have been made of A. atrorubens (mostly as A. triphyllum) and A. Dracontium. ‘The factors influencing the sex of the flowers in the inflorescence of A. atrorubens apparently are en- vironmental or nutritional. By manipulating the growth conditions it is possible to control the sex of the flowers that will be produced. Thus, in rich soil with an abundant water supply, carpellate flowers are produced, while staminate flowers are produced with the reduction of the root system and leaf surface of the plant. REFERENCES: ALEXANDER, E. J. Southern or notes. Castanea 5: 91-93. 1940. [A. Steward- nii in Buncombe Co., N. C. ATKINSON, G. F. St ess on aac in plants. Bot. Gaz. 28: 1-26. 1899. [Meiosis in “A. triphyllum”; n = 16.| Bowpven, W. M. Diploidy, solyploidy: and winter hardiness relationships in the flowering plants. Am. Jour. Bot. 27: 357-371. 1940. [Includes A. Dracon- tium, A. quinatum, A, atrorubens, A. concinnum. BruckMan, L. Fasciations in Arisaema, Rudbeckia and Viola. Torreya 6: 193— 195. 1906 1960] WILSON, GENERA OF ARALES 61 FERNALD, M. L. What is see triphyllum? = ie spermatophytes of eastern North America. Rho 42: 54. Gow, J. E. Embryogeny of Piers triphyllum. a ‘an 45: 38-44. 1908. a 2 re dimorphism in Arisaema triphyllum. Am. Midl. Nat. 8: 41- Hurt Leston, D. G. The three subspecies of Arisaema triphyllum. Bull. Torrey Bot. Club me 407-413, 1949. MacDoveat, D. T. Seedlings of Arisaema. Torreya 1: 2-5. 1901. [A. triphyl- lum and A. Dracontium. | MaHEsSHWARI, S. C., and P. P. KHANNA. The embryology of Arisaema walli- chianum Hook. f. and the systematic position of the Araceae. Phytomor- phology 6: 379-388. 1956. MEEHAN, T. Notes on Arisaema triphyllum. Bot. Gaz. 11: 217. 1886. Mortier, D. M. On the development of the embryo-sac of Ariscema triphyllum. Bot. Gaz. 17: 258-260. 1892 PickETT, F. L. Length of life of eal triphyllum corms. Proc. Indiana Acad. Sci. 1912: 77, 78. 1913. | About 4 years. . The development of the sae sac of Arisaema triphyllum. Bull. Torrey Bot. Club. 40: 229-235. 1913 . The germination of seeds of Apiene. me Indiana Acad. Sci. 1913: 125-128. 1914. [A. triphyllum and A. Dracontium. ] A contribution to our knowledge of ae triphyllum. Mem. Torrey Bot. Club 16: 1-55. 1915. [A morphological study. | Pryit, L. vAN DER. On the flower biology of some plants from Java with general remarks on fly-traps. [Includes species of Arisaema.| Ann. Bogor. 1: 77- 99. 1953 Putnam, B. L. Determination of sex in Arisaema triphyllum. Asa Gray Bull. 6: 50-52. 1898. RENNERT, R. J. Teratology of Arisaema. Bull. Torrey Bot. Club 28: 247-250. 1901. [A. triphyllum, A. Dracontium. | . Seeds and seedlings of Arisaema triphyllum and Arisaema Dracontium. Ibid. 29: 37-54. 1902. Rogrinson, B. L. A notable monstrosity of Arisaema triphyllum. Rhodora 18: 222, 223. 1916. SCHAFFNER, H. Control of the sexual state in Arisaema triphyllum and Araema Dracontium. Am. Jour. Bot. 9: 72-78. 20. The sexual nature of vegetative or dichotomous twins in Arisaema. Ohio Jour. ‘Sci 22: 149-154. 1922 . Siamese twins of Arisaema triphyllum of opposite sex experimentally in- duced. /bid. 26: 276-280. 1926 STonE, W. Arisaema pusillum in Pennsylvania and New Jersey. Torreya 3: 171, 172. 1903. Wuerry, E. T., and J. E. Benepicr. Plant finds in ee 1939. Arisaema Stew- ardsonit. Cues 4: 137. 1939. [Watauga Co., N. C.] Subfam. PISTIOIDEAE Engl. 7. Pistia Linnaeus, Sp. Pl. 2: 963. 1753; Gen. Pl. ed. 5. 411. 1754. Floating, aquatic, stoloniferous herb with a short stem bearing a rosette of leaves and numerous branching, adventitious roots. Leaves sessile, with simple, ovate to obovate-cuneate, densely pubescent blades, 3—15 cm. long, 62 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI each surrounded at its base by a membranous sheath. Spathe 2—4 cm. long, tubular at its base, but free and spreading above, slightly constricted above the middle. Plants monoecious, the spadix both staminate and carpellate; spadix fused to the median line of the spathe, free at its distal end. Flowers without a perianth, the carpellate solitary on the lower part of the spadix, the staminate whorled above; stamens 2-8, appearing to terminate the spadix. Staminate flowers subtended by a membranous collar surrounding the free portion of the spadix, this in turn subtended by a short-stalked, cordate, membranous flap. Gynoecium 1-carpellate with a 1-locular ovary containing numerous orthotropous ovules. Fruit green, ovoid to ellipsoid, many-seeded, crowned by the persistent style. Seed with a rugose, thick testa, a minute embryo, and abundant endosperm. Typr AND SOLE SPECIES: P. Stratiotes L. (Generic name from Greek, pistos, liquid, in reference to the aquatic habitat.) — WATER-LETTUCE. Fic. 5. Pistia. a-g, P. Stratiotes: a, habit, X ™%4; b, inflorescence, X 3; c, inflorescence, lateral view, portion of spathe removed, x 3: Dees ver- tical section, showing ovules, X 4; e, evuie, 3c 12 4 mature frit, < 3% £ 6. € ) A genus of a single species, Pistia Stratiotes L. (2 n = 28), found float- ing in ponds and streams of tropical and subtropical Africa, Asia, and Amer- ica. In our area it occurs in Florida and the Gulf States where it frequently forms dense mats on the water. Pistia is of considerable biological im- portance not only because of its weedy nature, which frequently results in the clogging of waterways, but also because it may represent the evolution- ary line through which the Lemnaceae developed. The embryo-sac develop- ment is of the Polygonum (normal) type At least in Florida, Pistia does not seem to produce seeds, although the ovary tends to enlarge and to become somewhat inflated. When the plant is pollinated artificially, seeds and fruit develop normally. Whether the failure of seed development is a result of the absence of suitable pollinating agents must still be established, but the evidence at present seems to favor such an interpretation. 1960 | WILSON, GENERA OF ARALES 63 REFERENCES: ARBER, A. On the vegetative morphology of Pistia and the Lemnaceae. Proc. Roy. Soc. London B. 91: 96-103. 1919 Brooks, J. S. The cytology and moroholony, of the Lemnaceae. Thesis (Ph.D.), Cornell University. 1940. [Includes a morphological study of Pistia. | Fu, C. 8. Effects of light, temperature, oxygen and storage upon the seed germi- nation of Pistia Stratiotes L. (In Chinese.) K’é Hsiieh T’ung Pao (Scientia) 1958: 377, 378. 1958. [Light required; best results with water-stored seeds in continuous light of high intensity. ] GopINATH, K. On the weed-destroying habit of Etroplus suratensis (Bloch.). Jour. Bombay Nat. Hist. Soc. 43: 664, 665. 1943. [A fish; destroys other plants but shows a marked aversion to Pistia. | Iro, T. Floating-apparatus of the leaves of Pistia Stratiotes L. Ann. Bot. 13: 466. 1899. JAEGER, P. Observations sur la fleur du Pistia Stratiotes L. Compt. Rend. Acad. Sci. Paris 220: 467-469. 1945. Kuortzscu, J. F. Uber Pistia. Abh. Akad. Wiss. Berlin 1852: 329-359. pls. 1-3. 1853. Kocu, K. Ueber Pistia in Allgemeinen und Pistia Turpini Blume insbesondere. Bot. Zeit. 10: 577-585. 1852. SCHLEIDEN, M. J. Bemerkungen iber die Species von Pistia. Allgem. Garten- zeit. 6: 17-20. 1838. [A systematic account of the genus listing nine species. | Uncer, F. Beitrage zur Physiologie der Pflanzen. I. Bestimmung der in den In- tercellulargangen der Pflanzen enthaltenen Luftmenge. Akad. Wien. Sitz- . 12: 367-378. 1854.* LEMNACEAE (Duckweep FaMILy) Small aquatic plants, floating on or below the surface of the water, con- sisting of a highly reduced, fleshy or membranaceous plant body (frond), lacking distinct stems and leaves, and either rootless or with one or many roots. Plants monoecious, the flowers borne in marginal pouches or in pouches on the upper surface of the frond. Staminate flower consisting of a single stamen, borne singly or in pairs; anther 2-locular. Carpellate flowers solitary, consisting of a single carpel; ovary 1-locular with 1 or 2 basal ovules. Fruit a 1- or 2-seeded utricle This family, generally considered to be composed of four genera and 26 species, is distributed in streams and other bodies of fresh water in most parts of the world. All four genera occur in our area and are represented by a total of 11 species which are of considerable economic and biological importance. Species of the Lemnaceae have been reported to serve as an important food for wild fowl, marsh birds, and aquatic animals, such as fish and snails. The small plant body of the Lemnaceae is oval or elongate, and bears at its basal margin one or two reproductive pouches where the buds (and in two genera the flowers) develop. Reproduction is primarily vegetative; a single plant may multiply to several million in one season. It has been 64 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI estimated that compact floating mats of Spirodela contain from 100,000 to 200,000 plants per square yard of water, and mats of Lemna 300,000 to 800,000 plants, while those of Wolffia may have 1,000,000 to 2,000,000 plants per square yard. It is generally accepted that the Lemnaceae represent a group derived from the Araceae and that the genera exhibit a reduction and simplifica- tion series from Spirodela to Lemna to Wolffiella and Wolffia. Lawalrée, however, considered the Lemnaceae to be related to the Helobiales with a close affinity to the Najadaceae. The frond of the Lemnaceae has been interpreted as a stem functioning as a leaf, as a modified leaf, and as a structure partly foliar and partly axial. Arber compared the structure of Pistia with that of Spirodela and consid- ered the lemaceous frond to be axial at the base and foliar at the distal region. She interpreted the distal part of the frond as a “petiolar phyllode,”’ and considered the pockets of Spirodela on the lower side as being formed by “wings of the leaf-sheath, terminating in two minute ligular flaps” and on the upper side by the axis. A somewhat different point of view was ex- pressed by Brooks who interpreted the plant body of Spirodela as a single sympodial branch bearing a prophyllum, a leaf, and a bract and terminating in an inflorescence, all somewhat reduced and fused to each other. In the fall, modified fronds or “‘bulblets” (turions) are produced which sink to the bottom of the water where they overwinter, rise again in the spring, and begin to grow. These bulblets may also be produced during other unfavorable growth periods. The flowers of Spirodela and Lemna are enclosed in a membranaceous spathe within the reproductive pouch. This spathe is absent in Wolffia and Wolffiella. The flowers of a single plant mature at different times. When two staminate flowers and a carpellate flower are produced by the plant the sequence of their maturation is basipetal: the distal stamen matures first, then the carpel, and finally the second stamen. This difference in the time of the maturation of the flowers doubtless promotes cross pollination. Ap- parently pollination may result from direct contact of the flowers as the plants float and also by the agency of insects and other small aquatic ani- mals which are associated with these plants. The pollen grains are spherical, monocolpate and generally have spiny projections on the exine. These grains show some characters in common with certain types of the Araceae, while they are less similar to those of the Najadaceae. The ovule is bitegmic and the inner integument forms the operculum, which persists, resembling a stopper at the micropylar end of the seed. The endosperm is cellular and forms a chalazal haustorial process, the “caecum.” The monocotyledonous embryo has a short suspensor and, in some genera, lacks a radicle. These genera have been the subject of numerous laboratory experiments dealing with the physiology of flowering and the effects of minerals, light, temperature and other factors on growth and development. Very few physiological races have been found to be present within the Lemnaceae. The small number of races has been interpreted to be a result of the mor- 1960] WILSON, GENERA OF ARALES 65 phological specialization and antiquity of the Lemnaceae, their extreme reduction, and their ability to reproduce vegetatively. REFERENCES: ArRBER, A. The vegetative morphology of Pistia and the Lemnaceae. Proc. Roy. Soc. London B. 91: 96-103. 1919 ANEANA, L. H. Las Lemnaceae y las larvas de los mosquitos. Anal. Inst. Biol. Mexico 1: 33-37. 1930.* BLACKBURN, K. B. Notes on the chromosomes of the duckweeds (Lemnaceae) introducing the question of chromosome size. Proc. Univ. Durham Phil. Soc. 9: 84-90. 1933. Brooks, J. S. The cytology and morphology of the Lemnaceae. Thesis (Ph.D.), Cornell University, 1 From, F., J. M. VipAt, andl J. COLEMAN. os action of herbicides on Lemna- ceae. Proc. Penn. Acad. Sci. 23: 85-90. 1949.* GoEBEL, K. Zur Organographie der Lemnaceen. Flora 114: 278-305. 192 HeEcELMAIER, F, Die Lemnaceen. Eine Monographische Untersuchung. ee zig, 1868. Systematische Ubersicht der Lemnaceen. Bot. Jahrb. 21: 268-305. Hicks, L. E. Flower production in the Lemnaceae. Ohio Jour. Sci. 32: 115-128. 1932. [A report on a series of experiments to induce flowering. | . Ranges of pH tolerance of the Lemnaceae. /bid. 32: 237-244. 1932. . The Lemnaceae of Indiana. Am. Midl. Nat. 18: 774-789. 1937. Horen, F. vAN. On the hibernation of Lemnaceae. Jour. Bot. 8: 36-40. 1870. LANDOLT, E. Physiologische und okologische Ree as an Lemnaceen. Ber. Schweiz. Bot. Ges. 67: 271-410. 1957 LAWALREE, A. La position systématique des Lee et leur classification. Bull. Soc. Bot. Belg. 77: 27-38. MCATEE, - Eleven important wild- duck foods. U. S. Dep. Agr. Bull. 205: —25. [ Lemnaceae, 3-5. McCann, é ae ee on Indian duckweeds, Lemnaceae. Jour. Bombay Nat. Hist. Soc. 43: 148-162. MaHESHWARI, S. C. Sy vee mesiLibn of the family Lemnaceae. Proc. IX Internatl. Bot. Congr. 2: 246, 247. 1959. [ Most closely related to Araceae. | Mason, H. L. A flora of the marshes of California. Univ. California. Berkeley. 1957. [Lemnaceae, 327-343. Includes detailed illustrations of the vege- tative body, flowers, fruits, and seeds of Lemna trisulca, L. gibba, L. minor, L. perpusilla, L. minima, Spirodela oligorrhiza, S. polyrhiza, Wolffia colum- biana, W. arrhiza, W. cylindrica, and Wolffiella lingulata. | SaecerR, A. Manganese and the growth of Lemnaceae. Am. Jour. Bot. 20: 234- 245. ScHLEIDEN, M. J. Prodromus monographiae Lemnacearum oder conspectus generum atque eas Linnaea 13: 385-392. 1839. [A systematic account. | Tuompson, C. H. A revision of the American Lemnaceae occuring north of Mexico. Ann. Rep. Missouri Bot. Gard. 9: 21-42. 1898. Key TO THE GENERA OF LEMNACEAE A. Plant with 2 marginal reproductive pouches; segments with 1—many roots; inflorescence of 2 ye e and 1 carpellate flower surrounded by a mem- branaceous spa 66 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI B. Fronds with 2 or more roots; conspicuously 5-18 nerved. . — B. Fronds bearing a single root; obscurely 1-3 nerved. ........ . Lemna A. Plant with a single marginal reproductive pouch; segments a. oii escence of 1 staminate and 1 carpellate flower without a spathe be co elongate, thin; stipe-attachment on one side of the reproductive ahs ee a oe W ol ffiella. ee Frond ellipsoid or ovoid, thick; stipe-attachment within reproductive DOUG (Hol BVIGOUE).- 2ic.esdsccrcicneetpicekagheavrnewe 4 1. Spirodela Schleiden, Linnaea 13: 391. 1839. Plants solitary or in groups of 2-5. Frond round or obovate, thick, purple beneath and green above with S—18 palmately arranged nerves radiating from the node (centrum). Roots 2-16 to each frond. Each frond with 2 marginal reproductive pouches at the basal region. Inflorescence of 1 carpellate and 2 (rarely 3) staminate flowers, inclosed within a membra- naceous spathe. Stamens 2-locular, dehiscine longitudinally. Gynoecium with a 1-locular ovary and 1 campylotropous or 2 anatropous ovules. Utricle with winged margins. TypE species: S. polyrhiza (L.) Schleid. (Name from Greek, speira, a cord, and delos, evident, alluding to the con- spicuous roots. ) A genus of three species, represented throughout our area by S. polyrhiza (2n = 40) in aquatic habitats. While Spirodela is one of the most abundant and widely distributed of the Lemnaceae it seldom occurs in pure stands but is usually found growing in association with species of Lemna and Wolffia. Spirodela oligorrhiza (Kurtz) Hegelm. has been reported from two collections in Missouri and may possibly be found to occur in our area. It has oblong or narrowly obovate fronds with only four to six nerves and two or three (rarely more) roots. Spirodela, the least reduced of the Lemnaceae, frequently has been com- pared to Pistia, of the Araceae. Jacobs, in a morphological study of Spirodela polyrhiza, interpreted the lower flap of the reproductive pcuch as “bracts which were derived by phylogenetic reduction from leaves.” The plant was considered to be a “greatly reduced type of rosette in which the vertical axis has been re- duced to three highly compacted nodes.” Brooks has shown that the flowers of the plant are always produced in the larger, usually left-hand reproduc- tive pouch, while the smaller, right-hand reproductive pouch produces only vegetative buds. This led him to interpret the right-hand pouch as being formed by the prophyllum and the basal part of the foliage leaf, and the left-hand, flowering pouch by the bract and foliage leaf. Occasional plants may be found, however, in which this arrangement is reversed. The embryo sac was described by Brooks as being of the “Polygonum” (normal) type. This is the same type found in Pistia but differs from the other known types in the Lemnaceae, all of which are bisporic and of the “Allium” type. However, in contrast to Pistia, which forms four mega- spores, three of which disintegrate, Spirodela forms two megaspores and one dyad cell; the dyad and one megaspore degenerate. The mature ovule is 1960 | WILSON, GENERA OF ARALES 67 campylotropous when only one is present in the ovary; when two are pres- ent they are almost completely anatropous. In addition to its possession of numerous roots, Spirodela differs from the other genera of the Lemnaceae by the presence of vascular tissue in the roots, the formation of two ovules in the ovary (ZL. gibba L. is described as having as many as 7 ovules), and by the “Polygonum” type of development of the embryo sac. Brooks reports the presence of a root in the embryo of Spirodela, but this needs confirmation, since Lemna has recently been shown not to form one during its development within the seed. REFERENCES: Bitcover, E. H., and D. H. Siettnc. Effect of various factors on the utilization of nitrogen ae iron by Spirodela polyrhiza (L.) Schleid. Pl. Physiol. 26: 290-303. Sine Criark, N. A., and E. E. FrauM. Influence of auxins on reproduction of Lemna major. Ibid. 15; 735- Oe 1940.* [S. polyrhiza. | and LLER. The stimulation of Lemna major by organic matter meee gists and nonsterile conditions. Soil Sci. 31: 299-308. 1931.* [S. polyrhiza. | GEISSMAN, T. A., and L. Jurp. The aubaoey abi of Spirodela oligorrhiza. Arch. Biochent Biophys. 56: 259-263. 1955. Goruam, P. R. Growth factor studies | Spirodela polyrrhiza (L.) Schleid. Am. Jour. Bot. 32: 496-505. : HeceL_mater, F. Uber die Fructifikationstheile von Spirodela. Bot. Zeit. 29: 621-629, 645-666. 1871. Jacogs, D. L. An ecological life-history of Spirodela polyrhiza (Greater Duck- weed) with emphasis on the turion phase. Ecol. Monogr. 17: 437-469. LO KANESEKI, S. The influence of “ukikusa” (Spirodela polyrrhiza Schleid.) on the micrometeorology in the bare water field. (In Japanese.) Soc, Agr. Met. Japan J. Agr. Met. 11: 37-38. 1955.* ManHEsHwakrt, S. C. Spirodela polyrrhiza: the link between the aroids and the duckweeds. Nature 181: 1745, 1746. 1958.* SAEGER, A. Spirodela oligorrhiza collected in Missouri. Bull. Torrey Bot. Club 61: 233-236. 1934. SovEcEs, R. Embryogénie des Lemnacées. Développement de l’embryon chez le Lemna minor L. Compt. Rend. Acad. Sci. Paris 248: 1896-1900. 1959. STERKI, V. Winter-buds of Spirodela polyrhiza (L.). Ohio Nat. 10: 181, 182. 1910 2. Lemna Linnaeus, Sp. Pl. 2: 970. 1753; Gen. Pl. ed. 5. 417. 1754. Fronds solitary or in groups of 2—5, each frond with 1-3 nerves and a single root without vascular tissue. Marginal reproductive pouches 2, of unequal size, at the basal region of each frond, the flowers, when produced, always appearing in the smaller pouch, the vegetative buds in both. In- florescence of 1 carpellate and 2 staminate flowers surrounded by a mem- branaceous spathe. Anther 2-loculed, dehiscing transversely. Gynoecium with a 1-loculed ovary and 1 campylotropous ovule. Utricle ovoid and smooth. Type species: L. minor L. (Name from Greek, lemna, an ee plant, possibly Callitriche, mentioned by Theophrastus.) — Duck EED. 68 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI A genus of about ten species, five of which occur within our area. By far the most common is Lemna minor L. (2n = 40, 42), which frequently forms dense covers over quiet waters. Lemna trisulca L. (2n = 44), al- though less conspicuous because it floats beneath the surface of the water, occurs almost as frequently as L. minor. Three other species known from our area are L. valdiviana Phil., L. minima Phil., and L. perpusilla Torr. Several of the species of Lemna have been used extensively in physiolog- ical studies, although L. minor and L. gibba L. have been the subject of the greater number of the papers on the subject. (Reports of experiments on L. major Griff. refer to Spirodela polyrhiza.) Recently Kandeler has been able to induce flowering in Lemna gibba (2n = 64): under controlled conditions it has a day-length requirement of 12-14 hours. (See also Hill- man, 1959). Lemna usually is described as having one to seven ovules in the ovary: Hegelmaier described L. gibba as possessing one to six. It appears that our species have only a single ovule per carpel, but a careful investigation of the various species should be made to determine this accurately. During the development of the embryo sac most of the nucellus disap- pears, but a few cells persist as the nucellar or micropylar cap. The endo- sperm is cellular and has a chalazal haustorial outgrowth or “caecum.” According to Maheshwari, no radicle is formed in the embryo. However. both Lawalrée and Brooks describe the development of a root in the embryo. Brooks has also shown that the embryo of Lemna has only one pouch, the vegetative pouch. These embryological characters of Lemna seem to agree much more closely with those of Araceae than with those of the Helobiales, which differ in the nature of their endosperm, the absence of a caecum, and in the absence of a nucellar or micropylar cap. REFERENCES: Asupy, E. The interaction of factors in the growth of Lemna. III. The interre- lationship of duration and intensity of light. Ann. Bot. 43: 333-354. 1929, (See also, I. 42: 771-782. 1928; II. 43: 329-332; IV. 805-831. 1929.) and E. WANGERMANN. The effects of meristem ageing on the morphology and behavior of fronds in Lemna minor. Ann. N. Y. Acad. Sci. 57: 476— 483. 1954. and others. Studies in the morphogenesis of leaves. III. New Phytol. 48: 374-381. 1949; VII, Parts I & I. 50: 186-199, 200-209. 1951: VIII. 51: 355-358, 1952; IX. 52: 298-311. 1953; X. 54: 182-198. 1955. | Ex- perimental studies on L. minor. ] Beatson, M. E. Sub-fossil pollen of Lemna in Quaternary deposits. Data for the study of post-glacial history. XV. New Phytol. 54: 208. 1955. BIERHUIZEN, J. F. Observations on Beas oe in Lemna minor L. Meded. Landb. Wageningen 54: 311- Me 1954 BLACKMAN, G. E., and R. C. ieee N-CU HAME. The influence of pH on the phytotoxicity of 2:4- sdidilevosheneg eet acid to Lemna minor. New Phytol. 52: 71-75. 1953. Biopcett, F. H. Homologies of the frond in Lemna. Science 39: 291, 292. 1914 1960] WILSON, GENERA OF ARALES 69 . Development of the embryo and the germination in Lemna perpusilla. Ibid, 292. . Morphology of the Lemna frond. Bot. Gaz. 60: 383-390. 1915. ———. The embryo of Lemna. Am. Jour. Bot. 10: 336-342. 1923. BotToMLey, W. B. Some effects of organic growth-promoting substances (auxi- mones) on the growth of Lemna minor in mineral culture solutions. Proc. Roy. Soc. London B. 89: 481-507. 1917 BroncniartT, A. Note sur la structure du fruit des Lemna. Arch. Bot. 2: 97- 104. 1833. ae account of the structure of the seed and fruit in Lemna minor and L. CALDWELL, O. W. On the life-history of Lemna minor. Bot. Gaz. 27: 37-66. 1899. [Special attention is paid to the gametophytes and fertilization. | Duraitty, G. Sur la nature réelle de la “fronde” et du “‘cotylédon” des Lemna. ee Mens. Soc. Linn, Paris 1: 147-149. 1879. [Frond of ee regarded as “un sympode d’embryons disposés a la suite les uns des autres.” | Erickson, L. C., and R. T. WEppINGc. Effects of ozonated pos on photosyn- thesis and respiration of Lemna minor. Am. Jour. Bot. 43: 32-36. 1956. FERNALD, M. L. Lemna cyclostasa an invalid name. Rhodora 37: 75, 76. 1935 GorHaM, P. R. Heterotrophic nutrition of seed plants Hea particular reference to Lemna minor L. Canad. Jour. Res. C. 28: 356-38 950. . Measurement of the response of Lemna to pie promoting sub- stances. Am. Jour. Bot. 28: 98-101. 1941. Guppy, H. B. On the habits of Lemna minor, L. gibba, and L. polyrrhiza. Jour. Linn. Soc. Bot. 30: 323-330. 1894. Hittman, W. S. Experimental control of flowering in Lemna. I. General meth- ods. Photoperiodism in L. perpusilla 6746. Am. Jour. Bot. 46: 466-473. 1959. II. Some effects of medium composition, aoe agents and high temperatures on flowering in L. perpusilla 6746. Ibid. —495. Hopkins, E. F. Manganese and the growth of Lemna minor. Science 74: 551, 552, LOS: Jounson, E. L. Effect of X-radiation upon the growth of Lemna minor. Univ. Colo. Stud. D. 1: 165-175. 1941. KANDELER, R. Uber die Bliitenbildung bei Lemna gibba L. I. Kulturbedingungen und Tageslangenabhangigkeit. Zeitschr. Bot. 43: 61-71. 1955. LAWALREE, A. L’embryologie a Lemnaceae: observations sur Lemna minor L. Cellule 54: 305-326. LEELADRISHNAN, G. A eee of diatoms in the leaf of Spirodela polyrrhiza Schleid. Curr. Sci. 28: 116. 1959.* ee W. The influence of phosphate on the photosynthesis oF Lemna ninor L. Proc. Nederland. Akad. Vetensch. C. 54: 287-295. 1951 ae F. Uber die Eee a ier emcee einiger ipa eee und ihre Anpassungen an da as ts r und gewisse wasserbewohnende Insekten. Kosmos Stuttgart 10: aL * ManesHwarI, S. C. The ie and embryo of Lemna and systematic posi- tion of the Lemnaceae. Phytomorphology 6: 51-55. 1956 Prrson, A., and E. GOLLNER. Beobachtungen zur Beene sony eolonic der Lemna minor L. Flora 140: 485-498. 1953 ScoTLanp, M. B. The animals of the Lemna association. Ecology 15: 290-294. 1934. SAEGER, A. The growth of duckweeds in mineral nutrient solutions with and without organic extracts. Jour. Gen. Physiol. 7: 517-526. 1925. 70 JOURNAL OF THE ARNOLD ARBORETUM (YO, EI STEINBERG, R. A. Mineral requirements of Lemna minor. Pl. Physiol. 21: 42- . 1946. VintEJoux, C. Recherches sur la racine de Lemna minor L. (Lemnacées), Ann. Sci. Nat. Bot. XI. 19: 211-261. 1958. Witson, W. Lemna gibba. Remarks on the structure and germination. Hook. Bot. Misc. 1: 145-149. 1830. [Description of seedlings. ] ZOLLINGER, J. Lemna minor — die kleine Wasserlinse. Leben Umwelt 12: 193- 197. 1956.* [A morphological study. | ZurzyYCKA, A. The influence of the wave length of light on the mov ements of chloroplasts in Lemna trisulca L. Acta Soc. Bot. Polon. 21: 17-37. ZURZYCKI, J. Formative effects of various spectral regions of light on oer trisulca L. Meded. Landb. Wageningen 57(11): 1-14. 1957. 3. Wolffella Hegelmaier, Bot. Jahrb. 21: 303. 1895. Fronds thin, linear-attenuate or curved, solitary or in groups of several cohering generations, forming a star-shaped colony, rootless, each with a single, triangular vegetative pouch at the base from which arise the daugh- ter fronds; each frond held to its parents by means of a “stipe.” Flower- ing reproductive pouch developing on the upper surface of the frond to one side of the stipe, thus resulting in a tendency toward asymmetry of the frond. Vegetative fronds floating just beneath the surface of the water. but in flowering the inflorescence-bearing portion of the frond breaking the surface. Inflorescence proterogynous and consisting of a single stamen with a 2-loculed anther and a single carpel with one orthotropous ovule, without a spathe. Utricle spherical. Type sprctes: W. oblonga (Phil.) Hegelm. (Name a diminutive of Wolffa.) A genus of about five species, primarily of the New World. It is repre- sented in our area by Wolffiella floridana (J. D. Sm.) Thompson, which is known in the area from Michigan and Ontario south to Louisiana, Texas. Florida, and Mexico, and by W. lingulata (Hegelm.) Hegelm., which has been reported from Louisiana. The flowers of Wolffiella were first found and described in W. oblonga (Phil.) Hegelm., from Argentina, in 1935. In 1938, flowering plants of W. lingulata were discovered in California, and, in 1948, a report of the discovery of the flowers of W. floridana was published. The flowers of the three species are remarkably similar, and their resemblance to the flowers of Wolffia is also striking. The flowering plants of Wolffiella floridana are slightly shorter, but wider and thicker at the base, than the vegetative plants. Furthermore, individuals are separated from the colonies during flowering time. Stomata are found only on the emergent portion of the frond during flowering. The carpellate flower of both Wolfia and Wolffiella matures before the staminate flower. The position of the inflorescence and the structure of the flowers of the two genera are alike in all essential features. Thus, the primary basis for maintaining Wolffiella as distinct from Wolffia is the position of the vegetative shoot axis with respect to the vegetative pouch. In Wolffella the vegetative pouch is formed on one side of the stipe which 1960] WILSON, GENERA OF ARALES 71 attaches it to the parent plant, while in Wolffia the stipe and the bud de- velop on the lower side of the pouch near the apex of the cavity. Wolffia tends to be bilaterally symmetrical, while Wolffiella tends to be asymmet- rical. This vegetative difference is presumably a very fundamental one in the morphology of the two genera. REFERENCES: BLAKE, 5S. F. eee lingulata in Louisiana. Rhodora 35: 226. GIARDELLI, M. s flores de Wolffiella oblonga. Revista Argent. Agron. 2 17-20. oe Hartman, R. T., and S. M. EnciisH. Wolffiella floridana in western Pennsyl- vania. Castanea 24: 45-47. 1959. [Includes drawing showing origin of new fronds. | Kurz, H., and D. Crowson. The flowers of Wolfiella ie GeDSm.) Thompson. Quart. Jour. Fla. Acad. Sci. 11: 87-98. 1948. Mason, H. L. The flowering of Wolffiella lingulata (Hegelm.) Hegelm. Madrono 4: 241- Zoe Pet Tuompson, C. H. The ligulate Wolffias of the United States. Ann. Rep. Mo. Bot. Gard, 7: 101-111. pls. 64-66. 1896 4. Wolffia Horkel ex Schleiden, Linnaea 13: 389. 1839. Fronds solitary or in pairs, globose or ellipsoid, without roots or veins. Frond with a single funnel-shaped, basal vegetative pouch from within which arise the young fronds, which soon become detached. Inflorescence of a single stamen with a 2- loculed anther and a single carpel containing 1| orthotropous ovule. Spathe absent. Utricle spherical and smooth. (Hor- kelia Reichb., 1830, not Cham. & Schlect., 1827; Bruniera Franch.; Bru- neria, Small.) Type species: W. Delilii Schleid., not Kurtz. = W. hyalina (Delile) Hegel. (Name in honor of J. F. Wolff, 1788-1806, who wrote on Lemna in 1801.) — WATER-MEAL. A genus of about 15 species, primarily of the tropics and subtropics, rep- fecentcds in our area by W. columbiana Karst., W. punctata Griseb. (2n = ca. 40), and W. papulifera C. H. Thompson. These are the smallest and simplest (by reduction) of the flowering plants, resembling small dots or grains floating in the water. Because of their minute size and the rarity of flowers these plants have been a subject of considerable botanical contro- versy. Dried, pressed plants-are usually so shriveled and distorted that their natural shape is impossible to determine, with consequent taxonomic diffi- culties. The dots which are described for W. punctata, for instance, are apparent in dried material, but not in living plants. It has been recom- mended that plants of Wolffia be preserved in formalin-acetic alcohol and placed in flattened glass ampules which can be mounted on herbarium sheets. Recent studies of the flowers and embryo development indicate that the anther is two- rather than one-loculed as has been previously reported. In Wolffia the carpel matures before the stamen, and pollination probably is effected through the agency of wind or rain. 72 JOURNAL OF THE ARNOLD ARBORETUM VOL, XLI REFERENCES: Camp, W.H. Distribution and flowering in Wolffia papulifera. Ohio Jour. Sci. oot 1e.. 1933, Dore, W. G. Wolffia in Canada. Canad. Field-Nat. 71: 10-16. 1957. | Includes chromosome number of W’. punctata. comparis on of W. columbiana, W. punctata, and notes on preparation of specimens Gupta, B. L. Studies in the development of the pollen grain and embryo sac of W olfiia arrhiga, Curr. Sci. Bangalore 4: 104, 105. 1935.* Jacogs, D. L. Wolffia papulifera in Florida. Am. Midl. Nat. 42: 110, 111. 1949. LaWALREE, A. La multiplication végétative des Lemnacées, en particulier chez W olffia arrhiza. Cellule 49: 337-382. Manesuwarl, S. C. The embryology of W ol fia. Phytomorphology 4: 355-365. 1954. . Endosperm and seed of Wolffia. Nature 178: 925, 926. 1956. Tuomas, R. F. Vascular plants previously unreported from Georgia. Castanea 16: 29-47. 1951. [Wolffia papulifera. | 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, IV i) STUDIES IN ARTOCARPUS AND ALLIED GENERA, IV. A REVISION OF ARTOCARPUS SUBGENUS PSEUDOJACA ? FrANcES M. JARRETT Subgenus Pseudojaca Trécul, Ann. Sci. Nat. Bot. IIT. 8: 117. 1847. Artocarpus section Pseudojaca Renner, Bot. Jahrb. 39: 368. 1907. Leaves alternate and distichous, simple, entire or nearly so; gland-hairs superficial or slightly immersed, heads 1(-—6)-celled; spongy mesophyll compact, lacking resin-cells. Stipules small, nonamplexicaul, scars lateral or intrapetiolar. Inflorescences without sterile, solid, elongate perianths. Male head, perianths 2—4-lobed or -partite, intermediates between perianths and interfloral bracts frequently present. Syncarp globose or + lobed, fleshy, the surface smooth, or papillate from perianth apices or from the clavate heads of interfloral bracts (series Clavati; the heads enlarged in A. styracifolius to form flexuous processes). LECTOTYPE SPECIES: Artocarpus lakoocha Roxb. In discussing the classification of Artocarpus in the previous paper in this series (Jour. Arnold Arb. 40: 125-127. 1959) it was pointed out that a marked contrast, which is reflected in the differing taxonomic treatments adopted, exists between the two rather distinct subgenera in the range of variation exhibited by their species. Subgenus Artocarpus, which was re- vised in that paper, was divided into two sections and a number of series based on a variety of morphological characters (e.g., orientation of the embryo, structure of the pericarp and position of the style, thin-walled vs. hypertrophied fruiting perianths, well-developed vs. vestigial interfloral bracts, and, in the leaves, presence or absence of a hypodermis and the shape of its cells, and the details of the gland-hairs). In subgenus Pseudojaca, a very different taxonomic pattern is found, due partly to a more reduced inflorescence structure, but also to the more re- stricted variation in vegetative characters and the occurrence of parallel evolution. Quantitative rather than qualitative characters have had to be used with much greater frequency than in subg. Artocarpus, both in dis- tinguishing and in arranging the species. Nineteen of the twenty species 1The preceding papers in this series will be found in Jour. Arnold Arb. 40: 1-29; 30-37; 113-155, 298-368. The loan of ‘material for study from two aes institutions is acknowledged ia Especial thanks are due to the Director of the Herbarium of the Department of Sys- tematics and Plant Geography of the Botanical Institute of the Academy of Sciences of the U.S.S.R., Leningrad, for the gift to the Arnold Arboretum of two 1 isotypes and one other duplicate from the collections made by the Sino- eee Expedition to Yunnan, 1956. 74 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI recognized here are assigned to section Pseudojaca, in which the species have been divided into two series on the basis of the shape of the inter- floral bracts. They have been arranged within the larger of these according to the length of the peduncles relative to the inflorescence heads and the shape of the male head, in conjunction with various other characters. There is a single anomalous species, Artocarpus altissimus, which appears to be referable to subg. Pseudojaca, but which is distinguished by the trinerved base and glandular-crenate margin of the leaves and by the deeply bifid styles. This is placed in a separate section, Glandulifolium, at the end of the paper and is further discussed there. Before proceeding to a discussion of section Pseudojaca, some general comments will be made on characters that have been of use in distin- guishing the species in this subgenus. In contrast to subg. Artocarpus, the internal structure of the syncarp has been found to provide cae ete ers of taxonomic significance only at the specific level. The fusion of the proximal regions of the perianths, the clear differentiation of an endocarp (as op- posed to the induration of the entire pericarp wall), and the development of fleshy fruiting perianths (slight as compared with subg. Artocarpus) may distinguish apparently allied species. However, with one exception, only the first of these characters is taken into account in this discussion since evidence is still inadequate concerning the others. It is probable that a larger number of species have somewhat fleshy fruiting perianths than is recorded here, since these can only be recognized in very well preserved syncarps. External characters of the female inflorescence are of consider- able value in providing specific distinctions (as in subg. Artocarpus). These include whether the surface is smooth or papillate at anthesis and maturity, the degree of exsertion of the styles, and whether or not the syncarp is lobed. The leaves also differ markedly from those of subg. Artocarpus in being rather uniform in their internal structure and in the details of the gland- hairs. However, the characters of the indumentum are of considerable taxonomic value, though once more primarily at the specific level. The hairs on the leaves and twigs may be straight, undulate or crisped, with either smooth or rough walls. In some species a varying proportion are rather stout, smooth-walled and hooked at the tip, and the presence of such hairs can be of assistance in identifying sterile material. In general, the hairs on the leaves are restricted almost entirely to the venation, an the prominence of the latter on the lower surface of the leaf is correlated with the density of the indumentum. However, in two of the species with prominent, pubescent venation, and in three others having the vena- tion not or slightly prominent beneath and subglabrous, nearly all the cells on the areolae on the lower surface may bear crisped hairs. The minute tomentum which is thus produced causes the leaves to appear glaucous beneath and seems to have been developed independently at least twice within the subgenus. There are six taxa in which the leaves are consistently glabrous or nearly so, and in five of these (4. gomezianus ssp. gomezianus, A. rubrovenius, A. nitidus, A. vrieseanus var. subsessilis and A. xantho- 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, IV is) carpus) a prominent reticulum is also lacking. With the exception of the last two, however, the similarity of the leaves in this respect is not, ap- parently, an indication of close taxonomic affinity. In A. reticulatus the subglabrous venation is acutely prominent beneath. The male inflorescences have been found to be of value in indicating wider affinities with the subgenus, but, unlike the female inflorescences, they quite often fail to provide satisfactory distinctions between the species. There is usually a characteristic shape for each species, but the variation in size may be considerable, so that the dimensions in allied species overlap. Specimens bearing male inflorescences, especially if immature, must often, therefore, be identified primarily by the use of vegetative characters, whereas specimens bearing female heads at any stage from anthesis to maturity usually can be identified chiefly on their characteristics. The species have been delimited primarily on the basis of inflorescence characters, and, in nearly all, one or more collections have been seen bear- ing both male and female inflorescences, enabling the characters to be correlated with certainty. However, since so many of the collections bear inflorescences of one sex only, or are sterile, vegetative characters have been of considerable practical importance in assigning material to the species recognized. (It may be remarked that the matching of collections on vege- tative characters was equally necessary in subg. Artocarpus, although the more clear-cut specific distinctions made the proceeding less critical.) In spite of the rather minor differences between some of the taxa in the characters of the leaves and twigs, it is possible to identify nearly all of the collections with male inflorescences and the majority of sterile collections. While this is partly a question of becoming familiar with the group, there are usually definable distinctions, and these are pointed out in notes to the species. For the purpose of identifying specimens on vegetative charac- ters, subg. Pseudojaca may be divided artificially into three groups of species characteristically having ‘“‘pubescent,” “glabrous” and “glaucous” leaves respectively. (Two species have subspecific taxa in both “pubescent” and ‘“olabrous” groups, and in at least two of the ‘‘pubescent”’ species the leaves may also be glaucous beneath). The majority of the specimens can be as- signed readily to one of these groups, and the notes give the distinguishing characters for all the species in the appropriate group that have overlapping ranges.” The two series recognized in section Pseudojaca are distinguished on the basis of a difference in the shape of the heads of the stalked interfloral bracts. Most of the species have bracts with peltate, discoid, ciliate heads, and are placed in series Peltati. There are, however, three species occurring in northern Indochina and southern China, in which the bracts have clavate or spathulate heads. These are assigned to series Clavati, which is further * Notes on “pubescent” oe will be found for India under Artocarpus gomezi- anus ssp. zeylanicus (31), for southeast Asia under A. tonkinensis (30), for western Malaysia under A. dadah (32), A. fulvicortex (40) and A. tomentosulus (41), and for a Rae eaors and eastern Malaysia under A. vrieseanus (33) and A. fretessit (35) ; notes “glabro species are given under A. nitidus (43) and on “glaucous” species under “a nee ae (45). 76 JOURNAL OF THE ARNOLD ARBORETUM [Vebe SL1 distinguished by a small but significant difference in the length of the anthers, these measuring 0.3-0.5 mm., as compared with 0.15—0.2 mm. in series Peltati. The dense pubescence on the bracts gives the surface of the male head a characteristic appearance; in series Peltati the surface is not conspicuously pubescent and the discoid heads of the bracts can be distin- guished under the binocular microscope. In two of the species in series Clavati, Artocarpus hypargyreus and A. petelotii, the heads of the bracts form papillae on the surface of the syncarp (not to be confused with the papillate apices of the perianths found in some members of series Peltati), but in the third, A. styracifolius, they are enlarged to form numerous flexuous, cylindric processes. The proximal region of the perianths is free in A. Aypargyreus, partially fused in A. petelotit and completely fused in A, styracifolius. However, in spite of the very distinctive syncarp, this last species shows a relationship to A. Aypargyreus in having leaves that are tomentulose on the areolae and hence glaucous beneath, with sub- glabrous main veins. In A. petelotii, on the other hand, the leaves are similar to those of most pubescent-leaved members of series Peltati in hav- ing the venation pubescent beneath, but the areolae glabrous. In series Peltati, as already indicated, it has been found that the most “natural” arrangement of the species apparently can be achieved by sub- dividing them according to the relative length of the head and the peduncle in the inflorescences at anthesis. The sixteen species then fall into two groups, in the first of which the peduncle is nearly always as long as, or longer than the head, and in the second of which it is less than half the length of the head, at least in the male inflorescence. The length of the peduncle in the female inflorescence seems to be inherently more variable than in the male inflorescence. Its relationship to the size of the head does not, in any case, hold beyond anthesis, owing to the enlargement of the latter as the syncarp matures, although the absolute length of the peduncle is usually greater in the first group. In this group, four pairs of species can be recognized (defined by a combination of characters), but their interre- lationships are not clear, and there are also two species showing reticulate alliances. The order adopted for the species is, in consequence, partly one of convenience. In the second group, the six species can again be assigned to three pairs showing a more or less close affinity to each other. An in- formal review of series Peltati, which is intended to indicate the type of variation that is found, and the reasons for the arrangement adopted, follows. In the first group of species, Artocarpus longifolius (Borneo) and A. ovatus (Philippine Islands) are distinguished by the long-pedunculate, rather small syncarps, the usually narrowly obovoid male head and the rela- tively narrow, pubescent leaves with short petioles. The syncarp has the perianths completely fused in A. longifolius, but free proximally in 4A. ovatus, and in the latter there are abundant persistent bracts on the sur- face. Except for A. tonkinensis the remaining species have globose or pul- vinate to short-obovoid male heads often overlapping in size and shape in the different species. 1960} JARRETT, ARTOCARPUS AND ALLIED GENERA, IV fii) Fic. 17. Distribution of some species of subg. PSEUDOJACA. oes ses de longifolius; 29, A. ovatus; 30, A. tonkinensis; 31, A. gomezi ssp. ee) omitting sneer from Assam and Cagay an Sulu), b, a ne 32, A. dad The next pair of species consists of A. gomezianus (ssp. gomezianus from Burma and Indochina to western Malaysia, and ssp. zeylanicus in southern India and Ceylon) and A. dadahk (Siam, Tenasserim and western Malay- sia), both of which have smooth, globose syncarps with few or no inter- floral bracts and free, fleshy fruiting perianths. The former is a species of monsoon regions and the latter chiefly of everwet forests, so that they are distributed differently in western Malaysia. Artocarpus dadah is also dis- tinguished from ssp. gomezianus by having pubescent, not glabrous leaves, and from ssp. zeylanicus by the colour of the hairs, which are red-brown, not greyish. Artocarpus vrieseanus (Mindanao to New Guinea and Melanesia) and A, xanthocarpus (northern and central Philippines) have syncarps that are very similar externally to those of the two preceding species, but are dis- tinguished from these and the syncarps of the following species-pair by the complete fusion of the perianths. Artocarpus vrieseanus is a very vari- able species in which the proportions of the head and peduncle character- istic of this group do not always hold; of the four varieties recognized, three have more or less pubescent leaves, but one has glabrous leaves and con- sistently short-pedunculate inflorescences. Superficially, this species often bears a strong resemblance to A. dadah or to A. fretessii, which is placed after A. xanthocarpus, although it is distinguished from the former by the frequently crisped indumentum on the twigs. This similarity has deter- 78 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI mined the arrangement adopted for these species, and the syncarp charac- ter, although a convenient distinction, may not have any great taxonomic significance. Artocarpus xanthocarpus is superficially similar to A. nitidus, a species belonging to the second group, in its small, glabrous leaves and short-pedunculate male inflorescences. However, the male head is often little or no longer than the peduncle and this, with the fused perianths in the syncarp, suggests that A. xanthocarpus is best placed here. Fic. 18. Distribution of some species of subg. PsEuDoJAcA. 33, Artoc iin vrieseanus; 34, A. xanthocarpus; 35, A. fretessii; 36, A. reticulatus: OT Gas subrotundi folius. Artocar pus fretessii (Borneo and the Philippines to western New Guinea) and A. reticulatus (Celebes and Moluccas) have female inflorescences that are distinguished by being distinctly papillate at anthesis and usually strongly lobed at maturity with rather few bracts. The former has a dis- tribution overlapping that of A. vrieseanus and is often difficult to distin- uish when sterile, the hairs likewise being crisped on the twigs. Artocarpus reticulatus differs from A. fretessii in being larger in all its parts and in having nearly glabrous leaves which, as noted above, have an acutely prominent reticulum, unlike those of the other glabrous- leaved species. There remain two species assigned to this group which show reticulate interrelationships. Artocarpus tonkinensis (Indochina and southern China) is placed after A. longifolius and A. ovatus, since it resembles both in having a relatively narrow, obovoid or ellipsoid male bead and the latter in hav- ing numerous persistent bracts on the syncarp. It differs from these species, however, in having crisped hairs, as in A. vrieseanus and A. fretessii. The 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, IV 79 perianths are fused in the syncarp, and the leaves have fewer lateral veins than in any other species of this group. Artocarpus subrotundifolius (Philippine Islands), which is placed last in this group, does not show any very clear alliances with the other members. The length of the peduncles, although exceedingly variable, as is the size of the very large male head, suggests that the species should be placed in this group; it also resembles A. fretessti and A. reticulatus in having a distinctly papillate female head at anthesis, with the perianths free proximally, although the syncarp is only shallowly lobed and the indumentum of the twig is patent. However, in all these characters, and in the long-exserted styles (also found, however, in A. reticulatus) and the broad, long-petiolate, pubescent leaves, often with an oblique base, it also resembles A. lakoocha, placed first in the following group of species. The species of the second group may be characterized by the consistently short male peduncle and show no obvious alliances with members of the first group, with the exception of A. subrotundifolius. There is a consider- able range of variation and, unlike the members of the first group, the species are most readily defined on characters of the male inflorescences and the vegetative parts, although syncarp characters can also be used. The first four species have rather large, broad leaves, with long petioles and an often oblique base, but fall into two quite distinct pairs. Artocarpus lakoocha (India to Indochina and Yunnan) and A. rubrovenius (Philippine Islands) both have rather large, obovoid to clavate (rarely ellipsoid) male heads, but differ considerably in their leaves, those of the former being pubescent on the prominent reticulum, and those of the latter being gla- brous, without a prominent reticulum. The second pair of species, A. fulvi- cortex (Malaya and Sumatra) and A. tomentosulus (Borneo), both of which are newly described, are distinguished from the other pubescent- leaved members of subgenus Pseudojaca by having the areolae on the lower surface, between the markedly prominent, pubescent reticulum, frequently tomentulose, and by the twigs being merely puberulent when young. They also differ from the other species of this second group in having globose male heads, but, whereas the syncarps in A. fulvicortex are subsessile, like the male heads, in A. tomentosulus the single syncarp seen has a peduncle 25 mm. long. In the former the syncarp has completely fused perianths; in the latter the structure is unknown, but in the remaining species of this group the perianths are free proximally. Finally, Artocarpus glaucus (western Malaysia) and A. nitidus (Assam to southern China, western Malaysia and the Philippines), have obovoid, clavate or cylindric male heads rather smaller than in the first pair of species, and leaves that are also smaller with shorter petioles. Artocarpus glaucus is distinguished by the very long, narrow male head, and, as the name suggests, by the leaves, which are tomentulose on the areolae be- neath, while the main veins are subglabrous. Although this indumentum resembles that found in two species of series Clavatz, there is no other indi- cation of an alliance between these species and A. glaucus, nor is there any evidence of a close relationship with the preceding pair of species. The *80 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI last species in this group is A. nitidus, which has a smaller male head than A. glaucus. It is widely distributed, and five subspecies are recognized on the basis of differences in the size and indumentum of the syncarp, associ- ated with slight, but characteristic, variations in the glabrous leaves. The species of section Pseudo jaca are thus separated by a variety of rela- tively minor characters, so that, although they are readily distinguishable (given adequate material), they require somewhat complex definition. There has also been some parallel evolution, especially, it would seem, in the loss of the indumentum, or the development of a tomentum, and in the complete fusion of the perianths in the syncarp. The distribution of some types of hairs among the species, especially between A, fretessia and A. vrieseanus, suggests the occurrence of some introgression. In view of all this it does not seem desirable to discuss in further detail interrelation- ships or evolutionary trends among the species, nor is it possible to define satisfactorily any supra-specific taxa beyond the two series here recognized. key which follows is based, as far as possible, on “natural” charac- ters, but practical considerations have, where necessary, taken first place in its construction. It is hoped that it will make possible the identification of most collections with male or female inflorescences at any stage from anthesis to maturity. (Where a dichotomy has had to be based on the in- florescences of one sex only, the specimens bearing inflorescences of the other sex ave eliminated, if possible, further on in the key under the first alternative.) For sterile specimens, or those with very young inflor- escences, the use of the footnote given above as a guide to the notes on vegetative distinctions is suggested. It should perhaps be pointed out that, while the species may be classified as characteristically “Dubescent,” “gla- brous” or “glaucous,” there are sufficient inconsistencies in the first two of these groups to limit the value of this distinction in preparing the key. The numbering of the species follows on from the previous paper in this series. KEY TO THE SPECIES OF ARTOCARPUS SUBGENUS PSEUDOJACA 1. Styles bifid; leaves glandular-crenate, base trinerved. .... 47. A. altissimus. Styles simple; leaves not as above. 2. Leaves densely greyish tomentulose beneath, except for the subglabrous main veins. 3. Syncarp peduncle to 5 mm.; male peduncle to 3 mm.; leaves with 8- US. pairs Jabeta) Veins. sacisyuedes te etisee wie 42. A. glaucus. 3. Syncarp peduncle 10-50 mm.; male peduncle 5-25 mm.; leaves with 4-9 pairs lateral veins. 4. Syncarp covered with flexuous processes; male head 4-7 mm. across; leaves with lamina decurrent on petiole, reticulum not 1 46, A. s Bae Ww <¢ =) a) 3 uo) © %. = 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, IV 81 2, Leaves not tomentulose beneath, or main veins and reticulum also patent- pubescent. 5. Peduncles in inflorescences at anthesis at least as long as the head, or, if shorter in the male inflorescence, at least 7 mm.; in mature syn- carp at least 13 mm 6. Male peduncle re 3 mm., head to 7 mm. across; leaves glabrous, reticulum not cee poets 7. Styles exserted to 0.4 mm.; male head with bracts eae stalked; leaves rounded or suriculat at the base, not dec rent; (Gren DCCUMELE TON 7 Sit) er rasta ee Cote Ae Gc 1: ai ean cu ea a A. vurieseanus var. subsessilis, 7. Styles exserted to c. 1 mm.; male head with bracts slenderly ae, leaves usually cuneate and slightly decurrent at the ase; (syncarp peduncle to 11 mm.). 34. A. xanthocarpus. ne peduncle at least 7 mm. and/or leaves with a reticulum prominent ae usually pubescent beneath. 8. Petiole 3-8 3.5-5 mm., base of leaf cuneate or auriculate; syncarp ae 30-60 mm.; male head obovoid, 8-20 « ae 12 mm., peduncle 12-30 mm. ............. 28. A. longifolius. 8. Petiole not so short and aie or base of leaf cordate. 9. Female head at anthesis with styles exserted to at least 1 mm. (a few usually ee on syncarp). 10. Styles eee —2.5 mm.; syncarp peduncle 40(?)- Ties a i 25-50 a 20-35 mm.; petiole (15-)25-65mm............. 7. A. subrotundifolius. Styles ees to 1-1.5 mm.; ae peduncle to 45 mm.; male head smaller. 11. Leaves glabrous or scabrid-puberulent on the prominent reticulum from hooked hairs, petiole 13-25 mm.; female head verrucose from conical papillae, or lobed at ay and smooth Me the ee peduncle 25-45 mm.; male head c. DCUO SS ct 5 su aei4 a ee 30 A: ee ee pubescent, without hooked hairs; female head papillate to nearly smooth. 12. Syncarp peduncle 35-40 mm.; leaves with 7- 11 pairs lateral veins; petiole to 18 mm.; sur- face of the male head densely pubescent from the Ligne (not peltate and discoid) heads of the bracts, oo 44. A. petelotii. Syncarp eed 8-25 mm.; leaves with 9- 18 pairs lateral veins, petiole 15-45 mm.; (male head 12-23 & 10-18 mm., surface cov- ered by legeee discoid, ae Roa of bracts 2-5 mm. A. lakoocha. 9. Female head at Sahecdc _ a pee te 0.7 mm. (a few usually persisting on sync 13. Syncarp subglobose or Ria lobed, numerous dis- os 2 — re — bo ? * There are a few species in which the length of the peduncle in relation to that of the head is variable, or in which the proportions differ in the male and female in- florescences; these are brought down on both sides of this dichotomy, the characters of the discordant inflorescences being given in parentheses. 82 JOURNAL OF THE ARNOLD ARBORETUM [VOL XE me wW coid heads of bracts persisting on surface, often com- pletely covering head at anthesis; male head obovoid to ellipsoid, at least 8 mm. across, peduncle at least 7 mm e “Leaves glabrous, reticulum not prominent n@ath,.¢..0.04 31. A gomezianus ssp. gomezianus. rai pubescent, reticulum distinctly prominent enea 15 a aie 15-45 mm.; (male head to 7 mm. across, peduncle to 2 mm.). .............. ee ee tor Tee 41. a tomentostulus. Petiole to 15 m 16. Syncarp ond (40-) 70-80 mm., peri- anths free proximally ; male peduncle 20- 40 mm.; twigs with patent hairs; leaves with 11-20 pairs lateral veins. ........ sie peers betas oaee yes 29, A. ovatus. Syncarp peduncle 30-40(-65) mm., perianths completely fused; male pe- duncle 7-12 mm.; twigs with appressed and crisped hairs; leaves with 6-9 pairs lateral veins. ...... 30. A. tonkinensis. Syncarp subglobose or lobed, bracts lacking at least on lobes, often nearly all deciduous before anthesis; male head pulvinate, globose or obovoid, if obovoid less than 8 mm. across 17. Young twigs greyish pubescent, or, if glabrous, leaves without prominent intercostal veins beneath. den oy ene ee deine ea 1. A. gomezianus. 17. Young twigs red-brown to yellow pubescent, or, if glabrous, leaves with prominent intercostal veins eee Syncarp with well-defined, subglobose lobes (unless many seeds are formed), the surface between papillate and often with bracts per- aa perianths free proximally; male heads ross, often borne on short-shoots od, bracts slenderly stalked; twigs usually with subappressed and crisped TYALTSE: -seetecw ted aes china a dndea he 35. A. fretessit. Syncarp globose or shallowly lobed. bracts oer nearly all deciduous; male heads to m. across, usually in leaf- axils. bracts ne stalked. 19. Syncarp yellow, subglobose and smooth, r (in var. papillosus) papillate at an- thesis and often shallowly lobed at ma- turity,.perianths completely fused; twigs usually with some or all of the hairs subappressed and crisped, varying sub- glabrous; larger leaves with up to 13 pairs lateral veins. ....... A. vrieseanus. —t way > — oo 1960 | JARRETT, ARTOCARPUS AND ALLIED GENERA, IV 83 5. Peduncles in inflorescences at ees i than half the length of the heads, in male inflorescence to 5 mm., in mature syncarp to 12 mm. 20. Leaves glabrous, Sa enien not or scarcely prominent beneath. bo a Male head 15-45 X 10-25 mm.; leaves often broadly aa tic to ovate, with a broad, pleas base, petiole 15-30 m EP ao ee Go ag ne re oe Male head to 12 & 7 mm.; petiole 5-25 22. Female head with perianths free See ity) styles ex- alee to 0.5 mm.; male head cae cylindric or GIA ALON ites oa hc dh ae rn A. nitidus, Dee Renate head with eas completely eee male head short-obovoid to globos 23. Styles exserted to 0 oA mm.; male head with bracts stoutly stalked; sag rounded or auriculate at the base--not dectrrent: .. .4 ee Ree oo. I wack oe Aci fc A, vrieseanus var. subsessilis. Styles exserted to c. 1 mm.; male head with bracts slenderly stalked; leaves Ree cuneate and slightly decurrent ab the bas Fee pai, 2s % ty yy 2s, a A. xanthocarpus. ath, no Ww 20. Leaves pubescent, reticulum distinctly prominent bene 24. Petiole to 15 mm. ........ 33. A. vrieseanus. 24. Petiole 15-45 m 25. Male head 12- 25 X 10-18 mm.; styles exserted to 1— 1.5 mm.; twigs pubescent; leaves never with the inter- aa tomentulose beneath. ...... 38. A. lakoocha. 25. Male head to 7 & 6 mm:; styles exserted to 0.5 mm. ; twigs subglabrous; leaves often with the intervenium tomentulose beneat 26. Leaves with 6-10 pairs lateral veins and few inter- costals; syncarp peduncle 5 mm. De ae ee eS RL 40. A. fulvicortex. 26. Leaves with 10-14 pairs lateral veins ne numerous intercostals; (syncarp peduncle c. 25 m cigs ic es hal Ge . Az; Pee te Section Pseudojaca Folia margine non glandulifera. Capitula feminea stylis simplicibus. Series Peltati Jarrett, ser. nov. Inflorescentiae bracteis interfloralibus peltatis. Capitula mascula cellis antherum 0.15—0.2 mm. longis 28. Artocarpus longifolius Becc. For. Borneo, 629. 1902, “longifolia.” Holotype, Borneo, Teysmann HB 11312 (¥1 FI), Trees, height to 21 m. Twigs 4-6 mm. thick, subglabrous to densely 84 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI pubescent, hairs red-brown to yellowish, patent, straight or some longer and hooked at the tip. Leaves 9-33 * 4.5-11 cm., obovate-elliptic or ellip- tic, with an acumen to 2 cm. long, base cuneate, often auriculate, margin entire; main veins and reticulum very prominent beneath; glabrous above except for the pubescent main veins, venation beneath pubescent, hairs colourless, straight, varying to subglabrous; lateral veins 9-14 pairs, curved; intercostals parallel; green, drying pale or reddish brown, lighter beneath, venation concolorous; petiole 3-8 mm. long, stout. Inflorescences solitary in leaf-axils. At anthesis: male head 8-20 * 5-12 mm., obovoid; perianths of c. 3 free segments 0.4 mm. long; stamen 0.5 mm. long, filament broad, contracted above, anther-cells globose, 0.15 mm, long; bracts slenderly stalked, heads peltate, to 0.3 mm. across, these and perianths minutely pubescent; peduncle 12-30 « 1 mm., pubescent, hairs rufous or whitish; female head with styles exserted to 0.2 mm. through low papillae emerging between peltate bracts. Syucarp (submature) 2 cm. across, ellipsoid, drying fulvous, the surface nearly smooth, very shortly pubescent, with scattered persistent bracts; proximal region of perianths fused; peduncle 30-60 x 2-4 mm., pubescent, hairs rufous or whitish. DISTRIBUTION: endemic to Borneo. Borneo. SARAWAK. Kuching, Haviland 2189 (BM, kK, SING, 6, 2), Haviland & Hose 3205 (pM, & 2), 3206 (kx, 6, 9). West Borneo. Landak, Teysmann HB 11312, 1875 (rt, 6, 2). SOUTH AND SOUTHEAST BorNEO. Puruktjahu, bd 11109 (so). EAST AND NORTHEAST BorNEO. W. Kutei: near Lahum, Endert 1840 (Lt, 6). E. Kutei: Samarinda, Blajan River, Loa Lampong, Nedi 769 (ccE, On ie The material of Artocarpus longifolius is limited and the variation in the indumentum is considerable, but the collections agree in the characters of the inflorescences and in the distinctive shape of the leaves and petioles. The large stipules (5-15 mm. long), which are frequently persistent, and the short, stout petioles are unique in subg. Pseudojaca. The subglabrous collections may be from saplings or sucker shoots, since the leaves and internodes are somewhat longer than in the pubescent collections. The cuneate base of the leaf and the shorter petiole, in addition to the smaller inflorescences and the completely fused perianths in the syncarp, distin- guish this species from Artocarpus ovatus. 29. Artocarpus ovatus Blanco, FI. Filip. 666. 1837, “Arctocarpus ovata,” non Noronha, 1790 (nomen nudum), ed. 3. 3: 73. plate. 1879; Vidal, Revis. Pl. Vasc. Filip. 254. 1886; Ahern, Timber Tree Sp. Philip. 21. plate. 1901. Neotype, Luzon, Merrill SB 254 (Bm). Artocarpus cumingiana Tréc. Ann. Sci. Nat. Bot. IIT. 8: 119. ¢. 4, figs. 117, 118. 1847, “Cummingiana”’; Fern.-Villar, Noviss. App. 203. 1880; Merr. Publ. Gov. Lab. Philip. 27: 80. 1905, Sp. Blancoanae, 125. 1918, Enum. Philip. Pl. 2: 40. 1923; Renner, Bot. Jahrb. 39: 370. 1907; Elmer, Leafi. Philip. Bot. 2: 620. 1909; Whitford, Bull. Bur. For. Philip. 10(2): 28. ¢. 6. 1911; Brown, ibid. 22(2): 70. fig. 23. 1921. Holotype, Cebu, Cuming 1784 (P); isotypes (BM, CGE, K). 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, IV 85 Artocarpus acuminatissima Merr. Philip. Jour. Sci. 18: 49. 1921, Enum. Philip. Pl. 2: 40. 1923. Holotype, Luzon, De Mesa & Rosario FB 22777 (PNH, de- stroyed) ; isotypes (K, US); lectotype (kK). Trees, height to 25 m., bark brown. Twigs 2.5—4.5 thick, smooth or finely rugose, short-pubescent, hairs red-brown to pale yellow, patent, straight, or some longer and undulate. Leaves 15-33 & 6-16 cm., oblong to obovate- oblong, varying elliptic, with an acumen to 2(—4) cm. long, base cordate, varying rounded, rarely cuneate, margin entire; main veins and reticulum prominent beneath; glabrous or nearly so above except for the pubescent main veins, venation beneath thinly to densely pubescent, hairs colourless, straight or undulate, some on main veins stout and with hooked tips; lat- eral veins 11-20 pairs, curved, basal 2-3 pairs crowded; intercostals par- allel or reticulate; deep green above, paler beneath, drying red-brown to blue-grey above, pale or reddish brown beneath, venation usually concol- orous; petiole 8-15 mm. long. Inflorescences solitary or paired in leaf-axils, the male heads also on short shoots on older wood. At anthesis: male head 12-32 * 10-15 mm., obo- void (rarely subglobose); perianths 2- or 3-lobed, divided nearly to the base, 0.5 mm. long; stamen 0.8 mm. long, filament flattened, tapering above, anther-cells ellipsoid, 0.15 mm, long; bracts stoutly stalked, heads peltate, to 0.4 mm. across, these and perianths sparsely ciliate; peduncle (15—)20- 40 2 mm., indumentum as twigs; female head with styles exserted to 0.5 mm. through a dense covering of peltate bracts. Syncarp to 3 cm across, subglobose, shallowly lobed, yellowish white, drying brown, the surface smooth, very shortly pubescent, with numerous persistent bracts; wall c. 2 mm. thick; proximal region of perianths free, fruiting perianths c. 4-6, thin-walled, “seeds” (endocarps) subellipsoid, 12 « 9 mm.; core c. 8 mm. across; peduncle (40—)70-80 2.5 mm., indumentum as twigs. VERNACULAR NAMES: cubi, anubing. Uses: provides a hard timber; the wood of other species of subg. Pseudojaca in the Philippines is sold under the same vernacular names. DIsTRIBUTION: in forests to 2500 ft.; northern and central Philippine Islands. Philippine Islands. Minporo. Medina FB 24178 (a, 2). Mansalay: Barrio Manaul, Sulit PNH 17175 (k, L, PNH, 3); Mt. Vagaw, Conklin PNH 17464 (xk, PNH, 2). Luzon. Ahern 119 (Bo, 8 2), 786 (Bo, 2). Abra: Manabo, Paraiso FB 31103 (sinc, 6). Mountain Province: Kalinga Subprov., Lubuagan, Celes- tino PNH 7846 (a, 6 @). Benguet: Leano FB 24712 (us, 2). Isabela: San Mariano, Clemens 17032 (c), Ramos & Edano BS 46810 (c, sinc, 6 9 ). Nueva Viscaya: near Dupax, McGregor 11495 (c, sinc, 8). Zambales: Mt. Pinatubo, Fox PNH 4677 (pnw). Bataan: Alambra FB 25303 (stnc, 6). Nueva Ecija: Alvarez FB 22143 (L). Bulucan: Angat, Vidal 3841 (x, 2). Rizal: Ahern FB 2905 (BO, K, P, SING, 6, 2), Maneja FB 23978 (a, Bo, L, 3), Reillo BS 15173 (Bo, L, P, 8, 2); Antipolo, Merrill 1680 (x, 8), SB 254 (A, BM, BO, GH, Oe 6); Bosoboso, Merrill 2822 (a, BM, K, 2); Montalban, Loher 6948 (x, 6); Mt. Angilog, Lopez FB 42035 (A, BM, kK, SING, @, 9). Laguna: Los Banos, Elmer 8175 (Bo, K, 8), Holman 93 (a, 8); Los Banos, Mt. Makiling, Banaga 86 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI PNH 33385 (Lt, 6), Elmer 18467 (a, BM, K, L, 6), Rivera PNH 9481 (BM, L, PNH, 6), Sulit & Paa 38 (a, 2); Santa Maria, Maritac, Curran FB 10040 (Bo). Tayabas: Cailipan FB 26019 (Bo, L, 3, 2), Manaiel FB 27456 (pe); Ainoboan, De Mesa & secaale FB 22777, Sept. 1913 (kK, us, 2); Langumanoc, Ware FB 7 (po, 6, 2); Lucban, Elmer 9163 (A, BO, K, L, 3). Camarines: Alambra 28085 (A, 2), eee FB 10426 (Bo, 4); Mt. Bagseay Ramos & Edano BS 33922 (A, K, 6, @); Mt. Isarog, Curran FB 10444 (po, 2). Sorsogon: Curran FB 10523 (a0, 7 Trosin, Mt. Bulusan, Elmer 14424 (A, BM, K, L, 6, 2), 14458 (A, BM, K, 2), MARINDUQUE. Rosenbluth BS 12152 (no, 4). CATANDUANES. Ramos & Bdane BS 75302 (sinc, 6, @). StBuyAN. Magellanes, Mt. Giting- Giting, Elmer 12332 (A, BM, BO, K, L, 6). Maspate. Merrill 3078 (A, BM, K. cao. Vidal 3840 (a, K, $). CreBu. Cuming 1784, 1841 (BM, CGE, kK, p, 6). Necros. Cardona FB 24219 (A, us, 2), Everett 7301 (Bo). a Dein FB 15036, Everett 4305 (Bo, 6); Cadiz, Oliveros FB 29891 (xy, Oriental: Dumaguete, Cuernos Mts., Elmer 9890 (A, BM, BO, K, L, ¢). ee llo-ilo, Vidal 596 (K, L, 4). Blanco’s description of Artocarpus ovatus is clearly identifiable, but his name was reduced by Fernandez-Villar (1880) to A. cumingiana Tréc. (the misspelling of which as “cummingiana” arose from a misprint on the labels in the Paris herbarium). Merrill (1905) also used Trécul’s name since he found (presumably from ‘Index Kewensis”) that Noronha had published an Artocarpus ovatus in 1790 and was unaware that the latter was a nomen nudum. Artocarpus ovatus is quite distinct from the other Philippine species of subg. Pseudojaca in its long-pedunculate, relatively small inflorescences. Nevertheless, it has frequently been confused with A. fretessa. Under the latter are given characters distinguishing these two species and A. subro- tundifolius, another pubescent-leaved species which occurs in the Philip- pines. 30. Artocarpus tonkinensis A. Chev. ex Gagnep. Bull. Soc. Bot. Fr. 73: 90. 1926: Gagnep. in Lecomte, Fl. Gén. Indoch. 5: 737. 1928; Merr. Lingnan Sci. Jour. 6: 275. 1930, 7: 303. 1931; Metcalf, Jour. Arnold Arb. 26: 198. 1945. Syntypes, Indochina, Balansa 2486, Castellini 111, Poilane 1859, Service Forestier s.n. (Pp); lectotype, Poilane 1859 P). Artocarpus pauican A. Chev. Bull. Econ. Indo-Chine 20(no. 132): 861. 918, nomen nu Artocarpus sp., ve ie Sci. Jour. 5: 63. 1928. Trees, height to 15 m. Twigs 1.5-2.5 mm. thick, reddish brown, short- pubescent or puberulent, hairs usually appressed and crisped. Leaves 9— 23 «& 4-10 cm., elliptic, obovate- or oblong-elliptic, or obovate-oblong, acuminate, base cuneate or rounded, margin entire; juvenile leaves with margin shallowly dentate towards apex; main veins and reticulum prom- inent beneath; glabrous above or the main veins puberulent, venation beneath thinly pubescent, hairs colourless and undulate; lateral veins 6—9 1960 | JARRETT, ARTOCARPUS AND ALLIED GENERA, IV 87 pairs, curved; intercostals parallel; deep green above, paler beneath, usu- ally drying greyish green with reddish main veins and reddish or straw- coloured reticulum; petiole 4-12(—20) mm. long. Inflorescences solitary in leaf-axils. At anthesis: male head 10-25 x 8- 15 mm., obovoid to ellipsoid; perianths of 2 or 3 free segments 0.5 mm long; stamen 0.7 mm. long, filament tapering above, anther-cells ellipsoid, 0.2 mm. long; bracts slenderly stalked, heads peltate, to 0.4 mm. across, these and perianths short-ciliate; peduncle (5—)7-12 & 1.5 mm., velu- tinous; female head with styles exserted to 0.7 mm. through a dense covering of peltate bracts. Syucarp to 6.5 cm. across, subglobose, shallowly lobed, yellow, drying rufous, the surface smooth, pubescent, with scattered persistent bracts; wall 5-8 mm. thick; proximal region of perianths fused, fruiting perianths 6-12, thin-walled, ‘“‘seeds” (endocarps) subellipsoid, 12-15 & 9-12 mm.; core c. 10 mm. across; peduncle 30-40(-65) * 2.5 mm., velutinous. Uses: cultivated for the edible fruit. DISTRIBUTION: in forests to 4000 ft.; Indochina and southern China. Indochina. CampopiA. d’Alleizette, 1909 (Pp, 6). ANNAM. Bu Khang, near Vinh, Poilane 16644, 16647 (P, 2). Thank Hoa Prov.: Na bam, Poilane 1859 Garp os TONKIN. Bac-giang Prov.: Bac le, Serv. For. (2, 2). Ha-coi: Sai Wong Mo Shan, Tsang 30266 (a, 6); Taai Wong Mo Shan, Tsang 29321 (A, @). Song Hoa, Castellini 111, Oct. 1905 (Pe). China. YUNNAN. Szemao, Henry 13015 (a, K, 6). KweicHow. Bua-li, Chenfeng, Teng 91028 (a, 9). Kwancst. Ping Nam Hsien, Wang 40423 (a, 2). Kwantunc. Fan Ch’eng district, Kung P’ing Shan, Tsang 26749 (a, 8); Sunyi district, Sie River, Tsiang Ying 2677 (K, sING, 6). Hanan. Wang 33671 (a, P, 6, 2); Ching Mai district, Pak Shik Ling, Lei 830 (A, K, L, SING, 2); Lai area, Hung Mo Shan, Tsang & Fung 295 (A, BM, K, NY, P, 6, 2), 661 (K, Ny, US, 6); Lam Ko district, Siu Shui Hang, Lin Fa Shan, cia 250 (A, BM, K, @); Ling Shui district, Chim Shan, Fan Maan Ts’uen, Fung 20192 (A, BM, K, NY, P, US, ¢); Lokwui, How 72358 (A, P, 6); Ngai district, Chiu Sam heuer Lau 395 a BM, K, P, 8): Po-ting, How 72008 (A, BM, P, 3), 72552 (A, BM), 73076 (A, P, 3), 73252 (A, SING, 3), 73471 o BM, Deo) Taam-chau district, Shan Tong To, Tsang 706 (A, BM, K, 2); Wik Tsok ate McClure 9752 (K); Yaichow, eae 62632 (A, NY, US, Q). Cultivated. InpDocHINA. Tonkin, Ké S6, Balansa 2486, May 1886 (pP, 2). CHINA. Kwangtung, Loh Kong Tung, Taai Shaan, McClure 13564 (ieee) Chevalier, in 1918, gave details of the wood only under the name Arto- carpus tonkinensis, and this was not validly published until 1926, when Gagnepain supplied a botanical description. In the same area there are two other pubescent-leaved species with which A. tonkinensis might be con- fused. In Artocarpus lakoocha, the styles are exserted to 1-1.5 mm., the peduncles are shorter (the male to 5 mm., the female to 15(—25) mm.), the leaf has more numerous lateral veins (9-18 pairs), and the petiole is longer ((10—-)15-45 mm.). In A. petelotiw the interfloral bracts are clav- ate, and the surface of the syncarp hence more or less papillate, while the indumentum of the twigs and leaves is longer and not crisped. 88 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI 31. Artocarpus gomezianus Wall. ex. Tréc. Ann. Sci. Nat. Bot. III. 8: 118. 1847, “Gomeziana”:; Kurz, For. Fl. Burma 2: 433. 1877; King in Hook. f. Fl. Brit. Ind. 5: 544. 1888; King, Ann. Bot. Gard. Cal- cutta 2: 15. 1889, pro parte, excl. t. 14A et spec. King 4189, 5078, 7535, 8838, Maingay 1486 (A. nitidus ssp. griffith) et Wallich 4658A (Ficus callophylla Bl.); Parkinson, For. Fl. Andaman Is. 253. 1923; Kanjilal et al. Fl. Assam 4: 269. 1940. Holotype, Burma, Wallich 4660 (c, not seen, photograph in A); isotypes (CAL, CGE, K). Artocarpus gomeziana Wall. Cat. no. 4660. 1831, nomen nudum. Artocarpus petiolaris Miq. Fl. Ind. Bat. ue 422. 1861. Holotype, Sumatra, Teysmann HB 752 (uv); isotype (Bo eae pomijormis Teysm. & en Natuurk. iene me Tad 25: 400. ; Koord. & Val. Bijdr. Boomsoort. Java 11: _ J. Smith, . ae i, a 87. t. 235, 1907; Backer, Beknopte Fl. ee : Artocarpus lakoocha Roxb. var. B gomeziana (Wall.) Trimen, ib Fl. Ceylon 4: 99. 1898, quoad nomen, non quoad plantam. Artocarpus masticata Gagnep. Bull. Soc. Bot. Fr. 73: 88. 1926; Gagnep. in Lecomte, Fl. Gén. Indoch. 5: 739. 1928. Holotype, Annam, Poilane 5492 (Pp); isotypes (Kk, P). ssp. gomezianus Evergreen trees, height to 40 m., bark dark grey-brown. Twigs 2-4 mm. thick, smooth or finely rugose, appressed-puberulent, soon glabrescent. Leaves 11-25 7-16 cm., oblong, varying to elliptic, short-acuminate, base broadly rounded, varying to cuneate or subcordate, glabrous, margin entire or shallowly crenate; main veins prominent beneath; lateral veins 10-15(—20) pairs, straight or slightly curved; intercostals parallel; deep. green above, paler beneath, main veins greenish white, usually drying pale greenish or greyish brown with straw-coloured main veins; petiole 15-30 mm. long. Inflorescences solitary in leaf-axils. At anthesis: male head 10-25 mm. across, obovoid to subglobose; perianths 2- or 3-lobed, divided nearly to the base, 0.5 mm. long; stamen 0.6 mm. long, filament stout, abruptly con- tracted above, anther-cells globose, 0.2 mm. long; bracts stoutly stalked, heads narrowly peltate, to 0.4 mm. across, these and perianths with sparse, short hairs; peduncle 7-17 * 1 mm., puberulent; female head with peltate bracts being shed or already fallen and styles exserted to 0.3 mm. through perforations in the surface. Syucarp to 8 cm. across, subglobose, yellow with pink flesh, drying brown or black, the surface smooth, velutinous; wall c. 8 mm. thick; proximal region of perianths free, fruiting perianths several, fleshy, ‘“‘seeds’” (indurated endocarps) ellipsoid, 12: 10 tno.5 core c. 20 mm. across; peduncle 15-45 « 8 mm., velutinous. Uses: there is one record of the fruit being edible (Poulo Condore, Harmand 700); the roots are chewed with betel (Annam, Poilane 5492). DISTRIBUTION: in evergreen and semi-evergreen forest to 2000 ft. in regions with a distinct dry season; Assam ?, Burma, Andaman Islands, Siam, southern Indochina, Malaya, Sumatra, Java (western and central provinces), Philippine Islands (Cagayan Sulu, ? introduced). 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, IV 89 Burma. Pinmona [? Pyinmana], Huk, Aug. 1890 (p, sING). TENASSERIM. Tavoy: Wallich 4660 (leg. Gomez), Nov. 1827 (CAL, cGE, K, 6). Mergui: Vic- toria Point, Po Khant 11343 (pp, 9). Andaman Islands. King 416 (x, oD Siam. Ko Kahdat, Schmidt 556 (c, 8); Muak Tek, Sanaburi, Kerr 9064 (pM, 8); Chantaburi, Jeppeson, May 1947 (c); Sraburi, Phu Khae, Smitinand 1549 (cox). PENINSULAR SIAM. Bangtapan, Kerr 1436 (sae, 2s Krabi, Ao Luk, Kerr 18608 (pM, 6); Surat, Kaw Tao, Kerr 12700 (eM, kK, L, P, 8); Trang, Chum Het, Kerr 15214 (BM, K, P, 2); Trang, ae Tiang, Pox 3845 (sinc, 4, 9). Indochina. CaAmsBopia. Béjeaud 640 (NY ?). ANNAM. Nhatrang rove Phu-hu, between Nhatrang and ae Poilane 5492, ae 1923 (kK, P, 3). PouLo ConDoRE. Harmand 700 oe Malaya. Novadens Hill, Nees 1486 (GH, K, L, 6). Kepau. Baling, os dullah KEP 27356 (Ker); Kokmoi For. Res.. re KEP 42427 (KEP, 2); Songsong, ie -June 1890 (Pp); Sik, Mohamed KEP 11327 (KEP, @). ae TAN. Tum SGomer SFN 33529 (SING, 3), Ngadiman SFN 33698 (sING, 2). PAHANG. se Clough For. Res., KEP 78746 (KEP); Bentong, Ulu Benus, Symington KEP 51821 (KEP, 9). JoHore. Sungei Rhu Riba, Jason sae Corner (sinc, 6). LANKAw1 Istanps. P. Butong, Curtis 906 (K, sING, $). PENANG. Road ca Balik Pulau, Curtis 2462 (kK, sinc, 2). SINGAPORE. Didrichsen 4412 (c, 6). P. Tloman. Telok Paya, Nur SFN 21742 (so, ad SING, ¢ Sumatra. ATJEH. Seulimeum, Lam Temot, bb 587 6 (Bo, L). TAPANULI, Padang eS Se ee bb 6186 (Bo). West Coast. Teysmann HB 798 (Bo, U, 2); Baros, Teysmann HB 721 (1), 752 (Bo, vu, 6). East Coast. Huta eine Estate, near Kisarin, Krukof 225 (Bo, 6). BENKULEN. Redjang, a marindu, bb 8861 (Bo); Redjang, Tabah ae bb 2822 (Bo, L). P. W Baloken, Koorders 10578, 10579, 10580 (xo, Java. WesT JAVA. Buitenzorg: Aik Kulon, Tjikankung, Tjidjaringao, NIFS Ja 1250 (Bo, 2). Preanger: Palabuanratu, Sukabumi, Koorders 8687 (Bo, 6). CENTRAL JAVA. Pekalongan: Mayasari, Burger 3340 (Bo, 6); Prupuk, Wind 26 (Bo, L, 2). Semarang: Koorders 9419 (Bo, L); Kedungdjati, Koorders 8674 (0; L), 8676 (Bo, L, 6), 8677 (Bo, 6, 2), 8690 (Bo, L, 6), 25251 (BO, OD) ee 322, 20159" (Ae BOL), 33701 (B0,-K.20 6. ee Kedung- djati, pee Koorders 8675 (Bo, L, 6, 2), 8691, 9271, 25319 (zo, L); Kedung- djati, near Gunong Kenting, Roorders 24957 (Bo, L, ¢); Kedungdjati, Prizi, Koorders 25475 (Bo, L); Kedungdjati, Pungge, Koorders 8678 (xo, P), 8679 (Bo, L); Kedungdjati, Trimze, Koorders 8689, 8692 (Bo, L, P), 8693 (Bo); Telawa, Jansen 4637 (Bo, 6). Jogjakarta: Zuidergebergte, Teysmann (BO). eee poe, CaGAYAN Sulu. Antonio FB 30634 (ny), Merrill 5304 (BO, L, P, ?); Buluan Ranch, Santos 4772 (1, ¢, 2). Cultivated. Java. Bee Tate eae HB 7289 (Bo, Pp), VIII B 41 (1). The synonyms listed above all refer to Artocarpus gomezianus ssp. gomezianus, but the name A. lakoocha Roxb. var. gomeziana (Wall.) Trimen was used by Trimen and later authors in the sense of A. gomezianus ssp. zeylanicus. The variety griffithi, based on Malayan material, which was described under Artocarpus gomezianus by King in 1888 and 1889, is here trans- ferred to A. mitidus as a subspecies; the distinguishing characters and the confusion that has occurred between the entities in Malaya are discussed there. All other references in the literature to A. gomezianus as occurring 90 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI in Malaya are based on A. nitidus ssp. griffithii, Artocarpus gomezianus ssp. gomezianus is mainly restricted to the north of Malaya and the east coast, where there is a distinct dry season, although a single specimen has been seen from Singapore The distribution of ssp. gomezianus in western Malaysia is apparently somewhat discontinuous, corresponding to its climatic requirements. How- ever, no significant variation is shown within this subspecies and the maintenance of A. pomiformis Teysm. & Binnend. as a distinct species in Java presumably arose from its comparison with King’s description and plate, which were largely (the latter solely) based on material which should have been referred to his var. griffithii. The subspecies is apparently ab- sent from Borneo and it is possible that the specimens from Cagayan Sulu were taken from a tree, or trees, that had been introduced, although there is no evidence for this. Kanjilal et al. (1940) also record A. gomezianus for the Lakhimpur district of Assam. ssp. zeylanicus Jarrett, ssp. nov. Artocarpus lakoocha auct. non Roxb., Thwaites, Enum. Pl. Zeylan. 262. 1861; Beddome, For. Man. 219. 1873, pro gee Trimen, Handb. Fl. Ceylon 4: 99. 1898; Cooke, Fl. Bombay a 657. 1907, p.p.; Bourdillon, For. Trees Travancore, 371. 1908, p.p.; Talbot, ei Fl. Bombay 2: 529. LOT Wipes incl. fig. 533; Fischer in Ge. Fl. Madras 3: 1369. 1928, p.p.; Macmil- lan, Trop. Pl. & Gard. ed. 4. 250. bee Abeyes. & Rosayro, Checklist Ceylon, 48. 1939; Watanabe, Ic. Econ. Pl. S. Asia 2: 527. 1945. la ak lakoocha Roxb. var. comesana (Wall.) Trimen, Handb. F1. 9. 1898, quoad descr. et spec. Artocarpus gomezianus Tréc., W oes ae Trees. 420. 1959. Differt ab typo ramulis junioribus saepe appresse pubescentibus, pills undulatis, cinereis, raro flavis, foliis ovatis, ovati-lanceolatis, ovati-ellip- ticis vel elliptici-oblongis, acumine ad 2 cm. longo, basi subcordata, late rotunda vel late cuneata, costa nervis lateralibusque subtus prominentibus, venulis prominulis, pubescentibus, pilis canescens undulatis, vel glabris, nervis lateralibus utrinque 10-13, petiolo (7—)13—25(-30) mm. longo, capitulis masculis 8-11 mm. diametro, globosis, pedunculis 5-18 mm. longis, syncarpiis 3 cm. diametro, flavis, carne flava, pedunculis 13-15 (- 20) mm. longis. Hototyrre: Madras, Wight 2717 (kK); isotypes (Cc, GH, L). VERNACULAR NAMES: ofamb, watamba, lowi, southern India; kana- gona, Ceylon. Uses: the fruit is eaten and the tree appears to be planted in both Ceylon and southern India. STRIBUTION: in: evergreen forests to 3000 ft.; in the wetter parts of ee Ghats and Ceylon India. Cutahey, Buchanan, Mar. 1801 (Bm, ¢); Malabar, Concan, Stocks & Law (GH, P); Kanara, Vacombi, Cooke, Feb. 1893 (xk, 4, Q). BOMBAY. ee Kanara, Kar arwar, Fernandes 119 (Ay oye Dankly [7 Sangli], Stocks (x, 3); Supa, Fernandes 1143 (A, BLAT, @), Ritchie 1381 (K, ¢). Mysore. Marnhalli, 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, IV 91 ? Meebold 8435 (x, 8). Coorc, Near Mercara, Hohenacker 557 (BMS Ce kert, 6). Mapras. Mangalore, Wee 2717, Mar. 1852 (c, cH ) ) K, i Ceylon. Macrae 692 (3M, , 6), Thwaites CP 2232 (p., CGE, GH, P, ¢), 2463 a 3); Galle, Oe (CGE, ?); Galle, Batapola, Worthington 2462 (BM); of Ginganga, Hiniduma, Worthington 2312 (sm, ¢): Kadugannawa, Kolugala, Hancock 1379 (Bm); Kadugannawa, Udawela Fort Jungle, Worthington 1426 (pM, 2); Narawella [= Gear Eliya], Champion (ccE, 8); Ratnapoora, Thwaites CP 2831 (BM, CGE, GH, K, P, 2). Cultivated. CEyion. Mellue, Gardner (K, 2); Peradeniya, Hort. Bot., W ee 6744 (cGE, 9), Worthington, Rebs 1955%(CGE. 6): Artocarpus gomezianus ssp. zeylanicus, which is described here from the Western Ghats and Ceylon, has not previously been distinguished from A. lakoocha. However, within the Indian subcontinent, the latter appears to be restricted, as an iidige sous tree, to the north and east. The new en- tity is differentiated by the smaller, globose head (8-11 mm. diameter vs. 12-25 & 10-18 mm.) and the longer peduncle (5-18 * 1 vs. 2—5 eee, mm.) of the male inflorescence, by the smooth surface of the female head at anthesis and the shorter exsertion of the styles (to 0.3 mm. vs. 1-1.5 mm.) and, apparently, by the smaller size and smoother surface of the mature syncarp. In addition, the collections seen from India have rather distinc- tive ovate-lanceolate leaves, which are densely greyish pubescent beneath. In those seen from Ceylon, on the other hand, the leaves vary to elliptic- oblong in outline and these and the twigs are often subglabrous, so that specimens are not clearly distinguishable from A. gomezianus, sensu stricto. Since the dimensions of the male inflorescences, although usually smaller in the western entity, also overlap, it seems best to treat the latter as a sub- species of A. gomezianus. The nature of the variation found in Ceylon is not entirely clear, but it appears to occur between different trees and not merely to represent sapling and adult stages. Subglabrous collections from Ceylon were identified with A. gomezianus by both Thwaites and Trimen, and were described by the latter under the name 4. lakoocha var. B gomez- jana. 32. Artocarpus dadah Miq. Fl. Ind. Bat. Suppl. 420. 1861, Ann. Mus. Lugd.-Bat. 3: 213. 1867; S. Moore, Jour. Bot. 63, Suppl. 112. 1925: Corner, Gard. Bull. Singapore 10: 282. 1939, Wayside Trees, 653. 1940. Holotype, Sumatra, Tevsmann HB 4391 (uv); isotypes (zo, 7s lace mollis Miq. Fl. Ind. Bat. Suppl. 420. 1861, non Wallich, 1831 (nomen nudum), Ann. Mus. Lugd.-Bat. 3: 211. 1867. Holotype, Sumatra, Be HB 4211 (L); isotype (Bo). Gone rufescens Miq. Fl. Ind. Bat. Suppl. 420. 1861; Renner, Bot. Jah 1907. Holotype, Sumatra, Teysmann HB 379 3 (u); isotypes a K, Artocarpus tampang Miq. Fl. Ind. Bat. Suppl. 421. 1861, Ann. Mus. Lugd.- Bat. 3: 211. 1867. Holotype, Sumatra, Teysmann HB 3997 (wv); isotypes (Bo, K, L). 92 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI Ficus tampang Mia. FI. Ind. Bat. Suppl. 425. 1861. Holotype, Sumatra, Teys- mann HB 710 (wv); isotype (Bo). Ficus inconstantissima Mig. F1. Ind. Bat. Suppl. 431. 1861. Holotype, Sumatra, Teysmann HB 3529 (vu); isotypes (Bo, CAL, L). Artocarpus inconstantissima Mig. Ann. Mus. Lugd.-Bat. 3: 211. 1867. Artocarpus dadah Miq. var. pubescens Miq. Ann. Mus. Lug.-Bat. 3: 213. 1867. Holotype, Sumatra, Korthals s.n, (1). Artocarpus erythrocarpa Korthals ex Miq. Ann. Mus. LiedeBat.. 3s 213, 1867, pro syn. Artocarpus lakoocha Roxb. var. malayana King in Hook. f. FI. Brit. Ind. 5: 544. 1888: King, Ann. Bot. Gard. Calcutta 2: 15, excl. spec. De Fretes, Amboina. Syntypes, Malaya, King 1640, 4187, 5653 (CAL, not seen; dupli- cates examined, k, etc Artocarpus reniformis Becc. For. Borneo, 631. 1902. Syntypes, Sarawak, Beccari PB 3107, 3551 (81); lectotype, Beccari PB 3107 (¥1). Artocarpus peltata Merr. Jour. Str. Br. Asiat. Soc. 85: 166, 1922. Holotype, British North Borneo, Villamil 168 (PNH, not seen, photograph in A); iso- type (Bo). Artocarpus lakoocha auct. non Roxb., King in Hook. f. Fl. Brit. Ind. 5: 543. pro parte; King, Ann. Bot. Gard. Calcutta 2: 14. 1889, p.p., quoad spec. Griffith 4666, Maingay 1479; Ridley, Fl. Malay Penin. 3: 355. 1924. Artocarpus dasyphylla auct. non Miq., Merr. Pl. Elmer. Born. 46. 1929. Deciduous trees, height to 35 m. Twigs 2.5-5 mm. thick, rugose, densely short-pubescent, hairs red-brown or fulvous, straight, varying undulate, a few sometimes longer and hooked at tip. Leaves 10-30 5-17 cm., obo- vate- or elliptic-oblong, varying to ovate-elliptic, acute or acuminate, base rounded, varying to broadly cuneate or shallowly cordate, margin entire; juvenile leaves with the lamina pinnatifid or reduced to a nar- row sinuate wing along the midrib; main veins prominent beneath, retic- ulum less so; glabrous above, or nearly so, except for the short-pubescent main veins, venation beneath densely to thinly pubescent, hairs rufous to colourless, patent, straight or slightly undulate (in Borneo, except British North Borneo, hairs often inserted along edges of reticulum and appressed over areolae), subdeciduous, the lower surface scabrescent, varying per- sistent; lateral veins 10-20 pairs, curved; intercostals parallel ; dark green, drying dark or reddish brown to blue-grey above, pale or reddish brown beneath, the areolae sometimes greyish and, with the colourless hairs, ap- pearing glaucous, venation concolorous or often nigrescent; petiole 5-20 mm. long. Inflorescences solitary in leaf-axils. At anthesis: male head 8-15 mm. across, globose or pulvinate; perianths of 2 or 3 free segments 0.3 mm. long; stamen 0.5 mm. long, filament cylindric, abruptly contracted above, anther-cells globose, 0.2 mm. long; bracts stoutly stalked, heads narrowly peltate, to 0.4 mm. across, these and perianths short-ciliate; peduncle 8— 15 (Borneo, 8-20) X 1 mm., indumentum as twigs; female head with pel- tate bracts already fallen (densely covering young head) and styles ex- serted to c. 0.5 mm. through low papillae. Syncarp to c. 5 cm. across, sub- globose, green with deep pink flesh, the surface smooth, velutinous; wall c. 7 mm. thick; proximal region of perianths free, fruiting perianths sev- 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, IV 93 eral, fleshy, “seeds” (indurated endocarps) ellipsoid, 12 * 8 mm.; core c. 5 mm. across; peduncle 12—25(-40) (Borneo, 23—45(—80)) *& 4 mm., indumentum as twigs. VERNACULAR NAMES: tampang (Malay), Malaya, Sumatra, Borneo; dadah, Sumatra (recorded only twice). DISTRIBUTION: in evergreen forest to 3000 ft.; Tenasserim, Siam, Ma- laya, Sumatra, Simalur, Banka, Borneo. Lower Burma. TENASSERIM. Mergui: Mergui range, sine nom. 421 (pp, 2); Thamihla Chaungbya, Po Khant oe (K). Siam. Phetchbun, Lom Kao, Ban Phuhee, Smitinand 2587 (cGE). PENINSULAR SIAM. Kaw Pa-ngan, Kerr 1187 (BM, 2); Ranawng, Nok Nang, Kerr 16835 (em, 2); Satul, Klawng Ton, Kerr 14594 (pm, 2); Trang, Chawng, Kerr 15170 (pm, 6), Put 2371 (cceE); Trang, Kuantan, Kerr 17488 (pm, 2); P. Terutao, Kerr 14229 (gm, ¢). Malaya. pee Meh CF 21885 (strc); Gunong Raya, Dolman CF 21496 (SING, Sy Prov. WELLESLEY. Krian, Ridley 9385 (CAL, SING, 9); Kubang Ulu, Curtis, rae! 1890 (sING, 2); Tasek Gelugur, Ridley 6978 (3M, 6), 6980 (sING, é, 9), PERAK. Scortechini 37b ie Me BM, K, P, SING, 6, 2); Larut, King 2678 (kK, L, 2), 4187, Apr. 1883 (x, arab Chandering, King sad Mar. 1884 (BM, K, L, P, 2); Larut, @oning: en 6098 (BM, BO, K, P, 2); Sun Larut, Wray 2479 (SING, $): Trong, Wray 3176 (sING, 3, 9): Waterfall Hill, Wray 2560 (stnc, 9). KELANTAN. Walton KEP 32681 (KEP). TRENGGANU. Bukit Kajang, Ulu Bendol, Kemaman, Corner, Nov. 1935 (SING). PAHANG, Cameron Highlands, Batten Pooll, Nov. 1939-Jan. 1940 (sinc); Gali near Raub, Burkill & Haniff SFN 16918 (stnc, 6); Kuantan, Lamban CF 2707 (x, SING, 2 ungei Rompui, Bidin CF 15658 (sinc, @). SELANGOR. Klang Gates, Murdoch 57 (pM, 6); Kuala Lumpur, Mohamor, June 1890 (stnc, @); Kuala Lumpur, Weld Hills For. Res., Ahmad CF 2949, 5025 (snc, 2), Guard CF 863 (sinc, 8). NEGRI eee Triang Res., Tahir CF 607 (stnc, 2). MALACCA. Maingay 1479 (Bo, GH, K, L, 6, @); Ayer Punnas, Griffith 4666 (kK, P, 2). JouHore. Kluang For. Res., Holttum SFN 9204 (xk, sinc, 6); Sungei Berassau, Mawai-Jemalaung road, Comer Jan. 1936 (sinc); Sungei Kayu, Mawai-Jema- laung road, Kiah SFN 32185 (Bo). LAaNKAwr IsLtanps. Wvyatt-Smith KEP FILLOTEO ae 6, @); near Kuah, Curtis (sinc); P. Butong, Curtis 906 (x, SING, 2). PENANG. King 1532 (kK, 6), 1640, Aug. 1881 (Bo, CGE, K, 2), Wal- lich 4658B en Government Hill, Curtis 1222 (sinc, 6); West Hill, Curtis 1251 (SING, 2), 1743 (CAL, SING, 3). SINGAPORE. Cantley ae (SING, 3); Bukit Kalang, Ridley, 1892 (snc): Bukit Timah, Ngadiman SFN 34682 (a, BO, K, P, SING, 6), Ridley 4722 (BM, K, L, P, SING, 9); Chan Chu Kang, Ridley (SING, 2); Changi, Ridley 3358 (CAL, K, SING, 6, 2), 4724 (x, 3); Govern- ment House Domain, Best SFN 25995 (sinc, 2); MacRitchie Reservoir, Thomp- son Road end, Corner, June 1937 (sinc); Mandai road, Corner SFN 32545 (x, sinc), Ridley 4130 (BM, SING, 2); 11% miles Mandai road, by shore of Seletar Reservoir, Sinclair, Mar. 1953 (x); Pulau Ubin, Ridley 4721 (stnc, 2); Reser- voir Jungle, Corner, Feb. 1937 (sinc, 6). Sumatra. Grashoff 1032b (Bo, L, 2), Korthals (1, 2). TapaANuLt. Sibide, Parduaan, Rahmat si Boeea 6096 (1, 6); Padang Lawas, Purbasinamba, bb 6202 (Bo). West Coast. Fort de Kock, Teysmann 710 (so, vu); Mt. Sago near Pajahkumbuh, Meijer 7197 (L); Priaman, Diepenhorst 2186 (Pp). East Coast. Amplas, Jochems 3168 (Bo); Asahan, Aek Salabat, Rahmat si Boeea 9622 (a, t, 6); Asahan, Huta Padang, Krukoff 4368 (A, BO, L, NY, SING, 6); Asahan, 94 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLI Simpang Toba, bb 6344, 7184 (Bo, L); Beneden Langkat, Alur Gusta, bb 16375 (A, BO, 1); Langkat. Sungei Sedapan, bb 9364 (Bo); Lubuk Mambang, Koorders 10453 (Bo, - Sibolangit, Lérzing 5151 (Bo, L, 29). Dyampr. Danau Lama, bb 13640 (Bo, 9). PALEMBANG. Banjuasin and Rupuctreken, NIFS T 65 (no, i U2. 777 ABO, Eo, Sine. 2 one eee NIFS T 902 (xo, L, é); Batu Radja, Teysmann HB 3529 (xo, _L, u); Dermo Enim, etre HB 3793 (Bo, K, L, u, 6, 2); Komering Uh, Crashof 579 (BO); Lematang llir, Darmo, bb 8720 (Bo): Lematang llir, Gunong Megang, N/FS T 889 (Bo, L, 3), 1207 (Bo); Lematang Ulu, Grashoff 179 (Bo, L), Lambach 1200 (xo, L, Muara Dua, Grashoff 436 (Bo, L, 9), Teysmann HB 3856 (Bo, P); Muara Dua, Kisau, bb 9234 (Bo); Mulak Ulu, Grashoff 330 (Bo, L, 6); Pandananan, Oganulu, Teysmann HB 3742 (pP), 3997 (Bo, K, L, U); River Ruput, W. Suka Radja, Forbes 2948a (CAL, L); Tandjong Ning, River Bliti, Forbes 2789 (BM, L, SING, 2). LAMPONGS. ets ee HB 4211 (wo, L); Mangala, Teysmann HB 4391 (BO, K, L, 9). SrmaLur. Achmad 1259 (Bo, K, L, P, SING, U, 2); Landschap Tapah, Defajan, Mehmed 1805 (BO, K, L, SING, Q). BANKA. Blinju, Teysmann HB 7249 (Bo, K, L, 2), Vordermans 41 (Bo, 2); Djebus, Teysmann HB. f2a9 10. GP): Pangkalpina ang, Teysmann 7 6842 (BO, K, L, 6856 (BO, C, K, Lb, P) 6): Bunge: Liat, Teysmann HB 7256 (1, P); Zapadong, Teysmann HB 3200 (p), BILLITON. Mendljungpandan, oo (po). Riouw ARCHIP. P. Bintang, ea HB 7285 (Bo, K, L, P, 2 Borneo. SARAWAK. Kuching, Beccari PB 3107, Nov. 1867 (FI, K, 2), 3551, June 1867 i. FI, 6); Nanga Temulan, Daud & Tachun 35690 (SING, 9). West Borneo. Sambas, Perigi Limus, bb 7064 (Bo, 2); Sungei Sambas. Hallier 1153 (Bo, L, sING, ¢). SOUTH AND SOUTHEAST BoRNEO. Martapura, Djungur, bb 10377 (Bo, 6, 2); Tanah Bumbu, core Baru, bb 13308 (Bo, L), 13366 (Bo, K, L, 9). EAST AND NORTHEAST Re Balikpapan: Lam 3871 (Le Mentawir, Sauveur 97 (K, 6, 2). Berouw: Domaring, bb 18865 (A, BO, L, SING ~ Inaran, bb 12175 (Bo). W. Kutei: “Longe bb 16133 (Bo, L), 16145 (a, - Tandjong Isui, Endert 1945 (a, 6, 9); Upper Mahakam . Udjoh Bilong, bb 20612 (A, Bo, L). E. Kutei: ie Djanan, w. of Samarinda, ‘Koster- mans 6644 (L), 9970 (L, 6, 2); Tandjong Bangko region, near mouth of Maha- kam River, Kostermans 7018 (L). BritisH NorTH Borneo. Agama 484 (A, K 9); Elphinstone prov., Tawao, Elmer 21110 (A, BM, K, L, 2); Mt. Radabalek Dallas, Clemens 26322 (A, BM, K, L, 6, @), 27444 (A, BO, BM, K, L, ?); Sanda- kan, Melegrito 9044 (stNc), Ramos 1904 (A, BM, K, L, on Villamil 168, Mar. 1916 (Bo, PNH, 2); Sandakan, Ebpura, Beatrice Road. Alendre A 3252 (1, SING, ?); Sandakan, Leila For. Res.. Bukit Makara, Wood SAN A 3476 (A, K, L, SING, 6). This common and variable species was described under six different specific names by Miquel in 1861. Of these, the one chosen here is Arto- carpus dadah. This has been the name most widely used on herbarium specimens and Corner, in 1939, correctly identified as A. dadah the Malayan entity which had hitherto been regarded as representing 4. lakoocha. The latter is a species of the monsoon forest which has not been found south of the Siamese border, and the error arose from its treatment y King in his monograph. He described under A. lakoocha a variety malayana, citing three collections here referred to A. dadah, and giving A. tampang and A. rufescens of Miquel correctly as synonyms. However, under the type itself two further collections of A. dadah were listed and 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, IV 95 the description was partly based on these. As a result, the variety was not recognized by later workers and all Malayan material was identified as A. lakoocha. The fourth collection cited by King under var. malayana, De Fretes sn, from Ambon, is a sterile specimen which may be referable to A. urieseanus. The variation shown by Artocarpus dadah occurs chiefly in the length of the peduncles and in the indumentum. It has some geographical basis, but separation into geographical units is not possible, since the variation is continuous and is not parallel in the different characters. Collections seen from Borneo tend to have inflorescences with longer peduncles, as is indi- cated above in the description. In this area the indumentum of the leaves is also usually persistent, whereas elsewhere the leaves generally become subglabrous and slightly rough beneath. Some specimens from Sumatra and Malaya are, however, densely and persistently pubescent, and this character is often associated with a rather more prominent reticulum. A peculiarity found in material from Borneo (but apparently not in that from British North Borneo) is that hairs are inserted along the edges of the veins of the reticulum on the lower surface and appressed over the areolae. This is not to be confused with the minute tomentum developed from the epidermis of the areolae in some other species of subgenus Pseudo- jaca. The characters distinguishing Artocarpus dadah from A. fulvicortex and A. tomentosulus are noted below under those species. The only other species occurring within its area with which A. dadah might be confused is A. fretessii, which extends into the eastern part of Borneo. This differs in the more prominent, straw-coloured reticulum of the leaves, in the paler indumentum, crisped on the twigs, and also in the smaller male inflores- cences and the lobed syncarp. Characters distinguishing A. dadah from A. vrieseanus are given below. 33. Artocarpus vrieseanus Miq. Ann. Mus. Lugd.-Bat. 3: 212. 1867. Artocarpus cumingiana auct. non Tréc., Diels, Bot. Jahrb. 67: 176. 1935. Trees, height to 28 m. Twigs 2-4 mm. thick, pubescent to subglabrous, hairs red-brown to yellowish, + appressed and crisped or undulate, vary- ing patent and straight or hooked at the tip. Leaves 7-30 * 3-15 cm obovate-oblong to elliptic or ovate-elliptic, varying narrowly oblong, oblong-lanceolate or elliptic, acute or with an acumen to 2.5 cm. long, base narrowly to broadly rounded or shallowly cordate, margin entire; main veins prominent beneath, reticulum not at all to markedly prominent; glabrous above except for the usually pubescent main veins, venation beneath mod- erately to sparsely pubescent, varying subglabrous, hairs colourless, weak and undulate or crisped, or some, rarely almost all, stouter and hooked at the tip; lateral veins 8-14 pairs, curved; intercostals parallel, varying reticulate; dark green, drying brown, grey-green or greenish, venation con- colorous, varying straw-coloured; petiole 5-18 mm. long Inflorescences solitary in leaf-axils or the male heads sometimes borne 96 JOURNAL OF THE ARNOLD ARBORETUM (VOR: 01 on short-shoots on older wood. At anthesis: male head (3—-)5-15 mm. across, globose to short-obovoid; perianths of 3 or 4 segments, free or fused for half their length, 0.4-0.5 mm. long; stamen 0.75 mm. long, filament tapering above, anther-cells globose, 0.2 mm. long; bracts stoutly stalked, heads peltate, to 0.6 mm. across, these and perianths puberulent; peduncle 2-15 & 1 mm., velutinous, hairs usually crisped, varying subglabrous; female head with peltate bracts nearly all shed and styles exserted to 0.4 mm. through perforations in the surface (in var. papillosus through papil- lae). Syncarp to 6 cm. across, subglobose, yellow, drying orange to brown or black, the surface smooth, velutinous, often glabrescent, varying sub- glabrous and then clearly areolate; wall c. 2 mm. thick; proximal region of perianths fused, fruiting perianths numerous, thin-walled, “seeds” (in- durated pericarps) ellipsoid, variously compressed, 11 % 8-10 mm.; core 10-15 mm. across; peduncle 3-35 % 3 mm., velutinous, hairs usually crisped, varying subglabrous. DISTRIBUTION: in rain forest to 6000 ft.; Philippines (Mindanao), Celebes (Manado) ?, Moluccas, New Guinea, Bismarck Archipelago, Japen, Salawati, Aru Islands, Louisiade Archipelago, Solomon Islands. This widely distributed species exhibits a range of variation in the type and abundance of the indumentum, the length of the peduncles, the shape of the leaves and the prominence of their venation which is greater than in any other member of subgenus Psewdojaca. Four variants can be distin- tinguished, each with a consistent geographical distribution, and, al- though three of these are based primarily on vegetative characters, it seems desirable to describe them all at the varietal level, and so to bring some order into the rather chaotic variation presented by this species at first sight. The most widely distributed of the varieties is var. refractus, which ex- tends from Mindanao through New Guinea to the Louisiade Archipelago, and has fairly broad, obovate-oblong (varying to elliptic) leaves with moderately prominent venation and a variable indumentum, a relatively large male head (5-15 mm. across), and variable peduncles (3-15 mm. in the male inflorescence and (?2—)15-30 mm. in the syncarp). Specimens from a limited area in the mountains of eastern New Guinea, which have rather small, narrow leaves with a more strongly developed reticulum, are assigned here, but are discussed further under the variety. In the Solomon Islands, var. papillosus is distinguished primarily by the papillate surface of the female head at anthesis, but also by the often ovate-elliptic leaves. The indumentum is variable, as is the length of the peduncles (3-5 mm in the male inflorescence and 5-35 mm. in the syncarp), and the male head is rather smaller than in var. refractus (3-8 mm. across). The two other varieties recognized in Artocarpus vrieseanus have ranges overlapping that of var. refractus in New Guinea, but are distinguished by the thinly pubescent to subglabrous leaves, which are often rather thinly coriaceous and narrow in outline, and by the small male heads (4-7 mm. across). The type variety extends from northeastern New Guinea west- 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, IV SH ward to the Moluccas and perhaps Celebes (with one doubtfully identified collection from Bougainville). The leaves have a slender, slightly prominent reticulum, and these and the twigs are thinly pubescent or subglabrous; except in a few almost entirely glabrous specimens some of the hairs are consistently hooked at the tip (such hairs being found otherwise only in a few collections of var. refractus). The peduncles are rather long, meas- uring (4—)6-17 mm. in the male inflorescence and 15—35(—65) mm. in the syncarp. Finally, var. subsessilis, which is found in New Guinea, the Bis- marck Archipelago and Bougainville, has glabrous leaves lacking a prom- inent reticulum. Only the young twigs are puberulent, with straight or crisped, but not hooked, hairs, and the inflorescences have short peduncles (2-3 mm. long in the male inflorescence and 3—7(-13) mm. in the syn- carp). The biological significance of these varieties, apart from var. papillosus which has presumably arisen through geographical isolation, is not clear. The three others appear to be genuinely distinct, and not merely adult or sapling forms. Whereas var. refractus occurs widely in New Guinea, var. vrieseanus and var. subsessilis have not yet been recorded from the south- ern regions, and these entities may be separated in northern New Guinea by slightly different ecological requirements, although no evidence in sup- port of this can be gained from field notes. The species that appears to be most closely related to Artocarpus vrieseanus is A. xanthocarpus, from the northern and central Philippine Islands, and the characters distinguishing the latter are discussed below, under that species. Some of the varieties of A. vrieseanus may, however, bear a strong superficial resemblance to two apparently less closely allied species, A. dadah and A. fretessii, and the variability of A. vrieseanus is such that only the complete fusion of the perianths in the syncarp provides an entirely satisfactory taxonomic distinction. The west Malaysian A. dadah, in addition to having the proximal region of the perianths free, differs in the colour of the syncarp, which is green with deep pink flesh (in- stead of yellow), in the straight hairs on the twigs (sometimes hooked at the tip), and in the more numerous lateral veins on the larger leaves (to 20 instead of 13(—14) pairs). The range of Artocarpus fretessii, which extends from Borneo and the Philippines to the Vogelkop of New Guinea, overlaps that of A. vrieseanus to a considerable extent, but the syncarps of these two species, also, can usually be distinguished externally. In A. vrieseanus the syncarp is sub- globose, with numerous seeds, but in A. fretessi it is distinctly lobed, each lobe usually containing one of the relatively few (up to 12) seeds. There is a further difference between the species in the interfloral bracts of the male heads: in the former rather few of these have well-developed, peltate heads, with stout stalks expanding gradually into the head, and there are numerous intermediates to the perianth segments, whereas in the latter there are many peltate bracts with slender stalks and well-defined heads, and few intermediates. These distinctions, and also the complete fusion of the perianths in the syncarp, have been confirmed for all the varieties 98 JOURNAL OF THE ARNOLD ARBORETUM [| VOL. XLI recognized in A. vrieseanus. No satisfactory vegetative characters can be found serving to distinguish this species and A. fretessi over the whole of their ranges, but notes are given below, under the three varieties with a range overlapping that of A. fretessi, that will, it is hoped, aid in the iden- tification of sterile material. KEY TO THE VARIETIES OF ARTOCARPUS VRIESEANUS 1. Leaves without a prominent reticulum, glabrous; young twigs puberulent, the hairs da or crisped; male peduncle to 3 mm., —_ peduncle to 7(-13) m GOR fp .o5 ae ere ee a ew ee we es r. subsessilis. Leaves ae the reticulum slightly to markedly prominent Sens reticulum and twigs usually pubescent, or if subglabrous, some of the hairs hooked at the tip. 2. Leaves rather thinly coriaceous, often narrow, the reticulum slightly prominent beneath; reticulum and twigs subglabrous to thinly pubescent, except when nearly glabrous some of the hairs hooked at the tip; male peduncle (4—)6-17 mm., syncarp peduncle 15-35(-65) mm. long. ...... Sib ao ada bead yee ba beeen ban ea ee ae oa var. urieseanus, Leaves moderately to thickly coriaceous, the reticulum distinctly prom- inent beneath; reticulum and twigs Puree hanes subglabrous, the hairs undulate or crisped, rarely a few hooked at the t 3. Female head papillate at anthesis ee papers gate eeaaat peal gees eoriay ds aed RNa A a edie doe hay ae Ge aes var. papillosus. 3. Female head smooth at anthesis. .................. var. refractus ay i) var. refractus (Becc.) Jarrett, stat. nov. Artocarpus refracta Becc. For. Borneo, 630. 1902. Holotype, Aru Islands, Beccari s.n. (¥1); isotypes (F ) Twigs pubescent or rarely subglabrous, the hairs appressed and crisped, varying patent and straight or undulate, rarely a few stouter and hooked at the tip. Leaves obovate-oblong, varying obovate-elliptic, elliptic or oblong, sometimes narrow; main veins prominent beneath, the reticulum moderate- ly, sometimes markedly, prominent, pubescent or rarely subglabrous, the hairs undulate or crisped, rarely a few stouter and hooked at the tip. /n- florescences at anthesis: male head 5-15 mm. across, peduncle 3-15 mm. long; female head with the surface smooth. Synucarp, peduncle (?2—) 15-30 mm. long. DISTRIBUTION: in primary and secondary forest from sea level to 6000 {t.; Mindanao, Moluccas, New Guinea, Salawati, Aru Islands, Louisiade Archipelago. Philippines. MINDANAO. Ahern 676 (Bo, 2). Bukidnon: Mt. Katanglad, Sulit PNH 9951 (A, L, 6). Davao: De Mesa FB 27486 (kK, 9); Todaya, Cope- land 1244 (p, 6); Todaya, Mt. Apo, Elmer 10932 (A, BM, BO, K, L, 2). Zambo- anga: San Ramon, Hallier 4671a (1). Moluccas. Opt. Atasrip 45 (L ae New Guinea. VoceLKop. Manokwari: Bostuin Tafelberg, Siena BW 3830 (L, 2). Sorong: Sausapor, Versteegh BW 4619 (1, 6). DutcH NortH New GuINEA. Idenburg River, Bernhard Camp, Brass & Versteegh 14004 (A), 1960 | JARRETT, ARTOCARPUS AND ALLIED GENERA, IV 99 14021 (A, 2); Mamberamogebiet, Pionier Bivac, Lam 730 (Bo, K, L, U, 6, DutcH SoutH New GUINEA. Bian River, Branderhorst 275 (80, K, 9): Sungei Aéndua, Mimika, bb 32900 (x, L). Papua. Central Division: Tawarere, Brass 674 (A, K, 6); Mafulu, Brass 5184 (a, Bo, us, 2), 5394 (A, BO, K, US, 2); Sogeri, Forbes 86 (BM, L, 6). Northern Division: Isuarava, Carr 15351 (A, BM, KU SING. 0p) 0592 (pM, K, L, SING, $), 15762 (BM, K, L, SING): Cie omen. of Saiho along road to Divinikoari, Hoogland Macdonald 3492 (A, K, L, 2); Tufi papeeae halfway between Wanigela and Itoto, Hoogland 4822 (a, BM, 9). Western Division: lower Fly River, e. bank, opposite Sturt Island, Brass ee (A, L, 6, 2). MANDATED TERRITORY OF New GUINEA. Madang District: Kani- pebirce! Schlechter ee (A, K, L, 6, 2). Morobe District: Matap, Clemens 11174 (a, 2); Ogeramnang, Clemens 4550, 5138 (a, 6, 2); Quembung mission trail to Sattelberg, eee 1244 (a, 2); Sattelberg, Clemens 7584 (A). SALAWATI. Kaloal, Koster BW 4253 (1). Aru IsLANps. Giabu- lengan, Beccari, May 1873 (F1, 6, 2). LourstapE ArcutP. Rossel Island, Jinju, Brass 28909 (ACL, 6): With the type of Artocarpus refractus, which came from the Aru Islands, Beccari cited an additional specimen, PB 74, from Kapaor, Papua Onin, but this is apparently not in the Herbarium Universitatis Florentinae and has not been found elsewhere. Quite a wide range of variation is found within this variety, particularly, as noted above, in some collections from the mountains of eastern New Guinea. Elsewhere in New Guinea, var. refractus has not been seen from above 3000 feet (although one of the collections from Mindanao was made at 5500 feet), but from the Central and Northern Divisions of Papua, and the Madang and Morobe Districts of northeast New Guinea several collec- tions from an altitude of 3000 to 6000 feet have been seen. Some of these (Brass 5184, 5394, Forbes 86, Carr 15351, 15552, 13762 and Clemens 11174) have rather small, narrow leaves characterized by a very strongly developed, netted reticulum and could, perhaps, be regarded as represent- ing a distinct entity. However, the remaining four (Schlechter 17065 and Clemens 4550, 5138 and 7584) are intermediate to the typical form, having rather broadly elliptic leaves with a netted, but less prominent, reticulum. It is not possible to give satisfactory characters for distinguishing male or sterile collections of Artocarpus vrieseanus var. refractus from A. fretes- sii, apart from the difference in the bracts mentioned above. However, in the former the male heads are usually somewhat larger, with shorter peduncles relative to the size of the head; they are also only rarely borne on short-shoots, whereas this is very common in the latter. On the whole, A, fretessii tends to have a smaller leaf, with a more distinct, often straw- coloured reticulum, and somewhat more ascending lateral veins, but rather similar leaves are found in the mountain form of var. refractus. In the eastern Moluccas and New Guinea, A. fretessii is also distinguished by the rather markedly cordate base of the leaf. var. papillosus Jarrett, var. nov. Ramuli juniores pubescentes, pilis appressis crispatisque, vel patentibus et rectis undulatisve. Folia ovati-elliptica, elliptica vel obovati-elliptica; 100 JOURNAL OF THE ARNOLD ARBORETUM [VOL.. XLI costa nervi lateralesque subtus prominentes, venulae prominulae, pubes- centes, pilis plus minusve undulatis, ad subglabri. Jnflorescentiae ad anthesin: capitula mascula 3-8 mm. diametro, pedunculis 3-5 mm. longis: capitula feminea superficie plana. Syncar pia pedunculis 5-35 mm. longis. Hototype: Solomon Islands, Kajewski 2360 (a); isotypes (BM, kK, L, P). DISTRIBUTION: in rain forest from sea level to 4000 ft.: ; Solomon Islands. Solomon Islands. BoUGAINVILLE. Kugumaru, Buin, Kajewski 1920 (a, kK, 6). GUADALCANAL. Berande, Kajewski 2436 (A, BM, K, P, 2); Vulolo, Tutuve Mt., Kajewski 2501 (a, BM, P, 9). Matarra. Quoimonapu, Kajewski 2360, Dec. 1930 (A, BM, K, L, P, 6, 2), SAN CristopaL. Balego-nagonago, Brass 2826 (A, K, L, ee Seed formation in this variety is often somewhat irregular, so that the syncarp is slightly lobed, while the surface between the lobes remains papil- late as in A. fretessii; the lobes, however, differ from those found in the latter species in being shallow and ill defined. var. vrieseanus Artocarpus vrieseana Mig. Ann. Mus. Lugd.-Bat. 3: 212. 1867; Renner, Bot. Jahrb. 39: 369. 1907. Syntypes, Batjan, De Vriese s.n., Celebes, Manado, De Vriese sn. (L); lectotype, Batjan, De Vriese s.n. L gees antiarifolia Becc. For. Borneo, 630. 1902. Holotype, Jobi [ Japen ], Beccari sn. (¥1); isotype (F1). oe ee Tréc. var. stenophylla Diels, Bot. Jahrb. 67: 177. 1935, Holotype, northeast New Guinea, Ledermann 12863 (B); isotype (K). Twigs thinly pubescent to subglabrous, except when nearly glabrous some hairs patent or recurved and hooked at tip, the rest patent to crisped. Leaves oblong-elliptic to elliptic or obovate-elliptic, often narrowly so; main veins prominent beneath, reticulum slightly so, indumentum as on twigs. Inflorescences at anthesis: male head 4-7 mm. across, peduncle (4—) 6-17 mm. long; female head with the surface smooth. Syncarp, peduncle 15— 35(—65) mm. long. DISTRIBUTION: in primary and secondary forest from sea level to 3000 ft.; Celebes (Manado) ?, Moluccas (Batjan), New Guinea. aor (?) NortH PENINSULA. Manado, De Vriese (L, 2). Moluccas. Bat- . De Vriese (L, 6, 2). w Guinea. Duress NortH New Guinea. Geelvink Bay, Nabire, Kanehira & eae 11538 (A, Bo, 6), 11584 (a, Bo, 9); Sarmi, Tor River, Dirdjan. Leden BW 5364 (1, 6, 2), MANDATED TERRITORY OF New GUINEA. Madang District: near the Gogol River, near Mawan village, Hoogland 4928 (a, kK, L, 8, 2); Kani-gebirge, Schlechter 17854 (A, K, L, 4); Wobbe, Schlechter 16439 (A, K, L, 6). Sepik District: Kaiserin Augusta [Sepik] River, Felsspitze, Leder- mann 12863, Aug. 1913 (BM, K, 6). JAPEN. Ansus, Beccari s.n., Apr. 1875 (#1, é); near Senki Kaunda, Aet & Idjan 568 (1, 8); Serui, Watibu bb 30249 (BO, L, SING, The type material of Artocarpus vrieseanus in the Rijksherbarium, 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, IV 101 Leiden, consists of six sheets purporting to come from both Celebes and Batjan, but comparison of these indicates that only three gatherings are involved and that there has been some confusion in the labelling. The two sheets labelled as coming from Celebes, one of which has an attached submature syncarp (peduncle 55 mm.), match exactly the leaves and de- tached syncarp (peduncle also 55 mm.) of one of the sheets stated to be from Batjan. Since the other two gatherings are both labelled as from Batjan the record from Celebes appears somewhat doubtful at present. The second gathering consists of two sheets with immature male inflor- escences attached and in an envelope (the head to 7 mm. across and the peduncle to 17 mm. long). A twig on one of these sheets, which bears a leaf and a syncarp comparing closely with the first gathering, was prob- ably mounted here by mistake, and the other sheet is therefore chosen as the lectotype. The third gathering, which is very similar to the first, is represented by a sterile sheet from Batjan (matched by another sheet from Hasskarl’s herbarium not annotated by Miquel). All the material is un- doubtedly referable to the entity under consideration, since the perianths in the syncarp are completely fused, the bracts in the male head are stoutly stalked, and the leaves are thinly coriaceous with a slightly prominent reticulum. No hooked hairs were found on the leaves or twigs, but this is true of a few other almost completely glabrous collections of var. vriese- anus, such as Kanehira & Hatusima 11584, which is a good match for the first gathering mentioned above. A collection from Kugumaru, Buin, Bougainville, Kajewski 1940 (a, M, K, L, P, SING, 6, @), may be referable to this variety; the leaves are rather broadly elliptic, and these and the twigs are definitely pubescent, but many of the hairs are hooked at the tip and the young syncarps are smooth, not papillate as in var. papillosus. A sterile collection from Ambon, De Fretes sn. (CAL, GH, L, U), which likewise has pubescent leaves but abundant hooked hairs, may also be referable here since the leaves are, in addition, scarcely cordate (see below); it is, however, discussed further under A. fretessii. The latter may be distinguished vegetatively by the definitely pubescent leaves with a more prominent reticulum, the absence of hooked hairs (occasionally a few present in specimens from the Vogel- kop), and, in the eastern Moluccas and New Guinea, by the rather markedly cordate leaf base. var. subsessilis Jarrett, var. nov. Ramuli juniores, puberulentes, pilis appressis et rectis crispatisve, mox glabrescentes. Folia elliptica vel obovati-elliptica, saepe angusta; costa nervi lateralesque tantum subtus prominentes et puberulentes vel glabri. Inflorescentiae ad anthesin: capitula mascula 5-7 mm. diametro, pedun- culis 2-3 mm. longis; capitula feminea superficie plana. Syncarpia pedun- culis 3-7 mm. (Floyd 3453 ad 13 mm.) longis. Hootype: northeast New Guinea, Hoogland 4999 (a); isotypes (K, L). DISTRIBUTION: in primary and secondary forest from sea level to 5500 ft.; New Guinea, Bismarck Archipelago, Solomon Islands (Bougainville). 102 JOURNAL OF THE ARNOLD ARBORETUM [VOL Sees New Guinea. VoGELKop. Steenkool, road to Tembuni, van Royen 3591 (1). DutcH Nortu New Gurnea. Hollandia, Holtekang, Brouwer BW 1539 (1, 2): Wissel Lake region, Lake Tigi, Eyma 4883 (1, 2); Wissel Lake region, foot of Mt. Bubiro and Enarotali, Eyma 5124 (1, 2). Papua. Milne Bay District: Cape Vogel Peninsula, Menapi, Brass 21660 (A, 2). Northern Division: Kokoda, Carr 16420 (1); Tufi Subdistrict, near Koreaf village, Hoogland 4813 (a, BM, K, L, ?). Manpatep Territory or New GUuINEA. areal District: Gogol River valley, near Jal village, Hoogland 4999, July 1955 (a, K, L, 6, 2). Morobe Dis- trict: Lae, Botanic Gardens (“indigenous”), ca NGF 9079 (A, L, 2) New Britain. Keravat, Floyd 3453 (A, BM, K, L, 2). Solomon Islands. Boucatn- VILLE. Teop Island Waterhouse 46 (x, 4). This variety can readily be distinguished from Artocarpus fretessii by the almost entirely glabrous leaves lacking a prominent reticulum. we a Artocarpus xanthocarpus Merr. Publ. Gov. Lab. Manila 17: 10 1904, “‘xanthocarpa,” Philip. Jour. Sci. 1, Suppl. 43. 1906: Elmer, Leafl. Philip. Bot. 2: 626. 1909. Holotype, Luzon, Whitford 367 (PNH, destroyed); isotypes (kK, P, us); lectotype (Pp), Artocarpus lamellosa Blanco, nomen dubium, Elm. Leafl. Philip. Bot. 2: 625. 1909 Artocarpus lanceolata auct. non Tréc., Merr. Enum. Philip. Pl. 2: 42. 1923. Artocarpus rubrovenius auct. non Warb., Merr. Philip. Jour. Sci. Bot. 3: 401. 1908 Small trees, height to 8 m. Twigs 1.5-3 mm. thick, smooth or finely rugose, appressed-puberulent, soon glabrescent. Leaves 5.5-20 « 2.5-9 cm., obovate-elliptic, varying obovate- or elliptic-oblong, with an acumen to 3 cm. long, base cuneate, varying narrowly rounded, glabrous, margin entire; main veins prominent beneath, intercostals slightly so; lateral veins 6-11 pairs, curved; intercostals few, not parallel; green, drying brownish or greenish, venation straw-coloured, reddish or nigrescent in young leaves: petiole 5—23 mm. long. Inflorescences solitary or paired in leaf-axils. At anthesis: male head 3— 6 < 3-4 mm., globose to obovoid; perianths tubular, bilobed above, 0.5 m. long; stamen | mm. long, filament cylindric, tapering above, anther- cells ellipsoid, 0.2 mm. long; bracts slenderly stalked, heads peltate, to .3 mm. across, these and perianths ciliate; peduncle 2-3 « 0.5 mm.., velutinous; female head with peltate bracts mostly shed and styles exserted to 0.8-1 mm. through low papillae. Syncarp to 5 cm. across, subglobose, shallowly lobed, yellow, drying pale or reddish brown, the surface smooth or nearly so, velutinous, with a few persistent bracts; wall c. 3 mm. thick: proximal region of perianths fused, fruiting perianths several, thin-walled, “seeds” (pericarps with an indurated endocarp) ellipsoid, 8 « 6 mm.;: core c. 4 mm, across; peduncle 6-11 * 3 mm., velutinous. DISTRIBUTION: in forest to 1300 ft.; Mangsi Islands, northern and cen- tral Philippine Islands. Borneo. Mancsi Istanps. Wilkes (GH, ¢, 2). Philippine Islands. M1nporo. 1960 | JARRETT, ARTOCARPUS AND ALLIED GENERA, IV 103 Merritt FB 9894 (Bo, us, 6); Mansalay, Mt. Yagaw, Conklin PNH 17465 (a, L, PNH, 6). BATANES IsLaNps. Fenix BS 3581 (Bo, 6), 3814 (Ny, US, 2); Mt. Traya, Ramos BS 80012 (k, Ny, 6, 2), 80305 (kK, NY, 2). Luzon. Benguet: Leano FB 24715 (ny, us, 2). Cagayan: Klemme FB 6670 (x, Ny, us, ¢). Bataan: Lamao River, Borden 183 (A, Ny, 6); Lamao River, Mt. Mariveles, Whitford 367, June 1904 (xk, p, us, 2). Laguna: Majayjay, Curran & Merritt B 8055 (xy, Pp, us, 6). Tayabas: Labitag FB 25414 (A, K, P, US, Camarines: Aguilar FB 14345 (us, 6); Camarines Sur, Iriga, Vidal 1539 (A, K, 3); Mt. Bagacay, Ramos & Edano BS 33925 (SING, Q). Sorsogon: Irosin, Mt. Bulusan, Elmer 16247 (BM, GH, K, L, 6). BoHoL. Ramos 42581 (BM, BO, P, SING, 2). SIQUIJOR ISLAND. Piper 398 Ck Pe Oe Artocarpus xanthocarpus is, as noted above, apparently most closely allied to A. vrieseanus, from which it differs in the longer styles, exserted to c. 1 mm. at anthesis instead of only 0.4 mm., and the slenderly stalked bracts in the male head. Otherwise it is rather similar to A. vrieseanus var. subsessilis, but it may be distinguished by the base of the leaf, which is usually cuneate and slightly decurrent, instead of rounded or auriculate. The small, entirely glabrous leaves, lacking a prominent reticulum, give A. conthocaerer a strong superficial resemblatice to A. nitidus ssp. nitidus, which also occurs in the Philippines, and distinguishing characters are given below, under the latter entity. The type material of A. xanthocarpus bears mature syncarps only, but styles 1 mm. long have been found persisting on the specimen at the Muséum National d’Histoire Naturelle, Paris. The Wilkes expedition col- lection from the Mangsi Islands off the northeastern tip of Borneo repre- sents, at present, a rather outlying locality for the species. However, al- though the specimen is in rather poor condition, it, too, has the long styles on the syncarp and the characteristic, very small male inflorescences. The collections from the Batanes Islands are all distinguished by having a very long, slender acumen to the leaf. 35. Artocarpus fretessii Teysm. & Binnend. in Hassk. Abh. Naturf. Ges. Halle 9: 189. 1866, “Fretissi’’; Merr. Interpr. Rumph. Herb. Amb. 191. 1917. Metrosideros spuria Rumph. Herb. Amb. 3: 26. ¢. 13. 1743. Antiaris fretessii Teysm. & Binnend. Cat. Hort. Bog. 84. 1866, nomen nudum. Artocarpus dasyphylla Miq. Ann. Mus. Lugd.-Bat. 3: 212. 1867; Renner, Bot. Jahrb. 39: 369. 1907; J. J. Smith, Ic. Bogor. 3: 83. 1907. Syntypes, Celebes, Riedel HB 5841, Teysmann HB 5787 (wv); lectotype, Teysmann HB 5787 (u Artocarpus erythrocarpa Teysm. ex Miq. Ann. Mus. Lugd.-Bat. 3: 212. 1867, pro. syn. Bae rumphiana Becc. For. Borneo, 636. 19 Artocarpus dasyphylla Miq. var. flava J Je ne Ic. Bogor. 3: 85. ¢. 234. 907. Artocarpus leytensis Elm. Leafl. Philip. Bot. 1: 279. 1908, 2: 622. 1909; Merr. Enum. Philip. Pl. 2: 42. 1923. Holotype, Leyte, Elmer 7243 (PNH, de- stroyed): isotypes (A, BO, K); lectotype (k). 104 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI Artocarpus paloensis Elm. Leafl. Philip. Bot. 1: 280. 1908, 2: 621. 1909. Holo- type, Leyte, Elmer 7244 (pNH, destroyed); isotypes (A, Bo, K); lectotype K). Artocarpus rotundifolia Elm. ex Merr. Enum. Philip. Pl. 2: 42. 1923, pro syn. Trees, height to 40 m., buttresses small or none, bark pale brown, peel- ing off in flakes. Twigs 1.5-4 mm. thick, pubescent, hairs pale yellow to rufous, usually subappressed and crisped. Leaves 6-29 « 3-12 (-32 16) cm., obovate-oblong to elliptic, acute, attenuate or acuminate, base broadly cuneate to shallowly cordate (varying to deeply so in the Moluccas and New Guinea), margin entire; juvenile leaves pinnatifid; main veins and reticulum distinctly prominent beneath; glabrous above or nearly so ex- cept for the pubescent main veins, venation beneath moderately to sparsely pubescent, hairs colourless, straight or slightly undulate; lateral veins 9- 13 pairs, curved, basal 2—4 pairs crowded; intercostals parallel; dark green above, pale green or greyish beneath, usually drying greyish or brownish, paler beneath, venation concolorous, varying straw-coloured; petiole 5— 15 mm. long. Inflorescences solitary or paired in leaf-axils, or more frequently on short-shoots on older wood. At anthesis: male head 3-7 mm. across, sub- globose or obovoid; perianths of 2 or 3 free segments 0.4-0.5 mm. long; stamen 0.8 mm. long, filament slightly flattened, tapering above, anther- cells subglobose, 0.15 mm. long; bracts slenderly stalked, heads peltate, to .3 mm. across, these and perianths sparsely ciliate; peduncle 3-7 mm. long, short-pubescent; female head with peltate bracts mostly shed and styles exserted to 0.5 mm. through papillae. Syncarp to 4 cm. across, with one to several subglobose lobes, yellow, or dark red to purple, drying olive- brown to rufous, the surface smooth over the lobes, papillate between them, short-pubescent, with a few persistent bracts; wall c. 2 mm. thick over lobes; proximal region of perianths free, fruiting perianths l-c. 12 (New Guinea, 1 or 2), thin-walled, ‘“‘seeds” (thin, horny pericarps) subglobose, 8 X 7mm.; core c. 4 mm. across; peduncle 20-25(—30) mm., short-pubes- cent; VERNACULAR NAMES: maumbi, kelembi, Celebes; taewan, Ambon. DISTRIBUTION: in forest up to 2000 ft.; eastern Borneo, Philippines, Celebes, Moluccas, New Guinea (Vogelkop). Borneo. SOUTH AND SOUTHEAST BorNEO. Martapura, Tewingan, Boschbouw- prufstation 2108 (Bo, 6); Martapura, Twang Bangkal, bb 2475 (Bo, L, AND NORTHEAST BorNEO. Loa Haur, w. of Samarinda, ea Magee 6902 (1, 6, 2). BrittsH Nortu Borneo. Kamang- sian, Goklin 1302 (x, Philippine Islands. PALAWAN. Brooks Point, Addison Peak, Elmer Ae (A, BM, K, L, 6, 2); Puerto Princesa, Mt. Pu lear, Elmer 12944 (A, BM, K, L, Q). LEVTE. Franco FB 26409 (p), W enzel 811 (BM, GH, 2); Abuyog, Lake Danao, Krukeberg, Sept. = . 6); Palo, Elmer 7243, Jan. 1906 (A, BO, K, 2), 7244, Jan. 1906 (a, BO, K, 9). BULIRAN. McGregor BS 18663 (A, BO, P, SING, @). Panay. Ilo-ilo: oe Vidal 3833 (A, K, 6,2). MINDANAO. Surigao: Runos & Pascasio BS 34350 (BM, L, NY, SING, 9). CAMIGUIN DE MINDANAO. Ramos 1196 (BM, P, U, US, @). 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, IV 105 Celebes. NORTH PENINSULA. Gorontalo: Molinggapoto, bb 18022 (A, Bo, L). Minahassa: Manado, Koorders 19061 (Bo, L), Riedel HB 5841 (Bo, L, U, 2), 7264 (BO, P), Teysmann HB 5723 (Bo), 5863 (Bo, L, U); Sondaho, bb mies} BO, L); Tana Wangko, Teysmann HB 5787 (xo, v, 2): Kajuwatu, Koorders 19039, 19048, 19055 (Bo, L), 19058 (Bo, L, 2), 19431 (Bo); Lubu, Amurang, 19045 (Bo, L); Pinamorangan Mts., Koorders 19054 (Bo, Kk, L, 2); Ratahan, Koorders 19043 (Bo), 19060 (Bo, L, 2), 19306 ee Sembolei, Koorders 19062 (so, 2); Tondano, Koorders 19312 (Bo, 2). RAL CELEBES. Malili, V/FS Cel./V—216 no. 226 (Bo, L); Malili, Toli Tol, ne Cel./V-216 no. 109 (x0, 6); Malili, Usu, M7JFS Cel./III-55 no. 81 (Bo, 9), no. 238 (Bo, L, 6); Palu, Tomado, bb 28212 (Bo, L); Poso, Tokosondo, bb 17977 (A, BO). SOUTHWEST PENINSULA. Baleh-Angien, Teysmann HB 12359 (Bo, L, 2), 12481 (Bo); Bon- thain, Saluang, NJFS Cel./I-19 (so, L, 2); Maleku, bb 23909 (Bo, L). SouTH- EAST PENINSULA. Kolaka, Parso, bb 32510 (aA, Bo, L, 2). P. Muna. Labunti, bb 6041 (Bo, L, U, 2) Moluccas. TALAUD IsLANDS. Karakelang, e. of Beo, Lam 2626 (1, 2). HALMA- HEIRA. Djailolo, Tuguair, bb 23734 (A, BO, L); ee Beguin 1867 (Bo, L, 6, Ope aWe Cae eee 2303 (BO, K, L, SING, 2). BATJAN. Nanggapil, seme nom. TXCRGRO 3G) A ISLANDS. Mangoli, bb 29771 a BO, L); Sanana, Kali Waj Gaj, bb 28811 We BO, L, SING, 6); Taliabu, n. of Samia bb 29937 (A, BO, L, SING, 6). Buru. Wae Ula, bb 22801 (Bo, 3). New Guinea. VocELKop. Manokwari: Momi, bb 33417 (A, Bo, K, os Oransbari, Brouwer BW 2512, 2576, 2594, Mangold BW 2133 (1); Been Se 554 (L); Ransiki, Mioswaas, Koster BW 1270 (L); Ransiki, Warsuwi, Koster- mans 87 as K, L, SING, 9): Sidai, 65 km. w. of Manokwari, Koster BW 4450 (Lt, 6); Warnapi, bb 33627, 33629 (Bo, K, L). Sorong: Warsamson, 25 km. e. of Sorong, Schram BW 5904 (1). Cultivated. Mazaya. Singapore, Hort. Bot., Cantley 136 (kK, 6, 2), Ridley 3359 (BM, CAL, K, SING, 6, 2). Java. Bogor: Hort. Bot., VZJ G 105 (Bo, L, 2), Vill B 5 (1, &), Sutrisno 82 (cult. sub. VII G 105; origin Celebes), Oct. 1957 (L, $); Tjiliwung river, Kostermans, Aug. 1953 (, L). The nomenclatural type of Artocarpus fretessii is Metrosideros spuria [i] Taewan mas of Rumphius, for which Hasskarl provided an identifica- tion in 1866 in his key to the ‘Herbarium Amboinense” with the phrase “Artocarpus Fretissi T. & B. Teysm. in litt.” The identity of Rumphius’ Taewan mas with the species under consideration can be established with certainty from the plate (¢. 134), which shows the characteristic mature syncarps with rounded lobes and between these the papillate unexpanded surface. The artist has added a ‘‘calyx” to one of the fruits, but the like- ness is unmistakable and extends to the vegetative characters, while the description is also in agreement. Hasskar] offered no identification for the second plant, Taewan femina, treated by Rumphius under Metrosideros spuria, but, from the plate (¢. 13B), it appears to represent the same species at anthesis, when the female heads are globose with the entire surface papillate. In 1754, Linnaeus, in his key to Rumphius (Herb. Amb. 11 [resp. O. Stickman]), had incorrectly identified Metrosideros spuria as Ochna jabotapita L. (see Sprague, Proc. Linn. Soc. 165: 151-156. 1955, for a discussion of Linnaeus’ later treatment of this species). Merrill, in 106 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI his consideration of the “Herbarium Amboinense” (1917), mentioned Teys- mann and Binnendijk’s name, but was unable to refer Rumphius’ plant to any known species of Artocarpus. However, on the Bogor sheet of Beguin 1876, which was collected in 1921 in Halmaheira, there is a note to the effect that Beguin thought he had found Metrosideros spuria. Artocarpus fretessii antedates by one year and must replace the name A. dasyphylla Miq. (1867) under which the species has been known in In- donesia. Teysmann and Binnendijk’s epithet was based on the name of De Fretes, a resident of Ambon who presumably sent them the living ma- terial that appeared in their catalogue (1866) of the Hortus Bogoriensis under a nomen nudum, Antiaris fretessii, with the source given as Ambon and the vernacular name as faewan, The assumption that this plant was identified with Taewan mas and the generic name corrected to Artocarpus in Teysmann’s communication to Hasskarl is supported by a specimen at Leiden from the Hortus Bogoriensis, collected from a plant numbered V// G 105, which is referable to this species and bears the determination “Arto- carpus fretessti,’ with another note referring to Antiaris fretessii of the catalogue.* (A recent collection from this tree, however, gives the source as Celebes.) There is also in existence a sterile collection by De Fretes from Amboina, HB 5562, which bears on the sheet in the herbarium at Utrecht a reference to Metrosideros spuria. However, as already noted under Artocarpus vrieseanus, the identity of the collection is doubtful and it is perhaps better referred to that species. It was determined only as Artocarpus but was cited as A. lakoocha by Miquel (Ann. Mus. Lugd.-Bat. 3: 313. 1867) and under A. lakoocha var. malayana (= A. dadah) by King (Ann. Bot. Gard. Calcutta 2: 15. 1889 Artocarpus fretessii again shows considerable variability in both inflor- escences and vegetative characters, particularly in the colour of the syn- carp. Red or purple fruits are reported from Celebes and the Sula Islands, while yellow ones are recorded from Borneo, Celebes and the Moluccas. Although it would seem, from the frequent absence of syncarps from collec- tions for which their colour is reported, that the yellow male inflorescences may be mistaken for them in the field, the variation is undoubtedly gen- uine. Artocarpus dasyphylla was described by Miquel as having red fruits, and J. J. Smith (1907) described the yellow-fruited form as var. flava from a plant of unknown provenance in the Hortus Bogoriensis. There appears to be no other difference correlated with that of the fruit colour, nor is there any geographical separation of the two forms. If the two were to be main- tained as distinct, a new name would be required for the red form, since in Rumphius’ description the fruit was stated to be yellow. There does not, however, seem to be any real justification for erecting a new variety for such a colour variant. Variation in leaf shape, on the other hand, shows a definite pattern of distribution in which the leaf tends to become markedly cordate (and often *Tt seems best to take fretessi as the form in which Teysmann and Binnendijk in- tended to ee the name and to treat the spelling found in Hasskarl as an ortho- graphic erro 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, IV 107 somewhat elongate) in the Moluccas and New Guinea, and this is accom- panied by a reduction in the number of lobes on the syncarp (each of which usually contains a single seed). The distinctions between Artocarpus fretessii and A. vrieseanus are dis- cussed fully above, under the latter species. In the Philippines A. fretessii overlaps the ranges of two further species with pubescent leaves. The first of these, A. ovatus, may be distinguished by the relatively longer and nar- rower leaves with more numerous lateral veins (11-20 vs. 9-13 pairs), and by the patent hairs on the twigs. The second, A. Oe ear ie may be differentiated from all these species by the broad leaves with long petioles (25-35 mm. vs. 5-18 mm. 36. Artocarpus reticulatus Miq. Ann. Mus. Lugd.-Bat. 3: 213. 1867, “reticulata,” non Heyne ex Wallich, 1831 (nomen nudum), nec Hunter ex Ridley, 1909; Renner, Bot. Jahrb. 39: 369. 1907; Koord. Suppl. Fl. N. O. Celebes 2: ¢. 3, 3: 1. 1922. Syntypes, Celebes, Teysmann HB 5272 (vu), Ternate, De Vriese & Tevsmann s.n. (L); lectotype, Teysmann HB 5272 (v). Trees, height to 30 m., with small buttresses, bark rough, grey. Twigs 3-6 mm. thick, rugose, puberulent with hooked hairs, or glabrous. Leaves 15-35 & 8-16 cm., oblong-elliptic, acute or acuminate, base rounded or broadly cuneate, margin entire; main veins and reticulum prominent be- neath, the reticulum rather acutely so; glabrous, or the main veins above and the venation beneath scabrid-puberulent, the hairs hooked at the tip; lateral veins 12-15 pairs, curved, basal pairs slightly crowded; intercostals parallel; green, paler beneath, drying brown to yellow-green, usually paler beneath, venation concolorous; petiole 13-25 mm. long. Inflorescences solitary or paired in leaf-axils. At anthesis: male head (immature) c. 15 mm. across, obovoid; perianths of 3 or 4 free segments 0.2 mm. long; stamen . . . ; bracts slenderly stalked, heads peltate, to 0.3 mm. across, short-ciliate; peduncle 13 % 1 mm., short-pubescent; female head with peltate bracts mostly shed and styles exserted to 1.5 mm. through prominent papillae. Syncarp to 6 cm. across, subglobose or lobed, orange, drying brown, the surface verrucose throughout from conical papil- lae, or becoming nearly smooth over the lobes, pubescent, with scattered persistent bracts; wall c. 5 mm. thick over lobes; proximal region of peri- anths free, fruiting perianths several, somewhat fleshy, ‘‘seeds” (indurated endocarps) ellipsoid, 12 & 8 mm.; core c. 15 mm. across; peduncle 25-45 * 4 mm., velutinous. VERNACULAR NAMES: maumbi, maumbi sela, Celebes. DISTRIBUTION: in forest to 2000 ft.; Celebes, Moluccas (Ternate). Celebes. NORTH PENINSULA. Minahassa: Karoa, near Papo, ieee 19051 (Bo, L, 6); Lolombulan, near Pakuere, Koorders 19049 (Bo, L); Amurang, sine nom. et num. (Bo, 9); Amurang, Lobah Kolai, bb 17125 (A, a Lemo, bb 7536 (Bo); Ratahan, Koorders 19053 (Bo, L), Teysmann HB 35272 (BO.GAL, L, 108 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XLI ?); Pingsan, near Kajuwatu, Koorders 19046, 19047 (Bo, L). CENTRAL CELERES. Malili, Kawata, N7JFS Cel./V-88 (Bo, K, L, 6, 9). P. Muna. Lam- iko, bb 21754 (A, BO, L, @), Moluccas. TERNATE. De Vries & Teysmann (1). The inflorescence characters of Artocarpus reticulatus are intermediate between those of A. fretessii and A. subrotundifolius, but provide clear dis- tinctions from each of these species. The syncarp is usually strongly lobed as in A. fretessti, but the surface is more prominently papillate, and the styles are long-exserted as in A. subrotundifolius (to 1.5 mm. instead of to 0.5 mm.), though fewer seeds are formed than in the latter species. The male inflorescence, although it has been seen only when immature, ap- parently lies between the two other species in both the size of the head and the length of the peduncle. Vegetatively, A. reticulatus may be distin- guished from A. fretessii by the glabrous or nearly glabrous leaves and twigs, and by the rather acutely prominent reticulum. From A. vrieseanus var. urieseanus it is less readily distinguished, but the leaves are usually larger with more numerous, parallel intercostals, and a longer petiole (13- 25 vs. 5-15 mm.) Ww ~y Artocarpus subrotundifolius Elmer, Leafl. Philip. Bot. 1: 281. 1908, “subrotundifolia,’ 2: 619. 1909; Merr. Enum. Philip. Pl. 2: 43. 1923. Holotype, Leyte, Elmer 7265 (pNu, destroyed); isotypes (A, BO, K); lectotype (kK). Artocarpus nitidus auct. non Tréc., Ahern, Timber Tree Sp. Philip. 35. 1901, tabula sola Trees, height to 25 m. Twigs 4-6 mm. thick, smooth or rugose, red- brown to fulvous pubescent, hairs straight or undulate, subappressed or patent, slowly glabrescent. Leaves 18-36 11-22 cm. , broadly elliptic to obovate- oblong, short-acuminate, base cordate, varying broadly rounded and slightly oblique, margin entire; main veins and reticulum prominent beneath; glabrous or nearly so above except for the pubescent main veins, venation beneath pubescent, hairs colourless, straight, or some on the main veins stout and hooked at the tip, old leaves glabrescent, slightly scabrid: lateral veins 10-14 pairs, basal 2-3 pairs crowded; intercostals parallel; bright green above, duller beneath, drying red- brown to blue-grey above. paler beneath, venation usually concolorous: petiole (15—)25-35(-—65) mm. ong. Inflorescences solitary or paired in leaf-axils. At anthesis: male head (20—)25-50 & 20-35 mm., obovoid, ellipsoid or subglobose; perianths of 3 or 4 free segments 0.6 mm. long; stamen 1 mm. long, filament tapering slightly above, anther-cells subglobose, 0.2 mm. long; bracts rather slender- ly stalked, heads peltate, to 0.5 mm. across, these and perianths ciliate; peduncle (7- )15-22 * 3 mm., indumentum as twigs; female head with peltate bracts oe a and styles exserted to 1-2.5 mm. through conical papillae. Syncarp to c. 6 cm. across, globose, shallowly lobed, drying brown, the surface ee or becoming nearly smooth, pubescent, with scattered persistent bracts; wall c. 2 mm. thick; proximal region of peri- 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, IV 109 anths free, fruiting perianths numerous, thin-walled, “seeds” (indurated endocarps) subglobose, 14 mm. across; core c. 30 mm. across; peduncle 40(?)-75 & 5-7 mm., indumentum as twigs. DISTRIBUTION: in forest to 1000 ft.; Philippine Islands. Philippine Islands. Luzon. Zambales: Mt. seme Fox PNH 4677 (A, 9). Camarines: Alvarez FB 21235 (kK, us, 2). Sorsogon: Curran FB 10541 (Bo, K), 10542 (ny, us, 2), Vidal 3838 (x, od Tee Mt. Bulusan, Elmer oe (A, BM, K, L, 6, Q), Samar. Lasquety FB 27032 (Bo), Ramos 1604 (pM, GH, ZA P, SING, s), ce 25774 (ep, 2). Leyte. Rosenbluth FB 12789 (x, NY, us, 6, 2), Wenz Ae (BM, GH, 6, ?), 1576 (A, BM, GH, 8); near Gacao, Glassaan 794 (A, d, 2); Palo, Elmer 7265, Jan. 1906 (A, BO, K, 3). MINDANAO. Surigao: Wenzel Pe (A, BO, GH, K, ¢), Ramos & Pascasio BS 34757 (Ny, ¢). The maximum size attained by the male inflorescence in this species is considerably larger than in any other species of subgenus Pseudojaca, but the dimensions of both the head and the peduncle are rather variable. How- ever, the characteristic, broad, long-petiolate leaves enable male and sterile collections to be assigned to this species with certainty. The well-exserted styles at anthesis and the large syncarps, usually with long peduncles (but measuring only 7 mm. at anthesis in Glassman 794, cf. the male inflores- cence), are equally distinctive. Older leaves and twigs are often almost completely glabrescent, and, since growth usually occurs in flushes, there may be a considerable difference in appearance between young and old shoots. [To be concluded | Be - -_ JOURNAL OF THE ARNOLD ARBORETUM VoL. XLI APRIL 1960 NUMBER 2 STUDIES IN ARTOCARPUS AND ALLIED GENERA, IV. A REVISION. OF ARTOCARPUS SUBGENUS PSEUDOJACA * FRANCES M. JARRETT 38. Artocarpus lakoocha Roxb. FI. Ind. 3: 524. 1832, “Lakoocha’; Graham, Cat. Bombay Pl. 193. 1839; Wight, Ic. Ind. Or. 2: ¢. 681. 1843; Tréc. Ann. Sci. Nat. Bot. III. 8: 117. 1847; Dalz. & Gibson, Bombay Fl. 244. 1861; Brandis, For. Fl. N.-W. & C. Ind. 426. 1874, pro max. parte; Kurz, For. Fl. Burma 2: 433. 1877; King in Hook f. Fl. Brit. Ind. 5: 543. 1888, pro parte; King, Ann. Bot. Gard. Calcutta 2: 14. 1889, p.p., quoad ¢. 13, excl. spec. Griffith 4666, Main- gay 1479; Renner, Bot. Jahrb. 39: 370. 1907; Kanjilal, For. FI. Siwalik & Jaunsar, 379. 1911; Duthie, Fl. Upper Gang. Plain 3: 141. 1915; Troup, Silvicult. Ind. Trees 3: 883. fig. 326. 1921; Par- kinson, For. Fl. Andaman Is. 253. 1923; Haines, Bot. Bihar & Orissa 5: 824. 1924: Parker, For. Fl. Punjab ed. 2. 487. 1924; Osmaston, For, Fl. Kumaon, 504. 1927; Fischer in Gamble, Fl. Madras 3: 1369. 1928, p.p., quoad spec. Gamble et Lushington; Parker & Singh, Common Indian Trees, 26. ¢. 20. 1933; Kanjilal et al. FI. Assam 4: 268. 1940; Benthall, Trees Calcutta, 401. 1946. Holo- type(?), Bengal, Roxburgh s.n., Sept. 1812 (BM). Artocarpus lacucha Roxb. Hort. Beng. 66. 1814, nomen nudum Artocarpus bengalensis Roxb. ex Wall. Cat. no. 4655C. 1831, nomen nudum. Artocarpus reticulata Heyne ex Wall. Cat. no. 4655D. 1831, nomen nudum. Artocarpus mollis Wall. Cat. no. 4661. 1831, nomen nudum. Artocarpus yunnanensis H. H. Hu, Bull. Fan Inst. Biol. Peking Bot. 8: 32. 1937. Holotype, Yunnan, Wang 77078 (PE, not ca oe (A). Artocarpus ficifolia W. T. Wang, Acta Phytotax. Sin. 6: osad OSS RPA eek Holotype, Yunnan, Exped. Biol. Sino-ross. ad ee 676 (PE ?, not seen) ; isotype Gop Deciduous trees, height to 20 m., bark rough, grey or brown. Twigs 3-6 mm. thick, shallowly rugose, densely pubescent, hairs yellow to rufous, patent or subappressed, straight. Leaves 13-37 6-21 cm., ellip- tic, oblong or ovate, short-acuminate, base broadly cuneate, rounded or * Continued from volume XLI, p. 109. £12 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI subcordate, often oblique, margin entire or denticulate towards the apex; juvenile leaves shallowly pinnatifid; main veins and reticulum prominent beneath, the areolae often slightly bullate; glabrous or nearly so above, venation beneath pubescent, hairs colourless to pale rufous, undulate; lateral veins 9-18 pairs, straight or curved; a. parallel: green, drying greyish, greenish or pale brown, venation straw-coloured, small- est meshes of reticulum nigrescent; petiole (10—)15-45 mm. lone: Inflorescences solitary in leaf-axils. At anthesis: male head 12-25 X (7-)10-18 mm., ellipsoid, obovoid or clavate; perianths of 2(or 3) free segments, 0.5 “ns long; stamen 0.9 mm. fone filament tapering above, ie cells ellipsoid, 0.15 mm. long; bracts rather stoutly stalked, heads peltate, to 0.5 mm. across, these and perianths ciliate; peduncle 2- 2 mm., short-pubescent; female head with styles exserted to 1—-1.5 mm. through low papillae emerging between peltate bracts. Syncarp to 6 cm. across (to 12 cm. fide Winit 301), subglobose, shallowly lobed, yellow, drying brown, the surface irregularly papillate, pubescent, with numer- ous persistent bracts; walls c. 3 mm. thick; proximal region of perianths free, fruiting perianths several, fleshy, “seeds” (pericarps with a thin, horny endocarp) ellipsoid, 10 & 6 mm.; core c. 10 mm. across; peduncle 8-15(-25) & 4 mm., short-pubescent. VERNACULAR NAMES: lakuch (Sanskrit), barhal (Hindi), dahu or dheu ery India; myauklok, Burma; hat lom, hat non (Lao), Siam. Uses: the tree is often planted, especially in northern India, for its edible fruit. DISTRIBUTION: in evergreen, semi-evergreen and moist deciduous for- est to 6000 ft., in areas with a distinct dry season; eastern and northern India (Madras, Orissa and Bihar to Assam and Chittagong, and west- ward along the sub-Himalayan tract to East Punjab), Burma, Andaman Islands, Siam, Indochina, Yunnan; cultivated through much of its range and south to Bombay in India, sparingly introduced elsewhere in the tropics. India. Not localized: ‘East India,” Roxburgh (x, 29); Himalaya, Tikoli Val- ley, Edgeworth 216 (x, &); Nouholly, Hooker f. & Thomson, Dec. 1850 (x, L, U, P); Punkabarry, Gamble 1179A (x, 6); Tenga Ghats, sine nom. 1204 (kK). East Punyas. Kangra, Bhadwar, Koelz 4367 (A, 6). HrMAcCHAL PRADESH. Lower Kanaor [= Kunawara], Royle (GH, P, 9). UTTAR PRADESH. Kumaon: above Kota, Strachey & Winterbottom (GH, K); Outer Hills, Strachey & Win- terbottom 18 (GH, K, 4). Mapras. Ganjam District, Khalingia Ghat, Gamble 13699 (K); Vizagapatam, Rangalu Shola, Lushington, June 1914 (kK, 9). Brear. Hazaribagh, Meebold 3873 (x, 6, 2); Singbhum, Karampoda forest, Haines 636 (K, @). Orissa. Nilghiris, Pierre 6 (e). Sikkim: Hooker f. (K, 2°); below Kasseong, Hooker f., Apr. 1850 (k, 8); Rangit River, Clarke 27255 (kK), King, July 1876 (xk, 2). BENGAL. Roxburgh, Sept. 1812 (pm, ?), Voigt (c, ¢), sine Luckhipoor, Clarke 7018 (A); Haflong Hill, Blatter 29156 (K, 2); Khasia Hills, Hooker f. & Thomson (c); Naga Hills, Prain, 1886 (ccE); Naga Hills, Lang, 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, IV ini) Meebold 7156 (kK, 6). Manrtpur. On the way to Chamoo [? = Tamu], Watt 5167 (CAL). CuHITtTaconc. Hill Tracts, King 340 (sinc, 9), 417 (L, sING, 3), g.m., 1882 (L); Khana, Hooker f. & Thomson (xk). Burma. Diamond Island, Prain, 30 Nov. 1889 (caL); Myaungmya, Labwuta, OCeeeL OTL (Kee) Phanac,. fl. Salua n [? = Salween River], Wallich 4661 (cGE, K, P, 6, 2); Rangoon Buchanan Hamilton (pM), Dickason 6947 (A, 6); Ran- goon, Kamayut, Dickason 8256 (A, 6, 2); Sandoway, near Taungup, Lace 2930 (K, 6). TrNaAssERIM. Amherst: Moulmein, Falconer 991 (L). Mergui: Griffith 1053 (xk, @). Andaman Islands. Ali Masfid Reserve, Parkinson 385 (x, 2); Baradang, Parkinson 205 (pp, 6); Boru-Lung-Da, Parkinson 928 (pp,2). Siam. Dan Sai, Kao Keo Kang, Kerr 5804 (pm, 6); Me Kok, Muang Fang, Kerr 5158 (3m, 6); Me Lee, rea init 301, 302 (BM, Dy PENINSULAR Sram. Trang, Chawng, Buncoed 20 (cc Indochina. Laos. Pac Bac, near see Prabang, Poilane 20478 (Pp, 2); be- tween Phinh Ha and Lao Phu Tai, Poilane 25926 (p, 6); Vientiane, Poilane 20782 (Pp, 2). Tonxin. Lao Kay prov., between Nam Long and Phouy Tho, iss 25496 (p, 6). China. YUNNAN. Anderson, 1875 (CAL), Forrest 12252 ‘K, 6); Chen-Kang Hsien, Wang 72645 (a, ¢); Chin- -ping, Meng-la, Exped. a Sino-ross. 676, Apr. 1956 (a, 3); Fo-Hai, Wang 74902 (a, 2), 77078, June 1936 (A, 2); Lan-Tsang Hsien, Wang 76647(a, @); Mienning, Nanya, Yu 18102 (a, é): Shunning, Hila, Yu 16805 (a, 2); Szemao, Henry 11746 (A, kK, $, 2). The distribution of Artocarpus lakoocha has generally been given as India, Ceylon, Burma and Malaya, but three corrections to this must be made. Firstly, it was shown by Corner in 1939 (Gard. Bull. Singapore 10: 282) that the entity hitherto identified in Malaya as A. lakoocha in fact represented A. dadah, a species of western Malaysia. As already noted, the misidentification originated in King’s description of this taxon as A. lakoocha var. malayana, which he did not distinguish satisfactorily from the type. His account of the latter is a mixture of the two species, but the plate was based on Roxburgh’s original drawing of A. lakoocha (as was the plate in Wight’s “Icones”). No collections of this species have been seen from farther south than peninsular Burma and Siam. Secondly, it has been found that in India, also, two different taxa have been confused under the name Artocarpus lakoocha. This species appears to occur as an indigenous tree only in northern India and along the East- ern Ghats, although it is planted more widely. Several collections have been seen from the Bombay area, but none, apparently, is from a wild tree. This is supported by a statement by Graham in 1839 (the sole in- formation traced in the literature) to the effect that A. lakoocha was only found in a cultivated state in Salsette and the North Concan. On the Western Ghats and in Ceylon there occurs another entity which has usually been identified as A. /akoocha. However, it has been found that it is quite distinct and it is described above as A. gomezianus ssp. zey- lanicus, the differentiating characters being given there. Artocarpus la- koocha was described from Bengal and the type and drawing leave no doubt as to the application of Roxburgh’s name. The references in which the two entities have been confused are indicated above as far as possible; — 114 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XLI Fic. 19. Distribution of some species of subg. PsEupoJAcA. 38, Artocarpus ae 39, A. rubrovenius; 40, A. fulvicortex; 41, A. tomentosulus ; 42, A. Jae "43, ‘A. nitidus, a, SSp. nitidus, b, ssp. lingnanensis, c, ssp. hu milis and eee d, ssp. griffithii, dots, records not identified to subspecies; 44, ri petelotii: 45, A. hypargyreus; 46, A. styracifolius; 47, A. altissimus., only those which were based primarily on A. gomezianus ssp. zeylanicus are cited under the latter. Finally, the area of Artocarpus lakoocha must be extended eastward through Siam to Indochina, and to Yunnan from which it has been de- scribed twice without reference to Roxburgh. Most of the Indochinese collections were made after Gagnepain wrote up the Moraceae for the “Flore Générale de L’Indochine” (1928), and he did not mention 4. lakoocha. The distinctions between this species and A. tonkinensis are given above, under the latter entity. 39. Artocarpus rubrovenius Warb. in Perkins, Frag. Fl. Philip. 166. 1905, “rubrovenia”; Merr. Philip. Jour. Sci. 1, Suppl. 43, 1906, Enum. Philip. Pl. 2: 43. 1923; Elmer, Leafl. Philip. Bot. 2: G22 1909; Brown, Useful PI. Philip. Ls 470, 1941. Holotype, Tucon Warburg 12949 (B). Trees, height to 15 m., bark mottled grey and brown, Twigs 2.5-5 mm. thick, appressed-puberulent, soon glabrescent. Leaves 11-26 & 5-14 cm., ovate to elliptic, with an acumen to 2.5 cm. long, base broadly rounded 1960 | JARRETT, ARTOCARPUS AND ALLIED GENERA, IV 115 to broadly cuneate, often oblique, glabrous, margin entire; main veins prominent beneath, intercostals slightly so; lateral veins 8-13 pairs, curved; intercostals parallel; green with yellowish-white main veins, dry- ing brown or blue-grey above, paler brown beneath, venation reddish or nigrescent in young leaf, straw-coloured when mature; petiole 15-30 mm. ong Inflorescences solitary or paired in leaf-axils. At anthesis: male head 15-45 & 10-20 mm., obovoid to clavate; perianths of 2-4 segments, free or fused at the base, 0.5 mm. long; stamen 0.7 mm. long, filament cylin- dric, abruptly contracted above, anther-cells ellipsoid, 0.2 mm. long; bracts stoutly stalked, heads peltate, to 0.3 mm. across, these and perianths pubescent; peduncle 1.5-3 *& 1.5 mm., velutinous; female head with styles exserted to 0.5—-0.8 mm. through papillae emerging between peltate bracts. Syncarp (submature) to 3 cm. across, subglobose or shallowly lobed, drying brown or cinereous, the surface nearly smooth, short- pubescent, with numerous persistent bracts; proximal region of perianths free; peduncle 5-10 & 3 mm., velutinous. DISTRIBUTION: in forest to 1200 ft., Luzon (? also in Mindoro). Philippine Islands. Luzon. Haencke 433 (Ny, 2). Quezon (Tayabas): Baler, Merrill 1034 (x, us, @), eee PNH 2511 (A, PNH, SING, 2); Casiguran, Ramos & Edano BS 45226 (Bo, Ny, 6); Lagumanoc, Mer Hlieco90 (ke US, 6): Lucban, Elmer 9128 (A, BO, K, L, 4), McGregor BS 47395 GNAp SOs): Sapelor Warburg 12949, Mar. 1888 (B, 6, 2). Bataan: Curran 5439 (us); Lamao River, Mt. IMennelcn Borden FB 2562 (BO, K, SING, 2), 2915 (Ny, us, 3 2498 (BO, K, P, SING, 6). Pampanga: Camp Sess Mt. Pinatubo, Binicr 22030 (BM, , 6, 2). Rizal: Ahern FB 3197 (Bo, K, SING, US, 2); Morong district, Vidal one (x, us, 2); San Mateo, Vidal 1548 a Eee, ay; Bata angas: Ramos & Deroy BS eae 6 (A, BM, GH, K, L, P, SING, ¢). Gina Alvarez FB 21440 (BM, Bo, kK, P, 6, @), Ramos Loaf (aur. BO, GH, L, P, SING, De Albay: Banao, @umepatan: ieee PNH 18505 (1, PNH, 2). Sorsogon: Cua FB 10524 (k, sINc, 2), Vidal 3839 (x, 3); Irosin, Vidal BSS0 (Re eo) rosin, Mt. Bulusan, Filmer 15598 (2 BMT Osan), Sudit PNH 2753 (A, BO, PNH, SING, 2). Although, as was pointed out in the discussion of section Pseudojaca, Artocarpus rubrovenius shows a relationship with A. lakoocha, it is read- ily distinguished by the glabrous leaves lacking a prominent reticulum, and, at anthesis, by the shorter styles. The differences from the other glabrous-leaved members of subgenus Pseudojaca that occur in the Philippines, A. xanthocarpus and A. nitidus ssp. nitidus, are given under the latter entity. Merrill stated in 1923 that Artocarpus rubrovenius occurred on Min- doro and Cagayan Sulu, in addition to Luzon. The former record may be based on the two collections cited by Merrill that have not been seen, Darling FB 14704 and Merrill 1517. It is presumed, from the identifica- tion on the herbarium sheet, that the latter is based on Merrill 5304, although the number was not cited. This collection is here referred to A, gomezianus ssp. gomezianus. 116 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XLI 40. Artocarpus fulvicortex Jarrett, sp. nov. Artocarpus sp., Corner, Wayside Trees, 658, ¢. 197. 1940. Ramuli juniores puberulentes; folia late elliptica vel subrotunda, nervis lateralibus utrinque 6-10, nervis transversalibus paucis, costa nervis lateralibus venulisque subtus valde prominentibus, pubescentibus, inter- venio saepe minute pubescenti; inflorescentiae subsessiles; capitula mas- cula subglobosa, 4-6 mm. diametro; syncarpia globosa, superficie plana, pubescentia, bracteis persistentibus. Arbores [ad 35 m. altae|, deciduae, cortice nova fulva vel rufi-brunnea. Ramuli juniores 4-8 mm. crassi, subrugosi, puberulentes, cortice in sicco mox squamis decidua; [ramuli juveniles pubescentes|. Folia c. 15 X 10 [10-20 x 7-14] cm., late elliptica vel subrotunda, obtusa vel breviter acuminata, basi cuneata [vel rotunda], integra, supra subglabra, costa nervis lateralibusque pubescentibus exceptis, subtus costa nervis laterali- bus venulisque valde prominentibus, pubescentibus, pilis rufis, interdum sparsis, intervenio minute pubescenti vel puberulenti, [foliis juvenilibus subglabro], supra saturata viridia, subtus subglauca, in sicco utrinque rufi-brunnea vel caesia; nervi laterales utrinque 7 [6-10], curvati; nervi transversales pauci, paralleli vel reticulati; petiolus 20 [15— 05) mm. longus. Inflorescentiae axillis foliorum solitariae vel geminae. Ad anthesin: capitula mascula 4-7 * 4-6 mm., globosa vel obovoidea; perianthia pro- funde bifida, 0.7 mm. longa, breviter ciliata; stamina 0.8 mm. longa, filamentis supra attenuatis, cellis antherum ellipsoideis, 0.15 mm. longis; bracteae crasse stipitatae, capitibus peltatis, ad 0.5 mm. latis, breviter ciliatis; pedunculus 0.5-1.5 & 1 mm., breviter pubescens; canine femt- nea stylis inter bracteas peltatas crebras 0.5 mm. exsertis. | Syncar pia ad 5 cm. diametro, globosa, fulva, carne lutea, in sicco rufi-brunnea, superficie plana, pubescenti, bracteis numerosis persistentibus, inconspicuis, in indumento immersis; stratum externum syncarpii c. 5 mm. crassum; “semina” (endocarpia cornea) plura, obovoidea, 10 X 5 mm., perianthiis omnino conjunctis inclusa; axis syncarpii c. 15 mm. dieeietio: peduncu- 4 mm., breviter pubescens. | (Inflorescentiae typi anthesin atque syncarpium maturum (Corner SFN 34663) ab eadem arbore collectum omnes spiritu vini conservae descriptae.) Hototryre: Malaya, Corner SFN 33686 (SING). DIsTRIBUTION: in lowland evergreen forest to 250 ft.; Malaya, Sumatra, Banka. Malaya. Perak. Batu Gajah, Corner, Aug. 1936 (stNG); Slim River, Corner, Aug. 1937 (sinc), PAHANG. Kemansul For. Res., Symington KEP 49822 (KEP). NecrRr SEMBILAN. Seremban, Gale KEP 63363 (KEP, 9). Mavacca. Main- gay 1485 (K). SINGAPORE. Cantley (sinc); Cluny Road, Ridley 4829 (BM, CAL, SING, 2); Chanchu Kang, Ridley 4129 (kK, SING, ¢, @): Krangl, Coodensivh 3379 (CAL, SING); ne, end of MacRitchie Resencnr Sma SFN 38916 (SING, L, 6); Reservoir Jungle, Corner, Jan. 1937 (sinc). Sumatra. PALEMBANG. Ban- 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, IV Li? juasin, Bentajan, NJFS T 1030 (Bo); Lematang Ilir, Gunong Megang, N/JFS T 364 (Bo, L, 2), 592 (Bo, L, vu, $, 2). Lamponcs. Kotabumi, bb 35005 (Bo); Tulang Bawang Ilir, Bandjar Agung, bb 8951 (Bo). BANKA. Blinju, Grashoff 20 (BO, L); Koba, Teysmann HB 7242 (Bo, c, CAL, K, L); Muntok, Batu Balai, Teys- mann HB 6843 (BO, CAL, K, L, P). Cultivated. Matava. Singapore, Hort. Bot. (all from tree in potting yard), Corner SFN 33686, Sept. 1937 (sinc, 6, 9), 34501, 34663 (SING, 2), Merah SFN 33549 (sinc). This species was described as Artocarpus sp., “Orange-barked Tam- pang,” in 1940 by Corner, who stated (p. 650) that the bark colour was distinctive, since it was grey or brown in other Malayan species of the genus. The broadly elliptic or subrotund leaves, with few lateral veins and a markedly prominent reticulum on the lower surface, have a charac- teristic appearance which enables this species to be identified readily when sterile. However, no previous description under a scientific name has been found, and specimens have been identified as either A. lakoocha or A. dadah. Artocarpus fulvicortex differs from the latter in the leaf characters given above, and also in the very short peduncles of the in- florescences, the numerous bracts persisting on the syncarp and the nearly glabrous young twigs. 41. Artocarpus tomentosulus Jarrett, sp. nov. Differt ab A. fulvicortice foliis ellipticis, elliptici- vel ovati-oblongis, nervis lateralibus utrinque 10-14, nervis transversalibus plurimis, syn- carpio pedunculo 25 mm. longo. Arbores ad 20 m. altae. Ramuli juniores 2.5-4.5 mm. crassi, rugosi, puberulentes. Folia c. 18 & 10 [11-23 %& 6-12] cm. elliptica ad elliptici- vel ovati-oblonga, acuminata [acumine ad 2 cm. longo], basi rotunda [vel late cuneata], integra, supra glabra, costa nervis laterali- busque pubescentibus exceptis, subtus costa nervis lateralibus venulisque prominentibus, pubescentibus, intervenio minute tomentoso [vel glabro], in sicco supra pallidi- vel rufi-brunnea, subtus pallidi-brunnea; nervi laterales utrinque c. 12 [10-14], curvati; nervi transversales plurimi, paralleli; petiolus 15{[-45] mm. longus. Inflorescentiae axillis foliorum solitariae. [Capitula mascula (imma- tura) ad 7 mm. diametro, globosa; perianthia tubulosa, 0.6 mm. longa, supra bilobata, minute pubescentia; stamina (nondum exserta) cellis antherum ellipsoideis, 0.15 mm. longis; bracteae tenuiter stipitatae, capi- tibus peltatis, ad 0.4 mm. latis, minute pubescentibus: pedunculus c. mm., brevissime pubescens.] Syncarpium (submaturum) 3 cm. diametro, subglobosum, fulvum, in sicco brunneum, superficie parum papillosa, pubescenti, bracteis peltatis plurimis persistentibus; pedunculus 25 X 3.5 mm., breviter pubescens. HototyPe: British North Borneo, Keith 7671 (SING). DISTRIBUTION: in forest to 2000 ft.; endemic to northeastern Borneo. 118 JOURNAL OF THE ARNOLD ARBORETUM [VObs, ut Borneo. EAST AND NORTHEAST BorNEO. Berouw: bb 18909 (a); Bekmuari, bb 19133 (A, Bo, L). Bulungan: Mara, bb 10806 (so). BritisH NorTH Borneo. Kabili-Sepilok For. Res., Keith 7671, Sept. 1937 (sinc, 2); Kalabakan, 30 miles wnw. of Tawau, Wood SAN A 3684 (A, KEP, L, SING, yi The material of Artocarpus tomentosulus is very limited, but the col- lections are well characterized by the leaves, which have fairly closely set, spreading lateral veins with numerous parallel, rather slender intercostals. The type, Keith 7671, bears a submature syncarp, and this specimen and 6b 10806, which is sterile, have the intervenium minutely tomentose. The remaining collections, Wood SAN A 3684 (with male inflorescences) , bb 18909 and bb 19133 (sterile), have a glabrous intervenium. It is pos- sible that the latter are taken from juvenile shoots, since the leaves are somewhat larger than in the previous collections. If these three collections are matched with the other two on the basis of the venation, as seems justifiable, the characters of this species correspond with those of 4. fulvi- cortex in the small, subsessile, more or less globose male heads, and in the combination of prominent, patent-pubescent venation on the lower sur- face of the leaf with a frequently minutely tomentose intervenium, although the syncarp peduncle is considerably longer (25 mm. instead of 4 mm.). These leaf characters are unique in the subgenus; in other species with a tomentose intervenium (A. glaucus, A. hypargyreus and A. styracifolius ) the venation is less prominent, and the main veins, at least, are subglabrous. Collections of A. dadak from Borneo (which can be distinguished by the narrower leaves and the pubescent twigs) may also appear to have hairs on the intervenium of the leaf, but these are nearly all inserted along the edge of the veins of the reticulum and appressed over the areolae. 42. Artocarpus glaucus Blume, Bijdr. 483. 1825, “olauca’,; Tréc. Ann. Sci. Nat. Bot. III. 8: 121. 1847; Miq. in Zoll. Syst. Verz. Ind. Archip. 2: 90, 96. 1854; Miq. Pl. Jungh. 293. 1854, FI. Ind. Bat. AZ) 258: 1859, Ann. Mus. Lugd.-Bat. 3: 212. 1867; Koord. & Val. Bijdr. Boomsoort. Java 11: 26. 1906; Backer, Beknopte Fl. Java 6: 16. 1948. Neotype, Java, Zollinger 704 (P). Artocarpus glaucescens Tréc. Ann. Sci. Nat. Bot. III. 8: 120. 1847; Miq. in Zoll. Syst. Verz. Ind. Archip. 2: 90. 1854; Miq. Fl. Ind. Bat. 1(2): 288. 1859, Ann. Mus. Lugd.-Bat. 3: 212. 1867; Renner, Bot. Janrd.. 392 369, 1907. Holotype, Java, Zollinger 704 (p); isotypes (BM, GH, K, L, P). Artocarpus zollingeriana Miq. in Zoll. Syst. Verz. Ind. Archip. 2: 90, 95, 1854; Miq. Fl. Ind. Bat. 1(2): 289. 1859. Syntypes, Java, Zollinger 512 (®), Bogor, Hort. Bot., Zollinger s.n. (u); lectotype, Zollinger 512 (P). Artocarpus biformis Miq. FI. Ind. Bat. Suppl. 419. 1861. Holotype, Sumatra, Teysmann HB 4444 (u); isotypes (Bo, L). Artocarpus tephrophylla Miq. Fl. Ind. Bat. Suppl. 422. 1861. Holotype, Sumatra, Teysmann 4504 (vu); isotypes (BO, L). Artocarpus inaequalis Teysm. & Binnend. Cat. Hort. Bog. 382. 1866, nomen nudum. 1960 | JARRETT, ARTOCARPUS AND ALLIED GENERA, IV 119 eee ae glaucescens Tréc. var. tephrophylla Mig. Ann. Mus. Lugd.-Bat. 3: 1867. ee denisoniana King in Hook. f. Fl. Brit. Ind. 5: 544. 1888; King, Ann. Bot. Gard. Calcutta 2: 14. ¢. 8B. 1889; Ridley, Fl. Malay Penin. 3: 355. 1924. Syntypes, ae King 10318, 10843, 10987 (CAL, not seen; duplicates examined, k, etc. erie glaucus Blume var. villosiusculus Warb. ex Renner, Bot. Jahrb. 39: 369. 1907, nomen nudum Evergreen trees, height to 40 m., buttressed or not, bark dark erey, peeling. Twigs 1. Ge 4 mm. thick, finely rugose, appressed- -puberulent, soon glabrescent. Leaves 4-33 2-16 cm. , elliptic to elliptic-oblong, varying to oblong, obovate, ovate, or ovate- lanceolate small leaves Sian rather narrow, acuminate or attenuate, base cuneate, varying to rounded, often slightly unequal and decurrent, margin entire; juvenile leaves lobed: main veins prominent beneath, intercostals slightly so; glabrous above, venation appressed-puberulent beneath, glabrescent, intervenium minutely tomentose (juvenile leaves subglabrous to short- -pubescent beneath) ; lat- eral veins 8-15 pairs, curved; intercostals parallel: old leaves becoming bullate above between venation; glossy dark green above, light green to dull grey beneath, drying yellowish to bluish grey above, greyish glau- cous beneath from the tomentum, young leaves brown with nigrescent main veins; petiole 7—25(—45) mm. long. Inflorescences solitary or paired in leaf-axils. At anthesis: male head 11-24 & 5-6 mm., narrowly oblong or clavate; perianths 2- or 3-lobed, divided nearly to the base, 0.6 mm. long; stamen 0.8 mm. long, filament cylindric, contracted above, anther-cells ellipsoid, 0.2 mm. long; bracts slenderly stalked, heads peltate, to 0.4 mm. across, these and perianths short-ciliate ; peduncle 1-3 & 1 mm., velutinous; leona head with styles exserted to 0.3 mm. through a dense covering of peltate bracts, later through papillae emerging between bracts. Syncarp to 3 cm. (fide Koor- ders & Valeton, 1906, to 7 cm.) across, subglobose, often shallowly lobed, yellow to orange-brown with light-orange flesh, drying dark brown to black, the surface papillate or becoming nearly smooth, short- pubescent, with numerous persistent bracts; wall c. 3 mm. thick: Bronal region of perianths free, ee perianths several, “‘seeds”’ (horny pericarps), ellip- soid, 10 & 7 mm.; core c. 10 mm. across; peduncle 2-5 mm., velu- tinous. VERNACULAR NAME: tiwu landu (Sundanese), Java. DISTRIBUTION: in evergreen forest to 2500 ft., Malaya, Sumatra, Simalur, Banka, Borneo, Java, Lesser Sunda Islands (Sumba, Roma, Tanimbar Islands). Malaya. KepaH. Katumbah, Meh 17874 (xk, stnc, 3). Perak. Ulu Bubong, King 10161 (A, SING, 2), 10318, June 1886 (BM, K, L, 6, 2), 10843, Aug. 1886 (BM, K, 2), 10987, 1886 (k, P, SING, 4, 2). Pan HANG. Bentong, imac CF 5052 (x, é); 3 miles s. of Kasia Lipis, Burkill & Haniff SFN 17168 (so, xk, SING, 6, 2). SELANGOR. Ulu Gombak, Murdoch 304 (SING, 6, 2). JoHore. 120 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI 1014 miles Kota Tinggi-Jemalaung road, Sinclair SFN 40348 (K, L, SING, ¢, 2). Sumatra. TAPANULI. Padang Lawas, Hatiran, Rahmat si Boeea 4883 (A, K, L, 2), 4920 (A, K, L, 6). West Coast. Priaman, Diepenhorst HB 1325 (xo, L, P, U). East Coast. Aer Kandis, near Rantau Parapat, Bila, Rahmat si Boeea 2467 (a, L, SING, 6, @); Sibolangit, Docters van Leeuwen-Reijnvaan 12711 (Bo, K, L, 6), Lérzing 5170 (Bo, 6). Dyampr. Danau Lama, bb 13643 (Bo); Simpang, bb 13115 (Bo), BENKULEN. Kroi, Gunong Nemala, bb 8758 (po); Ranaumeer, Talangtotong, between Simpang and Sepatuhu, ee pee 3378 (Bo, @). PALEMBANG. Lematang Ilir, Gunong Megang, N/JFS T 0) (Bo, L, 9); Musi Ilir, Muara Punojung, bb 9202 (Bo); Musi Ulu, es 159 (go). Lamponcs. Kalianda, Gunong Kadjabasu, De Wit 14 (Bo); Kalianda, Ketjapi, bb 8954 (Bo, L); Siring Kebou, Teysmann HB 4504 (po, L, u); Tara- bangi Ilir, Teysmann HB 4444 (Bo, L, U). SIMALUR. Achmad 966 (BO, K, L, SING, u, 6, 2); Landschap Tapah, Defajan, ie: 1405 (Bo, L, u, 6, @), 1675 (Bo, L, P, 6), 1721 (Bo, L, SING, 6, 2). BANKA. S. Banka, Rindik, bb 11585 (Bo). Borneo. SARAWAK. Sungei Sama, ae & Tachun SFN 35729 (SING, 2). WEST eos Ketapang, Riamdadap, bb 8306 (Bo); Sambas, Sai, bb 7086 (po, L, 6). SOUTH AND SOUTHEAST BoRNEO. Bandjarmasin, Motley 1279 (CGE, 2 ae Tewe, Dusun Ulu, Sungei Pararawen, Dachlan 2411 (1). BRITISH NortH Borneo. Wood 2159 (Bo, 6, 2); Beaufort south mile 18, Wood 2608 (stnc, 6); Tenom, Pangie, Angian Herb. For. Dept. B.N.B. 10482 (KEP, 6). Java. Hasskarl HB 1869 (1), De Vriese ‘< L, U), Zippelius 1332 (HB 7267) (so), Zollinger 512 (Pp, 6), 701 (K), s.m. (K, L, P, U). West JAva. Bantam: Gunong Kantjana, Koorders 41644 (Bo); Gunong Mung, Gunong Karang, Pulasari, Koorders 40118 (Bo, L); Tapos, oe HB 7268 (Bo); Tyibadui, Kosala, Forbes 450 (BM, BO, CAL, SING, 6, 2); Tjikoja, Zollinger 704 (BM, GH, K, L, P, 3), 706 (vu); Tjimara-Udjong Kulon, Gunong Marang, Koorders 8672 (Bo, L, P); Tjimara-Udjong Kulon, Gunong Rompang, Koorders 8671 (A, Bo, , P) 9932 (Bo). Buitenzorg: Leeuwiliang, Pasir Honje, Bakh. van den Brink 6953 (BO, K, L, P, SING, U, 6); Nangala, Bakh. van den Brink 7629 (Bo, L, 2); Nangala, Gunong Menteng, Bakh. van den Brink 7717 (Bo, K, L); Nangala, Gunong Tjiputih, Bakh. van den Brink 7726 (Bo, K, L, P, U, 6, %); Nangala, Tjilankop near Tjiputih, Bakh. van den Brink 7719 (Bo); Tjibinong, NIFS Ja 2702 (Bo, @). Preanger: Sukabumi, Tjisahong, Tjisalak, Koorders 39445 (no, K, L, P, @); Njalidung, UAl 6555 (Bo, 2). CENTRAL JAVA. Banjumas: Bandjar- negara, Pagedongan, NJFS Ja 2549 (A, Bo); Bandjarnegara, Pringombo, Koor- ders 8673 (Bo, L), 33881 (BO), 39184 (BO, K, L), 39194 (Bo, L, Pp). Pekalongan: Subah, Pesan, Koorders 27594 (Bo, L, P, 2). Semarang: Ungaran, Telomojo, Koorders 39234 (a, BO, K). EAST Java. Pastirany: Tangkil ee Koor- ders 23784 (BO). Lesser Sunda Islands. SumMBA. E. Sumba, Djuli, bb 15133 (Bo). Roma. Hila, bb 7219 iO). TANIMBAR ISLANDS. P. Jamdena, between Kampong Ilgnei and Otimmer, 2uwalda 4131 (K, L, 6, 2) = bb 24224 (A, BO, L, SING, 2). Cultivated. Bogor, Hort. Bot., Teysmann HB 7274 (no, P, 9), s.m., 1867 (BM, L, 2), Zollinger C0) No collections determined as Artocarpus glaucus by Blume himself have been found, but the description is clearly identifiable. The glaucous lower surface of the leaf distinguishes this species from any other member of the genus growing in Java, and, with the elongate, subsessile male head, 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, IV 121 from the rest of subgenus Pseudojaca. The leaves vary considerably, but continuously, in shape from tree to tree, and they also become markedly coriaceous with age. Several specimens bear what appear to be galled shoots. These are short and highly branched, and bear many inflores- cences, which are usually male, but are abnormally shaped. The heads are obovoid with peduncles to 5 mm. in length, and the flowers are often malformed. 43. Artocarpus nitidus Tréc. Ann. Sci. Nat. Bot. ITI. 8: 119. 1847. Evergreen trees, height to 35 m., buttressed or not, bark red-brown. Twigs 1-4 mm. thick, smooth or finely rugose, appressed-puberulent, soon glabrescent. Leaves 3.5-23 x 1.5—9 cm., elliptic, obovate- or oblong- elliptic, to obovate- or ovate-oblong, obtuse to attenuate or with an acumen to 2.5 cm. long, base cuneate or rounded, margin entire; main veins prominent beneath, intercostals slightly so; glabrous, or puberulent beneath on the main veins (rarely throughout in ssp. lingnanensis): lateral veins 5-15 pairs, straight or curved: intercostals few, usually parallel; dark green above, paler or yellowish green beneath; petiole 5—15(—25) mm. long. Inflorescences solitary or paired in leaf-axils. At anthesis: male head 3.5-12 XX 2.5-7 mm., oblong, obovoid or clavate; perianths of 2-4 seg- ments free or fused at the base, 0.5-0.7 mm. long; stamen 0.8-1 mm. long, filament broad, tapering or contracted above, anther-cells short- ellipsoid, 0.2 mm. long; bracts slenderly stalked, heads peltate, to 0.4 mm. across, these and perianths ciliate; peduncle 1-3 & 1 mm.; female head with the styles exserted to 0.5 mm. through a covering of peltate bracts, or through perforations or low papillae, the bracts scattered. Syncarp 1.5-6 cm. across, subglobose, the surface smooth, with scat- tered persistent bracts, the indumentum various; wall 1-5 mm. thick; proximal region of perianths free, fruiting perianths 1-12, thin-walled, “seeds” (pericarps with a horny endocarp), subglobose to ovoid, 8-10 « 7-8 mm.; core 5-8 mm. across; peduncle 1.5-4(-20) & 1.5-3 mm. ) DIsTRIBUTION: in forest or savannah to 5000 ft.; Assam, Burma, Siam, Indochina, southern China, Malaya, Sumatra, Borneo, northern and cen- tral Philippines. Four taxa which have previously been recognized as species (with sev- eral additional synonyms) are here reduced to subspecific rank under Artocarpus nitidus, since they are only separable from this species and from each other on the size and indumentum of the syncarp, and on slight differences in the shape and venation of the leaves. The type sub- species, which occurs in the Philippines, is readily distinguished by the very small, few-seeded, velutinous syncarp. The two subspecies placed following this, ssp. lingnanensis, which extends from southern China to peninsular Siam, and ssp. kumilis, which is restricted to Borneo, appear to be very closely allied to each other. They have larger, also velutinous t22 JOURNAL OF THE ARNOLD ARBORETUM [yOL, XL syncarps, and are distinguished only by minor, but characteristic, differ- ences in the leaves. However, in view of the variation found in the length of the female peduncle in ssp. Aumilis (discussed under the subspecies) , it seems best to retain them as distinct taxa of equal rank for the present. The remaining subspecies, which usually have more numerous lateral veins in the leaves than the three preceding entities, are ssp. borneensis, from Borneo, and ssp. griffithii, extending from Yunnan and Indochina to Sumatra and Borneo. They are likewise very closely related, differing only in the indumentum of the syncarp. In ssp. borneensis the surface is densely covered by minute hairs with readily deciduous, multicellular, clavate heads which, when fallen, often form a reddish powder around the dried syncarp. These hairs are presumably equivalent to the gland-hairs occurring generally within Artocarpus, although in subg. Pseudojaca the heads of the latter are usually unicellular. In ssp. griffithii, on the other hand, the syncarp is almost entirely glabrous. The occurrence of three subspecies in Borneo has led to difficulties in identification which are increased by the finding of specimens, all bearing very small, elliptic, long-acuminate leaves, which are referable on the characters of the syncarp to all three entities. Larger-leaved collections of ssp. Aumilis, lacking or bearing only male inflorescences, can be distin- guished on vegetative characters, but no such characters have been found that would separate ssp. borncensis and ssp. griffithii. Collections from Borneo which lack syncarps and which are referable to one or other of these subspecies, are listed separately under ssp. borneensis. ‘There are, in addition, a few male or sterile collections from Borneo and elsewhere that could not be identified to subspecies and have not been cited. On the evidence of specimens bearing syncarps (none seen from western Borneo), ssp. borneensis is relatively common in British North Borneo, whereas ssp. griffithii is absent there, though widely distributed outside Borneo. The provenance of the small-leaved collections appears to be variable, since Corner noted for ssp. griffithii (Wayside Trees, 654. 1940, as A, gomezianus) that such specimens came from the crowns of large trees, whereas Beccari derived the specific epithet for A. Aumilis, of which the type has small leaves, from its being a small tree. Three collections referable to Artocarpus nitidus have been seen from Burma and Assam, but they bear only male inflorescences and, from the shape of the leaves, might represent either ssp. /ingnanensis or ssp. grifithit. They are: Dickason 6938, 6981, Rangoon (A, 4); Kanjilal 4145, Makum Range, Barjan, Lakhimpur (x, ¢). A collection from the Lushai Hills, Sen Gupta 7643, is cited by Kanjilal et al., Fl. Assam 4: 269. 1940, as A, gomeziana Wall. var. griffithii King (= ssp. griffithit). KEY TO THE SUBSPECIES OF ARTOCARPUS NITIDUS 1. Syncarp to 1.5(-3) cm. across, velutinous, seeds 1-3(-6). ..... . ssp. nitidus. 1. Syncarp larger, seeds more numerous. 2. Syncarp velutinous; larger leaves with 5-11 pairs lateral veins. 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, IV 123 3. Leaves obtuse to shortly and bluntly acuminate. .. ssp. lingnanensis. 3. Leaves with an acumen to 2.5 cm. long, lateral veins often marked] SCEMCINOgmem eh) sae tem oe hs. 9 CU ee eee Peres ssp. ees 2. Syncarp not velutinous; larger leaves with 8-15 pairs lateral vein 4. Syncarp covered by readily deciduous, clavate hairs. .. ssp. Pees AP OV McAupesUDEADTOUSs ios. y s\n) ben Ee os ssp. griffithit. ssp. nitidus Artocarpus nitidus Tréc. Ann. Sci. Nat. Bot. III. 8: 119. 1847, “nitida’; Mig. Fl. Ind. Bat. 1(2): 288. 1859; Fern.-Villar, Noviss. App. 203. 1880; Vidal, Revis. Pl. Vasc. Filip. 254. 1886; Renner, Bot. Jahrb. 39: 368. 1907; Elmer, Leafl. Philip. Bot. 2: 624. 1909. Syntypes, Luzon, Cuming 1078, 1081 (p); lectotype, Cuming 1078 (Pp). Artocarpus lanceolata Tréc. Ann. Sci. Nat. Bot. III. 8: 121. 1847; Miq. FI. Ind. Bat. 1(2): 288. 1859; Fern-Villar, Noviss. App. 203. 1880; Vidal, Revis. Vasc. Pl. Filip. 255. 1886: Elmer, ea Philip. Bot. 2: 624. 1909. Holotype, Luzon, Callery 60 (Pp); isotype (P). Artocarpus lamellosa auct. non Blanco, Merr. Publ. Gov. Lab. Manila 27: 80. 1905, Sp. Blancoanae, 124. 1918, Enum. Philip. Pl. 2: 41. 1923. Leaves 4-13 & 2-5.5 cm., obovate-oblong, with an acumen to 1.5 cm. long, base rounded, varying broadly cuneate, margin entire; juvenile leaves to 16 XX 7.5 cm., denticulate towards the apex; lateral veins 6-9 pairs, curved; drying brown to blue-grey above, brownish or greenish below, the venation often straw-coloured, reddish in some young leaves. Male head 6-10 3 mm., oblong or clavate; peduncle 1-2 & 1 mm., velu- tinous. Syncarp to 1.5(-3) cm. across, drying light brown or cinereous, short-pubescent; seeds 1-3(—6); peduncle 2(-6) 1.5 mm., velutinous. DISTRIBUTION: in forest to 500 ft.; northern and central Philippine Islands. Philippine Islands. Luzon. Ilocos Norte: Burgos, Ramos BS 27283 (a, Bo, p, 6). Abra: Valera PNH 13846 (a, PNH, 2). Pangasinan: Lopez FB 24217 (A, 6, 2), Medina FB 13503 (us, 8). Rizal: Calawan, Callery 60, 1840 (p, 3). Cavite: Maragondong, Te 4167 (BM, L, P, US, 4). Batangas: Merrill SB 100 (A, BM, BO, GH, K, 6, 2); Lobo, Vidal 1540 (Ay Rolin cs © eeleagiuna: Mt. Makiling, Cs PNH ‘9802 (A, PNH, 3), Elmer 18279 (a, BM kor Lec. 2), Sulit PNH 16923 (PNH). ak Cuming 1078 (BM, CGE, K, L, P, SING, 4), 1081 (BM, CGE, K, L, P, SING, 6, 9). BLAS. Cortes & Rendal FB 17845 (kK, us, 6, 2). Cesu. Lopez & Reyes pp Fone (SING), Ramos BS 11022 (BM, x, 3). Panay. Miagao, Ilo-ilo, Vidal 3834 (a, Kk, 6, 2). Guimaras. Sulit P 11697 (A, L, PNH, 6, 2); Buenavista, Bo. Salvacion, So. Lande, Sulit PNH 11832 (A, L, 8, 2). Artocarpus nitidus was reduced by Merrill in 1905 to the earlier A. lamellosa Blanco which has, since then, been accepted as the correct name for the species (in the restricted sense of ssp. nitidus). In making this identification Merrill was following Fernandez-Villar, who in 1880 had also regarded the two species as synonymous, although he retained Trécul’s name. However, while Blanco’s description undoubtedly refers to one of the three glabrous-leaved members of subg. Pseudojaca that occur in 124 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI the Philippines, it does not agree with the entity under consideration. Artocarpus lamellosa was described as having a fruit the size of a hen’s egg, which was full of seeds, whereas in 4. nitidus ssp. nitidus the syn- carp does not usually exceed 1.5 cm. in diameter or have more than three seeds. The account could apply to either of the two other species, A. xanthocarpus Merr. (1904) or A. rubrovenius Warb. (1905), but no characters are given that would differentiate between them. In view of the long-standing misidentification it seems best, therefore, to reject A. lamellosa as a nomen dubium. Artocarpus nitidus ssp. nitidus has frequently been confused with A. xanthocarpus, but the latter is distinguished by the globose to obovoid, rather than oblong or clavate male head (3-6 & 3-4 mm. vs. 6-19 X Rs) mm.), by the longer styles (exserted to 0.8-1 mm. vs. 0.5 mm.), and by the larger syncarp (to 5 cm. vs. 1.5(-3) cm. across) with a longer peduncle (6-11 mm. vs. 2(-6) mm.) and several (instead of 1-3(-6)) seeds. In Artocarpus nitidus ssp. nitidus the leaves are usually obovate- oblong with a rounded base, and dry (when mature) blue-grey above and light brown beneath; in A. xanthocarpus the leaves are longer and narrower in outline, with a cuneate or sometimes narrowly rounded base, and they dry brown or greenish on both surfaces. Merrill, in 1923, re- duced A. xanthocarpus to A. lanceolata Tréc. but examination of Trécul’s type has shown that it is merely a collection of A. nitidus ssp. nitidus with rather narrow leaves, the male inflorescences being characteristic of the latter entity. Artocarpus rubrovenius can be distinguished from both these taxa by the larger male inflorescences and leaves, the latter often having a broadly rounded and oblique base. The specimen Borden FB 2946 (xo, sinc), which Merrill cited in his enumeration of the flora of the Lamao Forest Reserve (Philip. Jour. Sci. 1, Suppl. 43. 1906) under A. lanceolata (as a species distinct from A. nitidus and A. xanthocarpus) must be referred to Antiaris toxicaria Lesch. Elmer (1909) cited under A. lanceolata, with some doubt, Borden 1145 from the same locality, but, unless this is the field number of the same collection, no specimens have been seen in the course of this study. ssp. lingnanensis (Merr.) Jarrett, stat. nov. Artocarpus parva Gagnep. Bull. Soc. Bot. Fr. 73: 89. 1926; Gagnep. in Le- _ Fl. Gén. Indoch. 5: 735. fig. 90. 1928. Syntypes, Tonkin (culti- vated), Balansa 740, 4112, 4544, Bon 412, 466, Fleury 37765 (P); lecto- type, Balansa 4112 (®). Artocarpus sampor Gagnep. Bull. Soc. Bot. Fr. 73: 90. 1926, excl. syncarp.; Gagnep. in Lecomte, Fl. Gén. Indoch. 5: 738. 1928. Syntypes, Cambodia, Chevalier 31915, 36953, Fleury. 30049, Cochinchina, Pierre 1637 (excl. syncarp), Annam, Poilane 6879, Siam (cultivated), Thorel 2784 (p); lecto- type, Poilane 6879 (P). Artocarpus lingnanensis Merr. Lingnan Sci. Jour. 7: 302. 1931, 13: 56: 1934, Syntypes, Kwangtung (cultivated), McClure 13471 (uc, not seen; dupli- cates examined, K, P), Honam Island, McClure s.n., Sept. 1924 (uc, not seen). 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, IV 129 Leaves 6-14 X 3.5-8 cm., elliptic to oblong- or obovate-elliptic, the outline often somewhat irregular, obtuse or shortly and obtusely acumi- nate, base rounded or cuneate, usually slightly decurrent, margin entire or shallowly and irregularly crenate; lateral veins 6-10 pairs, straight or curved; drying brown or grey, venation concolorous or, in young leaves, nigrescent. Male head 4-12 3-4 mm., obovoid, varying oblong; peduncle 1-1.5 & 1 mm., velutinous. Syncarp to 5 cm. across, red with pink flesh, drying brown, enous: seeds c. 5-12; peduncle 1.5-3 « 2 mm., velutinous, VERNACULAR NAMES: ma hat, Siam: sampor, Cambodia; cay chay, Annam; hung kwai muk, kwai he southern China. Users: planted Por the ecible fruit; the bark and roots are also added to betel. DISTRIBUTION: in evergreen and mixed forest, and in savannah, to 3000 ft.; Siam, Indochina, southern China (Kwangtung, Hainan); culti- vated, especially in Tonkin Siam. Chantaburi, Klung, Kerr 17956 (pm, 2); Chantaburi, Ma Kaun Kaim. Kerr 493 (pm, 6); Kao Ngon, Loi, Kerr 8795 (pm, 6); Ko Chang, Schmidt T2OMGye erie Collins 892 (us, 2). PENINSULAR S1AM. Krabi, Tambon Kao Panom, Kerr 18647 (pm, 2); Satul, Kuan Po, Kerr 13809 (pm, 8°); Surat, Tha Chang, Ratn 2373 oe ae Indochina. CAmBopia. Dee (Pp, 2). Kompong Chnang Prov.: nea mpong Chnang, Chevalier 31915, Mar. 1914 (p, @): Kralanh For. Res., ee so Mar. 1914 (Pp, 2); Prey Chang Ka Tamau, Chevalier 36953, Jan. 1918 (p, 2); Prey Phum Laung, Chevalier 36903 (?, 6). Stung Treng Prov.: betwe een ao Pranap and Chéom Khsan, Poilane 14881 (p, 6). CocHincutna. Bien Hoa Prov.: Song Lu, Pierre 1637, Mar. 1877 (kK, P, US, all with Soy of A. gomezianus ssp. gomezianus [ 5-10 cm., ovate to ovate- or obovate-oblong, acuminate, base usually shallowly cordate, margin glandu- lar-crenate; juvenile leaves with 1-2 pairs lateral lobes; main veins and intercostals prominent beneath, reticulum slightly so; puberulent above with the main veins short- pubescent, rather thinly pubescent on the venation beneath, glabrescent; lateral veins 5-9 pairs, straight, basal pair usually rather strongly developed and leaf base hence trinerved; intercostals parallel, basal tertiaries strongly developed; green (turning orange-yellow on the tree), drying blue-grey to brown above, paler brown beneath; petiole 10-25 mm. long, often geniculate at the junction with the lamina. Inflorescences axillary on short-shoots borne on older wood before the new leaves appear. At anthesis: male head 5-8 2-3 mm., ellipsoid or cylindric; perianths of 4 free segments 0.7 mm. long, puberulent; (stamen not exserted), anthers oblong, 0.4 mm. long; bracts slenderly stalked, heads peltate, to 0.6 mm. across, ciliate; peduncle 5—7 0.5 mm., velutinous; female head 5 mm. across, globose, the surface velutinous, with deeply bifid styles exserted to 1 mm. through papillae, and numer- ous peltate bracts; proximal region of perianths fused; peduncle to 12 < 2 mm., velutinous. Syncarp VERNACULAR NAME: klutum, Sumatra. DIsTRIBUTION: in evergreen forest to 1800 ft.; Sumatra, west Borneo. Sumatra. ATJEH. Langsa, A. Tjanang, bb 14504 (Bo). Tapanuti. Angkola and Sipirok, bb 5626 (Bo, L). East Coast. Simelungun, G. Maligas, bb 20436 (go, L). PALEMBANG. Heyne (Bo, L, 6, @), Praetorius (L); Batu Radja, Teysmann HB 3734 (so, L); Muara Dua, Teysmann HB 3903 (Bo, 1); Muara Enim, Teysmann HB 4042 (Bo, L); Sekaju Munie, Teysmann HB 3972 (go, L). Lamponcs. Kebang, Teysmann HB 4214 (no, L). Borneo. WEST BoRNEO. Sanggau, oe 18789 (Bo, L). Cultivated. Java. Bogor, Hort. Bot., HB 7241 (Bo), VIII B4 The material available of Artocarpus altissimus is inadequate for the preparation of a full description of the species. The type is sterile, as are all the other collections cited, except for one made by Heyne in 136 JOURNAL OF THE ARNOLD ARBORETUM [VOL Li Palembang which has inflorescences at anthesis. These agree with the description, based on material from two trees in the Hortus Bogoriensis, that was published by J. J. Smith in transferring Miquel’s Morus altissi- mus to Artocarpus, Smith stated that the syncarps did not mature on these trees, but I am informed by Mr. E. J. H. Corner that he found fruit on them containing seeds measuring c. 7 * 5 mm. The germination of these was characteristic of Artocarpus, since the cotyledons remained enclosed in the seed-coat, while the epicotyl lengthened and the first pair of foliage leaves were opposite, the subsequent ones being arranged spirally. The specimens match closely in their vegetative characters, which are aberrant for the genus in several respects. The leaves are palmately tri- nerved at the base, usually with a strong development of the basal ter- tiary veins, and the petiole is frequently geniculate at the junction with the lamina. The margin is crenate-dentate, with evenly spaced patches of glandular tissue lying at the end of short nerves running out to the edge of the leaf. These patches are present in addition to the glandular hairs with unicellular heads that are characteristic of subg. Pseudojaca. The leaves have a superficial resemblance to some members of the Tili- aceae, and Miquel described another collection as Grewia subcordata, which was reduced to Artocarpus altissimus by Burret in 1926 The inflorescence characters are, nevertheless, those of Artocarpus, although the long-exserted, deeply bifid styles are again unusual, at least for subg. Pseudojaca. The tree is deciduous and the inflorescences are borne on short-shoots before the new flush of leaves appears. Smith, however, shows a young leaf with a crenate margin attached to the shoot- bearing inflorescences. These shoots are entirely leafless on the Heyne collection, but very young leaves dissected from the buds appear to show the glandular margin, while the twigs and buds agree with those of leafy shoots. This species thus appears to be correctly described and referred to Artocarpus. In the more or less distichous arrangement of the leaves on the ultimate twigs, and the nonamplexicaul stipules, as well as in the size and shape of the inflorescences and the completely divided male perianth, A. altissimus agrees with subg. Pseudojaca. However, there is no obvious affinity with any other members of the subgenus and, in view of the very aberrant vegetative characters, the species is assigned to a separate section Glandulifolium. SPECIES EXCLUDENDAE 5 ARTOCARPUS (?) AFRICANA Sim, For. Fl. Port. E. Africa, 102. ¢. 32. 1919 (holo- type, Mozambique, Sim 5999 (pre, not seen)) = Treculia cf. africana Decne ex Tréc. Ann. Sci. Nat. Bot. III. 8: 109. 1847. 5 The concluding portion of this paper refers to the genus Artocarpus as a whole, ie., to this and the preceding paper, Studies in Artocarpus and allied genera, III. A revision of ek subgenus Artocarpus. Jour. Arnold Arb. 40: 113-155, 298- 368. 1959. 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, IV 17 ARTOCARPUS BRACTEATA King - Hook. f. Fl. Brit. Ind. 5: 540. 1888; King, aan Bot. Gard. Calcutta, 2: 7. ¢. 1B. 1889 = Parartocarpus bracteatus (King) Bec ee CANARANA Mig. Verh. Ned. Inst. III. 5: 20. 1852 (holotype, renee! near Mangalore, Hohenacker 796a [leg. Metz] (vu); isotypes (c, K Hopea wightiana Wall. ex Wight & Arn. Prodr. 85. 1834. Metz’ ee em echinate insect galls that were mistaken by Miquel for the male inflorescences of Artocarpus. In describing this specimen as A. camarana he erroneously suggested that it might be the same as A. lanceifolius Roxb. (a species of western Malaysia described from Penang). The latter determination appeared on the herbarium labels and Thiselton Dyer, who cited the collection cor- rectly under Hopea wightiana in the “Flora of British India” (1: 309. 1874), quoted only this identification, on Miquel’s authority, without mention of A, canarana. This is the origin of the statement by Watt (Dict. Econ. Prod. India 4: 273. 1890), followed by Cooke (Fl. Bombay 1: 86. 1903), that Rox- burgh had described the galls of Hopea wightiana under the name Artocarpus lanceifolius. The same galls were the basis of the name A. ponga Dennst., and they are commented on further below. ARTOCARPUS CANNONI W. Bull in T. Moore, Florist & Pomol. 1875: 210. 1875; Van Houtte, Fl. Serres 21: 131. tab. 1875 = Ficus cannonii (W. Bull) N. E. Brown, Gard. Chron, III. 3: 9. 1888, “canon”: Solereder, Bull. Herb. Boiss. II. 3: 515. ¢. 3. 1903. See below under Appear laciniat Veitch, ARTOCARPUS CERIFERA Miq. Ann. Mus. Lugd.-Bat. 3: 212. 1867 = Parartocarpus venenosus (Zoll. & Mor.) Becc ARTOCARPUS ELONGATA Miq. Fl. Ind. Bat. Suppl. 419. 1861, Ann. Mus. Lugd.- Bat. 3: 213. 1867 (holotype, Sumatra, eee Teysmann HB 767) = Sloetia elongata (Miq.) Koord. Exkursionsfl. Java : ARTOCARPUS EXSCULPTA W. Bull, Gard. ae II. 10: 84. ¢. 11. 1878 (Ficus exsculpta in ind.) = Ficus exsculpta W. Bull, Cat. no. 154, 5. ¢. 7. 1879. This species was mentioned briefly in the Gardeners’ Chronicle of 1878 as having been exhibited recently by W. Bull at Preston under the provisional name of Artocarpus exsculpta. The figure given of a sterile shoot closely resembles the plate of a fertile shoot of Ficus exsculpta, which appeared as new intro- duction from the South Sea Islands in Bull’s Catalogue for 1879. The latter was also stated to have been exhibited at Preston, and the names may be assumed to refer to the same plant. This is Lapeorted by the apparent correc- tion in the index of the Gardeners’ Chronicle. ARTOCARPUS FORBESII King in Hook. f. Fl. Brit. Ind. 5: 539. 1888; King, Ann. Bot. Gard. Calcutta 2: 7. ¢. 1A. 1889 = Parartocarpus venenosus (Zoll. & Mor.) Becc. ssp. forbesii (King) Jarrett.® ARTOCARPUS FRUTESCENS Renner, Bot. Jahrb. 39: 367. 1907 = Prainea frutes- cens Becc. aro INVOLUCRATA K. Schum. in Schum. & Hollr. Fl. Kais. Wilhelmsland, 39. 1889 = Parartocarpus venenosus (Zoll. & Mor.) Becc. ARTOCARPUS LACINIATA Veitch, Gard. Chron. II. 4: 159. 1875; Solereder, Bull. Herb. Boiss. II. 3: 515. ¢. 3. 1903 = Ficus cannonii (W. Bull) N. E. Brown, ° This ae is to draw gees to the new ee which it is neces- sary to make here: Parartocarpus venenosus (Zoll. & Mor.) Becc. ssp. forbesii (King) Jarrett, ae et stat. nov.; basionym Pia forbesii King, Ann. Bot. Gard. Calcutta 2: 7. t. 1A. 1889. 138 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI Gard. Chron. III. 3: 9. 1888. Both Artocarpus laciniata and A. cannoni were introduced as seedlings from the Society Islands. They were valued for their dissected, brightly coloured leaves and the two species were based on differences in these. On flowering both proved to be figs and Solereder, who studied them in some detail, showed that they were conspecific, agreeing in characters of the fig and of leaf anatomy. The dissected leaves were juvenile forms, whereas the adult leaves were entire. Solereder considered that the differences in leaf colour between the two species and a variety of A. laciniata, described by Veitch as A. laciniata Cie might be due to cultural conditions. ARTOCARPUS LrMpATO Miq. Fl. Ind. Bat. Suppl. 421. 1861 = Prainea limpato (Miq.) Beumée ex Heyne. ARTOCARPUS PAPUANUS Renner, Bot. Jahrb. 39: 367. 1907 = Prainea papuana Becc. ARTOCARPUS PONGA Dennst. Schliissel Hort. Ind. Mal. 30. 1818, nomen illegi- timum; Hassk. Hort. Mal. Rheed. Clav. 34. 1867 = Hopea wightiana Wall. ex Wight & Arn. Prodr. 85. 1834. This species is based on Rheede’s Ponga (Hort. Ind. Malab. 4: 73. t. 35, 1683, as Pongu on the plate); I am indebted to the late A. H. G. Alston for its identification. Echinate galls bearing a superficial resemblance to the inflorescences of some species of Artocarpus frequently occur in Hopea wightiana and were shown in the illustrations of Wight (Ill. Ind. Bot. ¢. 37. 1840) and Beddome (FI. Sylvat. tf. 96. 1871). The appearance of these galls, and also of the leaves, agrees well with Rheede’s picture, and the galls, which are formed from terminal or axillary buds, sometimes occur in clusters, as shown there. Rheede described these structures as calices containing several seeds, which were presumably the pupae of the insects. Brandis (Ind. Trees, 68. 1906) records ila pongu as a Tamil name for Hopea wightiana. No precise determination for Ponga appears to have been published previously, although the suggestion that it might be a species of Fg ae was made by Hamilton in a key to the “Hortus Malabaricus” (Jour. Linn. Soc. 17: 195. 1835). Artocarpus ponga antedates Hopea wightiana, ae this will not necessitate any nomenclatural change, since Dennstedt’s name was based on a monstrosity (Int. Code, Art. 67. 1956) ARTOCARPUS RIEDELIL Mig. Ann. Mus. Lugd.-Bat. 3: 213. 1867 = Parartocarpus venenosus (Zoll. & Mor.) B ARTOCARPUS SCANDENS oe Bot. Jahrb. 39: 367. 1907 = Prainea scandens King. ARTOCARPUS TYLOPHYLLA Miq. Fl. Ind. Bat. 1(2): 289. 1859 = Parartocarpus venenosus (Zoll. & Mor.) Becc. ARTOCARPUS VENENOSA Zoll. & Mor. Natur- en Geneesk. Arch. Ned.-Ind. 2: 213. 1845 = Parartocarpus venenosus (Zoll. & Mor.) Becc. ARTOCARPUS VENENOSA Zo Mor. var. TYLOPHYLLA Mig. in Zoll. Syst. Verz. Ind. Archip. 2: 89, 95. 1854 = = Parartocarpus venenosus (Zoll. & Mor.) Becc. ARTOCARPUS WOODII a aes Jour. Sci. Bot. 3: 221. 1908 = Parartocarpus venenosus (Zoll. & Mor.) Bec ARTOCARPUS sp., iy Fi, Elmer. Born. 46. 1929 (Elmer 20978) = Aporosa nitida Merr. l.c. 143. Two nomina nuda are also to be excluded from the genus. 1960 | JARRETT, ARTOCARPUS AND ALLIED GENERA, IV 139 ARTOCARPUS BIFARIA Wall. ex Miq. Ann. ee Lugd.-Bat. 3: 213. 1867, pro syn. Sloetia sideroxylon Teysm. & Binnend. ex Kurz, Jour. Linn. Soc. Bot. 8 167. 1864, nomen superfluum = Sloetia ee (Miq.) Koord. Exkursionsfl. Java 2: 90. 19 ARTOCARPUS ? = FINLAYSONIANA Wall. Cat. no. 4662. 1831, nomen nudum Sloetia elongata (Mig.) Koord. The two collections ce were cited by Wal- lich under this number consist, in his herbarium at Kew, of specimens of Sloetia elongata. Wallich 4662A was collected by Finlayson and has abnor- mal, much branched inflorescences. Under the number 4662B there are two sheets collected from a plant in the Calcutta Botanic Garden which had been introduced from Penang. One bears a shoot with normal inflorescences. The other bears a single leaf and a label with the determination in Wallich’s hand “Morus ? ? bifaria Wall.” This suggests that the source of th Artocarpus bifaria Wall. mentioned above is another sheet of this collection which had been assigned to the latter genus. NOMINA DUBIA ARTOCARPUS DIMORPHOPHYLLA Miq. var. MACROPHYLLA Maiq. Fl. Ind. Bat. Suppl. 417. 1861. No specimens bearing this determination have been seen and the description is not identifiable. ARTOCARPUS LAMELLOSA Blanco, Fl. Filip. 667. 1837, ed. 2. 465. 1847, ed. 3. 3: 74. 1880; Ahern, Timber Tree Sp. Philip. - 1901, excl. tab. The reasons for treating A. lamellosa as a nomen dubium are discussed above, under A nitidus ssp. nitidus, which Merrill erroneously reduced to Blanco’s species in 1905 (Publ. Gov. Lab. Manila 27: 80). ARTOCARPUS PARVIFOLIA Voigt, Syll. Ratisb. 2: 53. 1828. The brief description of trilobed leaves, drawn up from a cultivated specimen from the garden of Belvedere, Thuringia, is not identifiable. ARTOCARPUS RETICULATA Hunter ex Ridley, Jour. Str. Br. Asiat. Soc. 53: 114. 1909. Ridley thought that A. reticulata was probably referable to A. lancet- ee and this seems likely, but the description is inadequate for identifica- feeneeen ROTUNDA (Houtt.) Panzer in Panzer & Christmann, Pflanzensyst. 10: 380. 1783; Merr. Jour. Arnold Arb. 19: 331. 1938. Merrill identified this plant with: Artocarpus rigidus and it has been discussed under that species (Jour. Arnold Arb. 40: 153. 1959). RADEMACHIA ROTUNDA Houtt. Nat. Hist. II. Pl. 11: 455. 1779. This is the basionym of Artocarpus rotunda (Houtt.) Panzer. NOMINA NUDA‘* ARTOCARPUS ANGUSTIFOLIA Roxb. Hort. Beng. 66. 1814. ARTOCARPUS BLUMEI Tréc. var. seaidesgier hs Handl. Fl. Ned. Ind. 3: 370. 1900; Merr. Bibl. Enum. Born. Pl. 21 ARTOCARPUS (?) FOENIFORMIS Eeden, een eal Oost.-Ind. 122. 1886, ed. 2. 242. 1905. ARTOCARPUS MADAGASCARIENSIS Bojer, Hort. Maurit. 290. 1837. “The following list of names, published without description and of undetermined application, is appended for the sake of completeness. 140 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI ARTOCARPUS NEO-CALEDONICA Linden, Cat. 52. 1871; Guillaumin, Bull. Soc. Bot. Fr. 90: 35, 1943. ARTOCARPUS OvATUS Noronha, Verh. Batavia. Genoot 5(5): 7. 1790. Noronha gives the Javanese name for this species as riandelica and this may be a error for mandelica, a name which is recorded for Artocarpus rigidus. THE ARNOLD ARBORETUM, Present Address: RoyAL BoTANIC GARDENS, HARVARD UNIVERSITY K EW, ENGLAND 1960] BAILEY, POLLEN MORPHOLOGY 141 SOME USEFUL TECHNIQUES IN THE STUDY AND INTERPRETATION OF POLLEN MORPHOLOGY I. W. BAILEY EACH OF THE VARIOUS TECHNIQUES that has been utilized in the study of pollen morphology has limitations as well as advantages. It is essen- tial, therefore, that the conclusions attained by the use of any single tech- nique be in harmony with those obtained by other techniques. This is particularly true where pollen is subjected to more or less drastic treat- ments during preparation for visual examination. In all such cases evi- dence should be obtained regarding the effects of each treatment upon the normal form and structure of living pollen. A large amount of research in recent years has been focused upon the visible structure of the exine of acetylated pollen. Comparatively scant attention has been devoted to the study of the intine and the protoplast. If the various layers of the pollen wall are to be accurately defined and classified, it is necessary that more be learned concerning their chemical composition and physical properties. Summations of accumulated cir- cumstantial evidence obtained by the use of diversified techniques can provide significant clues in this connection. In so doing, generalizations should be based, not solely upon the pollen of one or two randomly selected species, but upon investigations of a wide range of representative gymno- sperms and angiosperms. Mangin attempted to do this in a brief paper published in 1889. The significance of this paper cannot be adequately appreciated without read- ing Mangin’s other papers dealing with the differential staining, differ- ential solubilities, and other properties of the polysaccharides that occur in plant tissues in general. He recognized, not only that the intine differs markedly from the exine, but also that it is chemically hetero- geneous. He concluded that it is composed of a mixture of polysaccharides, in most cases largely of cellulose and “pectose,” the cellulose being con- centrated adjacent to the protoplast and disappearing in the external zone or zones and particularly in the so-called “Zwischenkorper” of Fritzsche (1837) which commonly occurs in the apertural regions of the pollen wall. It should be noted in this connection that the results of Mangin’s researches on pollen were published before he obtained a sample of ruthenium red (ammoniated oxychloride of ruthenium) from Joly (1892). He (1893) considered this dye superior, in the differential staining of pectic compounds, to such basic dyes as phenosafranin, methyl blue, etc., that he utilized so extensively in his earlier investigations. How- ever, he does not appear to have used ruthenium red at a later date in an extensive investigation of pollen. 142 JOURNAL OF THE ARNOLD ARBORETUM [YOL, 2a UTILITY AND RELIABILITY OF RUTHENIUM RED In microscopy, ruthenium red has the advantage of being soluble in water but insoluble in alcohols, anhydrous glycerin, chloroform, benzene, xylol, clove oil, and other reagents utilized in preparing sections for micro- scopic examination. However, in aqueous solution it has the disadvan- tage of deteriorating rapidly in the light at ordinary room temperatures. This difficulty can be overcome by keeping solutions in the dark in a re- frigerator. I have solutions of the dye (Edward Gurr, Ltd.) that are in good condition after more than a year. A more serious difficulty occurs in attempting to make permanent mounts of stained pollen, as also of sec- tions of plant tissues. In all cases thus far the red color disappears in time from stained parts of the preparations. There have been two extreme views regarding the utility of ruthenium red. At one extreme are those who assume that it is specific for pectic compounds, whereas at the other extreme are individuals who have shown that it stains a variety of chemical substances and therefore conclude that it is utterly unreliable and useless in microchemical investigations. In this connection, I have made in the past extensive observations upon the staining reactions in aqueous ruthenium red of a wide range of in- organic and organic substances of known chemical composition, Tests have been made not only with naturally occurring gums and mucilages and pectic compounds and “polyuronide hemicelluloses” extracted from plant tissues, but also, through the cooperation of Professors Ernest Anderson and Karl P. Link, with fractions of these substances of known chemical composition. It seems that galacturonic and glucuronic acids when methylated, as salts, or in intimate chemical association with hexose and pentose sugars stain characteristically in aqueous ruthenium red. Where the substances themselves are soluble in water, their staining re- actions may be studied by using a concentrated solution of the dye in admixture with a proportion of alcohol which inhibits or retards their rapid solution. In contrast to the general uniformity in the behavior of these naturally occurring substances, ees free acids when isolated and urified are very variable in their reactions in aqueous solutions of ruthenium red. When they stain, nes may give a yellow, rather than a red, coloration, and likewise may turn the color of the ruthenium solu- tion from red to yellow. The rapidity and the intensity of staining varies in buffers of widely different pH. This raises complex questions regard- ing the effects of reagents in buffers upon the dye and upon the chemical composition of the substances being tested. Preliminary treatments of cells and tissues with acids, alkalies and other reagents should be avoided if possible. The most significant results are obtained when living cells or freshly cut sections of living tissues are quickly immersed in a solution of ruthenium red in distilled water or in relatively pure spring water. Although the dye stains oxycellulose, hydrocellulose, the nucleus, proto- plast, and other organic substances, as well as pectic compounds, gums, mucilages, and ‘‘polyuronide hemicelluloses,’ there are significant dif- 1960] BAILEY, POLLEN MORPHOLOGY 143 ferences in its behavior in these cases. In my investigations of cell walls of the higher plants, I have found that those naturally occurring poly- saccharides which contain sugar acids (e.g., galacturonic, glucuronic) or their methylated or salt (e.g., calcium pectate) derivatives tend to stain very intensely and with extraordinary rapidity in dilute (1/5000) solu- tions of ruthenium red, whereas other naturally occurring organic sub- stances, if they stain at all, do so more gradually and commonly less in- tensely. Of course, it should be realized in this connection that the stain- ing may be inhibited or masked by lignification, cutinization, or suberiza- tion of cell walls. Thus, although the dye is not specific for pectic com- pounds and must be used with adequate precautions, its use does pro- vide a simple and rapid method of obtaining useful clues or leads in study- ing the occurrence and distribution of polyuronides in plant cells and tissues. Although not conclusive by itself, it becomes increasingly so when in harmony with evidence obtained by other oe e.g., differential solubilities, polarized light, electron microscopy, etc. The so- called intine of pollen commonly stains with remarkable clarity and rapidity when freshly collected pollen, viable dry pollen, dead pollen from herbarium specimens, and pollen preserved in alcohol are immersed in dilute aqueous solutions of ruthenium red. Frequently the differential staining is so rapid that it is clearly visible by the time that a mount can be made and examined under a microscope. In fact, it may sometimes be advantageous to retard the staining by using the aqueous solution in ad- mixture with a high proportion of glycerin. Such admixtures are relatively stable when kept in the dark at ordinary refrigerator temperatures. In- tense staining of the intine occurs first in the apertural region or regions of the pollen wall and subsequently in parts that subtend the exine. The exine does not stain, retaining its original greenish yellow or other colors. On the contrary the protoplast may in time develop a red coloration. The accelerated staining of thicker parts of the intine in apertural regions may be due to a higher concentration of stainable substances in such areas or to retarded penetration and diffusion of the dye through the exine of nonapertural parts. That the proportion of stainable substances varies is indicated by differences in the ultimate intensity of staining, not only s between the intines of different types of pollen, but also in different zones or layers of a single thick intine. That penetration of the dye is retarded or actually inhibited at times by the exine (or by external coat- ings of oily or resinous substances) is indicated by the fact that the intine of inaperturate pollen may not stain at times unless the exine is mechanically ruptured or abraded. Ruthenium-red staining of the intine of pollen from herbarium speci- mens, as well as of freshly collected living pollen, provides a simple, rapid, and reliable means of studying variations in the normal form of the intine in the various taxa of the angiosperms. It is particularly useful in demonstrating variations in the form and thickness of the intine in apertural regions of pollen. Although the spectacular differences in color between the intine and exine require illustration in color, the intense red 144 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI staining of the thicker parts of the intine can be reproduced in intense black by the use of a green filter as illustrated in PLATE I. The pollen of Calycanthus varies from monocolpate (Fig. 1) to zonaper- turate or belted (Fig. 3) to bicolpate (Fig. 2) and infrequently to tri- colpate. In each case the aperture or apertures are subtended by con- spicuously thickened parts of the intine which stain intensely and with remarkable rapidity in dilute aqueous solutions of ruthenium red. In these pollen grains, as in those of Jllicitum (Fig. 4) which have three nar- row grooves that extend from pole to pole,' the thicker parts of the intine are considerably broader than the transverse diameter of the furrows. This is in contrast to those forms of pollen (Fig. 5) in which the ex- cessively thickened parts of the intine, ‘“Zwischenkorper” of Fritzsche (1837), “oncus” of Hyde (1955), are but slightly broader than the diam- eter of the apertures. In the case of the Winteraceae, where the pollen occurs characteristically in tetrads, the New World representatives differ conspicuously from the Old World representatives of the family (Bailey and Nast, 1943). In the monoporate grains of the latter genera and species, the intine is somewhat thicker in the apertural part but does not protrude (Fig. 7), whereas in the New World representatives the proto- plast protrudes and is jacketed by a thick coating of intensely staining intine (Fig. 8). In a majority of the angiosperms the thickenings of the intine subtend the apertures of varying form, size, and number, but this is not invariably the case, as has been shown by Dressler (1957) in the Euphorbiaceae. For example, in the tricolporate pollen of Pedilanthus (Fig. 6) there are six riblike thickenings which extend nearly from pole to pole. These thickenings of the intine are lateral to the elongated colpae, rather than directly beneath them. OCCURRENCE AND DISTRIBUTION OF CELLULOSE IN POLLEN Although the intense ruthenium-red staining of the outer part of the intine, particularly in apertural regions of pollen, is not necessarily in- dicative of pectic composition, it suggests, in correlation with the iso- tropy of this part in polarized light, plasticity during pollen tube emer- gence, expansion and contraction during wetting and drying, and solu- bility in cold 4% sodium hydroxide and other reagents which do not dissolve cellulose, that the outer part of the intine commonly con- tains a large proportion of uronide polysaccharides in its chemical com- As regards the inner part of the intine, the differing conclusions attained by Mihlethaler (1953) and Sitte (1953) by the use of electron micros- copy raise an important question regarding the occurrence of cellulose in the inner part of the intine adjacent to the protoplast as hypothesized by Mangin (1893). In diversified representatives of both gymnosperms and angiosperms this part of the intine, although staining more or less For illustrations compare Wodehouse (1935), page 336, Fig. 92, or Erdtman (1952), page 256, Fig. C. 1960] BAILEY, POLLEN MORPHOLOGY 145 intensely in ruthenium red, exhibits anisotropy in polarized light. Further- more, it stains a characteristic blue when pollen grains are immersed directly in chloro-iodide of zinc or in 65% sulphuric acid following stain- ing in iodine. Where the intine is very tenuous and obscured by the exine, the cellulosic part may be retained and rendered clearly visible by removing the protoplast and the noncellulosic constituents of the intine in 3% sodium hydroxide at 56° C. (Figs. 9-11). Not only is the inner part of the intine insoluble in reagents which do not dissolve cellulose, but also it is removed by standard solvents of cellulose, e.g., 72% sulphuric acid, cuprammonium reagents, as well as during prolonged acetylation of pollen. Furthermore, its microfibrillar composition can be revealed by electron microscopy as demonstrated by Sitte (1953). In addition, it should be noted that when the noncellulosic constituents of the intine are completely removed the cellulose-containing inner layer no longer stains rapidly and intensely in ruthenium red. In general (with the exception of certain unusual forms of pollen, e.g., Eupomatia), the cellulose-containing inner layer of the intine of both eymnosperms and angiosperms is tenuous and of relatively uniform thick- ness. This is in contrast to the striking variations in the thickness of the outer part of the intine that occur so frequently in various taxa of the higher plants. During the disruption of pollen walls due to differences in the contraction or expansion of the protoplast and wall layers, the cellu- losic layer of the intine commonly tends to remain adjacent to the proto- plast. SOLUBILITY OF THE EXINE IN MONOETHANOLAMINE The exine of pollen, like the cuticle of plants, is generally considered to be relatively inactive chemically since it persists for such prolonged periods in geological strata and dissolves only after prolonged drastic treatments. Much to my surprise, therefore, I have found that exines of freshly collected living pollen at anthesis (e.g., of such gymnosperms as Taxus, Tsuga, Pseudotsuga, Pseudolarix, and Pinus and of such angio- sperms as Liriodendron, Magnolia, Asimina, Taraxacum, Coreopsis, Ostrya, Populus, and Calycanthus) dissolve in three hours or less when immersed in monoethanolamine at a temperature of 97° C. The only exceptions that I have encountered thus far in a preliminary investigation are pollen of Ephedra and Pinus strobus from old herbarium specimens and of Eupo- matia preserved for a long period in F.A.A. fixative (formalin-acetic acid-alcohol). The fact that the exine of freshly collected pollen of Pinus dissolves suggests that changes occur under certain conditions of prolonged drying which inhibit solubility. Unfortunately no freshly col- lected pollen of Ephedra and Eupomatia is available, as yet, for such comparative purposes. In the case of Liriodendron, it is possible, by adequately controlling temperature and time of treatment, to remove the thick exine (Fig. 12) leaving the protoplast, the layers of the intine and oil globules but slightly modified visually (Fig. 13). Of course, 146 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI this raises an important question, viz., whether dissolving the exine of a large amount of pollen in monoethanolamine might yield a solute capable of separation and analysis. Such analyses might afford significant clues regarding the chemical composition of the exine and deserve to be more intensively investigated. DISCUSSION Summations of evidence obtainable by the application of diversified techniques indicate that the cellulosic part of the intine of both gymno- sperms and angiosperms occurs in a comparatively narrow zone adjacent to the protoplast. The anisotropic cellulose occurs, however, in associa- tion with an isotropic polyuronide (or mixture of polyuronides and polysaccharides) which stains rapidly and intensely in dilute aqueous solutions of ruthenium red and which is readily removable by reagents which do not dissolve cellulose. The outer part of the intine, particularly in thickenings related to apertures, contains little, if any, cellulose and is composed of a polyuronide (or a mixture of polyuronides and poly- saccharides) which stains and dissolves as does the material associated with the cellulose in the innermost part of the intine. The noncellulosic constituent in thicker parts of the intine (related to apertures in the exine) is plastic (i.e., easily penetrated or pushed aside during emergence of the pollen tube) and contracts and swells, with corresponding invagina- tion and evagination, during drying and rewetting of living pollen. The consistency in the occurrence and behavior of this part of the intine in a wide range of taxa suggests that it serves two important functions, (1) to protect the protoplast in apertural parts of the wall and (2) to facilitate emergence of the pollen tube. Professor A. Orville Dahl and I plan to discuss this aspect of the intine in greater detail in a subsequent joint paper. We are also correlating evidence obtained by phase and electron microscopy and other diversified techniques in detailed studies of specific forms of pollen. The question of a revised terminology for wall layers, e.g., whether the use of intine should be restricted to the cellulosic layer and exintine or mesine used in referring to material that intervenes be- tween the cellulosic layer and the exine, had best be deferred until com- prehensive investigations of a wide range of representative pollen forms have been completed. ACKNOWLEDGMENTS This investigation was supported by a grant from the American Philo- sophical Society. I am much indebted to the curators of the Gray and Arnold herbaria for permission to collect pollen from a wide range of herbarium specimens, to Dr. A. C. Smith for assistance in obtaining pollen of the Winteraceae and other families, to Dr. C. E. Wood for pollen obtainable in his collection of flowers preserved in 70% alcohol, to Mr. L. J. Brass for collecting and preserving flowers of critical genera 1960] BAILEY, POLLEN MORPHOLOGY 147 in F.A.A. fixative, and to Dr. R. L. Dressler for living pollen of Pedilan- thus. I am also indebted to Professors A. Orville Dahl and Kenneth V. Thimann for kindly reading the manuscript of this paper and for making a number of helpful suggestions. LITERATURE CITED BarLey, I. W., and CuHartottTe G. Nast. 1943. The comparative morphology of the Winteraceae I. Pollen and stamens. Jour. Arnold Arb. 24: 340-346. Dress_er, R. L. 1957. The genus Pedilanthus (Euphorbiaceae). Contr. Gray He ee G. 1952. Pollen morphology and plant taxonomy. Chronica Botanica fees J. 1837. Ueber den Pollen. Mém. Sav. Etrang. Acad. St. Petersburg : 649-770. Hype, H. A. 1955. Oncus, a new term in pollen morphology. New Phytol. 54: 255-256. Jory, A. 1892. Composés ammoniacaux dérivés du sesquichlorure de ruthénium. Compt. Rend. Acad. Sci. Paris 115: 1299-1301 MAnGIN, L. 1889. Observations sur la membrane du grain de pollen Mar. Bull. Soc. Bot. France 36:274-283. . 1893. Sur l’emploi du rouge ae ruthénium en anatomie végétale. Compt. Rend. Acad. Sci. Paris 116: 653 WopveHousE, R. P. 1935. pollen a grains. McGraw-Hill Book Co. EXPLANATION OF PLATES PLATE I Pollen differentially stained in aqueous ruthenium red and photographed with a green filter at a magnification of 770. Fics. 1-3, 6, freshly collected living pollen; 5, 8, pollen from herbarium specimens; 4, pollen preserved in 70% alcohol; 7, pollen ples in FAA fixative. Fics. 1-3, Calycanthus sp. (cul- tivated, Cambridge, Mass.): 1, monocolpate grain viewed in optical section Parallel to long axis - colpus: 2, dicolpate grain viewed in optical section parallel to the long axis of the colpus; 3, zonaperturate grain viewed in diag- onal optical section. Fic. 4, Zllicium floridanum Ellis (cultivated, Henry Foun- dation, Gladwyne, Pa.), tricolpate grain, polar view. Fic. 5, Nowhuysia arfaken- sis (Gibbs) Steenis ae 2198 [A]), triporate grain. Fic. 6, Pedilanthus sp. (Dressler), tricolporate grain, polar view. Fic. 7, Zygogynum Baillonit Tiegh. (Buchholz, New Caledonia, 1947), each grain of the tetrad with a single more or less circular aperture. Fic. 8, Drimys granadensis L. f. var. mexicana (DC.) A. C. Smith (Hinton 14441 [cx]), characteristic tetrad of New World section Wuntera, each grain of the tetrad with a single aperture and protruding protoplast jacketed by a thick coating of stainable intine. PLATE II Fics. 9, 10, Liriodendron tulipifera L. (cultivated, Arnold Arboretum): 9, pollen after treatment in 3% NaOH at a temperature of 56° C. which removes the protoplast and the noncellulosic constituents of the intine — when im- 148 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI mersed in chloro-iodide of zinc the contracted cellulosic residue of the intine stains a deep blue, in striking contrast to the greenish-yellow color of the exine, < 1130; 10, characteristic anisotropy of two cellulosic residues in polar- ized light, & 400. Fic. 11, Pinus strobus L. (Norwell, Massachusetts), ia collected living pollen fllowae the same treatment as in Fic. 9, & 1130 PLATE III Pies, 12, 13, tte tulipifera heen Arnold Arboretum) : freshly collected living po in a mix ) and glycerine, viewed in optical section parallel to am long axis . the Len showing coarsely warty exine and conspicuous thickening of the intine in the apertural region, X 1130; 13, pollen grain after a brief treatment in monoethanolamine at a temperature C,—exine has been removed leaving the intine, protoplast, and oil globules in place, * 1130 Jour. ARNoLD Ars. VoL. XLI PLaTE I BAILEY, POLLEN MorpHOoLocy PuaTE II / Jour. ArnoLtp Ars. VoL. XLI BAILEY, POLLEN MorPHOLOGY Jour. ARNOLD Ars. VoL. XLI PLATE III BAILEY, POLLEN MorPHOLOGY 152 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI THE GENERA OF SARRACENIACEAE AND DROSERACEAE IN THE SOUTHEASTERN UNITED STATES ! CARROLL E. Woon, JR. SARRACENIACEAE Dumortier (PircHER-PLANT FAMILY) Perennial, rhizomatous insectivorous herbs with tubiform or pitcher- like leaves. Flowers bisexual, solitary [or few and racemose], nodding, the scape naked [or few-bracteolate|]. Flowers regular, 5[4,6]-merous, the insertion hypogynous. Sepals 5[4,6], free, imbricate at the base, alae sometimes colored [and petaloid]. Petals 5: tor none, tree, deciduous. Stamens numerous, free, the filaments filiform, the anthers versatile, extrorse, 2-locular; pollen grains single, polycolporoidate to 5—6-colporoidate. Gynoecium syncarpous, the style simple below, the apex [entire or oe or| peltate-dilate, the ovary 5[3]-locular, the placentation axile; ovules numerous, anatropous, 2-integumented. Fruit a 5[3]|-valved, jordiiaaal capsule. Seeds numerous, small, winged [or not|, with a undant endosperm, the embryo minute, near the hilum. TYPE GENUS: Sarracenia L. A small, distinctive family of disjunct distribution, with three well- marked genera: Sarracenia (eight species of eastern North America), Darlingtonia Torr. (D. californica Torr. |Chrysamphora californica (Torr.) Greene], 2x = 30, of northern California and southern Oregon), and Heli- amphora Benth. (six species described from the isolated table-top moun- tains of northern South America). The highly modified pitcher-leaves and the large, bisexual, pentamerous flowers with axile placentation are characteristic. The group has been allied with various families or combinations of families: Papaveraceae haglage for a biologically oriented generic flora of the southeastern United States, a joint project of the a Herbarium and the Arnold Arboretum made possible through the support of George R. Cooley and the National Science Founda- tion. The scheme follows that Pere at the beginning of the series (Jour. Arnold Arb. 39: 296-346. 1958). It may be remarked here, however, that the area covered is bounded by and includes North Carolina, Tennessee, Arkansas, and Louisiana, that descriptive material inapplicable to our species is included in brackets, and that references marked by an asterisk have not been oe Other published portions of these studies will be found in Jour. Arnold Arb. 40 1959, 1960 In connection with these families I am senegal Ree for various comments and helpful suggestions to Dr. Patricia R. Roberts and to Dr. C. R. Bell, upon whose respective work on Dionaea and Sarracenia I have drawn heavily. I am also grateful to my im oe colleagues for their continuing interest and advice and to i Brown, Jr., R. B. Channell, G. R. Cooley, D. Demaree, C. W. James, C. E. Jenner, J. Kucyniak, ce Renshaw, H. F. L. Rock, and A. J. Sharp for their help in vari- ous ways. As in previous papers in this series, the illustration is Dorothy H. Marsh. 1960] WOOD, SARRACENIACEAE AND DROSERACEAE 153 and Nymphaeaceae or these plus Ranunculaceae; Droseraceae or Nepen- thaceae, or both; Cephalotaceae; Francoaceae; Cistaceae. Although now placed with the insectivorous Nepenthaceae and Droseraceae by many authors, the true relationships are still in need of careful study in terms of the entire plant. REFERENCES: epee H. Nymphaeacées. Hist. Pl. 3: 77-104. 1872. [Includes Sarraceni- ee G., and J. D. Hooker. Sarraceniaceae. Gen. Pl. 1: 48. 1862; 965. 1867 Gray, A. Sarraceniaceae. Gen. Pl. U. S. 1: 105-108. pls. 45, 46. 1848. Lioyp, F. E. The Carnivorous Plants. xv + 352 pp. Chronica Botanica, Wal- tham, Mass.,.1942. [Chap. I. Heliamphora; Il. Sarracenia; II. Darling- tonia. Comprehensive review of structure and biology; extensive bibli- ography. | MAcFARLANE, J. M. Sarraceniaceae. Pflanzenr. IV. 110(Heft 34): 1-38. 1908. MArRKGRAF, F, Les homologies des feuilles et les relations ree ne les Sarraceniales. VIIT® Congr. Int. Bot. Rapp. Comm. I. Sect. 2, 4, 5, 50-52. 1954. [Interprets leaves of Sarracenia, Drosera, sen oe Dioncophyllum as showing basically the same morphology. Upuor, J. C. T. Sarraceniaceae. Nat. Pflanzenfam. ed. 2. Ub. 704-727. 1936. WUNSCHMANN, E, Sarraceniaceae. Nat. Pflanzenfam. III. 2: 244-252. 1891. 1. Sarracenia Linnaeus, Sp. Pl. 1: 510. 1753; Gen. Pl. ed. 5. 226. 1754. Perennial insectivorous herbs of bogs, seeping terraces and wet savan- nas. Plants rhizomatous, the 3-8 leaves clustered, erect and trumpet- shaped to decumbent and pitcher-shaped, liquid-filled to varying degrees, with a ridge or wing on the adaxial side and a terminally expanded hood; leaves green, yellow, maroon, or variegated; ensiform leaves sometimes present. Flowers solitary, nodding on erect scapes. Sepals 5, with 3 appressed, persistent bracts. Petals 5, yellow or maroon (or deep red), deciduous. Stamens 70-80 (arising in ontogeny in 10 groups). Gynoecium 5-carpellate; style simple below, the upper portion expanded into a large, persistent, 5-lobed umbrella-like structure, with a small stigma under each of the notched lobes; ovary 5-locular. Fruit 5-valved; seeds winged on one side. Chromosomes 2—4 microns long, similar, 2x = 26. TYPE SPECIES: S. purpurea L. (Named for Michel Sarrazin de )’Etang, 1659— 1734, physician at the Court of Quebec, who sent S. purpurea to Tourne- fort.) — PircHER PLANTS, TRUMPETS, BUGLE-GRASS, Boc-BUGLES, DuUMB- WATCHES, WATCHES, BUTTERCUPS, EVE’s CUPS, FROG-BONNETS. Eight species, all confined to the southeastern United States, with the exception of S. purpurea, the most widespread species, which ranges from western North Carolina northward along the mountains and from south- ern Mississippi eastward and northward along the Coastal Plain to New- foundland, Labrador, Mackenzie, and Saskatchewan. Form and habit (and in some instances, color and marking) of pitchers 154 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI are the most useful taxonomic characteristics, in conjunction with color and shape of petals, odor, curvature of sepals, and presence of ensiform leaves. Leaf size is not a constant characteristic. (See Bell for key, ranges, and comments. ) The six species of sect. ErecTAE Uphof have upright, tubular or trumpet- shaped leaves with variously shaped hoods which largely prevent the en- trance of rain. The group includes Sarracenia leucophylla Raf. (S. Drum- mondii Croom) and S. rubra Walt. with maroon petals, and S. minor Walt., S. flava L., S. oreophila Wherry, and S. alata (Wood) Wood (S. Sledgei Macfarl.) with yellow petals. The last three are closely related vicariads with yellow-green trumpet-shaped leaves bearing large hoods with more or less reflexed margins. They differ in ensiform leaves, hoods, petal shape and texture, and odor. Insofar as is known, the three are geographically separated: S. flava distributed on the Atlantic coastal plain (southeastern Virginia to western Florida and southern Alabama and a few disjunct stations inland in North Carolina); S. oreophila at the south- ern end of the Appalachians (northeastern Alabama to Elmore County, Alabama, westernmost central Georgia, and perhaps formerly Fentress County, Tennessee); and S. alata on the Gulf coastal plain (Mobile County, Alabama, to eastern Texas). Sarracenia rubra {, Jones (W herry ) Bell (S. Jonesii Wherry) has the fragrant maroon flowers, recurving sepals, and characteristically maroon-reticulate hoods of f. rubra and differs only in its larger leaves being more sharply expanded toward the leaf opening which is less closely covered than in f. rubra. Sarracenia purpurea and S. psittacina Michx., both with maroon petals and decumbent pitchers and the latter with the most highly modified leaves in the genus, constitute sect. SARRACENIA (sect. Decumbentes Uphof). Sarracenia purpurea supposedly is represented southward by var. venosa (Raf.) Fern. (ssp. venosa (Raf.) Wherry), a doubtfully distinct taxon based upon proportions of ‘‘pitcher” and hood Sarracenia is confined to wet, mediacid soils and hence occurs largely in bogs, wet savannas, at the edges of pocosins, and on springy, sandy slopes on the Coastal Plain. The greatest concentration of species and individuals is in the southern half of Georgia and Alabama and in north- ern Florida. Only S. minor is known from peninsular Florida (as far south as Highlands County). The ranges of S. purpurea, S. psittacina, S. leucophylla, and S. rubra overlap at least in part those of all species except S. oreophila, which is completely isolated geographicall The genus is a homoploid complex (2 = 26) of interfertile but mor- phologically distinct species. Sixteen of nineteen “geographically pos- sible” natural hybrids have been reported from the wild, and two others have been produced artificially. Artificial hybrids involving three or four species have been made. No hybrids of S. oreophila are known. Most hybrids are sporadic where two or more species occur, but around Mobile Bay, where S. leucophylla and S. alata commingle, much hybridization and backcrossing are evident. The extent to which introgression occurs is not known. Over most of their ranges the various species seem to retain 1960] WOOD, SARRACENIACEAE AND DROSERACEAE 155 their biological identities primarily through geographical and seasonal isolation. The remarkable leaves are provided with nectar glands and are some- times conspicuously colored (e.g., bright yellow-green to red in S. flava, the upper part of leaf and hood white, reticulate with red and green in S, leucophylla) or translucent spotted (S. psittacina, S. minor). Insects attracted by nectar or coloration may fall or crawl into the pitchers, and, if unable to escape, die and are digested. Only the pitchers of S. purpurea are water-filled; those of others which are provided with pro- tective hoods secrete varying quantities of fluid, however. The morpho- logical homologies of the leaves are still under discussion. Sarracenias are fed upon by a number of moths known from no other plants: Olethreutes feeds upon the flowers and seeds; Papaipema is a rhizome-borer; and three species of Exyra eat the leaves. Other insects, including a sarcophagan fly, a gnat, a sciarid fly, and two harmless mosquitoes dwell in the pitchers. The larvae of the last, members of the tropical genus Wyeomyia, inhabit the water-filled leaves of S. purpurea. The larvae are purely aquatic and overwinter in the pitchers, surviving freezing and thawing. Pupation with subsequent emergence as the adult does not occur (at least in the North) until the proper photoperiod and suitable temperatures for growth are reached in the spring. Although pitcher-plants in some areas are accused of harboring mosquitoes, larvae of other genera do not occur in the pitchers, and the adult Wyeomyia smithiu does not bite man. Sarracenias are highly ornamental and were popular horticultural plants in Great Britain from about 1880 to 1890. With care, all of the species succeed in cultivation in pots filled with fine, sandy, acid muck and standing in about an inch of water. Full sunlight with a bright southeast- ern exposure is desirable. A cool, winter dormant period is usual. REFERENCES: Many references prior to 1942 omitted; see review and bibliography in Liovyp, arnivorous : ArBER, A. On the morphology of the pitcher leaves of Heliamphora, Sarra- cenia, Darlingtonia, Cephalotus and Nepenthes. Ann. Bot. II. 5: 563-578. 1941 BELL, C. R. A cytotaxonomic study of the Sarraceniaceae of North America. Jour. Elisha Mitchell Sci. Soc. 65: 137-166. pls. 8-14. 1949. [Sarracenia, Darlingtonia. | Natural hybrids in the ae Sarracenia. I. History, distribution, and taxonomy. /bid. 68: 55-80. pls. 1-7. 1952. . Sarracenia ne eee Ibid., 70: 57-60. 1954. [S. leuco- phylla for S. Drummondii Croom, and S, alota (Wood) Wood replaces S. Sledgei tome | and IF. W. Case. Natural hybrids in the genus Sarracenia. II. Current notes on distribution. bid. 72: 142-152. 1956 Case, F, W., Jr. Some Michigan records for Sarracenia purpurea forma hetero- phyla. Rhtedora 58: 203-207. 1956. [A yellow-flowered form.] 156 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XLI Harper, R. M. The American pitcher-plants. Jour. Elisha Mitchell Sci. Soc. 34: 110-125. pls. 2-6. 1918. Hecner, R. W. The protozoa of the pitcher plant, Sarracenia purpurea. Biol. Bull. Woods Hole 50: 271-276. 1926. Heppurn, J. S., and F. M. Jones. Occurrence of antiproteases in the larvae of the Lae associates of Sarracenia flava. Contr. Bot. Lab. Univ. Penn. 4: 460-463. 1919. , ——and E. Q. St. Joun. The biochemistry of the American pitcher plants. Trans. Wagner Free Inst. Sci. 11: 1-95. 1927.* [Full bibliography. ] Hom, T. Contributions to the knowledge of the germination of some North American plants. Mem. Torrey Bot. Club 2: 57-108. pls. 5-19. 1890-1891. [S. purpurea, 63, 64, pl. 7, figs. 23-27. | Livincston, M. J. Microsporogenesis in Sarracenia minor Walt. Quart. Jour. Fla. Acad, Sci. 13: 61-71. 1950. MacrarLANnE, J. M. Observations on Sarracenia. Jour. Bot. 45: 1-7. 1907. [Includes history of name, notes on various spp. | Marxcrar, F. Les homologies des feuilles et les relations taxinomiques chez les Sarraceniales. VIII° Congr. Int. Bot. Rapp. Comm. Sect. 2, 4, 5, 6: 50-52. 1954. Nicuots, M. L. The dev elopment of the pollen of Sarracenia. Bot. Gaz. 45: 31-37. pl. 5. 1908. [Five species, two hybrids. Russett, A. M. The macroscopic and microscopic structure of some hybrid Sarracenias compared with that of their parents. Contr. Bot. Lab. Univ. Penn. 5: 3-41. pls. 1-5. 1919. Sureve, F. The development and anatomy of Sarracenia purpurea, Bot. Gaz. 42: 107-126. pls. 3-5. 1906. Sms, J. Sarracenia flava. Bot. Mag. 20: pl. 780. 1804. (See also S. minor. 21: I. 849, 1805; > purpurea. 41: pl. 1710. 1825; S. rubra. [S. leucophylla? ] 63: pl. 3515. 18 6.) SmiTH, J. B. se of the New Jersey State Agricultural Experiment Station upon the _ occurring within the state, their habits, life history, etc. v + 482 pp. Trenton, N. J. 1904. [342-353, description and life history of eae smithi Coquillet. ] Watcort, M. V. Illustrations of North American pitcher-plants. (Including E. Wuerry, Distribution of North American pitcher plants, pp. 1-25, and F. M. Jones, Pitcher plants and their insect Sere pp. 25 34.) 34 pp. 15 col. pls. Smithsonian fae Washington, D. C., 1935. Wattr, A. Determination of the nature of the volatile base Jet the rhizome of the pitcher plant ba pases Jour. Am. Chem. Soc. 67: 2271. 1945.* [For use in medicin Wuerry, E. T. Acidity ern of the Sarracenias. Jour. Wash. Acad. Sci. 19: 379-390. 1929. [Includes description of S. Jonesii.] The geographic relations of Sarracenia purpurea. Bartonia 15: 1-6. pl. 1. 1933. The Appalachian relative of Sarracenia flava. Bartonia 15: 7, 8. 1933. [S. oreophila Wherry. | DROSERACEAE Salisbury (SUNDEW FAMILY) Annual or perennial, rosulate or short-stemmed [or caulescent], in- sectivorous herbs of damp or wet soils [a few completely aquatic or of 1960 | WOOD, SARRACENIACEAE AND DROSERACEAE 157 dry soils], the leaves alternate, infolded or circinate in vernation, the blades modified as active traps (Dionaea, Aldrovanda) or equipped with mucilage-tipped irritable tentacles (Drosera, Drosophyllum). Roots mostly adventitious. Flowers bisexual, regular, the insertion hypogynous. Sepals and petals 5, free. Stamens 5—15(—20), extrorse; pollen in tetrads (except Drosophyllum). Gynoecium syncarpous; styles 3[5] and deeply bifid [or otherwise divided] or single and with fimbriate stigma; ovary 1-locular with 3[{5] parietal placentae or the ovules at the base of the locule; ovules anatropous, 2-integumented. Seeds small, numerous, the embryo minute, straight, embedded in abundant endosperm. (Including Dionaeaceae Small.) Type Genus: Drosera Four genera, Drosophyllum Link, Dionaea Ellis, Aldrovanda L., and Drosera L., the first three monotypic, Dionaea and Drosera in our area. The sensitive leaves, bisexual, usually pentamerous flowers, unilocular ovary, and parietal or basal placentation mark the family. Numerous reticulate interrelationships link all four genera and make separation into two families an arbitrary matter. addition to its present position, the family has been placed near the Violaceae and Ochnaceae in the Parietales (Violales) near the Saxi- fragaceae, and near the Lentibulariaceae. The validity of the present association with Sar and Nepenthaceae (the three constituting the Sarraceniales) is in need of oon examination, and the relationships of the family are still unsettled. REFERENCES: BAILLON, H. Droséracées. Hist. Pl. 9: 225-235. 1888. BENTHAM, G., and J. D. Hooker. Droseraceae. Gen. Pl. 1: 661-664. 1865. Darwin, C. Insectivorous plants. ed. 2. London, 1875. Diets, L. Droseraceae. Pflanzenr. IV. 112(Heft 26): 1-136. 1906. . Droseraceae. Nat. Pflanzenfam. ed. 2. 17b: 766-784. 1936. DrupeE, O. Droseraceae. Nat. Pflanzenfam. III. 2: 261-272. 1891. Gray, A. Droseraceae. Gen. Pl. U. S. 1: 191-200. pls. 83-86. 1849. [Drosera, Dionaea, see Lioyp, F. E. The Carnivorous Plants. xv + 352 pp. Chronica Botanica, Wal- tham, vos 1942. [Chap. X. Drosera; XII. Dionaea and Aldrovanda. Comprehensive review of the structure and biology of these plants; extensive bibliography. ] KEY TO THE GENERA OF DROSERACEAE Leaf blades 2-lobed, hinged lengthwise in the middle, sensitive and closing, the marginal bristles interlocking; stamens mostly 15 (10-20); styles united: -placentation- basal... meeps tre ee 1. Dionaea. Leaf blades linear to suborbicular, equipped on margins and upper surfaces with numerous gland-tipped irritable tentacles; stamens 5; styles 3, so deeply bifid as to appear to be 6; placentation parietal. .... 2. Drosera. 158 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI 1. Dionaea Ellis, Directions for bringing over seeds and plants from the East Indies and other distant countries, etc., 38. 1770; Nov. Act. Upsal. 1: 98. p/. 8. 1770. Rosulate, scapose, insectivorous herbs, the leaf blades hinged medially and forming a sensitive trap, the petioles winged (scarcely so in full sun in summer), their bases forming a subterranean bulb-like structure around the short, horizontal, sympodial rhizome, the roots adventitious. Inflorescence from the terminal bud, cymose-umbelliform. Sepals 5. Petals 5, white, marcescent. Stamens usually about 15. Gynoecium of 5 united carpels, style 1, the stigma with numerous papillae, the ovary 1-locular, the placentation basal. Capsule opening irregularly; seeds obovoid, numerous, black and shining. Embryo sac development normal (Polygonum type). TyPpE AND SOLE SPECIES: D. Muscipula Ellis. (Name from Greek, Dione, originally the mother of Aphrodite | Venus] but later used as ‘daughter of Dione,’ hence Aphrodite or Venus.) — VENUS’ FLY-TRAP, FLY-TRAP. Dionaea Muscipula (2n = 30 or 32?) is known to occur only on the Atlantic coastal plain from Beaufort County, inland to Moore County, HEP Pasta UTM Dionaea. D. Muscipula: a, habit, x ‘ b, mature leaf, x 1/2; c, young leaf, the marginal Bi a folded within ii blade, < 1/2; d, flower, X 1; e, pollen tetrad, & ca. 200; f, gynoecium, a portion of ovary wall removed to show basal Seer of ovules, semidiagrammatic, * 5: 2, open truit, show- ing irregular dehiscence, se eeds, and marcescent stamens ‘and petals, >< 2+ hi, ed, < 10; i, seedling (after Smith, 1931), x 8 1960] WOOD, SARRACENIACEAE AND DROSERACEAE 159 North Carolina, and southward to the south bank of the Santee River, Charleston County, South Carolina. Roberts and Oosting suggest that the total range is determined by gross soil characteristics (those of the St. Johns soil series), while local populations are restricted by the depth to which the surface soil desiccates in dry periods and by light intensity. Fire influences the latter and helps to maintain the limited habitat which is ecotonal between wet, evergreen-shrub bogs (pocosins) and the dry sands of Pinus australis-Aristida savannas. The slow seed ace (prob- ably by rain), the short period of seed viability, and the summer germina- tion under rather particular requirements of ae hi and loose organic matter on a sand surface are additional fac Dionaea is famous for its bear-trap-like hinged a Blidee each half of which is equipped with three sensitive bristles, which, when stimulated, cause the trap to close. Entrapped insects are prevented from leaving by overlapping marginal bristles and are digested by enzymes secreted by minute glands on the blade surface. Many problems remain in con- nection with the operation of the traps (which have been compared with those of Aldrovanda) and with the nutrition (both mineral and organic) of Dionaea. The occurrence of Dionaea, Sarracenia, Drosera, Utricularia, and Pinguicula together in soils low in mineral nutrients is suggestive. Although often occurring in widely scattered colonies, the species is locally abundant. Vast numbers formerly were sold as novelties, only to perish in cultivation, but the plant is now protected by law in North Carolina (General Statutes of North Carolina, Section 14-129.1). Dionaea can be cultivated for scientific purposes in constantly moist peat or peaty sand in full sunlight with a cool, dormant period in winter. If grown in shade the petioles become more broadly winged, successive traps diminish in size and the plant eventually dies. (Bright light is also important in flower production.) Propagation may be by seeds or by cuttings from the vegetative parts or the inflorescence before the flowers open. wering occurs from late May to mid-June (or July). Each pro- terandrous flower is open about three days. Seeds are shed by mid-July and germinate immediately. (Prompt storage under cold, moist condi- tions will maintain seed viability for up to a year, however.) Embryology and seed germination are similar to those of Drosera rotundifolia. In dry- ing, leaves and flowers tend to blacken, as with various hemiparasites. REFERENCES: Many references ee to 1942 omitted; see review and bibliography in Lioyp, Carnivorous Pla Crark, A. H. A new eee for Venus’ fly oe: (Dionaea Muscipula). Science 88: 188. 1938. [Se. corner of Moore Co., N. C. Coker, W. C. The distribution of Venus’ a trap (Dionaea Muscipula). Jour. Elisha Mitchell Sci. Soc. 43: 221-228. pl. 33. 1928. [Includes map, reprint of Ellis’ description. ] DuBois, C. G. Dionaea Muscipula. Bot. Gaz. 16: 200, 201. 1889. [Reports proterandry, duration of flowering, pollination. ] 160 JOURNAL OF THE ARNOLD ARBORETUM [VOL XLT Fratstapt, A. Anatomie der ie aes ie von Dionaea Muscipula Ell. Beitr. Biol. Pflanzen 2: 27-64. pls. 1-3. 1877. HARSHBERGER, J. W. An unusual aa e vegetative reproduction in Dionaea Muscipula. Bot. Gaz. 44: 382, 383. 1907. [Leafy branchlets in the inflores- nce; see also Contr. Bot. Lab. Univ. Penn. 1: 46-49. pls. 5, 6. 1892. | MacrartaneE, J. M. Contributions to the history of Dionaea Muscipula, Ellis. Contr. Bot. Lab. Univ. Penn. 1: 7-44. pl. 4. 1892. [Concerned mostly with leaf function. | Roperts, P. R., and H. J. Oostinc. Responses of Venus fly trap (Dionaea Muscipula) to factors involved in its endemism. Ecol. Monogr. 28: 193— 218. 1958. |The most comprehensive study. | SMALL, J. K. Dionaea Muscipula. Addisonia 10: 1, 2. pl. 321. 1925. [See also J. Stms, Bot. Mag. 20: pl. 785. 1804. | SmirH, C. M. Development of Dionaea Muscipula. I. Flower and seed. Bot Gaz. 87: 507-530, pls. 20-24. 1929. II. Germination and development of seedling to maturity. /bid. 91: 377-394. 1931. [Includes comments on Seeaepia See also Horm, T., Mem. Torrey Bot. Club 2: 71, 72. 1891, r seed germination. | a O. A physical analysis of the opening and ‘closing movements of the lobes of Venus’ fly-trap. Bull. Torrey Bot. Club 75: 22-44. 1948. [See also BRown and Suarp, Bot. Gaz. 49: 290-302. 1910. | e fundamental action potentials developed in the lobes of Venus’ ne (Dionaea Muscipula) due to a stimulus. (Abs.) Jour. Elisha Mitchell Sci. Soc. 66: 112-114. 1950. [A compound action potential of two components, “one attributed to a _ comparable to the neural C-action-potential of unmyelinated mammalian nerve fibers and the other to the movement of an eee yt prea with the hydrodynamic re- eur of the lobes in closure. E. B. DarpEN. The action potentials obtained from Venus’s-flytrap (Dionaca Muscipula). Science 111: 491, 492. 1950. & oO 2.. Drosera: lannaeus, Sp..PL-12 281. 1753 Gen, Fl, ed, 5.-136,.1754. Perennial or annual, insectivorous herbs, mostly of acid sands and bogs; plants rosulate, scapose [or sometimes caulescent and up to 1 m. high]. Leaves usually petiolate, the blades suborbicular to filiform, the margins and upper surfaces in ours with irritable, gland-tipped tentacles; stipules present or absent. Flowers 5-merous, the petals white, pink, rose-purple for yellow], marcescent. Stamens 5. Gynoecium of 3[5]| united carpels, the 3 styles deeply bifid in ours, the ovary 1-locular with 3 parietal placentae. Capsule splitting between the placentae, 3-valved, with numer- ous minute seeds. Embryo sac development normal (Polygonum type). Type species: D. rotundifolia L. (Name from Greek, droseros, dewy, from the glittering, glutinous droplets secreted by the glandular tip of each tentacle.) — SUNDEW, CATCH-FLY. More than 100 species in three subgenera, twelve sections, distributed primarily in the Southern Hemisphere (more than 50 species in Australia and New Zealand alone). Only sect. Drosera (sect. Rossolis Planch.) is represented in North America. Seven well-marked species occur in 1960] WOOD, SARRACENIACEAE AND DROSERACEAE 161 the United States, two, D. linearis Goldie and D. longifolia L. (D. anglica Huds.), entirely to the north of our area. Characters of seeds, stipules and leaves are especially diagnostic in Drosera. The genus is vegetatively plastic in response to environment, and, as a result, seeds are among the most dependable taxonomic criteria. Petal color, seldom noted, varies in some species: the petals of D. inter- media Hayne, white or pinkish in the North, are purple in South America, and those of D. brevifolia Pursh, usually white or pink, are pink to purple in Texas (and perhaps in South America). Drosera rotundifolia, of circumboreal distribution, extends southward on the Coastal Plain to southeastern Virginia (and Charleston, South Carolina?) and along the mountains to northern Georgia. Leaf shape, stipules, and pale sigmoid-fusiform seeds are characteristic. Drosera capillaris Poir. (D. tenella Willd.), vegetatively similar, but smaller, with ellipsoid seeds coarsely papillose-corrugated in 14-16 ridges, and with a very different range (Virginia to Florida and Texas; Cuba, Jamaica; northern South America; southern Mexico, British Honduras), is some- times confused with this species. Drosera brevifolia (Virginia to Florida, Tennessee, Arkansas, Texas, Cuba, northern South America (?), Para- guay, Uruguay, and northern Argentina) lacks stipules and has glandular scapes and sepals and obovoid seeds with crateriform markings. Drosera intermedia, with conspicuous stipules, oblong-spatulate leaf blades, scapes which curve outward at the base, and oblong, densely papillate seeds, is of wide distribution from Europe to Newfoundland and westward to Minnesota, southward along the Coastal Plain to Florida and Texas, and to Cuba, Hispaniola, and northern South America. Drosera filiformis Raf. var. filiformis, with erect, filiform leaves to 25 cm. long, and purple tentacles and petals, is confined to the Coastal Plain from Cape Cod, Massachusetts, southward to southeastern North Carolina (? South Caro- lina and Georgia). Varietas Tracvi Diels (D. Tracyi Macfarl.), of west- ern Florida and southern Georgia to Mississippi, seems to be separated geographically from var. filiformis, although it has been reported from South Carolina. It is a distinct taxon, differing in the larger leaves (to 40 cm. long), the greater size of the floral parts, and the green tentacles. Section Drosera is a polyploid group of about 40 species among which chromosome numbers of 2” = 20, 40, and 80 are known. The North American taxa are diploid, with the exception of D. longifolia (2n = 40) which appears to be of amphiploid origin. (Species of other sections have been reported with 2m = 28, 32, 34, and 60.) Three naturally occurring diploid hybrids and two triploids, all sterile, are known from this section. Chromosomal sterility is indicated in one of the diploid hybrids. In some areas, D. brevifolia, D. capillaris, D. filiformis, and D. intermedia occur in close proximity, but hybrids are unreported in our area, although that of the latter two has been found once in New Jersey. An artificial hybrid might show whether a genetic barrier exists between D, filiformis var. filiformis and var. Tracyi. The leaves of Drosera catch and digest various small animals (mostly 162 JOURNAL OF THE ARNOLD ARBORETUM [voL. XLI insects) which become entangled in the mucilaginous secretion of the tentacular glands. The tentacles bend inward, pressing the capture against the leaf blade which, in some species (e.g., D. rotundifolia, D. intermedia, but not D. filiformis), may also bend enfolding the prey. Both movements involve cell elongation: a single tentacle is capable of bending approxi- mately three times before the limit of growth is reached The genus exhibits remarkable regenerative powers, and, in some species (e.g., D. rotundifolia), asexual reproduction by adventitious plant- lets from inflorescences or detached leaves is common. Individual flowers of the scorpioid (rarely forked) cyme are open for only a single morning in sunlight and are self-pollinated upon closing if cross-pollination has not already occurred. Some of our species produce overwintering buds of tightly compacted leaves, but D. brevifolia apparently behaves as an annual in many areas and has been so described (D. annua Reed REFERENCES: See family references. Lroyp (1942) summarizes especially structure and func- tl leaves. Most references included by him have been omitted. Arisz. W. H. Absorption and transport by the tentacles of Drosera capensis. V. Influence on the transport of substances inhibiting enzymatic processes. Acta Bot. Neerl. 2: 74-106. 1953. [See his ee for I- Barctay, H. G. A veer report of the ecology of a Drosera aentens Proc. Okla. Acad. 18: 22-25. 1938. [D. brevifolia (as D. annua) in se. Okla. | Brat, L.. G. Kenpa, and F. WEBER. Redes fehlen den Schleisszellen von Drosera. Protoplasma 40: 633-635. 195 BRUMMER-DINGER, C. H. Notes on Giana Droseraceae. Acta Bot. Neerl. 4: 136-138. 1955. | Taxonomic notes; includes 2 or 3 of our spp. | CHURCHILL, J. R. An unusual form of Drosera intermedia, var. americana. Rhodora 2: 70, 71. pl. 15, fig. 8. 1900. | Caulescent form in deep water (= natans Heuser). | Favarp, A. Etude histologique de la formation du point végétatif dans le hee ees de{s] pétioles de Drosera rotundifolia L. Compt. Rend. Acad. Sci. Paris 237: 1753-1755. 1953. —. Formation et premieres différenciations au. systeme conducteur chez Drosera Labia L. au state jeune. /bid. 240: 338-340. 1955 ence dune différenciation précoce sur lévolution re systéme er an de Drosera Macon Lies L. lors de la germination. Avortement de la racine. [bid, 242: 405-407. ———. Evolution ae point végétatif A en rotundifolia L. Deuxiéme étape. foreh 2386-2 : Se cu du point végétatif en méristéme d ‘inflorescence chez Drosera rotundifolia L, Ibid. 243: 2124-2127. 1956. Et es inflorescences de Drosera les Hayne: la cyme uni- pare scorpoide et ses modifications. /bid. 246: 8—2511. 1958. . Etude cytologique du dév slopperent de Teena de Drosera rotundi- folia L.: ontogenése des méristémes primaires. /bid. 248: 1570-1573. 1959. FERNALD, M. L. A peculiar variety of Drosera rotundifolia. Rhodora 7: 8, 9. 1905. [Var. comosa Fern. See also M. Levine, Further crane ee on 1960 | WOOD, SARRACENIACEAE AND DROSERACEAE 163 chloranthy in Drosera intermedia. Gaz. 62: 389-399. pl. 13. 1916. Both instances a response to ene ane ecological conditions? | Hatt, B. A. The floral anatomy of Drosera and Begonia and its bearing on the theory of carpel polymorphism. Am. Jour. Bal 36: 416-421. 1949. Hamet, M. R. Observations sur le genre Drosera. Bulk soc, Bot. Pr. IV. 7: 26-38, 52-76. pl. 2. 1907. [Taxonomy; key to all spp.; includes drawings of seeds of some. | Hooker, H. D., Jr. Mechanics of movement in Drosera rotundifolia. Bull. Torrey Bot. Club 44: 389-403. 1917 Hutten, E. The amphi-Atlantic plants and their phytogeographical connec- tions. Sv. Vet-akad. Handl. IV. 7(1): 1-340. 1958. [D. intermedia, p. 56, map 38; dots in western Canada probably D. longifolia. | Leavitt, R. G. Seedlings and adventitious plants of Drosera. Torreya 9: 200— 203. 1909. ee M. Somatic and reduction divisions in certain species of Drosera. m. N. Y. Bot. Gard. 6: 125-147. pls. 16-19. 1916. ere J. M. Observations on some hybrids between Drosera filiformis and D. intermedia. Contr. Bot. Lab. Univ. Penn. 2: 87-99. pl. 12. 1899. NARASIMHACHAR, 8S. G. A contribution to the embrvology of Drosera bur- manni Vahl. Proc. Indian Acad. Sci. B. 29: 98-104. 1949.* —. A contribution to the embryology of Drosera indica, L. Ibid, 33: 290- 295. 1951 Paris, R., and J. C. Denis. Les droséras: leur ee tee dans divers ao eetre Ann. Pharm. Fr. 15: 145-159. 19 Pace, L. Parnassia and some allied genera. Bot. ae e| 306-324. 1912. [In- cludes development of embryo sac of D. rotundifolia. | Patankar, T. B. V. Further contribution to the embryology of Drosera bur- manni Vahl. Proc. Indian Acad. Sci. B. 43: 161-171. 1956. Raju, M. V. S., and T. B. V. Paranxar. Pollen morphology in three species A Drosera L. Grana Palynologica II. 1: 153-155. 1956. [D. Burmanniz, D. indica, D. peltata. See also G. Erdtman, Sv. Bot. Tidskr. 39: 1945.] Reep, E. L. Ecologic notes on Drosera annua. Torreya 16: 125-130. 1916. [See also ibid. 15: 246, 247. 1915.] RosENBERG, O. Cytologische und morphologische studien an Drosera longifolia x rotundifolia. Sv. Vet-akad. Handl. 43(11): 1-65. 4 pls. 1909. [Classic study of chromosomal behavior in a triploid hybrid; embryo sac and pollen ae t.] Sources, E. C. R. Embryogénie des Droséracées. Développement de l’embryon chez le Drosera rotundifolia L. Compt. Rend. Acad. Sci. Paris 202: 1457- 1459. 1936 TraNkowsky, D. A. La spermatogénese et la fécondation chez la Droseére (Drosera). (In Russian; French résumé.) Bull. Mosk. Obsch. Isp. Pri. Biol. 47: 104-112. 1938. Woop, C. E., Jr. Evidence for the hybrid origin of Drosera anglica. Rhodora 57: 105-130. 1955. [= D. longifolia L.; includes miscellaneous notes on U. S$. spp.; summary of chromosome numbers. Wynne, F. E. Drosera in eastern North America. Bull. Torrey Bot. Club 71: 166-174. 1944. [Key, descriptions, maps, illustrations of seeds of the North uae spp. ——. Drosera filiformis. Addisonia 22: 39, 40. pl. 724. 1945. [See also W. J. Hooker, re Mag. 63: pl. 3540. 1836. ] 164 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLI A REVISION OF THE GENUS CLETHRA IN CHINA SHIU-YING Hu TWENTY SPECIES OF CLETHRA have been recorded from China. Ninety per cent of these were described as new entities. These descriptions were prepared by British, French, American, Austrian, Dutch, and Chinese botanists, and their research was conducted in nine different European and American botanical institutions. In reviewing their work, it becomes apparent that some of them had no proper understanding of the variabil- ity of a species of Clethra. Consequently, repetition and confusion were created for the taxonomy of the Chinese species. Under these conditions the naming of a new collection of Clethra becomes a seemingly hopeless task. Even in the herbarium of the Arnold Arboretum where outstanding collections of Chinese specimens, including many types and photographs of types accompanied by fragments of the type-material, are deposited, one found it difficult to name an unidentified specimen of Clethra. A generalized treatment of the genus which contains a well-balanced key to the species and concise descriptions of them to aid in the identifica- tion of specimens is lacking. To supply such a treatise is the immediate aim of this revision. The genus Clethra has an Asiatic-American distribution. An analysis of the countries of origin of the species as listed in the /ndex Kewensts shows that China and Mexico have the largest numbers of species. Careful examination of the species described from these two countries and their adjacent land-masses indicates that morphological variations among the Chinese species are far more diverse than among their tropical American allies. Actually, all the outstanding morphological features of the Ameri- can species can be found among some of the Chinese species; yet the varia- tions of certain Chinese forms are not to be found among the New World species. For this reason, a systematic revision of the Chinese Clethra has a more far-reaching significance than the merely utilitarian aim mentioned above, It seems that the problems involved in the classification of Chinese Clethra constitute the core of investigations which may lead to an under- standing of the morphology, taxonomy, and geographical distribution of the genus as a whole. A proper knowledge of the Chinese species may throw some light on future studies of the American species. Dr. Robert C. Foster of the Gray Herbarium, Harvard University. has kindly read this manuscript and given many helpful suggestions. To him I wish to express my grateful appreciation. Unless accompanied by a note of explanation, all the specimens cited in this article are deposited in the herbaria of the Arnold Arboretum and the Gray Herbarium. 1960 | HU, CLETHRA IN CHINA 165 HISTORY The genus Clethra was established on a North American species, C. alnifolia, originally known to European naturalists from Carolina and Virginia. The name Clethra (derived from the ancient Greek name for the alder, klethra, in allusion to the resemblance of the leaves of the type species to those of Alnus) first appeared in Gronovius’ Flora Virginica, published in 1739. When Gronovius created this name he cited two earlier references, one from Plukenet’s PAytographia, published in 1691, and the other from Catesby’s The Natural History of Carolina, published from 1731 to 1743. He even quoted Catesby’s diagnostic comment on the taxon intact, “Alni folia Americana serrata, floribus pentapetalis albis, in spicam dispositis.” Linnaeus used Gronovius’ name, giving the same quotation. A specimen of C. alnifolia is preserved in the Linnaeum her- barium, number 567.1. The species was introduced into European gar- dens in the early eighteenth century. For a century and a quarter after the establishment of the genus, the classification of Clethra was based principally on American species. In 1839, De Candolle, in his Prodromus, placed Clethra in the tribe Andro- medeae of the family Ericaceae. His description of the anthers represented an erroneous observation. Apparently he mistook the developmental stage in the flower bud to be the mature anther and, though he did mention the inversion, he interpreted the cordate portion as the base and the mucronate and muticous end as the apex. In regard to the classification of the species, he arranged those known at that time in two sections, Euclethra and Cuellaria. He characterized his section Euclethra as having exserted stamens and pistils and deciduous leaves, and placed five North American species in this section. He characterized the section Cwedlaria as having the stamens and pistils included and assigned 17 species from Central and South America and a single species from the Madeira Islands to this section. Although Blume had published C. canescens Reinw. from Celebes and Borneo, De Candolle did not include this species in his sys- tem. In 1851 in an article, Studien iiber die natiirliche Klasse Bicornes, Klotzsch separated Clethra from the family Ericaceae and raised it to a family, Clethraceae, on the strength of its choripetalous corolla, corolla and stamens deciduous after anthesis, its obcordate anthers which are retro- flexed at first and then introrse and dehiscent by apical rimiform pores, and its bifid or trifid stigmas. The portion of the Bentham and Hooker’s Genera Plantarum that covers Clethra was prepared by Hooker. In this work Hooker placed Clethra at the end of the family Ericaceae as “Genus Anomalum.” By this time several species had been introduced into cultivation in England, e.g., C. arborea from the Madeira Islands, C. quercifolia from Jalapa in Mexico. C. tomentosa and C. alnifolia from eastern North America. With live plants as well as herbarium specimens at his disposal, he gave the most accurate description of the important characters of the genus. Re- 166 JOURNAL OF THE ARNOLD ARBORETUM feices rame aey garding the anthers, he pointed out that they are affixed at the middle of the back, are often obovate in shape and acute at the base, are extrorse at first then turning introrse, and dehisce above the middle by elongate pores. The Chinese Clethra were not known to botanists until 1883. In that year Hance published C. fabert from Loh-fau Mountain of Kwangtung Province. Two years before this, Franchet, in his Plantae Davidianae, had published Clethra scandens. However, this species is not a true Clethra, and since then has been made the type species of the genus Clematoclethra Maxim. When Hemsley prepared the enumeration of the plants from China, he saw Henry’s collections of Clethra from Hupei. He was indeed not critical in his observation. In publishing Forbes and Hemsley’s /ndex florae sinensis in 1889, he treated Henry’s collections as C. canescens Reinw., a species known only from Western Malaysia. In the same work he reduced C. faberi Hance (from Kwangtung), C. /ancifolia Turcz. (from the Philippines), and C. barbinervis Sieb. & Zucc. (from Japan) to C. canescens. In the same year O. Drude published his contributions on Clethraceae, Ericaceae, and other related families in Engler and Prantl’s Die natiir- lichen Pflanzenfamilien, In this work Drude accepted Klotzsch’s proposal of raising the genus to a family and in the description of the genus, fol- lowed Hooker. In the classification of the species, he adopted De Can- dolle’s two sections and placed the Asiatic species C. barbinervis in sec- tion Huclethra. He separated C. arborea Ait., a species from the Madeira Islands, from section Cyellaria and assigned it to an unnamed section which he placed between Exclethra and Cuellaria. Meanwhile, material collected by French missionaries who were sta- tioned in central and western China reached Paris. Franchet in 1895 published C. fargesi from eastern Szechuan and C. delavayi from western Yunnan. Incidentally, these species define the northern and western limits of the range of the genus in China. Bodinier, Cavalerie and Esquirol were French missionaries who were sent to Kweichow in later years. Their collections reached Léveillé who in 1912-3 published six more species of Clethra: C. bodinieri, C. cava- leriei, C. esquirolii, C. kaipoensis, C. lineata and C. pinfaensis. As indi- cated by a key to these species which he prepared in the Flore du Kouy- T'chéou, Léveillé observed several characters which are important for dis- tinguishing the Chinese species of Clethra. He observed that some species possess entire stigmas, while others have lobate or cleft ones. He also noticed that some species have elongate pedicels, while others have short ones. The arrangement of the racemes was also mentioned in his key, as some were umbellate while others were solitary. After E. H. Wilson returned from his third trip to West China, Rehder and Wilson worked cooperatively on Wilson’s numerous collections. In 1913, they described C. monostachya, a species characterized by its pilose styles. When Rehder worked over the ligneous plants described by Lé- 1960 | HU. CLETHRA IN CHINA 167 veillé. he was able to examine the types of Léveillé’s six species of Clethra which were described from Kweichow. In 1934 Rehder recognized three of them and merged C. esquirolii and C. lineata with C. cavaleriei and C. pinfaensis with C. kaipoensts. Aside from the species described by these workers, seven additional binomials were added to the genus by Handel-Mazzetti, Sleumer, Hao, Fang. and Li. None of these authors seems to have attempted to place his species in the existing taxonomic system, and some of these names are synonymous with earlier described species. TAXONOMIC CHARACTERS Habit of growth. The species of Clethra generally have the habit of growth common in the family Ericaceae. The branches have subverticillate branchlets with leaves crowded at the apex. The majority of the Chinese species are deciduous and two of them are evidently evergreen. In the deciduous species. each vegetative branchlet has a large, sub- sessile terminal bud and 1-3 subterminal axillary buds (Fic. 24, right). In the following year. the large terminal bud of a mature plant develops into a flowering shoot provided with a few smaller leaves. a bracteate raceme or panicle, and 1-3 axillary buds immediately below the inflores- cence (Fic. 24, left). The subterminal buds of a vegetative branchlet and the axillary buds of a flowering branchlet are of the same nature and function. They are all active during the growing season and consequently attain various stages of development before the leaves fall in the autumn. The more vigorously growing ones develop into stems 4-6 cm. long, each bearing 3 or 4 leaves crowded at the shoot apex around the terminal winter bud. The weaker ones generally develop into short stems 1-2 cm. long. each bearing 1 or 2 very small, bract-like leaves surrounding the terminal winter bud. Often the stem- -portion becomes so short that the bud appears as though it were a stalked winter bud. In this case. it bears no leaf, In the next growing season these winter buds either unfold and develop into short vegetative shoots with large leaves and strong, fat terminal buds ready to produce flowers the third vear. or they may de- velop into flowering shoots, depending upon the vigor of the plant. the environmental condition. and the heredity of the species. n an evergreen species, an average branchlet of a mature plant is usually a flowering branchlet (Fic. 25). From one to four axillary buds immediately below the inflorescence develop into leafy shoots, each bear- ing a terminal bud and several normal leaves crowded at the apex. Nor- mally. this terminal bud produces a flowering shoot during the next crowing season. The inflorescences of the Chinese evergreen species are all simple racemes. Schneider in 1910 observed the evergreen habit of C. arborea Ait. and used it as a key character in the classification of the then-known culti- vated species of Clethra. With the Chinese species the evergreen habit 168 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XLI is correlated with an entire stigma. It is an obvious key character for distinguishing species. Leaves. The leaves on the branchlets of a deciduous species are of two kinds. Those developed from a winter bud are the normal leaves and are usually the larger ones. Those developed from axillary buds which become active during the same season in which they are produced are always smaller and narrower. They often become bract-like and are more pubescent than the normal leaves on the same plant. In the study of Clethra, whenever the comparison of leaf-characters is employed, care must be exercised so that the comparison is made between leaves of the same origin. With the Chinese species, the normal leaves of a species seem to be quite stable in shape, indumentum, dentation and venation. Any variation occurring in these respects is usually correlated with some flower character, and, for this reason, the leaf-characters are used as auxiliary criteria for distinguishing species. Inflorescences. The Chinese species of Clethra are summer bloomers. The earliest flowering species, C. bodinieri, begins to flower in June and its fruits mature in early August. The latest flowering species, C. kaipoen- sis, is in full bloom in middle September. The majority flower in July and August, however. The inflorescences of most Chinese species of Clethra are either simple terminal racemes (Fics. 25, 26), or paniculate racemes. The panicles are sessile and subumbelliform (Fic. 27). The inflorescences of weak or old branchlets of a species normally bearing panicules may appear simply racemose. However, a good specimen of the species usually has paniculate inflorescences. The form of inflorescences, when correlated with leaf and flower characters, is a convenient and rather reliable character for dis- tinguishing species. For this reason, it has been used as a key character in most manuals. Rachis. The rachises of all the species of Clethra are densely pubes- cent. Those of the majority of the Chinese species have stellate hairs, but the rachises of C. bodinicri have simple, straight, appressed hairs. The type and density of the indumentum on the rachises is a useful aid in the recognition of species. Bracts. The bracts subtending individual flowers have been used by American botanists for distinguishing C. alnifolia and C. acuminata. Some botanists put special emphasis upon the relative length of the bracts and the flowers. Actually, the bract is a rather poor character to employ for distinguishing species, for the bracts of Clethra are caducous. In general, the first opened flower of a raceme is situated a little below the middle of the inflorescence and the other buds open progressively towards both ends. The bracts and the mature flowers are not present at the same time (Fics. 26, 27). It is only when the inflorescences are in the bud stage that the flower buds are subtended by bracts. The degree 1960 | HU, CLETHRA IN CHINA 169 of maturity of the inflorescence makes a great difference in the relative length of the bracts and the flowers. This character should be used with great caution. It can never be used as the only criterion for distinguishing species. Pedicels. The pedicels of Clethra exhibit a definite pattern of elonga- tion during anthesis. As illustrated by Fics. 1-6, which show stages from a fully grown flower bud to the dropping of the petals, the pedicels nor- mally increase three times in length. The lengths of the pedicels of indi- vidual flowers of a species at comparable stages of development are rather uniform, however, and, for this reason, the Chinese species may be divided into two distinct groups, the long-pedicellate and the short- pedicellate. Species in the long-pedicellate group have pedicels 5 mm. or more long immediately before anthesis (Fic. 1). When the petals begin to open, the length of the pedicels has doubled (Fic. 2), and by the time the petals drop off, the pedicel is three times as long as that of a mature flower bud (Fic. 3). After anthesis the pedicel elongates slightly; there- fore the fruiting pedicel of a species is longer than that of the flower. The pedicels of the short-pedicellate species are 1-2 mm. long imme- diately before anthesis (Fic. 4) and are normally shorter than the sepals. During anthesis, these pedicels elongate three-fold. As the sepals do not increase in size during this period, the length of the pedicels and that of the sepals are almost equal (Fics. 5, 6). The relative length of the pedicels and sepals is a reliable taxonomic character in Clethra. In employing this character one may use it only when the pedicels of the specimens are in comparable stages of development. Calyx. The Chinese species of Clethra are distinctly gamosepalous. The calyx is essentially patelliform, deeply divided into 5 lobes. The short united portion is truncate and ridged at the basal end. The imbricate lobes vary in shape and size. The calyx attains mature size early in the development of the flower and maintains its shape and size during an- thesis. Thus the calyx of a young flower bud and that of the young fruit of a species are about the same shape and size. The calyces of the Chinese species of Clethra differ in the degree of division and in the shape and size of the lobes. The calyx of C. bodinieri is cut half way to the middle and its lobes are ovate, only 3 mm. long (Fic. 7). The calyx of C. delavayi is deeply cut with lanceolate lobes (Fic. 9) while the calyx of C. Raipoensis is deeply cut with oblong lobes (Fic. 8). The characters drawn from the calyx and sepals are reliable and are usually correlated with the length of the pedicel, the shape of the petals, and the relative length of the filaments. They can serve as good criteria for distinguishing sections and series. Corolla. All the species of Cletkra are choripetalous. Like the sepals, the petals of different Chinese species of Clethra can be grouped into three types. In C. bodinieri the petals are oblong, ciliate along the margin, ven- 170 JOURNAL OF THE ARNOLD ARBORETUM [voL, xtt 21 —_ . a \ \ / 22 23 AS 20 Fics. 1-23. MoRPHOLOGICAL DETAILS OF CHINESE SPECIES OF CLETHRA., Fics. 1-6. Elongation of pedicels during anthesis, showing stages before, at, and after anthesis, < 1: 1-3, a long-pedicellate species, C. esquirolii (Tsoong ) 83440); 4-6, a short-pedicellate species. C. brammeriana (Wang 247). Fics. 7-9. Calyces, X 3: 7, shallowly divided calyx of C. bodinieri var. parvifolia (Tsang 22450) with ovate lobes; 8, deeply divided calyx of C. delavayi (Delavay 3319); 9, deeply divided calyx of C. kaipoensis (Tsang 27916) with oblong lobes. Fics. 10-12. Stamens. & 7: 10, C. cavaleriei (Tsang 21413), stamen from rather large flower with exserted stamens — note glabrous filaments, V-shaped elongate anther, slim thecae, elongate rimiform apical pores, and long basal process; 11, C. delavayi (Forrest 11590), stamen from large flower with in- cluded stamens —note villous filament, oblong-obovate anther, rather plump thecae, broad apical pores, short basal process; 12, C. brammeriana (Wang 247), stamen from small flower with exserted stamens — note glabrous filament 1960] HU, CLETHRA IN CHINA el trally barbate, and emarginate at the apex (Fic. 7). In C. delavayi the petals are large, almost entire, neither ciliate nor barbate, and are longer than the stamens (Fic. 8). In C. faberi and C. kaipoensis the petals are obovate, fringed, glabrous or slightly villose on the inside, and shorter than the stamens (Fic. 9). Petal characters, when correlated with char- acters of the pedicel, sepal, and stamen, can be used for distinguishing series. Stamens. A flower of Clethra has ten stamens disposed in two whorls, the outer opposite the petals, and with shorter filaments (Fic. 13). In C. bodinieri the sepalad anthers are slightly longer. In C. barbinervis the difference in the size of the anthers is not appreciable. The filaments of Clethra may be glabrous, hirsute, or villose. The in- dumentum of the filament has been used as a specific character. Before experimental evidence of the importance of hairs in the speciation of Clethra is available, this easily detected character is used for distinguish- ing species. In the historical review we have noted that Hooker in 1876 gave a very accurate description of the stamens of Clethra. He pointed out that the anthers of this genus are obovate and the thecae dehisce above the middle by elongate pores. It appears now that in some widely accepted manuals and text books on the taxonomy of vascular plants published in recent years, different descriptions of the anthers and their dehiscence are given. Some of these statements give a rather incomplete picture, while others involve inaccurate observations. For example, Fernald in 1950 stated, “anthers extrorse in bud and opening by pores at base and inverted in flower” and Lawrence in 1951 noted, ‘‘anthers 2-celled, extrorse, sagittate, inverted and inflexed in bud, dehiscing by apical pores.’ The determina- tion of the apex and base of an anther and the introrse or extrorse dehis- cence of its thecae depends upon the attachment of the anther to the fila- ment. The anther in Clethra is dorsifixed, often appearing versatile at abruptly enlarged at base, very short obcordate anther, rather plump thecae, small apical pores Fics. 13-16. Stages in maturation of stamens to show straightening of fila- ments in C. cavaleriei (H. H. Chung 2932), * 5: 13, position of young anthers i J id wa elongation of filament; 16, lateral view of stamen and petal at a more advanced stage, the filament almost straightened. Fic. 17. Fully opened flower of C. bodinieri (Tsang 22450), showing short stamens of one: whorl and long stamens of inner whorl, some petals and stamens remov Fics. 18- soe Pollen grains of C. delavayi, C. bodinieri, and C. faberi respec- tively. Fics. 21-23. Style and stigma types, X 10: 21, trifid style with three stigmas (C. cavaleriei): 22, undivided style with 3 apical lobes, 3 stigmas brammeri- ana); 23, undivided style with single punctiform stigma (C. bodinieri). 172 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI [es PRE) § ncn, KA is 45 Fics. 24—27. Habit and inflorescence types of Chinese species of Clethra, x VM. 24, Flowering branchlets of C. esquirolii (Steward & Cheo 394), showing formation of vegetative shoots from active axillary buds; 25, Fruiting branch- let of C. bodimieri var. parvifolia (Tsang 23822), showing leaves on growth of two years; 26, C. delavayi (leaf, Delavay 3319; flowering branchlet, Forrest 1960] HU, CLETHRA IN CHINA 173 anthesis. A mature anther is V-shaped (Fic. 10), obovate-oblong (Fic. 11), or obcordate (Fic. 12). The thecae dehisce by rimiform apical pores situated slightly on the ventral side of the anther. In a flower bud, the inverted position of the anther is caused by the outward folding of the filament which occurs at a point about one-fifth below the attachment of the anther (Fics. 13, 14). As the flower opens, the apical half of the filament increases in length several times faster than the basal half becoming geniculate in character (Fic. 15). The tension created by the elongating filament pushes the anther up (Fic. 16) and as the filament becomes straightened, the anther is set at its normal position with the apical pores pointing upward and inward (Fics. 1-12, 17). In 1952 Kavaljian published on the floral morphology of Clethra, and regarding the inversion of the anther he stated, “the morphological base becomes the apparent apex.” This statement is unnecessary because the rolling and folding of filaments in flower buds and the straightening of the filaments at anthesis are as common as the vernation and the un- folding of leaves. They are normal processes of growth. In Clethra the change of position of the anthers from bud to fully grown flower occurs in the filaments, not at the points of attachment of the anthers to the fila- ments. Therefore, what Kavaljian termed the “apparent apices” of the anthers are the “morphological apices.” Because of the change of the length of the filament during the short period of anthesis, specimens collected at different stages in the develop- ment of the flower may have different appearances. For this reason the nature of the flower in regard to its included or exserted anthers is not a very satisfactory criterion for distinguishing species. It can be used only as an auxiliary character. Variations in the size and shape of the anthers of different species are obvious. The longest anthers, such as those found in C. bodiniert and C. cavaleriei, are 2.5-3 mm. long, V-shaped in out- line, deeply parted above, with relatively slim thecae and elongated rimi- form apical pores (Fic. 10). The medium-sized anthers, such as those found in C. delavayi and C. monostachya, are 1.75—2 mm. long, obovate- oblong in outline, moderately parted above, with rather plump thecae and short and wide apical pores (Fic. 11). The short anthers, such as those found in C. kaipoensis and C. brammeriana, are only 1 mm. long, obcordate in outline, shallowly parted above, the thecae short, plump and with small apical pores (Fic. 12). Anther characters correlate with the size of flowers and the shape and sizes of sepals and petals. They can be used as auxiliary characters for distinguishing sections or series. Pollen grains of C. delavayi, C. bodinieri and C. faberi taken from herbarium specimens were examined. They differ in shape and structure in their apertures (Fics. 18-20). The grains of C. delavayi and C. faberi are tricolporate. The grains of C. faberi lack the protruding roof of the ora which Erdtman considered to be a characteristic feature of the pol- 15889); 27, C. fargesii (Henry 7270), showing subumbelliform racemes, small vegetative shoots developed from active axillary buds 174 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI len of the genus. The pollen of C. bodinieri seems to be porate and not colporate. The significance of pollen morphology in the taxonomy of in- frageneric groups of Clethra awaits further research and judgment by com- petent palynologists. Pistil. The pistil is tricarpellate. The ovary is always pubescent, while the style may be pubescent or glabrous, trifid or undivided at the apex. The presence or absence of hairs on the style has been used as a specific character, and this practice is adopted here. The trifid style has three terminal stigmas (Fic. 21). Most of the Chinese species of Clethra have trifid styles. The undivided style may have 3-lobed stigmas such as are found in C. brammeriana (Fic. 22), or it may have a single punctiform stigma such as is found in C. bodinieri (Fic. 23). The type of style and stigma may be used as an auxiliary character for distinguish- ing species in Clethra. Fruits. The fruit is a loculicidal capsule with persistent sepals and style. The relative length of the fruiting pedicels is a helpful character for distinguishing species. There seems to be a positive correlation between the size of the flower and fruit of a species. The fruits of the large-flowered species C. delavayi are 4-6 mm. in diameter, while those of the small- flowered species C. kaipoensis are 2.5-3 mm. in diameter. Most of the herbarium specimens which I have examined lack fruits, and their sig- nificance in the classification of the species is not emphasized. Seeds. Material with fully grown seeds is scarce in our herbaria. The seeds of C. bodinieri, C. barbinervis, C. cavaleriei, C. delavayi and C. monostachya were examined. Those of the first three species are irregu- larly angular, extremely reticulate and not winged. The seeds of C. delavayi and C. monostachya are slightly compressed, 1 mm. long, 0.5 mm. wide, reticulate and the cells along the margin are enlarged, but no wings are evident. The seeds of the Chinese Clethra are very different from the highly compressed and winged seeds of the North American species. ECOLOGICAL NOTES AND GEOGRAPHICAL DISTRIBUTION The Chinese species of Clethra are essentially mountainous forms. They occur in thickets or at the margin of woods in acid soil. The available material in our herbaria indicates that there are two distinct groups of species, the western and the eastern, which are separated by six hundred miles. No collection of Clethra has been recorded in the area in between. The western species constitute a small series, Delavayanae, of high mountain forms. They usually occur at altitudes of 2400-3800 meters. Clethra delavayi, C. yuiana, and C. monostachya belong to this group. In C. delavayi a glabrous variety has been recorded from a higher alti- tude than the pubescent variety. In this western range, C. monostachya occurs in the North and C. delavayi and C. yuiana in the South (Map 1). The eastern group involves a more heterogeneous assemblage of species. 1960 | HU, CLETHRA IN CHINA LS The area covered by these species is at lower altitudes, the majority of the species occurring between 600 and 1000 meters above sea level. A couple of species may reach an altitude of 1250-1600 meters. Clethra Raipoensis has been collected at altitudes of 1000-2100 meters. The distribution maps of different species involved in the eastern region indicate that the Nan-ling Range constitutes the center of the aggrega- tion of species. This range represents a chain of much-dissected ancient hills and mountains extending along the Tropic of Cancer from the Yun- nan-Kwangsi border, eastward to the Kweichow-Hunan-Kwangsi-Kwang- tung borders, and then northeastward to the Kiangsi-Fukien-Chekiang borders. Clethra bodimieri, the only species of section Monostigma, occurs prin- cipally on the southern flank of the Nan-ling Range, and from the west- ern end, its range extends southward to Hainan (Map 1). The distributions of three closely related species, C. cavaleriei, C. esqui- rolii and C. fargesii, form a U-shaped pattern with the base falling over the Nan-ling Range, the left arm extending along the Kweichow-Hunan border northward to the Szechuan-Hupei border, and the right arm ex- MOM AN it 3 ‘ 4 ua (am r., \ ae Si al = Pease BRS © cavaleriei & esquirolii ae @ fargesii . O monostachya am Saeeeee A 1 Sx nor eee, ail & bodinieri s per INDOCHINA @ = delavayi _ é } oi ‘ ‘a { oN N @ faberi “ewes O brammeriana, ma , 5 4 kaipoensis me " rar Maps 1-4. Distribution of some Chinese species of Clethra. 176 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI tending along the Kiangsi-Fukien border northward to southern Chekiang (Map 2). It is worthy of note that C. fargesii has its best development in the north on the western arm while C. cavaleriei is better developed in the north of the eastern arm. Map 3 shows the distribution of C. barbinervis in China and Japan. This species and C. fargesii are the eastern Asiatic representatives of series Alnifoliae. The other species of this series, C. alnifolia and C. acuminata, occur in eastern North America. The distribution of the species of Clethra in series Alnifoliae gives specific evidence to support the floristic relationship between the Lower Yangtze Region in China and eastern United States of America. Clethra faberi is a unique species. In geographical distribution its range does not overlap that of any other Chinese species (Map 4). In morphological characters it has the subcoriaceous leaves of the evergreen species in section Monostigma, the umbelliform-paniculate racemes and fringed petals of species in section Cyellaria, and the V-shaped anther of species in section Clethra. Evidently the distribution of C. fadert is on the northern periphery of the range of the series Faberianae. Clethra petelotii from Indochina, C. lancifolia from the Philippines, C. canescens from the Philippines, Borneo, and Celebes, and C. ledermanni from the western highlands of New Guinea are all related species. The Chinese species of section Cyellaria are concentrated in a more limited area of the middle Nan-ling Range (Map 4). It is noteworthy that all other species of this section are found in tropical America from Mexico to Brazil. In interpreting the distribution of the fresh-water fishes from tropical eastern Asia to the New World, Darlington in 1957 suggested that during different geological times, many kinds of fresh- water fishes evolved in the eastern Tethys region have radiated by way of the north temperate region to North and Central America. The dis- tribution of Clethra in geological time probably followed the same direc- tion. Clethra is an Asiatic-American genus. It is evident that C. arborea Ait., the single species described from the Madeira Islands, is out of the natural range of the genus. A comparative study of all the material in the Gray Herbarium and the Arnold Arboretum reveals that C. arborea from the Madeira Islands is closely allied to C. pulcherrima of the east coast of Sumatra. The two species are alike in their arborescent habit, large, oblong-elliptic, pubescent leaves, paniculate racemes, large fiowers, elon- gate petals, included stamens, hirsute filaments and V-shaped anthers. It is very clear that C. arborea is an introduction from the East Indies to the Madeira Islands in historical time. The small hairy capsules and the very minute seeds are adapted for attachment. It was probably acci- dentally introduced in connection with the spice trade, and was firmly established in the late eighteenth century when it first caught the atten- tion of British naturalists. Both C. pulcherrima and C. arborea are closely related to C. delavayi of western China. 1960] HU, CLETHRA IN CHINA Lah TAXONOMY Clethra Linn. Sp. Pl. 396. 1753, Gen. Pl. ed. 5. 188. 1754: Benth. & Hook. f. Gen. Pl. 2: 603. 1876; Rehder, Bibl. Cult. Trees Shrubs 502. 1949. TYPE SPECIES: C. alnifolia Linn. Evergreen or deciduous trees or shrubs 1-18 m. high. Leaves alternate, simple, exstipulate, chartaceous, rarely coriaceous or subcoriaceous, obo- vate, oblong, rarely lanceolate, base cuneate, rarely acute or rounded, apex acuminate, margin serrate, rarely subentire. Flowers in simple, paniculate, or subumbelliform racemes, rachis and pedicels stellate-pubescent, rarely covered by simple, straight hairs; calyx patelliform, with broad-ovate to lanceolate lobes; corolla choripetalous, emarginate, erose or fringed; sta- mens 10, in two whorls, the outer whorl opposite the petals, filaments glabrous or hirsute, reflexed in bud, straightened at anthesis, anthers 4- celled in bud, at anthesis V-shaped, oblong-obovate or obcordate, con- nective extending into a basal sterile process, thecae opening by terminal rimiform pores; pollen grains simple, tricolporate or triporate; ovary superior, pubescent, 3-locular, style trifid or undivided, 3-lobed or not lobed; stigmas 3, rarely 1; ovules 20-40 in each locule, on axile placentas. Fruit a subglobose capsule with persistent calyx and style, loculicidal. Seeds small, ovoid and irregularly angular or dorso-ventrally compressed, the seed coat of one layer of cells, extremely reticulate, in some species wing- like; endosperm fleshy and oily. About 120 species with Asian-American distribution. KEY TO THE SECTIONS, SERIES AND SPECIES oe Petals densely barbate on the inside; style undivided; stigma 1, punctiform; fruits globose, villose-hirsute; leaves evergreen; pollen triporate; hairs on the rachis and pedicels stianlty gray, appressed. Sect. 1. MonosTIcMA. 1. > . Petals glabrous or sparsely villose at the base on the inside; style trifid or undivided, 3-lobed with 3 stigmas; leaves deciduous (except C. smithiana) ; pollen Preoierate. hairs on the rachis and pedicels stellate or tufted, hire sute, usually ferrugineous. . Anthers V-shaped or oblong-obovate; pedicels longer than the sepals at anthesis; petals emarginate, rarely Peed: racemes various. ....... BOE Se PE ao ee eRe P RATES 2 ete, Oa ect. 2. CLETHRA. C. Flowers large, the petals 8-10 mm. long; stamens included; filaments hirsute, the hairs white; anthers oblong-ovate; racemes solitary. D. Style pubescent; petals papillose on the outside; leaves elliptic, attenuate at both ends, 6-14 cm. long, 2.5-5.5 cm. wide. Bey eso oh ain Gs he eR Se RS 2G, monostachya. D. Style glabrous; petals smooth outside; leaves various in shape and E. Leaves uniformly hispid above, pubescent beneath (except the 178 JOURNAL OF THE ARNOLD ARBORETUM SOL, XL1 glabrous variety), obovate or oblong. 7-23 cm. pee 3.5-9 cm Aes 4s cae deka ees bee Rate UA eS aes a E. Leaves glabrous above, inconspicuously oe -pilose on the cidub and lateral nerves beneath, lanceolate, ae 2 om, long, 3-4+ cm. wide. C. yuiana. . Flowers small. the petals 4- 6 mm. Jong: “stamens pete the fila- ments glabrous or villose with ferrugineous or nigrescent hairs; racemes various. F. Racemes solitary; sepals ovate-lanceolate, 4-5 mm. long: petals emarginate; anthers anaes 3 mm the base acute or obtuse... ee oe pene G. Filaments pubescent; pel half of the Betas glaucous; leaves ovate-elliptic. the base obtuse or rounded. .. 6. C. esquirolit. . Racemes paniculate or subumbelliform; sepals ovate-orbicular, 2-4 mm. long; petals more or less fringed. . Series C. ALNIFOLIAE. H. Fruiting pedicels 2-3 times as long as the diameter of the capsules: leaves chartaceous. I. Filaments glabrous; leaves obovate or broad-elliptic. the base cuneate; racemes paniculate; sepals ba ovate, 2=2.5 mim, Jone, ObIUse. .scyj.ckkseiaevnaids . C. barbinervis. Filaments mabeacents ‘Jeay es ovate-oblong, a base obtuse; racemes subumbelliform; sepals lanceolate, . mm. long. ee oak Ch a eed © Goes WH we get aoe RS "C. fargesil. H. Fruiting pedicels about as long as the eee of the cap- sules; leaves subcoriaceous. ........ Series D. FABERIANAE. as) _ B. Anthers cordate; wegiosls shorter than the sepals at smh: petals fringed; racemes subumbelliform. . Sect. 3. CUELLARIA. J. Style pubescent near the base. . oth babes WOR. pony aera: J. Style glabrous. K. Leaves deciduous. distinctly serrate. elliptic or oblong, chartaceous, 11-20 cm. long; 5-7 cm. wide; petals glabrous. L. Primary lateral nerves parallel- arcuate. 6-10 mm. apart: hairs on the lower leaf- ial or nerves stellate. 1 Filaments glabro N. Leaves cue hirsute on the nerves beneath, other- wise glabrous; margin sharply serrate. va Cea en eMeeduppslaeee au aun ewes aes Ei, 2G, eens N. Leaves uniformly velvety beneath; margin subentire. minutely mucronulate-ciliate or serrate. Ree ut oes 2... 12. C. brammeriana. M. Filaments elloses at the bas se. 13. C. Rwangsiensis. L. Primary lateral nerves parallel. 4— 6 mm. apart; hairs on the nerves beneath simple. appressed. _.. 14. C. polyneura. K. Leaves evergreen, subentire or remotely serrate near the apex. linear-lanceolate. 7-13 cm. long, 2-3 cm. wide. rn ere 15. Cz aniene 1960 | HU, CLETHRA IN CHINA 179 Section 1. MONOSTIGMA S. Y. Hu Clethra, sect. Monostigma, sect. nov. TYPE SPECIES: C. bodinieri Lévl. Frutices sempervirentes ramulosis hornotinis pilosis; foliis coriaceis; racemis solitariis, rachibus pedicellisque pilosis; sepalis ovatis; petalis intus barbatis, filamentis hirsutis; stylis glabris, apice punctiformibus One species, Kwangtung, Kwangsi, Hainan and Kweic {. Clethra bodinieri Lévl. Repert. Sp. Nov. Fedde 10: 475. 1912; nop ae Sinensia 5: 3. 1934; Merr. & Chun, Sunyatsenia 5: 156. var. bodinieri An evergreen shrub, 2—5 m. high, the branchlets ascending, the cur- rent year’s growth 2-3 mm. thick, inconspicuously pilose, the hairs simple. Leaves coriaceous, lanceolate, 5-9 cm. long, 1—-2.5 cm. wide, base cuneate to acute, apex caudate-acuminate, the acumen 1.5—-2 cm. long, glabrous above, inconspicuously pilose on the midrib beneath, angles of principal nerves slightly barbate; margin mucronate-denticulate, the basal half or one-third entire, lateral nerves 8, arcuate, obscure above, conspicuous be- neath; petioles 5-12 mm. long, pilose. Racemes solitary, 3-13 cm. long, peduncles distinct, rachis pilose, the hairs simple, gray; pedicels 7-8 mm. long at anthesis; sepals ovate, 2.5-3 mm. long; petals oblong, 5 mm. long, 2.5 mm. wide, densely barbate inside; filaments hirsute; style glabrous, the stigma entire. Capsules globose, densely villose-hirsute; persistent style 7-9 mm. long; pedicels 10-12 mm. long. Seeds brown, 0.5-1 mm long, angular-ovoid. Kwanctunc: W. T. Tsang 25451, 25620, 25828. Kwanost: R. C. Ching 5695, 5734, 5804, 7181; W. T. Tsang 22450, 24023; C. Wang 39602. Occurring in open thickets or woods at altitudes of 600-1600 meters; the white flowers with pinkish sepals and pedicels appear in June; very distinct species with no close relationship to any other Chinese species of Clethra var. latifolia, var. nov. Frutex sempervirens 4 m. altus, ramulis hornotinis incanis; foliis obo- vato-oblongis vel obovato-ellipticis, 7-9.5 cm. longis, 2.5-3 cm. latis, basi obtusis, apice acuminatis, acumine 5-10 mm. longo, margine serratis, supra glabris, subtus barbatis; racemis 6 cm. longis; filamentis dense villosis; stylo glabro. HaInAN: Loktung, in dense woods, S. K. Lau 27312 (type, a), June 27, 1936 (shrub 4 m. high, leaves green above). Occurring in dense woods; flowering in late June; distinguished from the typical C. bodinieri by its obovate-oblong leaves with obtuse bases. 180 JOURNAL OF THE ARNOLD ARBORETUM [voL. XL var. parvifolia, var. nov. Frutex 1.5—2.5 m. altus ramulis hornotinis sparse pilosis; foliis ellipticis, raro obovato-ellipticis, 3-5 cm. longis, 8-17 mm. latis, basi cuneatis et acutis, apice acuminatis, acumine 5-9 mm. longo, supra glabris, subtus barbatis aliter glabris; racemis 6-9 cm. longis; filamentis hirsutis; stylo Kwancsi: Shang-sze District, fairly common, in thickets, dry steep slope, W. T. Tsang 22450 (type, A), June 2 —7, 1933 (woody, 5 ft. high, flowers white, fragrant); W.T. Tsang 23822, 224506, - flowering in early June; distinguished from the typical C. bodinieri by its small leaves which seldom reach 5 cm. in length. Occurring in thickets and woods; Section 2, CLETHRA Clethra, sect. Clethra Clethra sect. ges a Prodr. 7: 588. 1839; Drude in Engler & Prantl, Pflanzenfam. IV. 1897. TYPE SPECIES: C. ee Linn. Deciduous trees or shrubs, branchlets stellate-pilose. Leaves char- taceous, rarely subcoriaceous. Rachis and pedicels stellate-pilose; racemes solitary or paniculate or subumbelliform, pedicels elongate; sepals ovate to lanceolate; petals glabrous inside or rarely slightly pilose near the base; style trifid or undivided and 3-lobed at the apex, stigmas 3. About 15 species, eastern Asia and eastern North America. Subdivided into four series. Series A. DELAVAYANAE 8. Y. Hu Clethra, sect. Clethra, series Delavayanae, ser. nov. Type species: C. delavayi Franch. Arbores vel frutices, foliis chartaceis, serrato-dentatis; racemis soli- tariis, pedicellis elongatis; staminibus inclusis, filamentis hirsutis; stylis trifidis; capsulis hispidis; seminibus dorso-ventralibus compressis. Three species in West China, the related species in Sumatra 2. Clethra monostachya Rehder & Wilson in Sargent, Pl. Wils. 1: 501. 1913; Rehder, Man. 675. 1927, ed. 2. 690. 1940, Bibl. 502. 1949; Chen, I]. Man. Chin. Trees Shrubs 942. 1937 Clethra delavayi sensu Stapf in Bot. Mag. 148: t. 8970. 1923; sensu Rehder, Man. 675. 1927, ed. 2. 690. 1940, non Franchet. A tall shrub or small tree, 2-8 m. high. Leaves elliptic, or ovate-elliptic, rarely lanceolate, 7-13 cm. long, 2.5—5.5 cm. wide, base cuneate, apex acuminate, acumen 1-1.5 cm. long, apiculate, margin sharply serrate, 1960] HU, CLETHRA IN CHINA 181 primary lateral nerves 16-21 pairs, both surfaces glabrous, the large nerves slightly pilose and their angles barbate beneath; petioles 1-2.5 cm. long, inconspicuously stellate-pilose in the grooves above; leaves pro- duced by active lateral buds lanceolate, 4-6.5 cm. long, 1.3-2 cm. wide, both surfaces subglabrous, the nerves and their angles pilose beneath. Racemes solitary, 7-18 cm. long, the rachis densely ferrugineous-hirsute, the hairs tufted; pedicels 6-10 mm. long at anthesis; sepals 4-5 mm. long, acute; petals 5-6 mm. long, papillose outside; filaments hirsute; style pubescent, apex 3-cleft. Capsules subglobose, 4 mm. in diameter; pedicels 15-18 mm. long; persistent style 7-8 mm. long. TUNNAN CerGhing2i4o2 iM kK. Li 2000" OVA 305. SIKANG (Seca on cee Tien-chuan, Y. S. Lin 1329; ae Yuen, Ta Hsiang Ling, . Y. Chiao 1616; Moupin, E. H. Wilson 1192 (paratype). West China, without ee locality, E. H. Wilson, Veitch Exp. 3927 (type). Occurring on the margins of woods and thickets at altitudes of 1700- 2800 meters in the mountains of the Yunnan-Szechuan-Sikang borders; a very distinct species closely related to C. delavayi; distinguished by its elliptic leaves attenuate at both ends, smaller flowers, papillose petals, pubescent filaments and styles. The specimens from Yunnan are not typical. The stamens are en- closed, as in C. delavayi. The material illustrated by Stapf has pilose styles, and it apparently belongs here. Rehder’s account in his Manual is based on Stapf’s illustration. The isotype of C. delavayi has glabrous stvles. 3. Clethra delavayi Franchet, Jour. de Bot. 9: 370. 1895; P. Dop, Bull. Soc. Bot. France 75: 731. 1928; Rehder, Man. 675. 1927, ed. 2. 690. 1940. Bibl. 502. 1949; Chen, Ill. Man. Chin. Trees Shrubs 942.1937; Merr. in Brittonia 4: 137. 1941. var, delavayi A tall shrub or small tree 4—5 m. tall, the current year’s growth densely stellate-hispid, 3-6 mm. thick. Leaves obovate or elliptic-oblong, 7-15 cm. long, 3.5-6 cm. wide, base cuneate, apex acuminate or acute, the acumen 5-10 mm. long, uniformly hispid above, tomentose beneath, margin serrate, lateral nerves 20 or 21 pairs; petioles 10-17, densely villose; axillary buds more or less stalked, usually bearing no leaves, the stalk 3-40 mm. long. Racemes solitary, 18-25 cm. long, the rachis densely ferrugineous-hispid, the hairs tufted; pedicels 6-7 mm. long at anthesis; sepals ovate-lanceolate, 5-6 mm. long; petals oblong-obovate, m. long, 5 mm. wide, apex emarginate, ciliate; filament hirsute at the base, the hairs white; style glabrous, the apex trifid. Capsules sub- globose, 4 mm. in diameter; persistent style 6-7 mm. long; pedicels 14—20 mm. long. UNNAN: Delavay 3319 (isotype); G. Forrest 11590, 15528, 15839, 17545; McLaren (C Collector 237); H. T. Tsai 58408; T. T. Yu 20979. 182 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. SUI Occurring in mixed forest at altitudes of 2400-3200 meters; a very distinct species, closely related to C. monostachya, distinguished by its obovate leaves which are uniformly tomentose beneath, large flowers with enclosed anthers, hispid filaments, and glabrous styles. var. lanata, var. nov. Arbor parva 4 m. alta; foliis obovatis vel oblongo-ellipticis, 11-12 cm. longis, 3.5—-5.5 cm. latis, supra hispidis, subtus dense lanatis; pedicellis fructorum 1.5—2 cm. longis. YUNNAN: Che-tse-lo, Pi-lo-shan, alt. 4000 m., H. T. Tsai 58263 (type, A), August 27, 1934 (tree, flower white); H. T. Tsai 58459, This variety occurs in the forest of Che-tse-lo of western Yunnan. The dense-lanate lower leaf-surface is very distinctive. var. glabra, var. nov. Arbor parva 3-5 m. alta; foliis oblongo-ellipticis vel obovato-oblongis, 13-23 cm. longis, 5.5—9 cm. latis, supra glabris, subtus glabrescentibus, ad costam et nervos lateralibus sparse pilosis; pedicellis florum 8-15 mm longis. YuNNAN: R. C. Ching 20478, 22107; K. M. Feng 3301; G. Forrest 8875; McLaren (D Collector 130); J. F. Rock 17111, 18461; H. T. Tsai 59905, 59951; C. W. Wang 63886, 68702; Wei-Hsi, Mount Shang-Ma-Kou, alt. 3400 m., J. F. Rock 17174 (type, A), August 1928 (tree, 3-5 m. high, flowers white). Occurring in mixed forests by streams at altitudes of 2800-3800 meters; the white flowers open in August; distinguished by its leaves which are glabrous above and pilose beneath on the principal nerves only. 4, Clethra yuiana, sp. nov. Arbor 7-10 m. alta, ramulis hernotinis 4 mm. diametro, glabrescenti- bus; foliis lanceolatis, 10-15.5 cm. longis, 3-4 cm. latis, utrinque acu- minatis, acumine 1 cm. longo, margine mucronato-serratis, supra elabris; subtus ad nervos primarios et costam pilosis, aliter glabris, nervis lat- eralibus utrinque 22-24, foliis ramulorum axillarium ellipticis vel lan- ceolatis, 1.5-2.5 cm. longis, 6-8 mm. latis, utrinque praesertim ad nervos pilosis; racemis solitariis, 22 cm. longis, densifloribus; rachibus stellato- pilosis, pilis brevibus; pedicellis florum 8-10 mm. longis; sepalis ovato- lanceolatis 5-6 mm. longis; petalis obovatis, 8 mm. longis, 5—6 mm. latis, apice emarginatis, utrinque glabris; filamentis hirsutis; stylis glabris, apice trifidis; capsulis ignotis. YuNNAN: Shunning, Hila, Wumulung, alt. 2750 m., among forest, common, T. T. Yu 1720 (type, A), July 13, 1938 (tree, 20-30 ft. high, flowers white). Occurring frequently in mixed forest at an altitude of 2750 meters; flowering in mid-July; closely related to C. monostachya, but distin- guished by its larger flowers and glabrous styles. — iF) ies) 1960] HU, CLETHRA IN CHINA Series B. ESQUIROLIANAE S. Y. Hu Clethra, sect. Clethra, series Esquirolianae, ser. nov. TYPE SPECIES: C. esquiroli Lévl. Frutices foliis ellipticis vel ovato-ellipticis, chartaceis, arguto-serratis; racemis solitariis, pedicellis elongatis; sepalis lanceolatis, staminibus ex- sertis, antheris V-formibus, elongatis; seminibus ovoideo-polygonis. Two species, southeastern China. 5. Clethra cavaleriei Lévl. Repert. Sp. Nov. Fedde 10: 476. 1912; Rehder, Jour. Arnold Arb. 15: 267. 1934; Hand.-Mazz. Beih. Bot. Centralbl. 56B : 449. 1937. Clethra lineata Lévl. Repert. Sp. Nov. Fedde 12:°534, 1913. Clethra longebracteata Sleumer, ibid. 38: 205. 1935. Clethra sinica Hao, ibid. 42: 85. 1937. A shrub or small tree 1-5 m. high, the current year’s growth densely pilose, the hairs short-stellate. Leaves elliptic, 6-10 cm. long, 1.5-4 cm. wide, base acute, apex acute or shortly acuminate, glabrous above, very sparsely pilose on the principal nerves beneath, margin serrate, lateral nerves 12 or 13 pairs, reticulations of veinlets distinct beneath; petioles 1-1.5 cm. long, appressed pilose and glabrescent; leaves produced by active axillary buds lanceolate. 2-7 cm. long, 1-1.5 cm. wide, both surfaces sparsely stellate-pilose. Racemes solitary, 9-15 cm. long. bracts longer than the pedicels; pedicels 7-10 mm. long at anthesis; sepals lanceolate, 5 mm. long; petals 6 mm. long; filaments glabrous; style glabrous, the apex trifid, Capsules subglobose, 4-5 mm. in diameter; persistent style 9 mm. long; pedicels 14-15 mm. long. FUKIEN: R. C. Ching 2250, 2300; H. M. Fan 9474, 9493, 9515; H. H. Chung 2923. Kwanctunc: R. Mell 883; W. T. Tsang 21413, 21661. K wanes: age AG Lau 28753, 28755, CHEKIANG: R. C. Ching 2101; Y. L. Keng 164 (isosyntype orc. acon 175 (isotype of C. longebracteata). KwetcHow: J. Cavalerie 5 (fragment and photo of type); J. Esquirol 3238 (photo and fragment of type of C. lineata); Handel-Mazsetti 172 = 10557. Hunan: Fan & Li 253. Occurring in woods and thickets at altitudes of 780-1250 m.; flowering in August; closely related to C. esquiroli, distinguished by its glabrous filaments. 6. Clethra esquirolii Lévl. Repert. Sp. Nov. Fedde 10: 475. 1912. Clethra cavaleriei sensu Rehder, Jour. Arnold Arb. 15: 267. 1934, p.p.; sensu Hand.-Mazz. Symb. Sin. 7: 760. 1936, p.p., non Lévl. A shrub 1-3 m. high, current year’s growth sparsely stellate-pilose, 3—4 mm. in diameter, subterete or angular. Leaves ovate-elliptic or ellip- tic, 6.5-10 cm. long, 2.5-4.5 cm. wide, base obtuse or rounded, apex short-acuminate, the acumen 1 cm. long, glabrous above, the nerves sparsely pilose and their angles slightly barbate beneath, margin serrate. 184 JOURNAL OF THE ARNOLD ARBORETUM | VOL. XLI lateral nerves 10-12 pairs; petioles 1-2 cm. long, hirsute, straight and stellate hairs intermixed; leaves produced on active axillary buds ellip- tic, 3-4.5 cm. long, 12-15 mm. wide, both surfaces stellate-pilose, the midrib and lateral nerves lanate beneath. Racemes solitary, 10-16 cm. long, the rachis ferrugineously stellate-hirsute; pedicels 9-11 mm. long at nay nea ovate-lanceolate, 4-5 mm. long; petals oblong, 6—7 mm. , 3-4 mm. wide, sparsely villose inside, basal half glaucous out- side; oe villose the hairs ferrugineous or nigrescent; style glabrous, trifid at the apex. Capsules subglobose, 4 mm. in diameter; pedicels 1.5—2 cm. long; persistent style 9-12 mm. long. Fuxien: R. C. Ching 2300. Kwanctunc: Lokchong, Y. Ysrang 1422. Kwancsi: Ling-Chuan, W. T. Tsang 27871; Tzu-yuen, T. S. Tsoong (= Z Chung) 83440; C. Wang 39565, 40068. Hunan: Sinning, Fan & Li 453: Wu-kang, Handel-Mazzetti 842 = 12397, Huper: A. Henry 2838. KwerIcHow: J. Cavalerie, July 19, 1898; J. Cavalerie 69 (E. Bodinier); J. Esquirol 651 (type, fragment and photo); Steward, Chiao & Cheo 394; Y. Tsiang 5460. Occurring in open ridges or ravines of the Nan-ling Range, its range extending from northern Kwangsi and Kweichow, eastward through Hunan to northern Kwangtung and northern Fukien, and northward to western Hupei; closely related to C. cavaleriei, distinguished by its ovate or ovate- elliptic leaves with obtuse or rounded base, its glaucous petals and pubes- cent filaments. Rehder and Handel-Mazzetti both interpreted this species as synony- mous with C. cavaleriei. The fragment of the type of C. esquiroli and Handel-Mazzetti 12397 both have pubescent filaments. In this respect, they are different from the type of C. cavaleriei. Clethra esquirolti is distinguished from C. fargesii by its larger flowers and solitary racemes. Series C. ALNIFOLIAE S. Y. Hu Clethra sect. Clethra, series Alnifoliae, ser. nov. Type species: C. alnifolia L. Frutices foliis chartaceis, obovatis vel lato-ellipticis, crasso-serratis; racemis paniculatis; Seana bus exsertis, antheris V-formibus; seminibus polygonis vel compressis. Five species, eastern United States of America and the Lower Yangtze Region of China. ~I Clethra barbinervis Sieb. & Zucc. Abh. Phys.-Math. Cl. Akad. Wiss. Miinchen IV. 3: 128. 1846; Gilg, Bot. Jahrb. Engler 34(Beibl. 75): 56. 1904; igen in Beih. Bot. Centralbl. 37(2): 164. 1919; Rehder, Man. 1927, ed. 2. 690. 1940, Bibl. 502. 1949; Kai, Pl. Sin. MI. 306: fig. 538. 1937. A shrub up to 10 m. high, the current year’s growth 3-4 mm. thick, subglabrescent or minutely stellate-pilose. Leaves obovate-elliptic, 7-14 cm. long, 3—-6.5 cm. wide, base cuneate and acute, apex abruptly short- 1960] HU, CLETHRA IN CHINA 185 acuminate, the acumen deltoid, 5-10 mm. long, glabrous above, midrib and principal nerves hirsute and their angles barbate beneath, the hairs simple; margin sharp-serrate; lateral nerves 12—14 pairs, arcuate; petioles 12.5 cm. long, pilose, the hair appressed; leaves produced by active axillary buds obovate, 3.5-6.5 cm. long, 1.5-2.5 cm. wide, sparsely stellate-pilose above, sparsely villose beneath, straight and stellate hairs intermixed. Racemes 3-6, paniculate, the rachis densely ferrugineous- hispid, the hairs tufted; pedicels 4-6 mm. long at anthesis; sepals sub- orbicular-ovate, 2—2.5 mm. long; petals obovate, 5-6 mm. long, 3-4 mm. wide; apex emarginate, and fringed; filaments glabrous; style glabrous, the apex trifid. Capsules subglobose, 4 mm. in diameter; persistent style 6-8 mm. long; pedicel 6-8 mm. long. TaIwaNn: R. Kanehira, Sept. 5, 1920. CHEKIANG: R. C. Ching 1642. ANHWETI: R. C. Ching 3237, 3277. SHANTUNG: Tsingtao, Zimmermann 432. Occurring in thickets on open hills at an altitude of 1000 meters; re- lated to C. faberi and C. fargesi through its paniculate inflorescences, short sepals, and fringed petals; distinguished from C. fabert by its obovate leaves and elongate pedicels and from C. fargesi by its glabrous filaments and obovate leaves cuneate at the base. Kanehira and Zimmer- mann’s collections are out of the normal range of the species. Probably these collections were made from cultivated plants. This species is com- mon in Japan, and it is most likely that this species was introduced into Tsingtao and Taiwan from Japan. 8. Clethra fargesii Franch. Jour. de Bot. 9: 369. 1895; Rehder & Wilson in Sargent, Pl. Wils. 1: 502. 1913; Rehder, Man. 674. 1927, ed. 2, 690. 1940, Bibl. 502. 1949; Chen, I]. Man. Chin. Trees Shrubs 942. fig. 835. 1937. Clethra canescens sensu Forbes & pines Jour. Linn. Soc. Bot. 26: 33. (Ind. Fl. Sin. II). 1889, non Reinw. ex Blum Clethra sleumeriana Hao, Repert. Sp. Nov. Fedde 42: 84, 1937. A tall shrub up to 4 m. high, the current year’s growth densely gl pilose, 3-5 mm. thick. Leaves ovate-elliptic, 7-14 cm. long, wide, base obtuse or rotundate, apex acuminate, the acumen 1-2 cm. long, glabrous above, along the midrib and the principal nerves stellate- villose and their angles barbate beneath, lateral nerves 16-17 pairs, arcu- ate; margin mucronate serrate; petioles 14-17 mm. long, sparsely stellate pilose; leaves produced by the active axillary buds lanceolate, 4-10 cm. long, 1.5-3.3 cm. wide, both sides stellate pubescent. Racemes 3-7, subumbelliform-paniculate, the rachis densely brown-hirsute, the hairs tufted; pedicels 6-10 mm. long at anthesis; sepals lanceolate, 4-5 mm. long; petals obovate, 5-6 mm. long; filaments hirsute near the base; style glabrous, the apex trifid. Capsules subglobose, 3 mm. in diameter; persistent style 5 mm. long; pedicels 12-13 mm. long. KiANGSI: H. H. Hu 1320. Hupet: Cheng & Hwa 821, 1068, 1110; H. C. 186 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI Chow 885, 1178; W. Y. Chun 3714, 3734; A. Henry 5818, 6407, 7220; E. H. Wilson (Veitch Exp.) 1326, 2222. Hunan: S.S. Sin 615 (type of C. sleumeriana. Hao, not seen). SzecHUAN: Farges 108 (type, not seen). Common in open woodlands and thickets in the middle Yangtze Re- gion; closely related to C. esquirolii, but distinguished by its umbelliform- paniculate racemes, smaller flowers and densely stellate-pubescent branch- lets Series D. FABERIANAE S. Y. Hu Clethra, sect. Clethra, series Faberianae, ser. nov. TYPE sPEciEsS: C. faberi Hance. Frutices foliis subcoriaceis, subglabris vel subtus stellato-pilosis, canes- centibus, serratis; racemis subumbelliformibus, pedicellis brevibus; petalis fimbriatis. About 10 species, China, Indo-China, the Philippines, Borneo, Celebes and New Guinea. 9, Clethra faberi Hance, Jour. Bot. Brit. For. 21: 130. 1838, “fabri”; Merr. Philipp. Jour. Sci. Bot. 13: 154. 1918; P. Dop, Bull. Soc. Bot, France 75% (32,1928, Clethra ag gee sensu Forbes & Hemsl. Jour. Linn. Soc. Bot. 26: 33. (Ind. i 889, p.p.; sensu Dunn & Tutcher, Kew Bull. Add. Ser. 10: 155,.( Fi, a Hongk.) 1912, non Reinw. Clethra liangit Li, Jour. Arnold Arb, 24: 449, 1943. A deciduous shrub 2—2.5 m. high, the current year’s growth sparsely pilose or glabrescent, 3 mm. in diameter. Leaves subcoriaceous, elliptic, or obovate-elliptic, rarely oblanceolate, 6-11 cm. long, 2—3.2 cm. wide, base acute, apex acuminate, the acumen 1 cm. long, margin sharply serrate, primary lateral nerves 10 or 11 pairs, parallel-arcuate, 6-8 mm. apart, impressed above, reticulations of the veinlets distinct beneath, lamina glabrous on both surfaces, on the principal nerves very sparsely pilose beneath; petioles 6-8 mm. long, sparsely stellate pilose; leaves produced by active axillary buds elliptic, 2.5-8 cm. long, 1—2.5 cm. wide, glossy, glabrous above, hispid on the nerves beneath. Racemes 2-7, paniculate, rarely on weak branchlets solitary, densely and ferrugineously caespitose- hirsute; many bracts persistent; pedicels 3-4 mm. long; sepals sub- orbicular-ovate, 2.5-3 mm. long, apex obtuse, mucronate; petals 4 mm. long, 2 mm. wide: inside sparsely villose below the middle. filaments glabrous, anthers acute at the base, divergent at the apex; style glabrous, 4 mm. long after anthesis, the apex subcapitate, lobed or shallowly cleft. Capsules 2.5 mm. in diameter. KWANGTU C. Ford 56; W. T. Tsang 26761. Kwancsi: H. Y. Liang 69645 (type “of CG. fe S. P. Ko 55527. YUNNAN: H. T. Tsai 60876, Occurring in densely shaded forest or on open, grassy slopes at altitudes of 970-1100 meters; the white flowers appearing in late July and early 1960] HU, CLETHRA IN CHINA 187 August; a very distinct species. Its short, obtuse sepals suggest a rela- tionship with C. darbinervis, but its short pedicels, subcoriaceous leaves and small anthers are very different. The isosyntypes of C. annamensis P. Dop (M. Poilane 7519, 7616) apparently belong here. Section 3. CUELLARIA (Ruiz & Pav.) DC. Clethra, sect. Cuellaria (Ruiz & Pav.) DC. Prodr. 7: 589. 1839, Cuellaria Ruiz & Pay. Fl. Per. Chil. Prodr. 59. ¢. 10. 1794, Syst. Veg. 103. 798 TYPE SPEcIES: C, ferruginea Ruiz & Pav. Trees or shrubs, branchlets densely stellate-pubescent. Leaves cori- aceous or chartaceous. Rachis and pedicels densely ferrugineous-stellate- pubescent; racemes paniculate or subumbelliform, pedicels very short; sepals oblong or ovate; petals erose or fimbriate; stamens exserted or in- cluded. Seeds winged, or polygonous. About 85 species, chiefly in tropi- cal America, 4 in China. 10. Clethra pinfaensis Lévl. Repert. Sp. Nov. Fedde 10: 476. 1912. Clethra kaipoensis sensu Rehder, Jour. Arnold ne 15: 268. 1934; sensu Hand.-Mazz. Symb. Sin. 7: 760. 1926, non Lév A tree up to 13 m. high, the trunk 30 cm. in diameter, the bark smooth, brownish. Leaves oblong-elliptic, 8-15 cm. long, 3- 5 cm. wide, base obtuse, apex acuminate, the acumen 1.5-2 cm. long, margin iene serrate; primary lateral nerves 14 or 15 pairs, parallel-arcuate, glabrous above, glabrescent beneath, the hairs on the nerves beneath simple, appressed: petioles 8-20 mm. long; leaves produced by active axillary buds 3-7.5 cm. long, 1.5-2.5 cm. wide, sparsely stellate-pubescent above. Racemes 4-6, subumbellate, 14-18 cm. long, rarely shorter, densely fer- rugineous- ee the hairs tufted; pedicels 2-3 mm. long at anthesis: sepals ovate, 2-3 mm. long; . petals 4 mm. long, 2 mm. wide; filaments glabrous; style pubescent at ‘the basal end. Capsules 3 mm. in diameter; fruiting pedicels 4 mm. long; persistent style 5-6 mm. long, the stigma trifid. Seeds varying in size and shape, compressed ovoid, triangular, or subcylindric, 1-1.5 mm. long, 0.5—1 mm. wide, reticulate, brownish: Kwancst: R. C. Ching 6026, 7000, 7116. KwrtcHow: J. Cavalerie 346 (photo and fragment of type); Handel-Mazzetti 254 = 10766, 325 = 10997. Occurring in woods at altitudes of 1300-1500 meters; flowering in late July and early August; closely related to C. kaipoensis, but distinguished by its styles being pubescent near the base. The style-character is clear in the fragment of the type. When Rehder and Handel-Mazzetti inter- preted C. pinfaensis and C. kaipoensis as conspecific, they ignored this distinct character. 188 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI 11. Clethra kaipoensis Lévl. Repert. Sp. Nov. Fedde 10: 475. 1912; Rehder, Jour. Arnold Arb. 15: 268. 1934, p.p. A deciduous shrub 1-3 m. high. Leaves oblong-elliptic, 13-19 cm. long, 4-9 cm. wide, base obtuse or rotund, apex acuminate, the acumen 2 cm. long, margin coarse- and sharp- serrate, primary lateral nerves 16-18, 7-13 mm. apart, glabrescent above, sparsely stellate- Asay along the nerves and barbate in their angles beneath: petioles 1.5-2.5 c long, densely ferrugineous-hirsute; leaves produced by the active anillary buds broad elliptic, 3-4 cm. long - 5-2 cm. wide, densely velvety on both surfaces, the hairs stellate. Racemes 4-8, subumbellate, occasionally soli- tary on weak branchlets, densely golden villose-hirsute; pedicels 3 mm. long; sepals deltoid-ovate, 3-4 mm. long; petals 4-5 mm. long; fila- ments glabrous, anthers 1.5 mm. long, base obtuse; style glabrous, the apex slightly a 3-lobed, after anthesis 5-6 mm. long. Capsule 4 mm. in diameter Kwancsi: Steward & Cheo 958; Y.W. Taam 24; W. T. Tsang 27900, 27916, 28425. KwericHow: J. Cavalerie 1221 (holotype, photo and fragment), Occurring in forest at altitudes of 1000-2100 meters in Kweichow and northern Kwangsi; closely related to C. kwangsiensis; distinguished by its unusually large leaves and its glabrous filaments. 12. Clethra brammeriana Hand.-Mazz. Anzeig. Akad. Wiss. Wien Math.-Nat. Kl. 58: 151. 1921, Symb. Sin. 761. 1936 A deciduous tree, the branchlets densely ferrugineous-hirsute, the hairs tufted, current year’s growth 5 mm. in diameter. Leaves oblong-elliptic or obovate-oblong, 11-15 cm. long, 4.5-8 cm. wide, base obtuse or ro- tundate, apex acute or shortly acuminate, margin subentire and minutely mucronulate-ciliate or aristate-serrate, glabrous above, evenly velvety beneath, primary lateral nerves 15 or 16, parallel-arcuate, 8-12 mm. apart; petioles 14-18 mm. long, ferrugineously or castaneously hirsute. Racemes 5-13, subumbelliform-paniculate, 10-18 cm. long, densely and ferrugine- ously lanate-hirsute, some bracts persistent; pedicels 2 mm. long, sepals ovate-deltoid, 3-5 mm. long; petals 4 mm. long, 2.5 mm. wide; filaments glabrous, anthers obcordate, exserted; style 5 mm. long after anthesis, the one undivided, 3-lobed. Fruit and seed not known. cst: T. H. Wang 247 (isotype). Kwanost: T. S. Tsoong (= Z. S. a 83414, Occurring in woods along ravines at an altitude of 600 meters; closely related to C. kaipoensis, but distinguished by its uniformly velvety lower leaf-surfaces. The flowers of Tsoong 83414, collected on July 15, are in very young bud. 13. Clethra kwangsiensis, sp. nov. Arbor, ramulis hornotinis teretibus, 5 mm. diametro, dense ferrugineo- 1960] HU, CLETHRA IN CHINA 189 pubescentibus pilis stellatis; foliis chartaceis, ovato-ellipticis, 16-17 cm. longis, 6—7.5 cm. latis, basi rotundatis vel obtusis, apice acuminatis, acu- mine 2 cm. longis, margine serratis, nervis lateralibus utrinque 20, arcu- atis, supra glabris, subtus sparse stellato-pilosis, ad nervis villosis; petiolo 1.5 cm. longo, dense hirsuto; foliis ramulorum axillarium breviter ellip- ticis, 5—5.9 cm. longis, 3 cm. latis, utrinque stellato-tomentosis, pilis supra rigidis; racemis 6, subumbelliformi-paniculatis, 15-17 cm. longis, dense ferrugineis, pilis stellatis; pedicellis 2-4 mm. longis; sepalis deltoideo- ovatis, 3.5-4 mm. longis, acutis; petalis 5 mm. longis, 2 mm. latis, fim- briatis; filamentis villosis; stylo glabro, apice trifido; capsulis ignotis. Kwancsi: Tzu Yuen District, in woods near a stream, T. S. Tsoong (= Z.S. Chung) 83570 (type, A), August 6, 1937 (tree, flowers white). Occurring in woods along a stream; the white flowers appearing in early August; closely related to C. kaipoensis, but distinguished by its filaments being villose on the basal half. 14. Clethra polyneura Li, Jour. Arnold Arb. 24: 449. 1943. Trees up to 18 meters high. Leaves oblong-lanceolate or oblong-elliptic, 9-15 cm. long, 3-6 cm. wide, base obtuse or rotund, apex acuminate, the acumen 1-2 cm. long, margin densely serrate, principal lateral nerves 18-25 pairs, parallel, 4-6 mm. apart, glabrous above, glabrescent be- neath, the midrib and primary nerves sparsely pilose, the hairs simple and appressed; petioles 1-2.5 cm. long, sparsely stellate-hirsute; leaves produced by active axillary buds lanceolate, 4-7 cm. long, 1.3—2 cm. wide, sparsely stellate above, only the nerves ferrugineously stellate-pubescent beneath. Racemes 6-8, subumbellate-paniculate, 10-14 cm. long, densely ferrugineous-hirsute, the hairs tufted; pedicels 1.5-2 mm. long at an- thesis; sepals ovate, acute, 2-3 mm. long; petals 3 mm. long, 1.75 mm. wide; filaments glabrous; style glabrous. Capsules 3 mm. in diameter, pericarp velvety; persistent style 6 mm. long, glabrous; fruiting pedicels 3—4 mm. long. Seed ovoid, reticulate, 1 mm. long, 0.5 mm. wide Kwancsti: S. K. Lau 28767 (type). HUNAN: Fan & Li 519. Occurring on hilly slopes at altitudes of 650 m.; closely related to C. J kaipoensis, but distinguished by its glabrous style and numerous parallel lateral nerves. 15. Clethra smithiana Fang, Contr. Biol. Lab. Sci. Soc, China. Bot. Set 2271 1.030: An evergreen shrub or small tree 6-8 m. high, branchlets glabrous, the current year’s growth 3-4 mm. in diameter. Leaves coriaceous, lanceo- late, 7-13 cm. long, 2—3.5 cm. wide, base cuneate, apex acuminate, the acumen I—2 cm. long, lateral nerves 15-17 pairs, arcuate, 8-13 mm. apart, obscure above, reticulations of veinlets conspicuous beneath; margin sub- entire, the apical half remotely sharp-serrate, both surfaces glabrous, the 190 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLI principal nerves very sparsely pilose beneath; petioles 10-15 mm. long, sparsely pilose. Racemes 2—8, subumbelliform, 8-12 cm. long, densely castaneous-hirsute, the hairs tufted; pedicels 1.5-2 mm. long; sepals ovate, 2 mm. long, obtuse or acute; petals 2.5-3 mm. long, 1.5 mm. wide; fila- ments glabrous, apex of the anthers divergent; style undivided, 3 mm. long after anthesis, the apex slightly enlarged, lobate. Fruit 3 mm. in diameter. Kwanesi: R. C. Ching 6854, 7209, YUNNAN: H. T. Tsai 51447 (type, not seen). Occurring in woods at altitudes of 1130-1600 m.; a very distinct species, closely related to C. faberi, but distinguished by its subentire leaves, very small flowers, short pedicels and its short, persistent style. 1960] SAX, POLYPLOIDY IN ENKIANTHUS 191 POLYPLOIDY IN ENKIANTHUS (ERICACEAE) HALty J. Sax ENKIANTHUS Is a comparatively small genus belonging to the Ericaceae. According to Rehder (1940, 1949) there are about ten species, although more than double that number have been reported. Many of the so-called species which were described in isolated instances were considered by Rehder (1940) and Wilson (1907, 1929) to be geographical forms of species already described Most species of Enkianthus are attractive shrubs or small trees growing wild in southeastern Asia, Japan, southern, central and western China, and the Himalayas. Fang (1935), in a preliminary study of the Chinese species, said that there is not a single species in the northern provinces of China. He commented that only six of the twenty-one species described are valid. The confusion in the classification of the genus is not surprising. J. D. Hooker (1879) wrote, “The Eastern genus Enkianthus presents four types of structure which almost indicate as many genera and would do so if the species had not been united by habit and if the characters were asso- ciated instead of applying to one species only.” The number of chromosomes in Enkianthus was studied to determine if there was polyploidy in the genus. The variety in the structure of the flowers in the small number of species of Exkianthus, as well as their lim- ited distribution, would offer possibilities of interesting combinations. Stomatal size and shape were also determined to learn if there was any correlation between stomatal size and chromosome number in the genus. Three species of Enkianthus with some varieties are growing in the Arnold Arboretum: E. campanulatus (Miq.) Nichols, with vars. albi- florus Mak. and palibinii Bean; E. perulatus (Miq.) Schneider; and £. subsessilis (Miq.) Mak. Some of these on Bussey Hill were planted as early as 1890. A more recent and successful planting of the same species was made on low ground along the meadow road to the administration building. Enkianthus cernuus (Sieb. & Zucc.) Mak. var. rubens (Maxim.) Mak. is now being propagated in the Arboretum greenhouse. The chro- mosome numbers of the species of Enkianthus in the Arnold Arboretum were studied from meiotic divisions in the pollen mother cells, from the divisions of somatic cells in the ovary and from regenerating parenchyma of the stem. Very good preparations of divisions in both metaphase and anaphase in the dividing egg cell were also obtained. The regenerating parenchyma is obtained according to the bark-peel method developed by Sax (1959). A longitudinal section of the bark is removed from the growing stem. The wound is covered with a strip of polyethelene film and left for several days. The newly developed tissue 192 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI is then sliced off and fixed. The anthers, ovules and bark tissue were fixed in alcohol-acetic solution for twenty-four hours. The material was then changed to 95 per cent alcohol. Pieces of these tissues were macerated and smeared on a slide and stained with acetocarmine. Camera lucida drawings were made from some of the slides. I wish to thank Dr. Karl Sax for making the bark peels and for checking most of the counts. The chromosomes were counted in the divisions of the pollen mother cells in Enkianthus subsessilis and several plants of E. campanulatus including the varieties albiflorus and palibinii. The chromosomes were also counted in the somatic tissue of the ovules. In the case of FE. perulatus, counts were made in the somatic tissue in the ovule where it was possible to obtain several very good division figures where the chromosomes could be counted. Counts were also made from the regenerating bark in F£. perulatus and E. campanulatus, where very good preparations of the divisions in the large parenchyma cells were obtained. In all cases several very good preparations were obtained. Enkianthus subsessilis was found to be a diploid with 11 chromosomes in the meiotic divisions in the pollen mother cells. They were very clear in the preparations and could easily be counted in both the metaphase and anaphase stages. The camera lucida drawings show clearly 11 bivalent chromosomes in the first metaphase of the dividing nucleus of the pollen mother cell (Fic. 1). The 11 anaphase chromosomes are very clear in Chromosomes of Enkianthus. 1, E. subsessilis, first meiotic co) h metaphase. All figures approximately <1 1960] SAX, POLYPLOIDY IN ENKIANTHUS 193 Fic. 2 just as they are oriented at the poles. There were 22 chromosomes in the somatic tissue of the developing ovule (Tare I). Enkianthus perulatus is also a diploid. The count in this case was made in the somatic tissue of the ovule and in the parenchyma of the regenerat- ing bark. Very clear counts in both metaphase and anaphase were ob- tained. There were 22 chromosomes in the somatic tissue, as is shown in Fic. 3 which was drawn from parenchyma tissue obtained by the bark- peel technique. Enkianthus campanulatus is a high polyploid — presumably an octo- ploid — with approximately 88 chromosomes. There were about 44 chro- mosomes in the divisions of the pollen mother cells of several plants labeled as this species. Multivalent chromosomes were present and an occasional univalent was seen. The divisions were regular, suggesting some degree of allopolyploidy. Only in a few cases were there any lagging chromosomes. Camera lucida drawings were made of the dividing nuclei from several of these plants; all show approximately 44 chromosomes (Fies. 4, 5). Enkianthus campanulatus var. albiflorus is also an octoploid with 2 = 44 at meiosis. Multivalent associations of three and four chromosomes were seen in some cases, but bivalents predominated. The divisions were regular except in rare instances (Fic. 6). The number of chromosomes in the regenerating bark and in the ovule was over 80, the counts varying from 82 to 89. The chromosomes were small and it was difficult to be exact with such large numbers. TABLE I shows the number in each case. The two diploid species belong in different sections of the genus, while the octoploid is in a third section. Polyploidy is not necessarily associated TABLE I. Chromosome Numbers in Enkianthus SPECIES ARBORETUM No. MEIOTIC SOMATIC ZONE * E. campanulatus 14528-1-D 44 -- 80-87 IV 14528-1-G 44+ 14528-1-C 88+ 23001-1-C 884 7692 44 507-58-D 88 507-58-E 78-87 var. albiflorus 14783-E 444 IV 7039-3-A 44+ 14783-G 88+ E. perulatus 20153-B 22 V 10128-B 22 3591 22 E. subsessilis 2449-B 11 22 Vv 2449 11 map of the climatic zones is found in Rehder (21). The zones are numbered irom De North Pole southward. They are irregular according to the prevailing clim 194 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI with taxonomic diversity. The fact that the octoploid species shows pre- dominantly bivalent chromosomes at meiosis indicates that it is an amphi- ploid with different genomes. Chromosome numbers have been reported for many genera of the Ericaceae (Darlington and Janaki-Ammal, 1945). Longley published the chromosome numbers of Vaccinium, a polyploid series with the basic number 12. Hagerup (1928) studied several genera: Calluna vulgaris, with 16; Kalmia, 24 and 48; Cassiope hypnoides, 48; Erica, 24; Chimaphi- la umbellata, 26; Ledum, 26 and 52; Ramischia, 38; and Pyrola, 46 and 92. Wanscher (1933) reported 24 chromosomes in Phyllodoce; Wulff (1939), 22 chromosomes in Moneses uniflora; Maude (1940), 24 in Daboecia cantabrica; Newcomer (1941), 24 in Gaultheria procumbens: allan (1941), 36 chromosomes in Bruckenthalia, 22, 44, 66 in Pernettya, 22, 44, 88 and 26 in Gaultheria, 24 in Pieris, 48 in Andromeda, 24 in Leucothoé, 24 in Erica, and 26 in Arbutus, Arctostaphylos, and Ledum. Hagerup (1941) counted 60 chromosomes in Enkianthus campanulatus. 26 in Chimaphila, and 52 in Ledum. Baldwin (1942) found 24 chromo- somes in Oxydendrum arboreum. Sax (1930) reported the number of chromosomes in Rhododendron as 26 for many species and 52 for two higher polyploids, the latter being native to North America. Nakamura (1931) found 26 to be the chromo- some number for nine more species of Rhododendron. Janaki-Ammal (1950) reported Rhododendrons with 4x, 6x, 8x, and 12x chromosomes in the high Himalayas in Eastern Asia, polyploidy increasing with altitude. Li (1957) also found 26 chromosomes in all the species of the “luteum” section of the azaleas, excepting the tetraploid R. calendulaceum and its hybrids and one triploi Thus, with few exceptions (Calluna vulgaris), the basic numbers of chromosomes are 11, 12, and 13. Polyploidy is prevalent in many genera, e.g., Vaccinium, Pe aie. Gaultheria, Pieris, and Ledum. As noted above, Hagerup (1941) reported 60 chromosomes for Enkianthus campanulatus, considering it a high polyploid, and suggesting 12 as the basic number both for nkianthus and for the family. Enkianthus has a very narrow distribution: southeastern Asia, central and western China, and Japan. The polyploid /£. campanulatus is more widely distributed than the diploids, which have a narrow range and are limited to the warmer parts of China and Japan. Yet, compared with the Rhododendrons, which are native to a large part of both hemi- spheres, the polyploid FE. campanulatus which extends from northern Japan to southern and central China and to the Himalayas in West China is a naturally limited species. However, when in cultivation it grows very well in the North Temperate Zone in Europe and the United States. The diploid species F. perulatus and E. subsessilis, as well as EF. cernuus var. rubens, are also able to grow fairly well when introduced as far north as Boston, Massachusetts. Stomatal size was also studied in three species of Enkianthus. The epidermis was peeled from the leaf and mounted on a slide in a drop 1960 | SAX, POLYPLOIDY IN ENKIANTHUS 195 of acetocarmine. Camera lucida drawings were made from the mounted epidermis. The diploids E. perulatus and E. subsessilis showed very little difference in size, but there was some difference in shape. Although no appreciable difference in size was found between the stomata in E. perula- tus and E. subsessilis, there is a striking difference in size between these diploids and the high octoploid E. campanulatus. The stomata of E. campanulatus are very much larger than those of the diploids, as was noted by comparing the figures in camera lucida drawings of the stomata of these species. There is a very definite correlation between stomatal size and number of chromosomes in the species of Enkianthus studied. Counts were made of the fertile and sterile pollen grains in the differ- ent species. Usually the fertility was high, around 85 per cent, but as low as 35 per cent was found. These latter cases were not consistent for any of the species studied and may not be typical, the aberrance having been caused, perhaps, by factors not studied. SUMMARY There are 11 meiotic chromosomes in Enkianthus subsessilis, 22 somatic chromosomes in E. perulatus and E. subsessilis, Both are diploid species. The meiotic chromosomes in Enkianthus campanulatus and var. albt- florus are 44, and there are about 88 + chromosomes in the dividing somatic cells. Enkianthus campanulatus and its varieties are octoploids. The stomatal size in Enkianthus perulatus and E. subsessilis, both diploid species, is about the same. The stomata of the two differ somewhat in shape. The stomata are much larger in E. campanulatus, the octoploid, and its varieties than in the diploids, showing a definite correlation be- tween stomatal size and chromosome number. Although there is some pairing in threes and fours in Enkianthus cam- panulatus, most of the chromosomes are bivalents in the polyploids. LITERATURE CITED BALDWIN, J. T. Cy Peo of Oxydendrum arboreum. Bull. Torrey Bot. Club 69: 134. 19 GALTANG EG, ay ane of Gaulthettya wisleyensis (Marchant) Rehder. Ann. Bot. 5: 579. 1941. DarLIncTon, D. C., and E. K. JANAKI-AMMAL. Chromosome atlas of culti- vated plants. 379 pp. G. Allen, London. 1945. Fanc, W. P. A preliminary study of the Chinese species of Enkianthus Lou- reiro. Contr. Biol. Lab. Sci. Soc. China HAGERUP, O. Morphological and cytological studies of Bicornes. Dansk. Bot. Ark. 6: 1. 1928 : Zytodkologische Bicornes Studien. Planta 32: 6. 1941. Hooker, J. D. Enkianthus himalaicus. Bot. Mag. III. 35: pl. 6460, 1879. JANAKI-AMMAL, E. K. Polyploidy in the genus Rhododendron. Rhododendron Year Book, London 5: 92-98. 1950 Lr, H. L. Chromosome studies in the azaleas of eastern North America. Am. Jour. Bot. 440: 8-14. 1957. 196 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI LoncLey, A. E. Chromosomes in Vaccinium. Science 66: 566. 1927. Loureiro, J. Enkianthus quinqueflorus. Flora Cochinchinensis 1: 277. 1790. MAUDE, - Chromosome numbers in some British plants. New Phytol. 39: 17. eo. i Cytology of Rhododendron. Trop. Agric. Tohoku Imp. Univ. 3: 103. 1931 Newcomer, E. ry Chromosome numbers of some Ping and varieties of ie and related genera. Proc. Am. Soc. Hort. Sci. 38: 4-68. 1951. ReEHDER, A. Enkianthus. In Manual of Cultivated ae and Shrubs. Ed. 2. 729-730. 1940. . Enkianthus. In Bibliography of cultivated trees and shrubs hardy in the cooler temperate regions of the Northern re Arnold Arbo- retum, Harvard University. Jamaica Plain, Mass Sax, K. oe stability in the genus Eee aia Am. Jour. Bot. 17: 247. : e cytogenetics of facultative apomixis in Malus species. Jour. Arnold Arb. 40: 289-297. 1959. [Bark-peel technique, p. WanscHER, J. H. Secondary (chromosome) associations in Umbeiliferae and Bicornes. New Phytol. 33: 58. 1933. Witson, E. H. The genus Enkianthus. Garden Chron. III. 41: 311, 344, 363. 1907. ——.. Enkianthus, aos 248 im China, Mother of gardens. 4 + 408 pp., map, 6 illustrations. Strat ord Co., Boston, Mass. 1929 . Enkianthus. In a I ae to nanke a garden. pp. 205-210. 1931 Wutrr, H. D. Chromosomenstudien an der ee Anglo- spermen Flora, IV. Ber. Deutsch. Bot. Ges. 57: 424 1960] WILSON, HYDROPHYLLACEAE AND POLEMONIACEAE 197 THE GENERA OF HYDROPHYLLACEAE AND POLEMONIACEAE IN THE SOUTHEASTERN UNITED STATES ! KENNETH A. WILSON HYDROPHYLLACEAE R. Brown ex Edwards (WATERLEAF FAMILY) Annual, biennial or perennial herbs [or rarely shrubs] with alternate or opposite, entire, pinnately or palmately divided leaves. Inflorescences terminal, axillary, or opposite the leaves, the flowers in cymes, panicles, or solitary. Flowers bisexual, regular, 5-merous. Calyx 5-parted, the sinuses sometimes bearing appendages. Corolla rotate, campanulate or funnelform, 5-parted. Stamens 5, epipetalous. Gynoecium syncarpous; styles 2 (rarely more) or 1, terminal: ovary superior, 1-locular with 2 parietal placentae or 2-locular and the placentation axile. Ovules 4—many, anatropous or amphitropous. Fruit a capsule. A family of about 18 genera and 250 species in three tribes, primarily of North America, but extending into South America, Asia, and Africa. The family is represented in the United ie by 15 genera centered in the Southwest; six genera occur in our ar The family is regarded as being Eee allied to the Polemoniaceae and the Boraginaceae. It may be distinguished from them by the combina- tion of bicarpellary 1- or 2-locular ovary (usually with numerous ovules) and the usually imbricate aestivation of the flowers which are generally borne on scorpioid cymes, Considerable variation is evident in the struc- ture and nature of the ovary. The gynoecium is 2-carpellate; the placen- tae may be parietal and fleshy or they may be narrow, extending into the ovary to meet without fusing and partition the ovary into two locules. In Hydrolea further elaboration of the placentae, accompanied by a fusion of the dividing wall, has produced a 2-locular ovary with axile placentation. The Hydrophyllaceae are cytologically perhaps the best known family of any size. Information on the chromosome numbers of the species in ‘Prepared for a biologically oriented generic flora of the southeastern United tates, a joint project of the Gray Herbarium and the Arnold Arboretum made pos- follows that outlined at the beginning of the series (Jour. Arnold eps 39: 296-346. 1958). Other published portions of these studies will be found in Jour. Arnold Arb. their assistance in de information, aiding in | the obtaining of specimens, and helping in various ways 198 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI the family has been accumulating, and, along with morphological studies, has already contributed to the understanding of the group. Cytological information, for instance, supports the recognition of subg. Cosmanthus of Phacelia, and also points to the relationship of Lemmonia to Nama. A comprehensive knowledge of the cytology of the members of this family will give a clearer indication of the value of chromosomal data in systematic studies. Presently the chromosomes of about 67 per cent of the species (about 200) have been counted, and counts are available for at least one species in each of the described genera. REFERENCES: BRAND, A. ge a oe ae pe Pflanzenr. IV. 251(Heft 59): 1-210. 1913. e Hydrophyllaceen der Sierra Nevada. Univ. Calif. Publ. Bot. 4: aa, 1912. Cave, M. S., and L. Constance. Chromosome numbers in the Hydrophyllaceae. Univ. Calif. Publ. Bot. 18: 205-216. 1942; II. 18: 293-298. 1944; III. 18: 449-465. 1947; IV. 23: 363-382. 1950; V. 30: 233-258. 1959. CONSTANCE, L. The genera of the tribe Hy ae ie the Hydrophyllaceae. Madrofio 5: 28-33. 1939. [Includes key to gene Hydrophyllaceae. Ju L. Abrams, Ill. Fl. ee States 3: 476-532. 1951. . Hydrophyllaceae. Inst. Paran. Catal. Gen. 4: 1, 2. 1956. Gray, A. A conspectus of the North American Hydrophyllaceae, Proc. Am. Aca d. Arts Sci. 10: 312-332. 1875 Peter, A. Hydrophyllaceae. Nat. Pdanzen tan: IV. 3a: 54-71. 1897, Svensson, H. G. Zur Embryologie der Hydrophyllaceen, Borraginaceen und Heliotropiaceen mit besonderer Riicksicht auf die Endospermbildung. Uppsala Univ. Arsskr. 1925. 2(Math. Nat. 2): 1-176. 1925. KEY TO THE GENERA OF HYDROPHYLLACEAE A. Leaves dissected, lobed or toothed; style 1, entire or 2-cleft. B. Placentae parietal, are ovary 1-locular . Flowers in scorpioid cymes; stamens exserted; leaves mostly basal and alternate; ee not cucullate: plants biennial or — eae Li Re Ga ea eee Bia sas ee ea a pa sarees _. Bydrophyllum. . Flowers solitary opposite the leaves or terminal a loose cyme; stamens included; leaves mostly cauline, all Rea or the lower opposite and the upper alternate; plants annual. D. Calyx sinuses auriculate; seeds od leaves - alternate ee Q? QD? (hadi cd eee eee cate be ee eee ee Nemophila. D, ee eee without auricles; seeds not cucullate; _ oppo- e belo IeTNATE ADOVE:, ( Veonnay Wullschlagel 804 (GoET). Coccoloba excoriata Linnaeus, Syst. Nat. ed. 10. 1007. 1759. Lindau’s treatment of Coccoloba excoriata (Bot. Jahrb. 13: 211-212. 1891) is confusing. Of the many specimens which he cited, the ones I have seen should be referred to C. venosa. I have not seen a Widgren specimen from Rio de Janeiro to which he referred, nor the one which Schomburgk Cosel to as “FI. : Faun. Br. Guy. 821.” This reference is to “Coccoloba nivea,”’ a synonym of C. venosa which applies to a plant under cultivation. Coccoloba venosa is a ane in the native flora of South America and will be discussed later. Coccoloba excoriata L., as correctly applied, is a synonym of C. tenuifolia L. (Howard, Jour. Arnold Arb. 38: 93. 1957), a native species of the Bahamas, Cuba and Jamaica, and has not been reported as either a cultivated or an indigenous plant of South America. Coccoloba fagifolia Jacq. Hort. Schonbr. 3: 55. pl, 352. 1798. Although Jacquin published an excellent plate of this species and at- tributed the plant to the vicinity of Caracas, the species has not been collected or recognized in recent years. I find Jacquin’s species difficult to place. In the Meisner herbarium, now at the New York Botanical Garden, there is a packet with three detached leaves bearing the legend, “Coccoloba fagifolia Jacq. H.S. (mann tremula, verisim. ipsins Jacquini patris script.) folia speciminis sterilis Herb. Jacquini in Hb. Mus. Vindo- 236 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI bon.” This appears to represent one of the specimens cited by Meisner in his treatment of the species for DeCandolle’s Prodromus (14: 165. 1856- “y.s.c. ex Schoenbr. in herb. Jacq. patr. et fil.”). These three leaves have long, thin petioles and blades of thin texture. They do not represent any species of Coccoloba known to me, and there is no evidence that they even belong to the genus. Moreover, these detached leaves do not compare favorably with the original illustration. Lindau (Bot. Jahrb. 13: 162. 1890) recognized the species and cited a specimen collected by Boos which was in the Vienna herbarium. How- ever, since the Coccoloba material in that herbarium was destroyed during World War II, the material which Lindau cited cannot be identified. Presumably the Boos specimen was a small one, for Lindau accumulated for the Berlin herbarium a large number of fragments from important species, but C. fagifolia is not represented in the material which I have on loan from that herbarium. No recent material has been assigned to Coccoloba fagifolia, so its interpretation must rest on the original illustra- tion, although a comparison of Jacquin’s plate with that of C. virens Lindley (Bot. Reg. 21: pl. 1816. 1836), which is C. coronata, shows many similarities. If Jacquin’s plate may be accepted as representing an accurate description, then it seems clear that C. fagifolia should be assigned to the synonymy of C. coronata Jacq. The proper identification of the Jacquin material in the Meisner herbarium remains unsolved. Coccoloba fallax Lindau, Bot. Jahrb. 13: 172. 1890. Coccoloba caurana Standley, Publ. Field Mus. Bot. 22: 73. 1940; Llewelyn Williams, Explorac. Bot. Guayana Venez. 189. 1942. Coccoloba fallax seems to be distinguished by the fascicled inflorescence of simple racemes, the conspicuously large ocreae, especially those clustered at the apex of the stem and around the racemes, and the strongly keeled midrib. The type collection of Coccoloba caurana has immature in- florescences, but comparable mature material has been found in Trinidad. There is no question that C. caurana is the same as C. fallax. The Trinidad specimens have been cited in a previous paper, in which the species was also recognized as occurring in Venezuela (Jour. Arnold Arb. 40: 79. 1959). The type of C. fallax is Crueger 114. The lectotype selected is the specimen in the Goéttingen Herbarium. The holotype of C. caurana is Llewelyn Williams 11366 (F). Coccoloba fasciculata Meisner in Warming, Symbol. Floram Bras. 6: £23, 2610; Lindau cited this epithet and reference in the synonymy of “C. longi- pendula” in his monograph (Bot. Jahrb. 13: 177. 1890). The original publication is a report on Warming’s collections from Brazil. One entry is “Coccoloba fasciculata Wedd., Meissner in DC lc. 166 and Mart. lc. 27,” with a citation of “Lagoa Santa: . . . Serra da Piedade legit Warm- 1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, Ix 250 ing Jan.-Febr. florentem.”’ This is obviously a misidentification and not a new entity, as Lindau implied. Lindau referred the epithet to the synonymy of C. longependula, which I cannot distinguish from the older C. sticticaulis (q.v.). Lindau also cited the Warming collections as follows: “Rio de Janeiro in silvis ad Lagoa Santa: Warming n. 129; ad radices montis Serra da Piedada: Warming n. 126.” I have seen specimens of each and have referred them to C. sticticaulis Weddell. Coccoloba fasciculata Weddell, Ann. Sci. Nat. III. 13: 258. 1849. I have discussed this species in an earlier paper (Jour. Arnold Arb. 41: 44. 1960) and have referred it to the synonymy of Coccoloba arborescens (Vellozo) Howard. Coccoloba fasciculata Weddell was based on Blanchet 796 collected in Bahia Province, Brazil. The type is in Geneva. Coccoloba fastigiata Meisner, Fl. Bras. 5(1): 34. 1855. Coccoloba goudotiana Meisner, Fl. Bras. 5(1): 35. pl. 13, fig. 1. 1855, not Weddell. In the original description Meisner described the species and a variety “8 glabrata,” and cited specimens to illustrate both taxa. This is con- trary to his usual practice, in the same volume and treatment, of con- sidering the species to consist of varieties for which specimens are cited. The variety glabrata Meisner, represented by “Schott 5537 (914)” and part of Schott 5540 (in Vienna), is better assigned to Coccoloba mosenii. For the type variety of the species, Meisner cited Schott 5540, collected in Rio de Janeiro Province, Brazil, and Schomburgk 1262, in the Berlin herbarium. The latter collection was supposedly made in British Guiana. Lindau, who worked with the Berlin material around the year 1890, mentioned the Schomburgk collection under the name Coccoloba fastigiata, but attributed it to Meisner, “pr. p. ex. Meissn.” I cannot determine whether Lindau actually saw this material, but I suspect that he did not, for in his citation of specimens seen (Bot. Jahrb. 13: 224. 1890) he listed Schomburgk 1262 as C. marginata. I have on loan the Coccoloba material from the Berlin herbarium and I fail to find the Schomburgk collection represented. I suspect that Meisner was in error in including the Schom- burgk specimen and that C. fastigiata should be typified only by Schott 5540. Meisner stated that the specimen he saw was in the Vienna herba- rium. Unfortunately, the Coccoloba material in this herbarium was destroyed during World War II. A photograph and a fragment of the holotype is at the Chicago Natural History Museum, and a more complete specimen is in Brussels. Therefore it seems wise to designate the specimen at Brussels as the lectotype. Coccoloba goudotiana Weddell was based on a Goudot specimen col- lected near San Luis, Colombia, and now in the Paris herbarium. Meisner elaborated on the original description and illustrated the species. In addi- tion to the Goudot specimen, Meisner also cited a collection by Pohl in 238 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI the Zuccarini herbarium at Munich. I have seen both of these specimens and it seems clear that the illustration in Flora Brasiliensis was made by combining features present in both specimens. Coccoloba goudotiana Weddell is C. obovata, while C. goudotiana as described and illustrated by Meisner is C. fastigiata. Brazil. Rro DE JANETRO: Aguacu, Schott 5540 (sr-—lectotype, aie and fragment of Vienna material); Mana, Glaziou 7888 (B, c, P); Realengo near Sapopemba, Glaziou 11441 (LE, P); without specific locality, ti Le ee 0 8 Without locality: Pohl s.n., in Herb. Zuccarini as cited by Meisner (™). Coccoloba fastigiata var. glabrata Meisner, Fl]. Bras. 5(1): 34. 1855. Meisner cited “Schott 5537 (914)” and, in part, Schott 5540 as repre- sentatives of his new variety with specimens in the Vienna herbarium. Lindau recognized the variety (Bot. Jahrb. 13: 172. 1890), citing Schott 5537 and Pohl 914 with specimens at Vienna and Berlin. The Vienna material has been destroyed and material from these collections is not included in the specimens of Coccoloba from the Berlin herbarium. A photograph of Schott 5540 and a fragment of this specimen from Vienna are preserved in the herbarium of the Chicago Natural History Museum and clearly identify the specimen as C. fastigiata. It appears that the fragment of Schott 5537 (F), all that remains of the collection, should be referred to C. mosenii Lindau (q.v.). Coccoloba ferruginea Endlicher, Catalogus Hort. Acad. Vind. 1: 274. 42; Ettingshausen, Blattskel. d. Apetalen 91. pl. 26. fig. 2. 1858. Lindau referred this epithet to the synonymy of his Coccoloba excoriata (g.v.). Endlicher cited the name without description as a plant which had been obtained from the Berlin botanic gardens and was cultivated in Vienna gardens. Ettingshausen illustrated a leaf skeleton, using the name Coccoloba ferruginea. He acknowledged that his material was from a cultivated specimen, but did not discuss the illustration or the species. The leaf skeleton is inadequate for positive identification, although it is probably C. venosa L.; none of the specimens I have seen ‘bears the name C. ferruginea. Coccoloba filipes Standley, Trop. Woods 40: 14-15. 1934. In the original description Standley wrote, “the distinctive characters of the species are the greatly elongate, many-flowered, lax racemes, with flowers on conspicuously long and slender pedicels.” The type, Dugand 380, was collected in flower at Santa Rosa, west of Barranquilla, Colombia, on March 13, 1933. The flowers, although borne singly at each nodule on the inflorescence axis, are apparently functionally staminate, producing pollen grains. The specimen, now at the U.S. National Herbarium, bears two terminal, thin-textured, darkened leaves, suggesting that the plant is deciduous. The leaves arise from above the base of the ocreae. 1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX 239 No additional material which can be referred to this species has been seen. It appears to be distinct among the Colombian species of the genus and, on the basis of the material available for study, is not closely related to or even similar to any other South American species thus far described. Although several species from Cuba or Puerto Rico have similar long pedicels, none are so tenuous. Coccoloba firma Martius ex Colla, Bee Pedem. 5: 48. 1836; Eyma, Meded. Bot. Mus. Utrecht 4: 3. The original description, while brief, is ae and clearly intended to be that of a new species. In a list of species Colla attributed the name to Martius. No more definite location than Brazil was given. Eyma also used the name attributed to Martius in comparing it with his proposed new species, Coccoloba micropunctata. Eyma cited specimens at Kew and the Natal Herbarium, Durban, which he said were listed under the name Coccoloba populifolia Weddell. The correct name for the latter species is C. alnifolia Casaretto. Eyma apparently recognized both C. firma and “C. populifolia.”” I have not seen the specimen from the Natal Herbarium, South Africa, nor have I been able to locate a specimen so named at Kew. There is a Martius sheet at the British Museum bearing the name C. firma, which I believe is properly C. alnifolia. No material of C. firma was found in the collections of Martius which I have on loan from Munich, Brussels and Leningrad. I am therefore reluctant to accept this unsupported name until the material which Colla saw, or at least that which Eyma cited, has been located. If this should prove to be the same as C. alnifolia Casaretto (1844), as I suspect, the older C. firma Martius ex Colla should be accepted as the correct name. Coccoloba gardneri Meisner, Fl. Bras. 5(1): 36. pl. 15. 1855; Howard, Jour. Arnold Arb. 41: 44. 1960 This species is based on Gardner 600 from Serra dos Orgaos, Brazil. In the original description, which is clearly based on the Gardner collec- tion, Meisner cited Polygonum frutescens Vellozo in synonymy and also Coccoloba vellosiana Casaretto, the latter with some doubt. I have pre- viously considered the identity of species described by Vellozo and Casa- retto and have also concluded that C. gardneri Meisner represents still a third species. Meisner’s species was placed in the synonymy of C. crescentitfolia by Lindau, but clearly it is not the same as C. arborescens (Vell.) Howard, doc. cit. I cannot adequately define the species on the basis of the single collection available and wish to call this problem to the attention of future collectors who may visit the area of Rio de Janeiro. Coccoloba glaziovii Lindau, Bot. Jahrb. 13: 163. 1890. Coccoloba cylindrostachya Lindau, Bot. Jahrb. 13: 163. 1890. 240 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI Coccoloba sublobata Heimerl, Denkschr. Akad. Wien. Math.-Nat. Klasse 79: 44. 1908. Glaziou 8089 is the only collection cited by Lindau in the original description. Of the specimens cited from four herbaria, I have designated that at Berlin as the lectotype. I am unable to distinguish between Coccoloba glaziovii and C. cylin- drostachya which Lindau based on Glaziou 8089 and 13135 respectively. Lindau distinguished between them by placing C. glaziovii in a group 0 species with the rachis glabrous and C. cylindrostachya in a contrasting group having the “rachis vario modo pilis instructa.’’ His species descrip- tions amplify this supposed difference and suggest others, e.g., the position of the petiole in relation to the base of the ocrea, which, however, is not borne out in a re-examination of the type specimens cited. I have seen all of the specimens which Lindau cited from both collections and can only conclude that one species is represented. Glaziou 3087 was annotated by Lindau as a possible new species. It is a vigorous shoot, probably adventitious in nature, with longer internodes, larger leaf blades and more pubescent parts. Lindau never published the name written on the sheet. I believe the specimen should be included in Coccoloba glaziovii. Another unpublished name honoring Lindau is found on the collection Schwacke 13673 with the author given as Schwacke. Although the specimens seen have more delicate branches, more tenuous and pendant inflorescences and more oblong leaves, I feel that this mate- rial is properly assigned to C. glaziovit. The Chicago Natural History Museum has a photograph of the holo- type of Coccoloba sublobata which was in the herbarium at Vienna. The type was lost during World War II and the only duplicate of this collection which I have seen is in the Berlin Herbarium. This Berlin specimen bears an unpublished name, attributed to Heimerl and referring to the subundu- late leaves, which is more appropriate than the one actually published. Although the type of C. glaziovii is a staminate plant, there is no question that Wacket’s collection represents the same species in fruit. Heimerl’s original description does not refer to the type collection by number. The number 12 cited below is legible in the photograph of the type and on the isotype. In his monographic treatment of Coccoloba striata (Bot. Jahrb. 13: 164— 165. 1890), Lindau cited numerous specimens from the Brazilian states of Bahia, Sao Paulo, Rio and Pernambuco. I have seen only one of these, Mosén 3664, which I believe to be more properly associated with C. glaziovit. Coccoloba striata, based on a Schomburgk collection from the Roraima area, appears to be a northern species. The remainder of the collections cited by Lindau should be examined for their correct relation- ship here. Brazil. Minas GEraAeEs: Itabira do Matto Ventro, Schwacke 13673 (8B, P). PaRANA: Volta Grande, Dusén 11966 (GH, NY). R10 DE JANEIRO: Rio de Janeiro, Glaziou 3087 (BR), 8089 (B-lectotype, BR, C, G, GH, LE, $), 13135 (type collection 1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX 241 of C. cylindrostachya, B, BR, G, LE). SAo Pauto: Serra do Cubatao, near Santos, Wacket 12 (B-isotype of C. sublobata) - Sorocaba, Mosén 3664 (p e Coccoloba goudotiana Weddell, Ann. Sci. Nat. III. 13: 260. 1849. This species is based on Goudot s.n. from San Luis, Colombia. The type specimen is in the Paris herbarium and the species is clearly synony- mous with Coccoloba obovata HBK. (1817). Coccoloba goudotiana Weddell sensu Meisner, Fl. Bras. 5(1): 35. i. 13. fig. 1. 1855 In the original publication Meisner did not intend to describe a new species, but only to list the Weddell species. However, he cited the Goudot material used by Weddell, as well as a Pohl collection from the Zuccarini herbarium which is not the same species. Meisner’s description and illustration combined features of both plants. I have seen the Pohl specimen, now in the Munich herbarium, and refer Meisner’s interpreta- tion of Coccoloba goudotiana to the synonymy of C. fastigiata Meisner. Campderia gracilis Meisner, Fl. Bras. 5(1): 23. pl. 6. 1855; DC. Prodr. 14: 170. 1856. Meisner described and illustrated this species in 1855. He cited a Spruce collection without giving a collector’s number from around Barra in the “Prov. Rio Negro.” The following year in the Prodromus he cited Spruce 958, reporting specimens to be in the DeCandolle Herbarium and the Herbarium Monacense (Munich). The illustration in Flora Brasiliensis appears to be a compilation of the two specimens. The original pencil sketch of the illustration of flowers and fruits is attached to the sheet at Munich and this sheet should be designated as the lectotype. However, I cannot determine the source of the fruiting material which Meisner illustrated, since comparable achenes do not appear on either sheet at the present time. Lindau referred Campderia gracilis to the synonymy of Coccoloba ovata Bentham. This appears to be the correct specific place- ment on the basis of the material I have examined, but there is a question as to whether C. ovata belongs in the genus Coccoloba. This matter will be discussed further under C. ovata. I have seen specimens of Spruce 958 from Munich (lectotype), Berlin, Geneva, the Gray Herbarium, Leningrad and Paris. Coccoloba gracilis HBK. Nov. Gen. 2: 176. 1818. Two specimens, one in the Willdenow Herbarium and one at Paris, can be considered to be authentic for this species. Both are relatively small and are characterized by slender, nearly geniculate inflorescences of func- tionally staminate flowers. The original description has information appli- cable to labels on both specimens but does not exactly match either. It 242 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI appears desirable to consider the Paris specimen as the lectotype since the only label on this sheet bears the number 3498 of Humboldt and the geographic location Rio Cachiyacu given at the time of publication, These two specimens are inadequate for accurate determination as to genus. The floral structure and that of the ocreolae and the bracts appear to bridge the few weak characteristics used to distinguish Coccoloba and Ruprechtia in staminate flowering condition. With the material available, the only possible course to follow is that taken by the previous authors and monographer who placed the species in Coccoloba. Additional material is needed to determine its correct position Macbride (Publ. Field Mus. Bot. 13: “460. 1937) assigned a fruiting collection, Weberbauer 6982, to this species in his treatment for the Flora of Peru. The Weberbauer collection bears no data regarding the plant or the place of collection beyond “Peru.” The old inflorescence axes are 4—8 cm. long and are erect. The fruiting peduncles are 2—2.5 mm. long. Fruits, one of which retains the fruiting calyx, are preserved in a packet. The perianth lobes are free nearly to the base in the fruit. The achene is only slightly triangular in outline, dark brown and shiny. This specimen is clearly a Coccoloba, though not necessarily the same as the authentic material of C. gracilis HBK. ru. Rio Cachiyacu, Humboldt 3498 (p—lectotype; without number or loca- tion [ Herb. Willd. 7701]). Coccoloba grandiflora Lindau, Bot. Jahrb. 13: 175. 1890. Lindau cited two collections in the original description, Glaziou 14217 and Miers 4657. The former should be selected as the type collection and the specimen at Copenhagen designated as the lectotype. This is the only sheet among those cited below on which the label states, “Coccoloba grandiflora Lindau n. sp.” Lindau’s work was based upon the material in the Berlin herbarium — the fragment of a branch together with one inflorescence having one oe and two attached leaves. I have not located the Miers collect Superficially, this mis resembles Coccoloba tenuiflora Lindau which is based on cultivated material of unknown origin. The type of C. tenut- flora was from a greenhouse plant, while C. grandiflora is from native material. I cannot determine whether cultivation (and, specifically, green- house conditions) could create the differences seen in these specimens. Coccoloba grandiflora has more conspicuous ocreolae which are membra- naceous, split longitudinally and flaring. In all other characteristics the species are similar. Both species are known only from flowering material. Brazil. Minas GERAES: Glaziou 20439 (B, LE, NY). R10 DE JANEIRO: Nuovo Freiburg, Glaziou 14217 (C-lectotype, B, K, LE, US Coccoloba grandis Bentham in Hooker, London Jour. Bot. 4: 624. 1845. This species is based on Schomburgk 825 collected on the Rio Branco, 1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX 243 British Guiana. Lindau placed the species in the synonymy of Coccoloba latifolia Lam. After having examined the type in the British Museum, I agree with this placement. Coccoloba guaranitica Hassler, Repert. Sp. Nov. 14: 161. 1915. Coccoloba guaranitica var. opaca Hassler, ibid. 162. Hassler compared his new species with Coccoloba schomburgku, but on the basis of the few specimens I have seen, it seems more nearly comparable to C. padiformis from Venezuela, although the leaves of C. guaranitica are smaller and more rounded at either end. The fruits of both species are known only from immature specimens, but in both the perianth lobes are conspicuous, imbricate and appear to surround only the upper half of the achene. Hassler did not select a type in the original description in which he mentioned two collections, Fiebrig 1429 and 1440. I have not seen the latter, but the former is a shoot of vigorous and rapid growth. The variety which Hassler described differs from the species in having smaller leaves, the margins of which are undulate-crenate. It also has a shorter inflorescence. The type of the variety, Fiebrig 4305, is a mature shoot system with many lateral flowering branches. It seems quite ap- parent that the specific name has been applied to younger and more vigorous specimens and that of the variety to the more mature branching specimens and thus the variety is not worthy of recognition. A specimen of Fiebrig 4305, the type of C. guaranitica var. opaca, was studied by Gross and annotated with both a specific name honoring Fiebrig and a varietal name referring to the crisp leaf margin. Neither name has been published, to my knowledge, although Gross published other epithets in the genus in small notes, often in obscure publications. Paraguay. Boquerén, Puerto Casado, Pedersen 4043 (a, c); Chaco, Fiebrig 1429 (m-isotype); between Rio Apa and Rio Aquidaban, Fiebrig 4305 (type collection of C. guaranitica var. opaca, B, GH, M, P); Bahia Negra, Rojas 13708 (BR, P, W). Coccoloba guianensis Meisner, Linnaea 21: 264. 1848. Several specimens were cited in the original description, but no type was selected. The specimens are obviously the same as the slightly anom- alous material described earlier by Bentham as Coccoloba marginata. The variation in material called C. guianensis and the nomenclature of the complex has been discussed in other papers (Howard, Jour. Arnold Arb. 40: 84, 85. 1959; 41: 45, 46. 1960). Coccoloba gymnorrhachis Sandwith, Kew Bull. 1932: 221. 1932; Eyma, Polygonaceae, Guttiferae & Lecythidaceae of Surinam, 111. 1932. Sandwith based the original description of this species on a flowering 244 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI specimen, Sandwith 168, from British Guiana. The plant was described as a ‘“‘bush-rope” and the type shows strong characteristics of a liana. The leaves, broadest above the middle and narrowed to the base, are rigidly coriaceous with the arcuate venation impressed above. The flowers are borne in clusters with tightly appressed ocreolae. Eyma supplemented the original description with a fruiting specimen from Surinam. The Richard collection cited below is also in fruit and is the same as the material which Eyma cited, but these two fruiting collections are not good matches for the type collection and may not belong here. The fruits are warty, as though insect-infested, but no evidence of insect larvae could be found. Additional collections are needed before the morphological characteristics of Coccoloba gymnorrhachis are fully understood and the species clearly defined. British Guiana. Essequebo River, Moraballi Creek near Bartica, Sandwith 168 (x-holotype, _ Surinam. Brownsbere: B.W. 6773 (x). oso unspecified: L.C. Richard s.n. (P). Coccoloba ilheensis Weddell, Ann. Sci. Nat. IIT. 13: 258. 1849. Coccoloba membranacea Klotsch, Linnaea 14: 289. 1840, nomen nudum. This is a poorly defined species requiring both field study and many more collections for an accurate interpretation. This species is similar to Coccoloba glaziovii, C. ochreolata and C. confusa. While collections representing the type of each of these species can be distinguished on sight, I cannot find any reliable morphological characteristics useful in separating them in a key. Brazil. Banta: Ilheos, Martius 1240 (p—-holotype, BR, G, GH, LE, M); Luschnath n. “1839” (LE). Without location: Luschnath 42 (LE). Coccoloba japurana Meisner, FI. Bras. 5(1): 25. 1855.4 This species is based on a Martius specimen from Ega in the Rio Negro of Brazil. Authentic material is in the Munich herbarium and is repre- sented in the collections of the Gray Herbarium by a photograph. The photograph is a montage of two herbarium sheets and two labels. One label has the hand-written annotations, “Coccoloba acuminata” and “Cocco- loba japurana” and on the same sheet is the annotation, “Alsodeia japu- rana Radlk.” One of the sheets has specimens with immature inflorescences. This almost completely overlaps the other sheet from which an infructes- cence protrudes. The fruit on this axis is clearly not that of a Coccoloba. The name Alsodeia japurana Radlk. (Sitz-ber. Math.-Phys. Klasse Akad. Miinchen 20: 182. 1891) is recorded in the first supplement of Index Kewensis as an observation. There is no reference to the name Coccoloba japurana Meisner in the article cited. However, Lindau at- Rinorea japurana (Meisner) Howard, comb. nov. Basionym: Coccoloba japurana Meisner in Martius, Fl. Bras. 5(1): 25. 1855. 1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX 245 tributed the transfer to Radlkofer in a list of excluded species and in a footnote in his monograph i.e., “C. japurana Mrtssn. = Alsodeia japurana Raprk.” (Bot. Jahrb. 13: 220. 1890). Alsodeia is a genus of the Violaceae for which most modern writers use the name Rinorea. Blake, who monographed the American species of Rinorea (Contr. U.S. Natl. Herb. 20: 317. 1924), listed Alsodeia japu- rana as a “doubtful species” with the comment, “This was described from specimens with very young flowers. It is said by Radlkofer to be allied to A. racemosa.” So many aspects of this misplaced epithet have been overlooked that I have made the new combination in Rinorea primarily to place the name in indices for the aid of future workers. An adequate interpretation of the photograph is impossible. Meisner’s original descrip- tion could well be a Coccoloba; Lindau saw the material now in the Munich herbarium and would surely have recognized a Coccoloba as to genus; Radlkofer worked on Alsodeia and would certainly have recognized that genus; Blake did not accept the species, but it is not clear what material he saw or to which description he referred. The fruit in a photo- graph of a properly labelled specimen, supposedly the type, 1s not a Coccoloba, but appears to be a Rinorea. The writer cannot solve the puzzle and may not have placed the specific epithet in its proper niche. Coccoloba laevis Casaretto, Nov. Stirp. Bras. 71. 1844; Lindau, Bot. Jahrb, 13: 186. fig. 40. 1890. Coccoloba cordifolia Meisner, Fl. Bras. 5(1): 37. 1855. Casaretto did not cite a specimen in the original description, but in the same publication he described other species based on his own collections. One sheet, Casaretto 2264, in the Turino herbarium, fits the description of Coccoloba laevis in all details and should be considered the holotype. I assume that Casaretto unintentionally omitted the citation of a specimen. In the original description of Coccoloba cordifolia, Meisner cited several specimens without selecting a type. He indicated the affinities of his new species with C. nivea, C. cordata, and C. candolleana and cited in synonymy “Coccoloba uvifera Salzmann Mss. in Herb. (non Linn.).” In his treat- ment for the Prodromus (14: 155. 1856) he cited the same specimens, but he did not repeat the suggested synonymy or the possible relation- ship. Instead he placed his species next to C. laevis, which he suggested might be identical with C. cordifolia or C. candolleana. Lindau was the first to place C. cordifolia Meisner in the synonymy of C. laevis Casaretto, a decision with which I agree. The Salzmann collection from Bahia which Meisner cited is represented in several herbaria and the collection at Kew shows the full range of variation from the small-leaved type of Casaretto’s species to the broader and more cordate leaved types found in Blanchet 3528. Lindau described and illustrated the fruit of Coccoloba laevis, but although I have seen most of the material he cited, I have not found a fruiting specimen, or even a single fruit, among them. If the figure pub- 246 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI lished by Lindau can be trusted, the possibility of a hybrid origin of this species should be examined. Coccol loba marginata or C. uvifera would seem quite likely as parents. At present C. laevis is an extremely variable species which is not clearly delineated. Brazil. BAHIA: peg on San Salvador, Dahlgren sn. (F); Ilha de Cal, Curran 106 (GH, NY, Y¥); Itaparica Island, Casaretto 2264 (to-holotype) ; Jacobina, ace ea Blanchet 100 (c, NY), 3528 (G, LE, P). PERNAMBUCO: pete Guillamin s.n,. (F). Rio DE JANEIRO: Rio de Tae: Glaziou 11445 (B, Salzmann 476 (c), sm. (kK, LE, P). Locality uncertain: Maceio, anes 1391 (kK) Coccoloba lanceolata Lindau ex Glaziou, Bull. Soc. Bot. Fr, IV. 11 (Mem. 3f): 573. 1911, nomen nudum. The collection Glaziou 19764 was cited by the author in the original publication with the brief description, “liane, fl. blanchatres, fruit noir.” The specimens seen are obviously from climbing plants. The leaves and infructescence are borne on short lateral shoots. While the epithet has no acceptable standing at the present time, I do not wish either to describe the plant more fully or to place the name in synonymy until further material from southern Brazil is available for study. It is probable that this collection should be assigned to C. salicifolia. The leaves of the Glaziou specimen, however, are more lanceolate-oblong in shape, less acu- minate at the apex and thicker in texture. The fruits match the illustra- tion given by Lindau for C. salicifolia. Brazil. Minas Gerars: Riacho das Varas, Glaziou 19764 (B, C, K). Coccoloba latifolia Lamarck, Dict. Encycl. 6: 61. dl. 316, f. 4. 1804. Coccoloba grandis Bentham in Hooker, London Jour. Bot. 4: 624. 1845. Lamarck described this species from material cultivated in the Jardin des Plantes, Paris. I have not seen authentic material, but his illustra- tion is of a single detached leaf which does not represent well the species as currently accepted. The description, although somewhat vague, seems applicable, but, since Coccoloba latifolia is similar to C. mollis, comparable field observations would be helpful. It differs in an almost complete lack of puberulence, in its much stouter and generally hollow stems and in having strongly bullate leaves. I have seen C. latifolia in Trinidad where it is a characteristic plant of savanna areas. Its habit is distinctive and this, together with the presence of many biting ants in the large ocreae, ake: it long remembered by collectors. Coccoloba grandis Bentham is based on Schomburgk 825. Lindau placed the species in the synonymy of C. latifolia, where it clearly belongs. Lindau referred three collections by Burchell from Sao Paulo and Para to this species. I have seen one sterile sheet of Burchell 3982 in the her- barium at Kew and feel that this sheet, at least, should be considered the adventitious leaf form of Coccoloba warmingi Meisner. 1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX 247 In addition to the localities listed below, the species is also known from Trinidad and its adjacent islands (Jour. Arnold Arb. 40: 81. 1959). Brazil. CEARA: Ceara, Curran 36 (GH). MaranHao: Campo de Boa Esperanca, Froes 1817 (A, NY). British Guiana. Mapenna, Courantyne River, B.G. Forest Dept. 2601 (A); Rio Branco, Schomburgk 825 (pm-—type of C. grandis). French Guiana. Cayenne, savannahs along St. Madeleine Rd., Broadway 750 (cH, NY); without locality, Barbier s.n. (a), Sagot 486 (a). Sather Koboerie, Herb. B.W. 5929 (A); without specific locality, Hostmann 682 (cH), s.n. (BR), Wullschagel $m. (M). Venezuela. DELTA AMACURO: Serrania Imataca, N. of Rio Guanamo, Wurdack & Monachino 39724 (a). Coccoloba laurifolia Jacquin, Hort. Schoenbr. 3: 9. pl. 267. 1798. This remains a troublesome name which I cannot place satisfactorily. Meisner recognized the species (DC. Prodr. 14: 165. 1856), noting that the type locality was Caracas, Venezuela, and he cited one specimen (7699) in the Willdenow herbarium. This specimen consists of two sterile shoots obtained from a plant cultivated in a botanic garden and certainly is not the Jacquin type. It is properly referred to Coccoloba diversifolia Jacq. In his monograph (Bot. Jahrb. 13: 158. 1890) Lindau also accepted Jacquin’s name and cited two specimens (without known collectors and from Caracas, Venezuela) to be found in the Delessert and Vienna her- baria. Lindau cited many additional collections from Florida, the Ba- hamas, Cuba, Hispaniola, Puerto Rico and the Virgin Islands. A specimen in the Prodromus herbarium at Geneva which Lindau saw, and the one I believe he cited, was probably collected by Bertero in Hispaniola. The Antillean and Florida material cited by Lindau has been referred to Coccoloba diversifolia Jacq. Although I have suggested that Coccoloba laurifolia Jacq. and C. diversifolia Jacq. may be the same (Jour. Arnold Arb. 40: 195-196. 1959), I am not entirely convinced of it. A re-examination of all material available to me from Venezuela has failed to reveal any plants which can be compared satisfactorily with the description and illustration supplied by Jacquin. The closest comparison in Venezuela would be with C. padi- formis Meisner based on the collection Moritz 377 from Caracas. Material from Central America which I have cited for C. padiformis (loc. cit. 210- 211) and additional collections to be cited in this study are not exactly comparable to Jacquin’s description and illustration. These differences at present are primarily in the venation as related to the texture of the leaf blade and in the shape of the leaf apex. A field study of Coccoloba plants in the vicinity of Caracas will be necessary to determine what species Jacquin had as a basis for his description and illustration of C. laurifolia. Coccoloba laxiflora Lindau, Bot. Jahrb. 13: 191. 1890. The holotype in the Berlin herbarium is Glaziou 11444 from Rio de 248 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLI aneiro, Brazil. This species is to be referred to the synonymy of Coccoloba ramosissima Weddell. Coccoloba lehmannii Lindau, Bot. Nat 29(Beibl. 49): 7. 1895; Howard, Jour. Arnold Arb. 40: 200. 1959. Coccoloba lehmanni Lindau, Repert. Sp. Nov. 1: 156. 1905. Coccoloba williamsii Standley, Publ. Field Mus. Bot. 11: 148. 1936. This species has been discussed in an earlier paper in which its range was extended to Central America and additional species from that area placed in its synonymy. The selection of a lectotype was also discussed at that time. Another similar species is Coccoloba lepidota A. C. Smith (g.v.). Additional material may show that this, too, should be placed in synonymy here. Two collections from Peru, Tessman 3896 and 5258 from Iquitos along the Amazon have been tentatively referred to this species. Several herbarium specimens have been seen of a collection made in 1871 from a cultivated plant in the Calcutta Botanic Garden. No collector or ata are given on the sheets, which have carried the name ‘“Coccoloba excoriata.” These are clearly to be referred to the present species. Colombia. ANTIoguUIA: Villa Arteaga, Lopez & Sanchez 40 (us); Cauca, Lehmann 7560 (p-lectotype). Meta: Puerto Lopez, E.L. & R.R. Little 8294 (N Venezuela. ANzoATEGUI: NE. of Bergantin, Steyermark 61217 (F); BaRINAS: Barinitas, Aristeguieta 1702 (us). MERIDA: Between Hacienda Agua Blanca, above La Azulita and Rio Capaz, Steyermark 56127 (F). Coccoloba lepidota A. C. Smith, Brittonia 2: 150. 1936. This species was distinguished by Smith by the “characteristic scales of the petioles and young stem parts.’’ Smith compared Coccoloba le pidota with C. ovata, which is clearly distinct. The separation of C. lepidota and C. lehmannii is more difficult and additional material may show that C. lepidota should be another synonym of that species. The lectotype of C. lehmannii, Lehmann 7560 (8), has shorter, more elliptic leaves and a predominance of simple pubescence. Coccoloba lepidota, as represented by the type collection, Krukoff 5660, has larger obovate-oblong leaves, broadest above the middle and tapering to an obtuse or truncate base. The young stems, petioles and ocreae are covered with lepidote scales and bear lesser amounts of simple hairs and resinous excretions. A tend- ency towards this development is found in the type collection of C. lehmannii and in the other collections cited below. It is probable that C. lepidota is an extreme variation of C. lehmannit. Brazil. AcRE: Near mouth of Rio Macauhan, Krukoff 5660 (xy—holotype, A, LE, M, W), 5659 (A, LE, M, NY Coccoloba leptostachya Bentham, Bot. Sulph. 59. 1856; Meisner, DC. Prodr. 14: 163. As has been pointed out in an earlier paper in this series (Jour. Arnold 1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, Ix 249 Arb. 40: 188. 1959), Bentham described this species, citing the type locality as ‘Libertad in Colombia.” The type is a Barclay specimen at Kew. I have studied this and have concluded that the specimen was collected in Central America. Recent collections from Libertad in El Salvador proved to match the Barclay collection well. I have seen no comparable material from Colombia. Coccoloba leptostachya Bentham is referred to the synonymy of C. barbadensis Jacquin (1760), which is known from Mexico, Guatemala and El Salvador. The species need no longer be considered in the South American flora. Coccoloba longependula Martius ex Meisner, Fl. Bras. 5(1): 27. pl. 9. 1855; Lindau, Bot. Jahrb. 13: 177. 1890. After an examination of the type of this species (Martius 759 from Minas Geraes, Brazil [m]), this has been referred to the synonymy of Coccoloba sticticaulis Weddell (q.v.). Coccoloba longiochreata Hassler, Repert. Sp. Nov. 14: 162. 1915. This species is clearly the same as Coccoloba cujabensis (qg.v.) and has been referred to synonymy there. Hassler cited two collections, Fiebrig 1284 and 1443, in his own herbarium. The collections were made in the Gran Chaco at Puerto Talavera, Paraguay. Coccoloba longipes S. Moore, Trans. Linn. Soc. II. 4: 446. 1895. Coccolobis padifolia Rusby, Mem. N. Y. Bot. Gard. 7: 235. 1927. In the original description Moore compared his new species with Cocco- loba laxiflora Lindau, which I now regard as the same as C. ramosissima Weddell. There is a striking similarity between these two species in the delicate inflorescences and the long peduncles. There are differences in the leaf size which should be re-examined when additional materials be- come available from southern Brazil. At present I distinguish C. longi- pes by the ovate-oblong leaf blades which taper from the middle to a blunt apex. Coccolobis padifolia Rusby was described without any comparison of other species. A study of the type collection indicates that it should be placed in the synonymy of C. longipes. Bolivia. Rurrenabaque, Rusby, Mulford Ex. No. 848 (ny-holotype of C. padifolia, GH). Brazil. Matto Grosso, S. Moore 577 (sm—holotype, B, NY). Coccoloba lucidula Bentham in Hooker, London Jour. Bot. 4: 627. Coccoloba sagotii Lindau, Bot. Jahrb. 13: 184. 1890. The two species treated here fall into widely separated portions of Lindau’s monographic treatment of Coccoloba, yet it seems clear to the writer that they are identical. Coccoloba lucidula was based on flowering 230 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI specimens with delicate, membranaceous, immature leaves which crinkled in drying, turned black and became lustrous on the upper surface. Ben- tham cited only one specimen collected by Schomburgk, “2nd Coll. 947 (1262). The species has not been collected again and no modern col- lections have been assigned to it. The mature foliage and fruiting speci- mens of C. sagotit would appear to belong in synonymy here, but since at present there is no comparable material for C. lucidula, there is need for additional mature specimens of the latter. Coccoloba lucidula is described as a woody vine by collectors of the specimens cited below. Perrottet 1820 (P) is a delicate vine tapering to a tenuous apex. On this specimen the immature condition of the leaves is clearly shown, from minute to fully expanded, though membranaceous, forms. Many of the mature leaves of other collections are folded, indi- cating that when fresh the midrib is sharply curved downward. The fruit is distinctive, being nearly spherical and smooth. A small stalk is dis- tinguishable at the base of the fruit and the apex is more or less obtuse, with very small, imbricate perianth lobes. Perrottet 83 from British Guiana which is referred here was cited by Lindau (loc. cit. 168) as Coccoloba racemulosa and, thus identified, was an important example of his Guiana— northern Brazil distribution (loc. cit. 116). Coccoloba sagotii was described by Lindau and was based on an unnum- bered Sagot collection from “Guyana gallica” near Cayenne. This is a fully matured branch of scrambling habit. The infructescence is old but fruits have been preserved. The leaves are coriaceous and shiny above. Lindau distinguished this species from C. lucidula by the glabrous branchlets and infructescence rachis, but close examination shows that in all reliable characteristics the type collections are similar. The pubescence present on material of C. sagotii was overlooked by Lindau. Lindau also referred to specimens in the herbaria at Berlin and Stockholm, but both of these are merely fragments. The most complete specimen of this collection is in the Paris herbarium. British Guiana. Coverden, Persaud 136 (F, K, NY); Demerara River, Jenman 6309 («); Ituni, south of Mackenzie, Cowan 39255 (K, NY); Roraima, Schom- burgk 947 (1262) (xK-type collection, BR). Without ae — Perrottet 83 (G), sm. (PB); Schomburgk 81 (8). French Guiana. CAYEN Martin s.n. Ce), Poiteau sm, (K), L.C. Richard s.n. (e), Sagot s.n. Gwe ie ar C. sagotii, B, P, S), Talbot sm. (K); Montagne de Kaw, Cowan 38798 (ny). Venezuela. Bolivar, Tumeremo, Stevermark 60942 (F). Coccoloba marginata Bentham, in Hooker, London Jour. Bot. 4: 626. Coccoloba guianensis Meisner, Linnaea 21: 264. 1848. Coccoloba martii Meisner, Fl. Bras. 5(1): 37. 1855. Coccoloba marti var. major Meisner, zbid. 38. Coccoloba martii var. minor ae ibid. Coccoloba nitida var. cordata Meisner, ibid. Coccoloba nitida var. rotundata Meisner, ibid. 1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX Zot Coccoloba trinitatis Lindau, Bot. Jahrb. 13: 182. 1890. Coccoloba douradensis Glaziou, Bull. Soc. Bot. Fr. IV. 11(Mem. 3f): 571. 1911 (provisional name with mixed type; see also C. densifrons). A discussion of this species is given in two earlier papers (Jour. Arnold Arb. 40: 84-85. 1959, and 41: 45-46. 1960). I have seen additional material (cited below) which extends the range of this species to Vene- zuela and possibly to the Brazilian states of Minas Geraes, Goyas, Bahia, Acre and Santa Catarina. Additional field study is needed to determine the variations in individual plants as these occur in South America. While I am following Lindau in considering Coccoloba marti a synonym of C. marginata, I wish to point out the possibility that C. marti more properly may be assigned to C. peltata Schott. Certainly the Salzmann collections from Bahia previously identified as “C. pendula” or C. nitida var. cordata are intermediate between material of C. peltata from Rio de Janeiro and material of C. marginata from the Guianas. At present Cocco- loba peltata may be represented only by anomalous material and there- fore the species may be incorrectly interpreted. Brazil. Acre: Rio Macauhan, Krukoff 5479 (c, K, M, w). AMAZONAS: Sao Paulo de Olivenca, Krukoff 9048 (F, K, NY); without eee locality, Ule 9347 (c, K). Banta: Chapada do Rio das Femmeas, Carrasco, Ilheos, Riedel 244 (LE, P), Blanchet 3049 (LE); Liitzelburg 516 (ar) : without specific locality, Salzmann 475 (p). Goyas: Chapada do Rio Preto, Liitzelburg 1304 (mM); Patavidado, Macedo 3.859 (K); without specific locality, Burchell 7768 (Pp). Minas GERAES: Caraca, Tavares 316 (Mm). British Guiana. Kaieteur Plateau, Maguire & Fanshawe 23316 (A, NY); Kaieteur Savannah, Potaro River, Jen- mann 831 (K); basin of Kuyuwini River, A.C. Smith 3030 (A); Waini River, e la Cruz 3712 (GH); without locality, Poiteau 179 (LE). Dutch Guiana. Paramaribo, Kappler 1620 (Pp), Wullschlagel 882 (Br); without specific locality, Hostmann 506 (Pe), Wullschlagel 992 (BR). French Guiana. Cayenne, Broadway 307 (cH); without specific locality, Lepriewr 187 (aA), Lequillon sn. (P), Melinon 252 (A). Venezuela. AMAzoNAS: Tamatama, Upper Orinoco, Llewelyn Williams 15233 (F); without specific locality, Gines 5105 (us). BoLtvar: Raudal Guaiquinima, Cardona 474 (us), 475 (us), Maguire 33134 (a, NY). Coccoloba martii Meisner, Fl. Bras. 5(1): 37. 1855. Coccoloba marti var. major Meisner, ibid. 38. 1855. Coccoloba martit var. minor Meisner, ibid. No type had been selected, but the species was considered to consist of its two varieties. An examination of the material cited led to the con- clusion that these taxa may be referred to the synonymy of Coccoloba marginata Bentham. There is a possibility, as was pointed out in the dis- cussion of C. marginata, that C. marti and C. peltata are the same. Coccoloba meissneriana (Britton) K. Schum. in Just, Bot. Jahresber. 28(1): 451. 1902. Uvifera meissneriana Britton in Rusby, Bull. Torrey Club 27: 129. 1900. 252 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. XLI This species is known from but two collections from the same area. It is similar to Coccoloba peruviana and eventually both may be included in C. obtusifolia (q.v.). At present it can be distinguished by the tomentum on the young stems and petioles, on the entire lower leaf surface and on the midrib of the upper leaf surface. The inflorescence is copiously pu- bescent, as well. The fruits are comparable to those of C. obtusifolia, hav- ing the achene surrounded by the imbricated lobes of the perianth. The Rusby collection is from a staminate plant and was in flower in May. The Bang collection, made in July, is in fruit. Bolivia. Guanai, Rusby 1918 (Ny—holotype, B, GH), Bang 1595 (A, GH, K, LE, M, NY Coccoloba membranacea Klotzsch, Linnaea 14: 289. 1840. This species is apparently based on a Luschnath collection from Bahia Brazil. The original description is brief, ““Arborescens, floribus lutescenti- viridibus.”” Lindau (Bot. Jahrb. 13: 165. 1890) regarded the original epithet as a nomen nudum and referred it to the synonymy of Coccoloba ilheensis. At that time he cited “Luschnath 42,’ a specimen of which is in the Leningrad herbarium. I have seen that sheet, but there is no anno- tation to indicate that it is the type of C. membranacea. The specimen is properly referred to C. ilheensts. Coccoloba microneura Meisner, DC. Prodr. 14: 163. 1856; Howard, Jour. Arnold Arb. 41: 42. 1960. This species has been discussed in the earlier paper cited above and referred to the synonymy of Coccoloba nitida HBK. The type was Purdie s.n., collected in the vicinity of Santa Marta, Colombia. Meisner reported the type specimen to be in the Arnott herbarium, but such a specimen cannot be found, although there is a specimen in the herbarium of the Royal Botanic Garden at Kew. Coccoloba microphylla Morong in Morong & Britton, Enum. Pl. Coll. Parag. 212. 1892; Ann. N. Y. Acad. 7: 213. 1893, not Griseb. 1866. This species was based on Morong 899, made along the Rio Pilcomayo in Paraguay. Because the epithet is a later homonym of Coccoloba micro- phylla Grisebach, Hassler renamed it C. morongii. An examination of the type collections shows that it should be referred to the synonymy of C. paraguariensis Lindau. Coccoloba micropunctata Eyma, Meded. Bot. Mus. Utrecht 4: 1. 1932. I am unable to accept Eyma’s criteria for distinguishing the material he cited as a species distinct from Coccoloba excelsa (q.v.), and so have referred his species to synonymy there. The type selected was Stahel 77 from Dutch Guiana. 1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX 253 Coccoloba mollis Casaretto, Nov. Stirp. Bras. 72. 1844. Coccoloba polystachya Weddell, Ann. Sci. Nat. III. 13: 261. 1850. Coccoloba paniculata Meisner, Fl. Bras. 5(1): 43. pls. 20, 21. 1855. Coccoloba polystachya var. mollis Meisner, DC. Prodr. 14: 151. 1856. Coccoloba polystachya var. glabra Lindau, Bot. Jahrb. 13: 133. 1890. Coccoloba polystachya var. pubescens Lindau, ibid. Casaretto cited no collection by number or name of collector in the original publication, so one must assume that he was referring to his own collection. Such a specimen, now in the Turino herbarium, was made on the island of Itaparica, near Bahia, Brazil, and the data on the label agrees in description and location with that published by Casaretto. The label on the specimen also stated the number of the collection as 2218 and the catalogue number as 80. Lindau cited “Casaretto 2218” and “Meisner 80.” These are one and the same sheet. This single sheet in the Turino herbarium, the holotype of this species, is a sterile specimen in poor condition consisting of two leafless twigs and five detached leaves, probably coming from an adventitious shoot since one twig is extremely pubescent. The conspicuous development of the pubescence can be matched on the collection Lutzelburg 295 in the Munich herbarium. The label states that this collection came from a tree 6 meters tall; however, the very large leaves, long internodes and copious pubescence all suggest ab- normal or adventitious growth. Other flowering collections by Litzelburg cited below from the state of Ceara appear to be transitional in pubescence, size and shape of leaves and length of internodes. Weddell described Coccoloba polystachya, being unaware of Casaretto’s name. He saw a living specimen which was given a catalogue number, and he also cited the collection “Martius 1242.” No specimens of the living plant appear to have been preserved, so the Martius collection may be taken as the type of C. polystachya. The specimens I have seen of Martius 1242 bear female flowers and a sheet from Leningrad has the fragment of a sterile shoot, but there is no doubt that the Leningrad collec- tion is the same as the more pubescent type of C. mollis Casaretto. In 1855 in the Flora Brasiliensis Meisner published a treatment of the genus Coccoloba. He accepted C. polystachya Weddell and referred C. mollis Casaretto to the synonymy of Weddell’s species, with a question. He cited specimens collected by Salzmann, Spruce and Weddell. The illus- tration given for this species is of a staminate plant and does not repre- sent the type of either C. mollis or C. polystachya. At the same time Meisner described a new species, Coccoloba paniculata, illustrated in two plates by staminate and pistillate plants. He cited an unnumbered collection by Pohl, the collection Poeppig 2649, and also Martius 1242, which Weddell had cited in the original description of C. polystachya. Meisner distinguished between C. polystachya and C. paniculata by the more abundant pubescence of the former. In 1856 Meisner repeated for the Prodromus the description of the two species and, in addition, listed Coccoloba polystachya var. mollis, based 254 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI only on C. mollis Casaretto and the collection “Casaretto 80.” It is of interest to note that Meisner cited Martius 1242 under both species, but indicated (with an exclamation point) only the citation of this collection as C. polystachya. In 1890 Lindau recognized Coccoloba polystachyva, with two varieties. Coccoloba polystachya var. glabra Lindau is based on C. paniculata Meis- ner, while C. polystachyva var. pubescens Lindau is based on C. mollis Casaretto, Lindau noted that the varieties were based on the amount of pubescence and that intermediates were to be found. Eyma (Meded. Bot. Mus. Utrecht 4: 4. 1932) recognized that the oldest name was Coccoloba mollis Casaretto and accepted this, including with it C. polystachya Weddell and C. polystachya var. pubescens Lindau. Eyma did not treat C. paniculata Meisner or C. polystachya var. glabra, thus implying his acceptance of them. Macbride’s treatment of the genus for the Flora of Peru (Publ. Field Mus. Bot. 13: 460. 1937) appears to be based on the work of Eyma, although no reference is given. I have seen only a few of the specimens cited by Eyma. In general the specimens from French and Dutch Guiana have a different aspect in the texture of the leaves and the color of the pubescence. Moreover, the petioles and branches of the inflorescence tend to be longer. The plants from this area may represent a geographic race, or perhaps even a distinct species. Ad- ditional material is needed for an understanding of the conditions seen in these plants. One collection of Krukoff from the Basin of the Rio Solimoes, also, is difficult to fit into the general pattern of Coccoloba mollis. This col- lection, Krukoff 8841, has leaves of still different texture and in this case the branches of the inflorescence are short, resembling those of Coccoloba dugandiana. The collection is in fruit and the samples opened, all sterile and hollow, are strongly triangular in outline and section. The lobes of the perianth are appressed against the apex, rather than coronate, as in the few fruits seen of typical Coccoloba mollis. At present the collection does not merit description as a new species. I have not accepted the glabrous variety created by Lindau, since ad- ditional field study of this species is needed to understand the variation in pubescence with the age and habit of the plant. The species seems clearly dioecious, the pistillate plants appearing to be more pubescent than the staminate plants. Collections made from the coastal areas are also more pubescent than those from inland stations in South America. The shape of the leaf, particularly the base, and the length of the petiole are extremely variable in the specimens cited. The species is easily recog- nized, since so few species of Coccoloba have paniculate inflorescences; owever, no existing description is adequate. I sincerely hope that some botanist in an area where this plant grows can make the necessary study of variations in C. mollis. Bolivia. SANTA Cruz: Sara, Bosques de Buenavista, Steinbach 6563 (A). Without specific locality: Yungas, Bang 299 (Gc, GH, K, LE, M, NY). 1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX 255 Brazil. AcRE: Rio Macauhan, Krukoff 5550 (A, M, NY); Seringoel Auristella, Ule 9346 (c, K). AMAZONAS: Ega, Poeppig 2649 (s, LE); Sao Paulo de Oli- venga, Palmares, Krukoff 8314, 8337 (A, BR, LE, NY), 8841 (A, BR, NY); Humayta near Tres Coe Krukoff 5550 (A, M, Ny). Banta: Camapuan, Riedel 628 (a, LE); Itaparica, Casaretto s.n. (ro-holotype); Sao Bento das Lages Littzelburg 295 (mM). CearA: Barxa d’Anta, Liitzelburg 26278 (m); Grangeiro, Liitzelburg 25800 (mM, w), 25838 (m); Soure, Drouet 2377 (cH); without specific locality, Gardner 1828 (Ny). Goyaz: Rio dos Alnas, Glaziou 21980 (a, LE); Tocanti- nopolis, Pires & Black 1650a (us); without specific locality, Burchell 7351-2 (GH). MaRANHAO: Loreto, Snethlage 656 (F); Maracassumé River, Froes 1811 (A, NY). ae Grosso: Cuyaba, Martius 1242 (m-holotype of C. Polvsiachya. BR, LE, NY). NAS GERAES: Paracatu, Riedel s.n. (LE). ParA: Barra do Rio Neato, eee sm, (Oct. 1850) (8, GH, LE, M, NY); Cassipa, Tapajos River re- gion, Krukoff 1246 (a, NY). PERNAMBUCO: Tapera, Pickel 2483 (GH). Piauny: Urussuhy, Suethlage 633 (F). Rio DE JANEIRO: Without specific locality, Burchell 5912 (GH). Locality not specified: Pohl s.n. (Br, M, NY). Dutch Guiana. Bradi- lifi, Matoela, Stahel 189 (A); Zanderij I., Herb. eee 189 (Ny), pire (A, NY). Peuador Manast: El Recreo, Balan, Eggers 14497 (A, B, LE, M), 15675 GH, K, LE, M, NY). French Guiana. Godebert, Wachenheim SM. (a): without locality, Melinon 106 (A). Peru. Loreto: Florida, Rio Putumayo at mouth of Rio Zubineta, Klug 1991 (a, GH, Ny); Rio Santiago, Tessmann 4372 (NY); Middle Ucayale, Tessmann 3195 (xy). Coccoloba monoica Ruiz ex Meisner, DC. Prodr. 14: 149. 1856. Meisner cited this name in synonymy as “Coccoloba monoica fl. peruv. Ruiz” and recorded seeing a specimen in the Berlin herbarium. Lindau (Bot. Jahrb. 13: 220. 1890) apparently saw the authentic material, since he referred the epithet to synonymy under Muhlenbeckia tamnifolia var. laxiflora Meisner. The type specimen was not located during a brief search in the Berlin herbarium several years ago. Coccoloba morongii Hassler, Repert. Sp. Nov. 14: 162. 1915. This was a new name, provided by Hassler for Coccoloba microphylla Morong (1893), not C. microphylla Griseb. (1866). The species is to be placed in the synonymy of C. paraguariensis Lindau. Coccoloba moritziana Klotzsch ex Meisner, Fl. Bras. 5(1): 28. 1955. Meisner published this epithet in the synonymy of his new Coccoloba moritzii var. opaca and reported that he found the manuscript name in the Berlin herbarium. I have not been able to locate such a specimen although Lindau (Bot. Jahrb. 13: 216. 1890) referred the name to the synonymy of C. ovata. Coccoloba moritzii Meisner, Fl. Bras. 5(1): 28. 1855. Coccoloba moritzii var. opaca Meisner, ibid. Coccoloba moritzii var. lucida Meisner, ibid. Lindau (Bot. Jahrb. 13: 216. 1890) has referred this species and its 256 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI varieties to the synonymy of Coccoloba ovata. The difficulty in typifying these names will be discussed under C. ovata. On the basis of the material I have seen, I believe Lindau’s action to be correct. In the original publi- cation Meisner attributed the name Coccoloba moritzii to Klotzsch and cited in the synonymy of C. moritzit var. opaca the manuscript name Coccoloba moritziana which he found in the Berlin herbarium. I have not been able to find the name published in any of Klotzsch’s writings. Coccoloba mosenii Lindau, Bot. Jahrb. 13: 173. 1890, “Moseni.” Coccoloba fastigiata var. glabrata Meisner, F). Bras. 5(1): 34. 1855. This species represents a climbing plant with the leaves borne on short lateral branches. The leaf blades have a characteristic shape, oblong- obovate to nearly lanceolate-obovate. The range of variation in habit, as well as in shape of leaf cannot be determined from the few specimens on hand. Additional material is needed for further study. The basis for the proper assignment of Coccoloba fastigiata var. glabrata to synonymy here has been discussed under that epithet. . SAo PauLo. Santos Lorosocaba, Mosen 3458 (s-lectotype, B, G, P), Loefgren 10432 (m). Without definite locality, Burchell 3844 (P). Coccoloba nigra Fawcett & Rendle, Jour. Bot. 51: 123. 1913; FI. Jamaica 3: 120. 1914; Howard, Jour. Arnold Arb. 38: 106. 1957. As I have discussed in an earlier paper, Fawcett and Rendle based this species on a collection annotated “Jamaica,” but without the collector’s name or number. The type in the Edinburgh herbarium has been studied, and it is certainly a fragment of Schomburgk 531, the type of Coccoloba ovata, from British Guiana. A specimen of the Schomburgk collection is also in the Edinburgh herbarium and the two sheets match, even to the lichens on the branches. The name Coccoloba nigra must therefore be assigned to the synonymy of C. ovata Bentham. Coccoloba nitida HBK. Nov. Gen. 2: 176. 1818. Coccoloba microneura Meisner, DC. Prodr. 14: 163. 1856. In the eighth paper of this series (Jour. Arnold Arb. 41: 41-42. 1960), I corrected an earlier mistake and correctly defined Coccoloba nitida as a species currently known only from Colombia. A lectotype (Humboldt 1627) was designated in the Paris herbarium. It was collected at San Bartholome on the Rio Magdalena. Coccoloba microneura is clearly the same species and was described from the Purdie collection, without number, from Santa Marta, Colombia. Coccoloba nivea Jacquin, Hist. Stirp. Am. 115. pl. 78. 1763. Several modern writers on South American vegetation have used this 1960] HOWARD, STUDIES IN THE GENUS COCCOLOBA, IX raSy| epithet without citing specimens which can be identified. Coccoloba nivea acq. is a synonym of C. venosa L. (1759), and specimens from South America will be cited under that name. Schomburgk referred to Coccoloba nivea under cultivation in his Flora and Fauna of British Guiana. Coccoloba novogranatensis ae Bot. Jahrb. 13: 192. 1890; Howard, Jour. Arnold Arb. 41: 40. I have referred this species to the synonymy of Coccoloba coronata Jacq. (q.v.). In an earlier paper (Jour. Arnold Arb. 40: 85-86. 1959) I selected as a lectotype one sheet of the collection Triana 978 in the herbar- ium of the Muséum d’Histoire Naturelle, Paris. The Triana collection was made between Anapoima and Apulo, Prov. Bogota, Colombia. Coccoloba nutans HBK. Nov. Gen. 2: 175. 1818. Authentic material of this species indicates that the specific name was proposed by Kunth. The original description is brief and reflects the in- adequate and immature condition of the specimens. Ocreae are not present on the stems and the inflorescence, described as nutant, is im- mature or abortive. Meisner (DC. Prodr. 14: 155. 1856.) repeated the original description, with minor changes, and reported, ‘Species non satis nota, nec in herb. Kunth, nec in Willdenowiano extans.”’ Lindau attributed the collection to Bonpland and cited from the Berlin herbarium a speci- men which consists of one detached leaf and the fragment of an inflores- cence 2 cm. in length. These fragments were obtained from the Paris herbarium. I have been able to examine the original collection in Paris which is, in turn, from the Bonpland herbarium, It currently consists of a short stem, without ocreae, and is broken at both ends. A single recurved lateral inflorescence having very immature flower buds is present. Poor though it is, this collection must be designated as the holotype. No recent collections have been assigned to this species. Macbride, who treated the genus for the Flora of Peru, saw no material. However, I believe that two of the collections he cited ander Coccoloba sphaerococca are more properly placed in the present species. A collection, Killip and Smith 29027 (Nv), made at Yurimaguas on the lower Rio Huallaga, Dept. Loreto, Peru, consists of the stem and leaves of a woody vine. One de- tached leaf is comparable to that of the Bonpland collection. Additional smaller leaves are present, but resemble the leaves of Coccoloba ascendens. Although Kunth did not record the height or habit of Coccoloba nutans in the original description, Meisner, Lindau and Macbride have referred to the plant as a tree. I believe that they are in error and that Coccoloba nutans is a woody vine usually with coriaceous, elliptical leaves rounded at the base, but that occasionally on vigorous shoots the oblong-obovate leaves with subcordate bases are produced. The species is similar to Coccoloba ascendens, but additional material is required for a proper understanding of the species. 258 JOURNAL OF THE ARNOLD ARBORETUM [¥OL, XL Killip and Smith 27958 and 29027 and that of Llewelyn Williams 3805, assigned in various herbaria to Coccoloba peltata Schott and Coccoloba sphaerocarpa Lindau, appear to be C. nutans. Coccoloba nymphaeifolia Schenk in Zittel, Handb. Palaeont. 2: 491. 1887; Lindau, Bot. Jahrb. 13: 181. 1890. Schenk used the name Coccoloba nymphaeifolia in comparing fossil leaf material to living species. At that time a plant called Coccoloba nym phaeifolia was under cultivation in the Leipzig botanical garden. A single leaf is preserved in the Berlin herbarium and bears the annotation, “Coccoloba nymphaeifolia de Jonge, H. Lips. Brazil.” Lindau correctly referred this specimen and name to the synonymy of C. peltata Schott. The epithet C. nymphaeifolia is a nomen nudum. To be concluded | 1960] SOLBRIG, LEAF VENATION IN RAOULIA 259 LEAF VENATION AND PUBESCENCE IN THE GENUS RAOULIA (COMPOSITAE) Orto T. Sorsric 4 MOopERN DICOTYLEDONS are characterized by a highly diversified pat- tern of reticulate foliar venation, and true open venation is a rare oc- currence. The persistence of presumably primitive dichotomous open vena- tion in Kingdonia (Ranunculaceae?) and Circaeaster (Ranunculaceae?, Circaeasteraceae?) as shown recently by Foster (1959a,b) suggests that reticulate venation has probably evolved from dichotomous venation. On the other hand, species of the genus Raoulia provide an example of an “open” venation which is derived, rather than primitive. Since the vena- tion of Raoulia has not been studied in detail previously, the present investigation is an attempt to provide further data which may contribute both toward a better understanding of the complicated picture of venation in angiosperms and toward a more natural classification of the genus Raoulia itself. Raoulia comprises a group of specialized plants which are typically found in dry, rocky habitats between sea level and 2000 meters in both the North and South Islands of New Zealand (to which the genus is restricted, with one possible exception in New Guinea). The habit varies from small, compact, creeping or tufted herbs to small or large cushion-forming plants of the type commonly known as “vegetable sheep.” The genus belongs to the family Compositae, tribe Inuleae, subtribe Gnaphaliinae, a taxonomi- cally difficult subtribe in which limits of genera are not always very Clear. Within the group Raoulia is, in fact, defined as much by its habit and geographical distribution as by any qualitative character (Cheeseman, 1925). As treated by both Beauverd (1910, 1912) and Cheeseman (1925), the genus is composed of 21 species grouped into two subgenera. These are further divided into two and three sections respectively, according to involucral and foliar characters. Material of seventeen species including representatives of each of the sections has been available for study. In the discussion which follows the nomenclature is that of Beauverd. 1 This investigation was suggested by Dr. Adriance S. Foster, who also encouraged and advised the author during its course. Sincere thanks are expressed for all the invaluable help received. Dr. John Rattenbury of Aukland University College, New Zealand, supplied dried material, and the University of California Botanical Garden grew and supplied living material of some species. Drs. Foster, Sherwin Carlquist, Lincoln Constance, and Carroll E. Wood kindly read the manuscript and made various suggestions. Finally, the help of the directors and curators of the Gray Herbarium (cu), the Arnold Arboretum (A), and the University of California Herbarium (uc) for allowing the use of herbarium material in their custody is acknowledged. 260 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI 12 Fics. 1-16. Leaf venation of species of Raozlia. Veins in black, mesophyll sclereids in longitudinal hatching. Differences in width of veins due to scleren- chyma. Each division of scale = 1 mm. 1, Raoulia Petriensis Kirk (Cheeseman n., GH); 2, R. glabra Hook. f. (Cheeseman s.n., GH); 3, R. subsericea Hook. f. (Petrie 193, GH). 4, R. tenuicaulis Hook. f. (Kirk s.n., cH); 5, R. Monroi 1960] SOLBRIG, LEAF VENATION IN RAOULIA 261 MATERIAL AND METHODS Most observations were made from dried herbarium specimens, for only two species, Raoulia glabra and R. australis, could be studied from living material. The leaves, bracts and flowers of all available species of Raoulia were cleared according to the technique of Foster (1955; see also Arnott 1959). In order to extract some dark material which remained after treatment with sodium hydroxide and chloral hydrate, leaves were bleached in a 50% solution of “Clorox” until reasonably clear, usually about five minutes, and then thoroughly washed to assure good subsequent staining. The leaves and stems of certain species of Raoulia, especially the cushion-forming ones, are heavily impregnated with tannins and other dark-staining substances. In order to remove these, long treatment with sodium hydroxide and “Clorox” is needed which damages the leaves or makes them so brittle that they disintegrate when the heavy coat of hairs is removed. Therefore, the clearing process often had to be interrupted as soon as a reasonably good view of the venation was obtained. Both surfaces of the leaves of all species of Raoulia are covered with a thick layer of long trichomes which entirely cover the leaf and which it was necessary to remove in order to study the venation. This was done by “shaving” the hairs off with a sharp scalpel after the leaves had been cleared and dehydrated. The smallness of the leaves (less than one centimeter in every case) and their fragility made this operation a tedious, time-consuming, and “nerve-shattering” operation and resulted, in spite of extreme care, in a high degree of tearing of the leaves. OBSERVATIONS AND RESULTS Leaf shape. The leaves of all species of Raoulia are small. The size of the mature leaves varies from 3 mm. in leaves of R. lutescens and R. Goyenii to nearly 8 mm. in R. grandiflora. Most species have leaves of 4—5 mm. in length. In general, the leaves are appressed to the shoots. The degree to which they are appressed varies greatly and is correlated with the general habit of the plant. In such creeping species as R. glabra or tufted ones as K. grandiflora, the leaves are appressed to the shoot only at their bases and are curved outward above. In these cases the leaves are somewhat keeled and divergent. The leaves of these species are generally broadly linear- lanceolate or subulate and are usually longer than the average for the Hook. f. (Kirk s.n., GH); 6, R. Parkii Buchanan (Anderson 171, GH); 7, R. bryoides Hook. f. (Cheeseman, 1911 cH); 8, R. Bacto: Beauverd (Anderson , GH); 9, R. Buchananii Kirk (Cockayne 7041, A); 1 australis Hook f. (Kirk s.n., GH); 11, oyenit Kirk (Kirk s.n., GH); is ee Hook (Cockayne 7048, GH); i R. mammillaris Hook. f. Rak 5 me GH arg lata Hook. f. (Petrie 149, GH); 15, R. Hectori Hook. f. (plies Deh GH): ); R. grandiflora Hook. f. (Petrie, 1889, GH). 262 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI On the other hand, in the cushion-forming species, such as R. bryoides or R. rubra, the leaves are very much crowded in a tightly appressed spiral, with the lower leaves covering most of the outer surface of those lying above. The leaves are rather flat, not keeled, and usually rectangular or nearly rectangular in shape. Aside from these two types, there are some intermediate ones, which are correlated with intermediate habit of growth (see TABLE I) Venation. Three main types of venation can be distinguished in Raoulia: the reticulate type (1), with three main veins; the semireticulate (II), with either three or more frequently one main vein; and the striate type (IIT). I. RETICULATE TYPE. The basic type is represented by Raoulia Parkii (Fic. 6). Three main veins enter the leaf base from the stem. The mid- vein branches two or three times and these branches in turn form lateral branches which anastomose with the two lateral veins, forming a simple reticulum. The following species can be arranged in a progressively simpler series. In R. tenuicaulis (Fic. 4) some of the secondaries do not connect and the leaf presents a “semiopen” pattern. Raoulia Monroi (Fic. 5) is very similar in venation to R. tenuicaulis. In R. subsericea (Fic. 3), the reduction is a little more accentuated: the laterals are less well marked, and the connections between the central and the lateral veins are usually unbranched. Some lateral open veinlets occur in this species. Finally, in k. glabra (Fic. 2) there are a main and two lateral veins with three to four connecting veins, and an occasional freely terminating veinlet. The pattern in R. glabra is scarcely reticulate. IT. SEMIRETICULATE TYPE. This type is not so clearly delimited as the preceding one, and, although a linear reduction series can be inferred, a subdivision into loosely connected subgroups (indicated by the letters a, b, c, and d) is probably more significant. a. Raoulia australis (Fic. 10) and R. Petriensis (Ftc. 1) have three main veins, which run unbranched for about three-fourths of the length of the leaves. These veins then branch forming a coarse meshwork with some unconnected veinlets. In R. australis the secondaries are relatively fewer than in R. Petriensis. These two species are the only ones of this type with three main veins and form a link with the plants of the pre- ceding type, particularly with R. Parkii, from which the leaves do not differ greatly. b. Raoulia lutescens (Fic. 8) and R. bryoides (Fic. 7) have basically the same or a similar arrangement of the end veinlets as R. australis, but, instead of three main veins, they only have a midvein running the length of the leaf. Two open veinlets which point towards the base of the leaf at each side,of the terminal reticulum seem to indicate arrested stages in the development of the two lateral veins, a situation similar to that in Minuartia aretioides (Caryophyllaceae) (Troll 1938, pp. 1085, 1086). Whether these veins have disappeared through reduction, or whether this is an indication of a first stage in development, cannot be answered with 1960] SOLBRIG, LEAF VENATION IN RAOULIA 263 certainty. However, the increasing specialization in other features of the plant, correlated with the simple leaf venation in this type, indicates that a reduction series is the more likely possibility. c, Raoulia Buchananii (Fic. 9), R. Goyenii (Fic. 11), and R. rubra (Fic. 24), in this order, show an increasing simplification in the pattern described in the above paragraph. The noteworthy features here are that there is no indication of vestigial lateral traces, and that the ultimate veinlets are dichotomously branched and terminate freely without anastomosing. The maximum degree of simplification is evident in R. rubra. d. Raoulia mammillaris (Fic. 13) and R. eximia (Fic. 12) also have an extremely simple venation which consists in these species of the mid- vein and, toward the tip of the leaf, two or three secondaries, which may or may not bifurcate in turn. A very interesting feature is that the mid- vein and the secondaries are very oddly shaped and enclosed by a sheath of sclereids. III. SrriaTe TYPE. Three species are of this type which is apparently not connected with the other two types. Raoulia grandiflora (Fic. 16) has the most elaborately developed venation of the three species. Three main veins enter the leaf. (Sometimes two of them are anastomosed at the very base of the leaf. Whether this indicates a possible one-trace condition at the node is not known.) The midvein bifurcates once or twice, the branches running parallel and finally anastomosing with the lateral veins towards the tip of the leaf. The lateral veins sometimes bifurcate, the branches of this division running parallel and then uniting again to form an elliptical loop. All three main veins meet at the tip of the leaf. One or more backward-pointing open veinlets are usually present at each side of the lateral veins. Short strands of vascular tissue unconnected with the system of veins are sometimes present. In R. Hectoru (Fic. 15) the pattern is essentially the same but much reduced, since usually neither the midvein nor the lateral veins branch. In R. subulata (Fic. 14) there is only a single unbranched midvein, the laterals having disappeared. The unique feature, aside from the venation, which unites these three species, is the extreme width of the vascular bundles which are covered by a thick sheath of sclereids. Sclerenchyma. The leaves of Raoulia present always a strong develop- ment of sclereids. This is more evident in the older leaves, in which may also be found parenchyma cells (and even epidermal cells) with thickened, lignified walls and forming a sheet that can occupy up to two-thirds of the leaf surface. The development of sclerenchyma fibers surrounding the vascular bundles varies from species to species. In R. glabra there is a relatively small development, but sclerenchyma usually forms an enveloping sheath at least as thick as the bundle proper. In the species with a striate type of venation (R. grandiflora, R. Hectoru and R. subulata) the develop- 264 JOURNAL OF THE ARNOLD ARBORETUM [VOL. II ment of a sheath of sclerenchyma fibers around the bundles is particularly notable (Fics. 14-16). Still, in other species, such as R. Buchananii, R. eximia, and R. mammillaris irregularly shaped sclereids are associated with the vascular bundles, particularly towards the end of the veins (Fics. 9, 12, 13). These sclereids account for the irregular diameter and shape of the bundles. In some species with “open” venation (e.g., R. Buchananii) the tip of the bundles may be totally devoid of sclereids. Sclereids are also present sometimes in the mesophyll. In these cases they are elongated polygonal cells with thickened walls. They are par- ticularly prominent in R. Petriensis, R. bryoides, and R. Buchananii (Fics. 1, 7, 9), where they are present in very young leaves. Neverthe- less, in most species some mesophyll sclereids are present, particularly in older leaves. Mesophyll sclereids usually form a broad sheet at each side of the mid- vein in the lower and middle portion of the leaf. They may also be present in isolated groups or forming small patches. The exact position of these sclereids is unpredictable in any particular leaf, but they are never present in the upper fourth of the lamina. Trichomes. As already mentioned, the surface of the leaves is covered by a thick cap of trichomes. These are long, uniseriate, multicellular hairs, of a type occurring commonly in Compositae (Metcalfe and Chalk, 1950). The type of hair is the same in all species, but there are differences in the general appearance and distribution of hairs on the leaf surface. The genus can be divided into two groups of species: 1) those which have their leaves covered by a thick layer of tightly interwoven trichomes forming a felt-like cap; 2) those that have their leaves with a cover of more or less straight and somewhat stiff hairs that usually point towards the apex of the leaves. Finally, Raoulia subulata has glabrous leaves, whereas R. glabra has only a few and widely spaced hairs on the leaf surfaces. Species of the first type usually have the entire leaf surface covered with a thick cap of interwoven hairs which may be as thick as or thicker than the leaf tissues proper. On the other hand, the species which have hairs of the “stiff” type, have these concentrated on the apical end and usually more densely on the adaxial surface in this region. These differences can be correlated with the pattern of apical growth. The species with “inter- woven” hairs have more elongated stems and reflexed leaves. On the other hand, the species with “stiff” hairs are true cushion plants. Their cauline leaves are numerous, imbricated, and closely appressed. As a consequence, only the upper third of the leaf is exposed. DIsCUSsSION Raoulia leaves present some of the few instances of open venation in dicotyledonous angiosperms. Although an open dichotomous venation is considered primitive in gymnosperms and vascular cryptogams, this con- 1960] SOLBRIG, LEAF VENATION IN RAOULIA 265 dition is undoubtedly secondary and derived in Raoulia, as already pointed out by Troll (1938). That this is so can be inferred from two sets of facts: 1) the correlation of open venation with specialization in other features of the plant; 2) the entirely different typology of the venation of Raoulia and of a typically primitive dichotomous venation as found in some gymnosperms and vascular cryptogams or in the primitive angiosperm Kingdonta uniflora (Foster 1959b). Further evidence is to be found in the venation of Ewartia, undoubtedly the genus closest to Raoulia (Beauverd 1910). The venation of both species of Ewartia investigated, E. catipes (Fic. 26) and E. nubigena (Fic. 25), is reticulate and of a type similar to that of Raoulia Parkii, although more elaborate. Since other Gnaphaliinae, including Gnaphalium itself, have a venation along the same lines as that of Ewartia, it may be concluded that the venation of Raowlia is atypical and specialized. At this point it may be of interest to see how the venation of Raoulia compares with that of Kingdonia uniflora, believed to be truly primitive (Foster 1959b). The nodal anatomy of Raoulia is not known, but the vascular supply of the leaf is formed by one to three traces. Kingdonia has a unilacunar node, with four bundles (two of which may appear as double strands) forming the vascular supply of the foliage leaf. More significant still is the total venation pattern. The veins in Kingdonia are dichotomously branched, the forks of the branches extending usually into the tips of the dentations. Occasional anastomoses, as well as “blind” endings unrelated to the dentations are not uncommon, but in general the pattern is quite regular and symmetric. In contrast with this, the pattern in Raoulia, even in those species with a very simple and “open” venation (e.g., R. eximia or R. rubra), is irregular, and of a type altogether dif- ferent. While Kingdonia has a dichotomous venation, Raoulia has a branching pattern with does not follow any regular established system. This is probably due to the connection of the “open” venation of Raoulia with the extreme reduction in the lamina and its derivation from a reticulate type. The “open” situation when present is possibly the result of arrested growth of the veinlets. In contrast, the regularity of the “open” venation in Kingdonia is the consequence of the full development of the leaf along an established pattern. It is interesting to note that in Kingdonia and in at least one species of Raoulia (R. grandiflora) short strands of vascular tissue unconnected with the system of veins are found. The reason for this remains unexplained. The position of these strands and their occurrence are rather erratic in both species. In brief we may say that, notwithstanding the fact that the types of venation present in Raoulia and Kingdonia have been sometimes classed together (Troll 1938), careful analysis reveals that they are very dif- ferent. Not only is one clearly derived and the other probably primitive, but there are also differences in number of traces, branching pattern, and 266 JOURNAL OF THE ARNOLD ARBORETUM yet: 201 4 regularity of veinlets. This once more demonstrates the necessity for a detailed study of the totality of the venation in angiospermous leaves, as has been repeatedly emphasized by Foster (1950, 1952, 1959b). From the morphological and anatomical point of view, the unique over- all pattern of venation in Raoulia is of interest. No other group of species so far investigated shows a similar venation pattern. The question of what determines the particular pattern of a leaf has occupied the attention of various workers, most of whom have tended to correlate “distribution of growth” with venation pattern. The correspond- ing bibliography has been reviewed critically by Foster (1952). The main conclusions which Foster draws from the literature are: 1) “that the ‘distribution of growth’ appears to vary widely in leaves, even between ones which are similar in form at maturity” and 2) “that the course of the main veins of the lamina ‘registers’ more or less faithfully the distribution of the longitudinal and transverse growth which occurred during ontogeny.” Foster’s own studies on Quiina pteridophylla show that this interpretation is rather “naive” and that intensive ontogenetic and comparative studies are needed before any major conclusion can be drawn. A similar conclu- sion is reached by Pray (1954) as a result of his studies on Liriodendron. No juvenile material was available for this study, and the extremely small size of the leaves and their heavy coat of hairs militated against the successful examination of leaf primordia. Also, lack of fresh material and of sufficient dried herbarium material of critical species precluded the use of sections. Nevertheless, the few data obtained seem to indicate that the venation pattern is influenced by — among other causes, undoubtedly — the pattern of growth of the leaf, which is, in turn, determined to some degree by the general organography of the shoot. In Ficures 17-20 are depicted four successively smaller primordia of Raoulia eximia and in Ficures 23-21 similar stages in leaves of R. Petriensis. Growth in width apparently takes place very early, while the primordium is very small (less than 2 mm.). At that time the pattern of the end “reticulum” is laid down and later growth in length has no im- portant effect on it. A full interpretation will have to wait until more physiological and ontogenetical studies have been completed. In brief, the venation types of Raoulia seem to be associated with the habit of the plant and the growth pattern characteristic of it. Simple venation types are correlated with extreme adaptive modification in other structures. A similar conclusion with respect to the node has already been pointed out by Bailey and Nast (1944). It may be significant that the venation pattern of the leaves of Raoulia is similar to that found sometimes in involucral and receptacular bracts of other Compositae. Miller (1938) has made an extensive study of vena- tion patterns in various types of phyllomes. She regards cataphylls, bracts, and perianth members as reduced structures corresponding to the bases of foliage leaves. Her evidence is the overall similarity of venation patterns in the various phyllomes in spite of differences in the nodal anatomy. 1960] SOLBRIG, LEAF VENATION IN RAOULIA 267 Post (1958) finds a similar situation in his investigations on the genera Frasera and Swertia (Gentianaceae). It might therefore be better from the morphological point of view to classify the foliage leaves of Raoulia as bracts, even though they carry the bulk of the photosynthetic activity. This points once more to the difficulty in defining the different types of phyllomes. The venation pattern in the involucral bracts of Raoulia, on the other hand, is similar in all species, consisting of a single vein which runs unbranched the entire length of the bracts (Fic. 27). differences in width due to ieee 25 Ee ee see (Hooker, 1839-43, GH); 26, E. nubigena Beauverd (Von Miiller s.n., GH). Fic. 27. In- volucral bract of Raoulia lutescens Beauverd Were 106, GH), showing mid- vein and sclerenchyma (hatching). Finally, one should mention the probable application of the results of these investigations to the taxonomy of the group. TABLE I shows various characteristics of the species here considered. The genus was divided by Beauverd (1912) into two subgenera, Raoulia (Eu-Raoulia) and Psychrophyton. This division, based mainly on char- acters of the pappus and number of flowers, is substantiated by the results of this study as far as the species here investigated are concerned. Sub- genus Raoulia is composed of compact, semicreeping plants which grow between sea level and 1800 m. altitude. The leaves are lanceolate to oblong- lanceolate in shape. All species but one, R. lutescens, have three traces 268 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI forming the vascular supply to the foliage leaf. The venation is either of the reticulate type or of the semireticulate type, but no species with “open” venation belongs to this subgenus. The pubescence is of the “interwoven” type, with exception of R. glabra which has very few hairs. Most members of subg. Psychrophyton are true cushion plants from TABLE I. Characteristics of species of Raoulia. ALTITUDINAL VENA- RANGE PUBESCENCE Malin TION SPECIES Hasit (meters) LEAF SHAPE TYPE VEINS TYPE Subg. RAouLia Parkii semicreeper 750-1800 lanceolate interwoven 3 I tenuicaulis semicreeper 0-1650 lanceolate interwoven 3 I Monroi semicreeper O-1150 lanceolate interwoven 3 I subsericea semicreeper 300-1600 lanceolate interwoven 5 I glabra semicreeper O-1500 lanceolate semiglabrous 3 I australis semicreeper O-1600 lance-oblong interwoven 3 Ila Petriensis semicreeper 1500 lance-oblong interwoven 3 Ila lutescens semicreeper 300-1800 lance-oblong interwoven 1 IIb Subg. PsycurRoPHYTON bryoides cushion 1200-2100 oblong straight 1 IIb eximia ushion 1350-2000 oblong straight 1 IIc mammillaris cushion 1300-2000 _— oblong straight 1 IIc Buchananii cushion 1200-1600 _— oblong straight 1 IId Govyenii cushion 400-1500 oblong straight 1 IId rubra cushion 1500 oblong traight 1 IId grandiflora tufted 1000-1800 lance-subulate interwoven 3 Ill Hectorii tufted 1200-1800 subulate interwoven 3 II subulata tufted 1200-2000 —_ subulate glabrous 1 III higher altitudes, usually above 1000 m., although R. Goyenii can be found at lower altitudes. R. grandiflora, R. Hectorii and R. subulata form small tufts, rather than appressed cushions. The leaf shape of species in subg. Psychrophyton is either oblong or subulate. The pubescence type is either straight, in the true cushion plants, or interwoven, in R. grandiflora, R. Hectori, and R. subulata. These last three species have a striate venation, while the rest of the species of this subgenus investigated have a semi- reticulate venation with only one trace entering the leaf from the node. Subgenus Psychrophyton was divided by Beauverd (1912) into three sections: Uninerves, those species with only one trace entering the leaf and a nontruncate leaf apex; Truncatae, species with truncate leaf apices, and Trinerves, species with three traces entering the leaf. As can be seen from the data of Taste I, this is an artificial division. A more natural arrange- ment is to divide the subgenus into two groups: one including R. grandi- flora, R. Hectorii and R. subulata, which share the tufted habit, the subulate or subulate-lanceolate leaves, the interwoven type of pubescence, and the striate venation; the other group comprising the cushion plants with oblong leaves, “straight” pubescence, and the semireticulate type of venation. 1960] SOLBRIG, LEAF VENATION IN RAOULIA 269 LITERATURE CITED Arnott, H. J. 1959. Leaf clearings. Turtox News 37: 192-194. BarLey, I. W., and Cuartotre Nast. 1944. The comparative morphology of the Winteraceae. IV. Anatomy of the node and vascularization of the leaf. Jour. Arnold Arb. 25: 215-220. BEAUVERD, M. G. 1910. Le genre Raoulia et trois genres nouveaux des Com- posées-Gnaphaliées de la flore océanienne. Bull. Soc. Bot. Genéve 2: 208- 252 . 1912. Nouvelles recherches sur le genre Raoulia. Ibid. 4: 41-55. CHEESEMAN, T. F. 1925. Manual of the New Zealand Flora. Ed. 2. 1163 pp. Wellington, N. Z. Foster, A. S. 1952. Foliar venation in angiosperms from an ontogenetic stand- point. Am. Jour. Bot. 39: 752-766. 1955. Comparative morphology of the foliar sclereids in Boronella Baill. Jour. Arnold Arb. 36: 181-198. 1959a. The Phylogenetic significance of Spgs venation in angiosperms. Proc. IX Int. Bot. Congr. 2: 119, 1959b. The procohclocical and taxonomic ees of dichotomous venation in Kingdonia uniflora Balfour F. et W. W. Smith. Notes Bot. Gard. Edinb. 23: 1-12 METCALFE, C. R., and CHALK, L. 1950. Anatomy of the dicotyledons. 2 vols. Clarendon Press, Oxford. Miier, EpirH. 1944. Die Nervatur der Nieder- und Hochblatter. Bot. Arch. 45: 1-92. Post, D. 1958. Studies in the Gentianaceae I. Nodal Anatomy of Frasera and Swertia perennis. Bot. Gaz. 120: 1-14. Pray, T. 1954. Foliar venation in angiosperms. I. Mature venation of Lirio- dendron. Am. Jour. Bot. 41: 663-670. pas W. 1938. Vergleichende Morphologie der héheren Pflanzen. 1(2): 1044-1094. Berlin. GRAY HERBARIUM, HARVARD UNIVERSITY. 270 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI THE GENERA OF MYRTACEAE IN THE SOUTHEASTERN UNITED STATES * KENNETH A. WILSON MYRTACEAE Jussieu (Myrtle FAMIty) Shrubs or trees with simple, opposite or alternate, coriaceous, glandular- punctate, exstipulate leaves, often with a continuous intramarginal vein. Inflorescence cymose, racemose, or paniculate or the flowers solitary. Flowers complete, regular, 4- or 5-merous. Calyx 4- or 5-lobed, distinct, or undivided in the bud and splitting irregularly in anthesis, or falling off as a calyptra. Petals 4 or 5 (attached at the apex of the hypanthium) or wanting. Stamens numerous, free or united into 5 bundles, the versatile anthers with gland-tipped connectives. Style and stigma 1, the ovary inferior, (1)2—5-locular, each locule with 2—many ovules. Fruit a 1—-many- seeded berry or a many-seeded capsule. Embryo variously shaped. A family of about 80 genera and 3000 species, primarily of the tropics of both hemispheres. The glandular-punctate leaves, inferior ovary, and numerous stamens usually are distinctive. Two subfamilies based primarily on the nature of the fruit generally are recognized: the Leptospermoideae, centered in Australia, are characterized by dry, dehiscent fruits; the Myrtoideae (with the single tribe, Myrteae) have fleshy, indehiscent fruits and occur in the tropics of both Old and New Worlds. The latter is divided into three subtribes based on the structure of the embryo. Subtribe Myrciinae Berg (cotyledons foliaceous, twisted and folded; the radicle elongate) is represented in our area by Calyptranthes. Subtribe Eugeniiae Berg (cotyledons fleshy, distinct, partly or completely fused, or conferruminate (closely adherent) ; the radicle very short) includes Eugenia and Myrcianthes, as well as the large Old World genus Syzygium. Sub- tribe Pimentinae Berg (cotyledons very short; embryo spiral or curved; radicle elongate) is represented in our flora by Myrtus, Psidium, and hodomyrtus. +Prepared for a biologically oriented generic aes of the vemos United States, a joint project of the Arnold Arboretum and the Gray Herbarium made possible through the support of George R. Cooley and fe ‘National Science Foundation, and under the direction of Reed C. Rollins te Carroll E. Wood, Jr. The scheme follows that outlined at the beginning of the series (Jour. Arnold Arb. 39: 296-346. 1958). e 1959, 1960. I am grateful to Rogers McVaugh who generously offered many com- ments and suggestions in connection with this family. I am also indebted to Leonard J. Brass, George R. Cooley, and Richard A. Howard for re aid in obtaining speci- mens and their help in various other ways 1960] WILSON, GENERA OF MYRTACEAE 271 Generic distinctions are difficult in the family, and, as a result, man different classifications have been proposed. Great emphasis has recently been placed on the structure of the embryo in determining various taxonomic groupings. Evidence from anatomy and pollen morphology has been used in connection with generic problems. The inflorescence of the Myrtaceae is very variable and has been inter- preted as resulting from various modifications of a potentially floriferous branch which develops in the axil of a leaf. The majority of the Myrtaceae which have been studied have a diploid chromosome number o Representatives of at least Eucalyptus L’Her., Callistemon R. Br., Syzygium Gaertn., Feijoa Berg, and Myrciaria Berg are cultivated in Florida as ornamentals or for their fruits. REFERENCES: ANDREWS, E. C. The development of the natural order Myrtaceae. Proc. Linn. Soc. New S. Wales 38: 529-568. 1913. [On the origin and distribution of the Myrtaceae. | ATCHISON, E. Chromosome numbers in the Myrtaceae. Am. Jour. Bot. 34: 159-164. 1947. [Includes all previously recorded chromosome numbers. | BARANOV, P. A. Coleorrhiza in Myrtaceae. Phytomorphology 7: 237-243. 1957 [1958] Berry, E. W. The origin and distribution of the family Myrtaceae. Bot. Gaz. 59: 484-490. 1915. Broc1i, B. Beitrage zur Anatomie der Myrtaceen Rinden. Thesis, Basel, 84 p. 1926.* BUSWELL, W. N. Florida myrtles. Am. Bot. 35: 138-143. 1929. [General de- scriptions and notes on the cultivated and native species, with some com- ments on their economic value. IncLe, H. D., and H. E. DApSwELL. The anatomy of the timbers of the south- west Pacific area. III. Myrtaceae. Austral. Jour. Bot. 1: 353-401. 1953. [Includes some American species. KAUSEL, E. Beitrag zur Systematik der Myrtaceen. Ark. Bot, II. 3: 491-516. 1956. [Proposes a classification of the seas on the basis of characters of the seed and embryo, and describes 5 new gene Myrtaceae. Inst. Parans Catal. Gén. 28: a “1957. [List of genera and synonyms of the Myrtaceae, excluding the “Leptospermaceae.’ McVaucuH, R. Nomenclatural notes on Myrtaceae and related families. Taxon 5: 133-135, 162-167. 1956. . Tropical American Myrtaceae. Notes on generic concepts and descrip- tions of previously unrecognized species. Fieldiana Bot. 29: 145-228. 1956. [Includes a study of the inflorescence. . Myrtaceae. Jz Flora of Peru. Field Mus. Nat. Hist. Bot. 13(4°): 569-818. 1958. [The most recent revisionary study. | Mauvritzon, J. Contributions to the embryology of the orders Rosales and Myrtales. Lunds Univ. Arsskr. II. Sect. 2. 35(2): 1-120. 1939.* MENNINGER, E. A. The cultivated eugenias in American gardens. Natl. Hort. Mag. 38: 92-104, 145-164. apes [Species of Syzygium, Acmena, Eugenia, Myrceugenia, Myrciaria, Pimenta, Myrcianthes, and Myrtus, both native and introduced; includes sone photographs. | ote JOURNAL OF THE ARNOLD ARBORETUM [VOL, XLI Mowry, H., L. R. Toy, and H. S. WotFe. Miscellaneous tropical and subtropi- cal Florida fruits. Revised by G. D. Ruehle. Fla. Agr. Ext. Serv. Bull. 156A: 1-116. 1958. [Includes species of Eugenia, Feijoa, Myrciaria, Psidium, Rhodomyrtus, and Syzygium. | NIEDENZU, F. Myrtaceae. Nat. Pflanzenfam. III. 7: 57-105. 1893. Prxe, K. M. Pollen morphology of Myrtaceae from the south-west Pacific area. Austral. Jour. Bot. 4: 13-53. 1956. [Includes some American species. | SMITH-WHITE, S. Cytological studies in the Myrtaceae. II. Chromosome num- bers in the Leptospermoideae and Myrtoideae. Proc. Linn. Soc. New S. Wales 73: 16-36. 1948. SOUBIHE SOBRINHO, J., and J. T. A. GurceL. Characteristics of the seeds of fruit-producing ener Revista Agr. Piracicaba 27: 83-90. 1952.* [In paaries ] Polyembryony and ab ec se embryony in citrus, Mangifera and fruit- bearing Myrtaceae. IJ. Dusenia 4: 421-428. 1953.* [In Portuguese. | Sturrock, D. Tropical fruits for edie Florida and Cuba and their uses. Publ. Atkins Inst. Arnold Arb, 1: 1-131. 1940. [ Myrtaceae, 81-92.]| WEBERLING, F. Untersuchungen iiber rudimentare Stiplen bei den Myrtales. Flora 143: 201-218. 1956. KEY TO THE GENERA OF MYRTACEAE > . Leaves alternate; inflorescence spicate; stamens united into 5 bundles; fruit capsular; bark white, spongy, peeling off in sheets. .......... 1 Mela euca. . Leaves opposite, indo seance paniculate, cymose, racemose or flowers glom- erate or solitary; stamens free; fruit a fleshy 1—-many-seeded berry. . Inflorescence a panicle; calyx undivided in bud, circumscissile, the lid usually ae attached at one side; petals absent; embryo with foli- aceous, twisted and folded cotyledons and an elongate ra jae . ce pada au eily eek ue as eh a ede he ain 4 ere ee ee 2 Calyptranthes. . Inflorescence a raceme or dichasium or the flowers solitary or glomerate; calyx-lobes, if developed, 4 or 5; petals white or rarely rose colored. C. Inflorescence racemose or the flowers glomerate or fascicled; calyx- lo 4; fruit 1(rarely 2)-seeded; embryo with fleshy, fused cotyle- spon ea eeeeeee eae Sa we adnan GE ee ee arenes 3. Eugenia. C. aie a dichasium or the flowers solitary. Calyx-lobes 4, distinct in the bud, petals 4, white, leaves pinnately veined. E. Flowers in dichasia, embryo with e distinct, fleshy cotyledons, radicle short; fruit 1(rarely 2)-seeded. 4, rcianthes. E. Flowers Sonar: embryo curved, i. cotyledons He radicle te; fruit many-seeded, .......64.0544 254: 55 rtus. D. Calyx-lobes 5, distinct in the bud, or the calyx closed in the bud and splitting irregularly j in anthesis, usually into 5 segments; petals > we iss) F. Petals white; leaves pinnately veined, calyx closed in the bud, splitting irregularly Itt ANEMERIS) 525444 4442c00nees 6. Psidium. F. Petals rose colored; leaves 3- veined calyx-lobes distinct in the DUES 2 ee nee ea eee he est ead yhoo 7. Rhodomyrtus. 1960] WILSON, GENERA OF MYRTACEAE 278 Subfam. LEPTOSPERMOIDEAE Niedenzu Tribe LEpTOSPERMEAE DC. 1. Melaleuca Linnaeus, Mant. Pl. Gen. 1: 14. 1767, nom. cons. Tree with alternate, 1-many-veined leaves. Flowers sessile in the axils of bracts, in dense or elongate spikes or occasionally solitary, the axis of the spike growing into a leafy shoot during or after flowering. Calyx-lobes 5, deciduous. Petals 5, white, orbicular, spreading in anthesis. Stamens numerous, united in 5 bundles opposite the petals. Ovary 3-locular, each locule with numerous ovules; hypanthium extending beyond the ovary. Fruit a loculicidal capsule dehiscing at the top, crowned by the hypan- thium. (Cajuputi Adans., 1763, nom. rejic.) Type species: M. Leucaden- dron (L.) L. (Name from Greek, melas, black, and leukos, white, alluding to the black trunk and white branches of one of the species.) — BoTTLeE- BRUSH. The subfamily Leptospermoideae, which is characterized primarily by the capsular fruits, has its center of distribution in Australia, and is repre- sented in our area by a single naturalized species, Melaleuca Leucadendron (L.) L. The bottle-brush occurs in southern Florida where it is both cultivated and naturalized. This tree is conspicuous because of its white, spongy bark which peels off in large sections. The chromosome number of three other Old World species has been determined as 2n = 22 The closely related genus Callistemon R. Br. is frequently cultivated in Florida and may be distinguished by its free stamens which are bright red in color. Various species of Eucalyptus L’Her. in the same subfamily are culti- vated in the warmer parts of our area. Although some of \the species have become naturalized in California, apparently none has become established in Florida. REFERENCES: ScHory, E. A. The cajaput tree in Florida. Carib. Forest. 19: 50-55. 1958. [M. Leucadendron, culture and uses. | SHARMA, V. N., and P. SincH. Preliminary chemical examination of the liquid exudate from abnormal growths in Melaleuca leucadendron Linn. Jour. Sci. Indus. Res. 15C(11): 256. 1956.* Subfam. MYRTOIDEAE Niedenzu Tribe MyrtTEAE DC. Subtribe Myrciinae Berg 2. Calyptranthes Swartz, Prodr. Veg. Ind. Occ. 79. 1788, nom. cons. Trees or shrubs with opposite leaves. Inflorescence a myrcioid panicle (e.g., with the principal branches opposite, elongate, and terminating in 274 JOURNAL OF THE ARNOLD ARBORETUM [VOI SoL5 single flowers or simple dichasia), the inflorescence axis usually abortive above the first node and the panicles appearing to be paired. Calyx un- divided in bud, circumscissile in anthesis but usually remaining attached at one side. Petals wanting. Stamens numerous, inserted on the margin of the hypanthium which is prolonged above the summit of the ovary; anthers versatile, splitting longitudinally. Ovary 2(seldom 3)-locular, each locule with 2 ovules. Fruit a 1- or 2(seldom 3)-seeded berry crowned by the hypanthium. Embryo with twisted and folded foliaceous cotyledons, and an elongate radicle. (Chytraculia P. Br., 1756, nom. rejic.; CAvytralia Adans., 1763, nom. rejic.) Typr species: C. Chytraculia (L.) Sw. (Name from Greek, kalyptra, veil, and anthos, flower, alluding to the lidlike dehiscence of the calyx.) — SpICEWooD, MYRTLE-OF-THE-RIVER. About 100 species native to tropical America and the West Indies; represented in our area by two species, Calyptranthes Zuzygium (L.) Sw. and C. pallens (Poir.) Griseb., both of which occur in southern Florida, as well as Cuba and other islands of the West Indies. Because of the hood- like dehiscence of the calyx, Calyptranthes is one of the easiest genera of the American Myrtaceae to recognize. However, many of the species are so variable in size and shape of leaves and in the characters of the flowers that they are by no means well marked. This genus, as well as all others in the family, is in need of a thorough revision which should be accom- panied by extensive field study. Subtribe Eugeniinae Berg 3. Eugenia Linnaeus, Sp. Pl. 1: 470. 1753; Gen. Pl. ed. 5. 211. 1754. Trees or shrubs with opposite leaves. Inflorescence racemose, the termi- nal flower of the axis usually wanting; axis sometimes extremely shortened, the inflorescences then resembling axillary fascicles, umbels, or glomerules [or flowers rarely solitary in the lowermost axils of otherwise leafy branch- lets]. Petals 4, orbicular, ovate or obovate, white, spreading in anthesis. Stamens numerous, free, the anthers versatile, splitting longitudinally. Ovary 2-locular, each locule with numerous ovules, the hypanthium ex- tending slightly beyond the ovary or not at all. Fruit a 1 (rarely 2)-seeded berry crowned by the persistent lobes of the calyx. Embryo apparently undivided, with thick, fleshy, fused cotyledons. Typr sprectes: EF. uniflora L., 2n = 22. (Named in honor of Prince Eugene of Savoy, 1663-1736, a patron of botany and horticulture.) — STopPeEr. A tropical genus of about 500 species represented in southern Florida by five indigenous species: Eugenia anthera Small, endemic, and E. myrtoides Poir. (FE. buxifolia (Sw.) Willd.), E. axillaris (Sw.) Willd., E. rhombea (Berg) Krug & Urban, and E. confusa DC., which occur also in the West Indies. Although some of the species of Eugenia are clear cut and easily recognizable, the great majority are ill defined and difficult to determine, being based on leaf or floral characters which are very variable. 1960] WILSON, GENERA OF MYRTACEAE 275 Eugenia has at various times included almost all species of the Eugeniinae, and the inclusive genus has been variously divided into smaller groups on the basis of the structure of the inflorescence and of the embryo, However, authors are by no means in agreement. Evidence from pollen morphology, wood anatomy, and anatomy of the bark supports the conclusion that the New World Eugenias are distinct from those of the Old World which are treated by the majority of recent authors as Syzygium Gaertn. Eugenia is said to differ from Syzygium in the pseudomonocotyledonous, apparently undivided embryo, and in the smooth seed coat which is free from the pericarp; the cotyledons of Syzygium are separate and the seed coat roughish, loosely or closely adhering to the pericarp. Among the Old World species which are cultivated in Florida are Syzygium Jambos (L.) Alston, the rose apple (2n = 28, c. 42, 46, c. 54); S. Cumini (L.) Skeels, the Java plum or jambolan (2n — 44,46); and S. malaccensis (L.) Merr. & Perry, the Malay apple (2n = 22). Eugenia Dombeyi Skeels, E. Luschnathiana Klotzsch ex Berg, E. uniflora L., Surinam cherry, and other species are know also in cultivation in Florida; the first two, however, are only sparingly grown, while E. uniflora is very popular both as an ornamental and for its edible fruits. REFERENCES: See family references, McVaucu (1956, Tropical American Myrtaceae, pp. 166- 169). Cuattaway, M. M. The anatomy of bark. VII. Species of Eugenia (sens. lat.). Trop. Woods 111: 1-14. figs. 1-10 [2 pls.]. [Supports the separation of the Old World eugenias from those of the New World.] DapsweELL, H. E., and H. D. INGLE. The wood anatomy of the Myrtaceae, I. A note on the genera Eugenia, Syzygium, Acmena, and Cleistocalyx. Trop. Woods 90: 1-7. pls. 1, 2. 1947. GAGNEPAIN, F. Classification des Eugenia. Bull. Soc. Bot. Fr. 64: 94-103. 1917. HENDERSON, M. R. The genus Eugenia (Myrtaceae) in Malaya. Gard. Bull. Singapore 12: 1-293. 1949. [Rejects the removal of Syzygium from Eugenia. | JavAWeEERA, D. M. A. Variation in the flower of Eugenia malaccensis Linn. Jour. Linn. Soc. Bot. 55: 721-728. 1957. [A study based on teratologies. | MerriLt, E. D., and L. M. Perry. The Myrtaceous genus Syzygium Gaertner in Borneo. Mem. Am. Acad. Arts Sci. 18: 135-202. 1939. [Includes notes on American species of Eugenia. | Prjt, L. vAN DER. Uber die Polyembryonie bei Eugenia. Rec. Trav. Bot. Néerl. 31: 113-187. 1934. Tiwary, N. K. On the occurrence of polyembryony in the genus Hugenia. Jour. Indian Bot. Soc. 5: 124-136. 1926. 4. Myrcianthes Berg, Linnaea 27: 315. 1854 [1856]. Trees or shrubs with opposite leaves, Inflorescence an axillary dichasium, the terminal (central) flowers usually sessile in the fork. Calyx 4-lobed [rarely 5-lobed], the lobes distinct, persistent. Petals 4, white, spreading in anthesis. Stamens numerous. Ovary 2-locular, each locule with numer- 276 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI ous ovules; hypanthium not extending beyond the summit of the ovary. Fruit a 1(seldom 2)-seeded berry crowned by the persistent lobes of the calyx. Embryo with distinct, fleshy cotyledons, the plumule shorter than the radicle. (Including Anamomis Griseb.) Type species: M. apiculata Berg (see McVaugh, Taxon 5: 143. 1956.) (Name from Myrcia, and Greek, anthos, flower, in reference to the resemblance of the flowers to those of the genus Myrcia.) — NAKEDWOOD. A genus of perhaps 50 or more species ranging from southern peninsular Florida to the West Indies, southward to Bolivia and Argentina, chiefly along the Andes, and southern Brazil. The genus is characterized by having solitary flowers or flowers in simple or compound dichasia, the hypanthium not extending beyond the ovary, the multiovulate ovary, and the embryo, as far as known, with two distinct cotyledons. Myrcianthes dicrana = occurs in hammocks along both coasts of southern peninsular Florida and on Key West and M. Simpsonii * in hammocks along the lower east coast of Florida and the Florida Keys. The latter differs in its larger flowers, many-flowered cymes, and the greater number of stamens. Myrcianthes has been said to differ from Anamomis in the 5-parted flowers and in the presence of a plumule in the embryo. In a study of the Peruvian Myrtaceae, McVaugh found no support for separating the two genera, and he reported that 4-merous species occasionally have 5-merous flowers, and that a plumule is present in all mature seeds of species of Anamomis which were examined. In other morphological details the two genera are very much alike. REFERENCES: See a family references, McVaucH (1956, Tropical American Myrtaceae, 9, 170; 1958, pp. 745-757.) SMALL, 7 K. The genus Anamomis in Florida. Torreya 17: 221-224. 1917 [1918]. Subtribe Pimentinae Berg 5. Myrtus Linnaeus, Sp. Pl. 1: 471. 1753; Gen. Pl. ed. 5, 212. 1754. Shrubs or small trees with opposite leaves. Flowers solitary, peduncled, in the axils of leaves or of bracts or reduced leaves at the lowermost nodes of an otherwise leafy branch. Calyx 4|or 5]-lobed, the lobes distinct, persistent. Petals 4, white, spreading in anthesis, Stamens numerous. Ovary 2- or 3-locular, each locule with numerous ovules; hypanthium not extending beyond the summit of the ovary. Fruit a many-seeded berry crowned by the persistent lobes of the calyx. Seed coat bony; embryo °Myrcianthes dicrana (Berg) K. A. Wilson, comb. nov. Eugenia dicrana Berg, Linnaea 27: 259. 1854 [1856], Anamomis dicrana (Berg) Britton, N. Am. Trees 728. rcianthes Simpsonii (Small) K. A. Wilson, comb nov. Anamomis Simpsonii Small, Torreya 17: 222. 1917 [1918] 1960] WILSON, GENERA OF MYRTACEAE Deh uncinate-curved, the cotyledons short and inconspicuous. (Mosiera Small.) Type species: M. communis L. (2n = 22). (Mvyrtus, the Latin name, from the Greek name, myrtos.) — SToppER (cf. Small) A genus of about 16 species, according to the most recent interpretation (Burret), with one species in Europe, one in Africa, and about 14 in Florida and the West Indies; in our area represented by one or perhaps two species. Myrtus verrucosa Berg (Mosiera longipes (Berg) Small) and M. bahamensis (Kiaersk.) Urban (Mosiera bahamensis (Kiaersk.) Small) occur in southern peninsular Florida and on the Florida Keys in pinelands and in hammocks. It is questionable that the two, which differ in size of flowers and to some extent in the habit of the plants, are distinct; if united, the correct name is M. verrucosa. The genus has been variously interpreted and has been considered to include as many as 50 species, most of which are now regarded as belong- ing to other genera. A study of Myrtus is badly needed and should take into account the American groups which have been allied with it. Such a study should establish the limits of the genus, and indicate whether our species rightfully belong to it. The nature of the inflorescence, fruit, seed, and embryo, clearly indicate that our species are not allied with Eugenia, and, in fact, belong in a different subtribe. REFERENCES: See also family references, McVauGH (1956, Tropical American Myrtaceae, pp. 173, 174. Burret, M. Myrtaceen-Studien. Notizbl. Bot. Gart. Berlin 15: 479-550. 1941. [Retains 2 Florida species in Myrtus. | SouEGES, R. Embryogénie des myrtacées. Développement de l’embryon chez le Myrtus communis L. Compt. Rend. Acad. Sci. Paris 210: 548-550. 1940. 6. Psidium Linnaeus, Sp. Pl. 1: 470. 1753; Gen. Pl. ed. 5. 211. 1754. Trees or shrubs with opposite, pinnately veined leaves. Flowers axillary and solitary or in 3-flowered dichasia. Calyx undivided in bud, splitting irregularly in anthesis, remaining attached to hypanthium. Petals 4 or 5, white, spreading. Stamens numerous, free, inserted on the hypanthium. Ovary inferior, usually 4- or 5-locular, each locule with numerous ovules. Fruit a many-seeded berry, crowned by the persistent segments of the calyx. Embryo curved, cotyledons short, radicle elongate. TypE SPECIES: P. Guajava L. (Name from Greek, psidion, pomegranate [Punica Granatum |.) — GUAVA, A genus of about 150 species of tropical and subtropical America. Psidium Guajava L., 2n = 22, originally introduced and cultivated for its edible fruits, has become naturalized in southern Florida as well as in many other warm parts of the world. The round or pear-shaped, yellow fruit is highly prized and is eaten raw or used to make jellies, preserves, or beverages. Many varieties have arisen in cultivation, but little taxonomic 278 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI work has been done on them. In view of the industry already built around the production of the guava, and its potentially even greater economic significance, a detailed horticultural and systematic study of this species and its varieties would be highly desirable. Several other species of Psidium, among them P. Cattleianum Sabine, the strawberry guava, 2” = 88, are in cultivation in Florida. REFERENCES: Kumar, L. S. S., and S. G. RANAbDE. Autotriploidy i 0 guava (Psidium Guajava, Linn.). Cut Sci. Bangalore 21: 75, 76. 1952. Lyncu, S. J., and R. O. NELSon. Current i en of vegetative ee of pies en lychee and guava in Florida. Ceiba 4: 315-337. 1956. Mustarp, M. Mangos and guavas as sources of ascorbic acid. an Fla. State Hort. Soc. 58: 187-190. 1946.* Rueute, G. D. Promising new guava varieties. Proc. Fla. State Hort. Soc. 59: 127-131. 1946.* Growing guavas in Florida, Fla. Sub-trop. Exp. Sta. Mimeog. Rep. 12. Homestead, 1947.* SCHROEDER, C. A. Priority of the species Psidium aga He Se Jour. Arnold Arb. 27: 314, 315. 1946. [P. ee Vee Pe WappincrTon, G., and F. M. Cist. The vitamin C content of Psidium Guajava. Proc. Fla. State Hort. Soc. 55: 110-112. 1943." 7. Rhodomyrtus (DC.) Reichenbach, Nomen. Gen. Pl. 177. 1841. Shrubs [or trees] with opposite, tomentose, 3-veined leaves. Flowers axillary and solitary or in 3-flowered dichasia. Calyx-lobes 5 [or 4], spreading. Petals 5 [or 4], rose colored, spreading. Stamens numerous, free, inserted on the hypanthium; filaments rose colored. Ovary [1-—]3- ocular with 2 rows of ovules in each locule separated by a vertical septum, thus appearing 6-locular; hypanthium not at all or only slightly prolonged beyond the ovary. Fruit a blue-black many-seeded berry crowned by the persistent lobes of the calyx. Embryo curved. TyPE sPEcIEs: KR. tomen- tosa (Ait.) Hassk. (Name from Greek, rkodon, rose, and myrtos, ancient Greek name for myrtle.) — DoOWNyY-MYRTLE, HILL-GOOSEBERRY. A small genus of about 25 species native to eastern Asia and Australia. Rhodomyrtus tomentosa (Ait.) Hassk. has escaped from cultivation and is becoming naturalized in southern Florida, especially along the Gulf Coast. The three-veined, opposite leaves, pubescent on the lower surface, are characteristic. It is cultivated as an ornamental shrub, and also for its small, globose fruits which are used in making jams and pies. REFERENCES: hao act H. On the cuticles of some recent and fossil Myrtaceae. Jour Soc. Bot. 48: 657-671. 1931. [Compares species of Rhodomyrtus ee Tristania. - JAYAWEERA, D. A. The morphology of the flower of Rhodomyrtus tomen- tosa, Wight. oe Jour. Sci, A, 13: 31-42. 1956. 1960] WEBSTER, WEST INDIAN PHYLLANTHUS 279 SUPPLEMENT TO A MONOGRAPHIC STUDY OF THE WEST INDIAN SPECIES OF PHYLLANTHUS Grapy L. WEBSTER WITH THE AID OF A GRANT from the National Science Foundation, the work on the West Indian species of Phyllanthus (Euphorbiaceae) earlier published in this journal (1956-1958) is now being expanded by use of various analytical techniques such as chromosome counts. During the summer of 1959, an extended tour through the West Indies made it possible to obtain more information on populations of a number of taxa. Certain species, especially those in sect. Xylophylla, will be treated in a future study dealing with intraspecific variation patterns. However, it seems appropriate to present here the observations that have accumulated on other species, as well as to note a few corrections. The numbers in brackets refer to the volume and page numbers of the monographic study in previous issues of this journal. Herbarium collections are those in the Stanley Coulter Herbarium, Purdue University (PuR), unless otherwise noted. [38: 51.] Sect. 2. Floribundi. 4. Phyllanthus tenellus Roxb. Add the following locality record: Puerto Rico: Bosque Estatal de Maricao, Hoconuco trail, weedy cleared area, Webster et al. 8901. This is the first record of this weedy species from Puerto Rico; it was not found elsewhere on the island. It may be expected to reach Cuba and Hispaniola, but thus far no specimens have been observed from either island. [38: 56.] Sect. 3. Kirganelia. According to Article 22 of the International Code of Botanical Nonien- clature (1956 ed.), Anisonema cannot be retained as the epithet of this section, since it contains the type species of subgenus Kirganelia. The type species, when the section is given the epithet Kirganelia, must be P. casticum Willem. instead of P. reticulatus. This ruling is unfortunate, in my opinion, since it fallaciously assumes a parallel between typification of a genus or species with that of a subgenus. The subgenus does not appear in the name of a species and is a purely intercalary category, as is the section. That the new rule does not really contribute to stability is shown in the present case. If subgenera are recognized in Phyllanthus, the epithet of this section must be Kirganelia; but if one chooses not to recognize subgenera then the same section must take the epithet Anisonema. 280 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI [38: 297.] Sect. 11. Phyllanthus. The synoptic key to taxa of sect. Phyllanthus requires two corrections: (1) the calyx lobes of P. amarus (no. 20) could often be described as apiculate, rather than acute; they are, in any event, distinctly more pointed than those in P. debdilis or P. fraternus. (2) A misprint has oc- curred in lead 5 under the second lead 3. It should read as follows: “5S. Style-branches not unusually modified; branchlets mostly with 15-30 leaves... .” [38:299.] Sect. 11. Phyllanthus, subsect. a. Niruri. 17. Phyllanthus mimicus Webster. The following additional col- lection has been recorded: Trinipap: without specific locality, Fendler 680 ex p. (pm). This collection resembles P. niruri in having only 20-25 leaves per branchlet rather than 35-45 as in the type collection of P. mimicus. However, the leaves are mostly only 4.5-5.5 mm. long, the male flowers are solitary and with calyx lobes 0.6-0.7 mm. in length, and the capsule is only c. 2 mm. broad. Since in all these respects Fendler’s col- lection agrees with P. mimicus, the breakdown in the leaves-per-branchlet character does not efface the distinctions between P. mimicus and P. niruri. Further collections of P. mimicus are highly desirable, and it was quite disappointing that the plant was not located on the expedition trip o Tobago. (38: 306.] Sect. 11. Phyllanthus, subsect. b. Swartziani. 18. Phyllanthus debilis Klein ex Willd. This species was recollected at Petit Bourg, Guadeloupe, after a 20-year interval (Webster et al. 8986), and, in fact, was locally common. It may therefore be considered as estab- lished, at least on Guadeloupe. So far, however, it has not been found on any other island. 21. Phyllanthus stipulatus (Raf.) Webster. Add the following locality record: St. Lucta: along trail, red clay soil, southeast slope of Piton Flor, Webster et al. 9291. This collection of P. stipulatus, the first recorded from St. Lucia, was interesting because of its small seed size. The mean of 30 seeds measures only 0.95 mm., and the range is from 0.93-1 mm. This is much smaller than measurements of a collection from Puerto Rico and two from Trinidad (Webster et al. 8954, 9888, 9924), which yielded mean seed lengths of 1.14, 1.06, and 1.12 mm. respectively. A collection of P. caribaeus (Webster et al. 9677) gave a mean seed length of only 0.9 mm. and a range of 0.85-0.95 mm. The St. Lucia collection is thus intermediate between the two species in seed size, but is closer to P. caribaeus and smaller than any collections of P. stipulatus seen else- where in Latin America. Nevertheless, in floral characters the St. Lucia collection conforms closely to P. stipulatus. It is possible that we are dealing here with an instance of introgressive hybridization. Although P. caribaeus has not been collected on St. Lucia, it might be expected to be 1960] WEBSTER, WEST INDIAN PHYLLANTHUS 281 there, for it occurs on both Dominica and St. Vincent. Further field studies of this interesting species-pair on all the central Lesser Antilles are most desirable. [38: 363.] Sect. 13. Conami. As with sect. Anisonema (see above), the epithet of this section must be changed from Nothoclema to Conami in order to conform with the current rules of nomenclature. [39: 50.] Sect. 15. Botryanthus. For the same reason, sect. Elutanthos must be replaced by sect. Botryan- thus. 45a. Phyllanthus nutans Sw. ssp. nutans. Additional field studies of this widespread Jamaican plant have only confirmed its previous sys- tematic disposition. Better material obtained of the population in the John Crow Mountains (Webster e¢ al. 8297) shows decisively that it can- not be given specific status. This population from the Ecclesdown area is distinguished by persistent stipules, tenuous pendent fruiting thyrses, and particularly by the large seeds. In this it resembles the population in the Cockpit country, as may be seen from the following table. No. SEEDS COLLECTION LOCALITY MEASURED MEAN Ss Howard et al. 14769 (A) John Crow Mts. 8 6.29mm. 0.42 Webster et al, 8297 « = a 15 6.22 0.25 8402 Cockpit Country 15 6.11 0.18 8477 i 5 6.83 0.26 8575 Seacoast near Lucea 4 4.39 0.16 The number of seeds available is small, so that a statistically satisfactory analysis cannot yet be made, but it would appear that populations of P. nutans at higher altitudes and in areas of greater rainfall tend to have larger seeds than those near the seacoast. A collection from Discovery Bay (Hunnewell 18844 [cH]) had seeds only about 4.1 mm. long, but the number of measurements was unfortunately not recorded. Collections from intermediate situations may have intermediate seed sizes (e.g., 5.29 mm. in Webster & Proctor 5516 from Big Level). Such variation as this in the Jamaican populations of P. nutans is perhaps correlated with ecotypic differences. It seems possible that in dried, open habitats there is a selection for smaller seed size than in moister, closed ones, as suggested by Salisbury in “The Reproductive Capacity of Plants” (1942). [39: 153.] Sect. 21. Epistylium. Field studies in Jamaica have clarified to some extent the differences in 282 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI inflorescence between the three species of this section. Although Phyl- lanthus cauliflorus could not be located, collections of P. cladanthus (Web- ster et al. 8279) and P. axillaris (Webster et al. 8464, 8472) were most helpful toward a better understanding of the morphology of these species. It now appears that the inflorescence difference is more absolute than was previously mentioned. In P. cladanthus the flowers are strictly cauliflorous, while in P. axillaris they are strictly axillary. Only in P. cauliflorus do both types of distribution occur. Each of the three species, therefore, has a distinct pattern of flower production. [39: 179.] Sect. 24. Xylophylla. Extensive collections of members of this group were made in Jamaica, Puerto Rico, and Guadeloupe. However, the analysis of several of the bi species (such as P. arbuscula and P. epiphyllanthus) is still not complete ee ! KK? | Bw. a Fic. 1. Vegetative branch of Phyllanthus eximius. 1960] WEBSTER, WEST INDIAN PHYLLANTHUS 283 and must be deferred to a separate paper. At this time the only addition to record is the following previously undescribed species. During a botanical trip to Jamaica in the summer of 1954, Mr. George Proctor guided Dr. Kenneth Wilson and the author on a hike into the little-explored John Crow range in the eastern end of the island. About halfway up the rugged slopes of this precipitous limestone massif we discovered in the rain forest what was obviously an undescribed species of Phyllanthus sect. Xylophylla. Unfortunately the plants were entirely barren, and publication of the species had to be deferred until flowering material could be obtained. During a return trip to the original locality in the summer of 1959, we were fortunate to encounter a considerable number of flowering individuals in the cloud-swept scrubby forest on the Fic. 2. Male flower and female flower of Phyllanthus eximius. upper edge of the plateau at an altitude of somewhat over 2000 feet. Plants previously sent to the Harvard University greenhouses by Mr. Proctor also came into flower in 1959, so that there is no longer a shortage of fertile specimens. 84. Phyllanthus eximius Webster & Proctor, sp. nov. Frutex vel arbuscula monoecia glabra 2-4 m. alta; phyllocladiis a plicibus rigidis spathulatis vel oblanceolatis 8-17.5 cm. longis, 1-4.5 ¢ latis, costa venisque conspicuis, marginibus ora cinctis. Floribus Ses vel rufidulis: flore masculo laciniis calycis plerumque 6, c. 1.1—1 longis, staminibus 3 filamentis ad basin connatis; flore femineo pedicello crasso c. 1.3—2.2 mm. longo, 0.4-0.6 mm. diametro, laciniis calycis 6, c. 1—1.2 mm. longis; disco inconspicuo; stylis dilatatis subliberis c. 1 mm. longis, c. 4-6-lobis; capsula c. 3.3-3.7 mm. lata. 284 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI Glabrous shrub or small tree becoming c. 2-4 m. high; branches of current year’s growth 3-7 mm. thick, reddish brown or often greyish or glaucous, smooth. Apical scaly cone irregular in outline, 8-12 mm. long, 5-10 mm. broad. Cataphylls persistent on current year’s growth but sub- sequently deciduous, pale (stramineous or light brown), indurate; stipules and blade free from each other, lanceolate, acuminate (sometimes at- tenuately so), mostly 3-6 mm. long, tips erect or squarrose. Phylloclades (modified branchlets) unbranched, becoming somewhat rigid, spathulate to oblanceolate, c. (8—) 10-15(-17.5) cm. long, 1-4.5 cm. broad, obtuse or rounded to emarginate at the tip, with (25—)35-55 conspicuously notched nodes, gradually tapering to a petiole-like base c. 1-2 cm. long; midrib prominent on both sides, lateral veins straight, conspicuous, each ending at a marginal notch; margins with distinctly differentiated rim running between the notches. Foliage leaves (euphylls) not seen; cataphylls of phylloclades blackish brown, trifid, the tips soon breaking off, less than 1 mm. long. Monoecious; cymules mostly bisexual, each with one female and several male flowers; flowers yellowish- or greenish-white, or reddish-tinged. Male flower: pedicel slender, c. 1-2 mm. long. Calyx-lobes 6 (rarely 5), thin, more or less unequal, spreading, elliptic to obovate, the inner larger ones c. 1.1-1.6 mm. long and 0.9-1.3 mm. broad, the outer smaller ones c. 0.8-1.1 mm. long and 0.6-1 mm. broad; lobes obtuse or rounded (occa- sional ones apiculate) at the tip, midrib unbranched. Disk-segments 6, not very massive, roundish or ellipsoidal, mostly 0.3—0.5 mm. long. Sta- mens 3; filaments 0.5—0.8 mm. high, united only at the very base (up to 0.1-0.2 mm.) into a massive pedestal, ascending-spreading, the anthers horizontal or deflexed; anthers emarginate, 0.3-0.4 mm. long, 0.4—0.6 mm. broad. Female flower: pedicel stout, straight, in fruit becoming 1.3—-2.2 mm. long and 0.40.6 mm. thick. Calyx-lobes 6, stiff (subcoriaceous), ascend- ing or spreading, unequal, c. 1-1.2 mm. long and nearly as broad, tips acute or obtuse, midrib not conspicuous. Disk inconspicuous, annular, not enclosing any part of the ovary. Ovary smooth; styles c. 1 mm. long, free except at the very base, spreading at anthesis (later more or less ascending), dilated distally, the style-tips cut into 4-6 slender subterete obes. Capsule oblate, rounded-trigonous, c. 3.3-3.7 mm. long; cocci c. 2.5 mm. long, reddish brown, not veiny; well-developed seeds not observed. TYPE COLLECTION: Jamaica, Portland Parish, John Crow Mountains, 2-2.5 miles SW of Ecclesdown, rain forest on top of plateau, alt. c. 2000-2500 ie 17 June 1959, Webster, Miller, Ellis, & Proctor 8320 (Holotype, PuR; isotypes to be distributed). An additional collection from the same locality but a lower altitude (c. 1500 ft.) is Webster & Wilson 5140 (PUR, GH). This striking and attractive species (whence the epithet eximius, extraordinary) appears to be strictly endemic to the rain forest on the middle and upper slopes of the John Crow Mountains. Ecologically, 1960] WEBSTER, WEST INDIAN PHYLLANTHUS 285 the plant belongs to that sizable class of endemic species of the John Crows adapted to growing on limestone rock almost bare of soil and under nearly continuous rainfall. The two collections cited are similar in most respects and are surely conspecific, but the payocues of the type collection are strikingly narrower than those of Webster & Wilson 5140, as appears from the following table. N ER OF COLLECTION SPECIMENS RANGE MEAN ny Phylloclade length 5140 25 88-173 mm. 125.8mm. 40.0 8320 50 81-165 mm. 130.4 mm. 25.1 Phylloclade width 5140 25 20-45 mm. 29.3 mm. 6.3 8320 50 11-25 mm. 18.7 mm. 4.2 Number of nodes 5140 25 25-56 43.4 8.1 8320 50 35-55 45.2 8.7 In order to accommodate PAyllanthus eximius in the treatment of sect. Xylophylla (Jour. Arnold Arb. 39. 1958), the key to the species of that section (p. 182) need be modified only at the last lead as follows: Phylloclades ca representing an entire branchlet, borne scattered directly _ ) 6. Cataph ils on main stem persistent, with stp UleS) ae blade fused (at least basally); phylloclades with indistinct lateral veins. .............. ee Re iy cp ik gd ee hoe RUSE AGRE P. epiphyllanthus. 6. Cataphylls on main stem more or less deciduous, ee a blade free; lateral veins of phylloclades conspicuous. .............. ha Cxtmius. Although it keys out next to Phyllanthus epiphyllanthus because of its unbranched phylloclades, P. eximius is much more closely related to P. arbuscula and probably represents a highly specialized local population derived from that species. The resemblance to P. arbuscula is especially marked in the long, pale cataphylls forming an irregular apical cone and in the prominent lateral veins and distinct marginal rims of the phylloclades. However, in addition to having simple rather than compound phylloclades, P. eximius also differs from P. arbuscula in having a shorter, thicker fruit- ing pedicel and in the lesser degree of fusion of filaments in the male flower The description of Phyllanthus eximius rounds out the preliminary descriptive work in sect. Xylophylla; it is almost certain that all the species in the group have now received names. There remains much work of in- terest to do, however, in analyzing the taxonomic complexities of three species (P. angustifolius, P. arbuscula, and P. epiphyllanthus). From an evolutionary point of view P. eximius is most interesting as a demonstration that the reduction of compound to simple phylloclades has occurred twice and independently within the section, the simple phylloclades of P. 286 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI epiphyllanthus representing a reduction from those of P. angustifolius, while the simple phylloclades of P. eximius have been derived from com- pound ones such as those of P. arbuscula. Field work during the summer of 1959 was supported by a grant from the National Science Foundation. Aid in obtaining collections was given by Mr. George Proctor on Jamaica; Mr. Roy Woodbury, of Rio Piedras, on Puerto Rico; Dr. Henri Stehlé on Guadeloupe; Mr. Peter Paul on St. Lucia; and the staff of the Imperial College of Tropical Agriculture (in- cluding P. T. Richards, W. D. Richardson, and N. W. Simmonds) on Trinidad. The illustrations of Phyllanthus eximius (Fics. 1 and 2) were drawn by Barbara Webster from the author’s preliminary sketches. DEPARTMENT OF BIOLOGICAL SCIENCES PURDUE UNIVERSITY West LAFAYETTE, INDIANA 1960] TOMLINSON, ANATOMY OF PHENAKOSPERMUM 287 THE ANATOMY OF PHENAKOSPERMUM (MUSACEAE) P. B. TomMiLinson IN A RECENT ACCOUNT of the anatomy of the Musaceae (Tomlinson, 1959) the anatomy of Ravenala is described. Although this dealt with only one species, R. madagascariensis, it is there implied that Phenakospermum (Ravenala) guianense (L. C. Rich.) Mig. falls within the range of anatomical variation of Ravenala. The anatomy of Phenakospermum has, owever, never been described in detail. Recently, thanks to the kindness of Dr. P. Campos-Porto of the Botanic Gardens, Rio de Janeiro, I received pickled material of parts of a mature plant, together with material of a juvenile plant, of Phenakospermum guianense. On first examination of this material it was immediately evi- dent that there were significant differences between Phenakospermum and Ravenala. More recently it has been possible to examine this material in detail, and the present account is based on these anatomical observations. From these results a comparison between Phenakospermum and other members of the Musaceae, particularly Ravenala and Strelitzia, has been made from the standpoint of systematic anatomy. Reasons for accepting Phenakospermum Endl. ex Miq. as a valid genus distinct from Ravenala Adans., within which it was formerly included, are given by Lane (1955). He also discusses the nomenclatural problems of Phenakospermum, and, since they are there satisfactorily resolved, it is not necessary to repeat them in the present paper. The chief diagnostic features which are considered by Lane to be characteristic of Phenako- spermum are the terminal inflorescence, five stamens in each flower, four or more rows of ovules in each loculus of the ovary, the bright red-orange, filamentous aril and straight embryo with the micropylar opening at the end of the seed. Ravenala, on the other hand, has a lateral inflorescence, six stamens in each flower, only two rows of ovules in each loculus of the ovary, while the aril is blue and membranaceous and the embryo is bent so that the micropylar opening is in the side of the seed. In addition, in the material of Phenakospermum which I received, there were short stolons covered by overlapping, distichous scales, evidently a means by which the plant could spread vegetatively. Ravenala also propagates itself by growth of basal suckers, but these are never at first stoloniferous. Apparently there are two distinct forms of the plant, the most obvious difference being that one has an erect, ligneous stem, the other having only an underground stem; small differences in the reproductive parts have also been recorded. Nakai (1948) has separated the form with a distinct trunk into a new genus, Musidendron. Lane suggests that the differences are not sufficiently constant to warrant this change. 288 JOURNAL OF THE ARNOLD ARBORETUM [¥OLs: Sit The same methods used in studying the anatomy of other members of the Musaceae were employed (Tomlinson, 1959). In sectioning the stem on the sliding microtome it was found necessary to soften the hard outer sclerotic layer by directing a jet of steam onto the cut surface during the sectioning process. ANATOMY OF THE VEGETATIVE PARTS Lamina dorsiventral. Hares absent. CuTicte forming a thick, abaxial, waxy layer, cracking and eroding from mature leaves although persisting as irregular tabular masses, often forming annular ridges above stomata; outer cutinized layer of epidermal cells rather thin. Epipermis with markedly sinuous anticlinal walls. Adaxial epidermis very uniform, outer wall thickened, made up of files of tabular cells; cells in surface view (Fig. 1) either square, or rectangular and longitudinally extended although end walls often rather oblique. Abaxial epidermis mostly with inner wall thicker than delicate, wholly cutinized outer wall; made up of narrow costal bands of shallow cells, costal cells more or less square in surface view (Fig. 2) and each often including a small silica body; alternating with wider intercostal bands of larger, usually longitudinally extended but rather irregular cells, those in the same file as stomata sometimes wider than cells elsewhere. StomatTa occasional adaxially; abaxially absent below both longitudinal and transverse veins and so restricted to short rectangular areas; not in regular files but 3-5 irregular series situated within each intercostal band. Terminal subsidiary cells usually short, wide but not otherwise different from other epidermal cells, Lateral subsidiary cells nar- row, deep, uniformly thin-walled; other epidermal cells sometimes modified adjacent to stomata but not uniformly arranged. Guard cells (Fig. 11) not sunken, each with 2 cutinized ledges. HypopEermis (Fig. 3) 2—3-layered adaxially, often 4-layered above main veins, l-layered abaxially. Outer- most adaxial hypodermal layer rather thick-walled, cells hexagonal and slightly transversely extended in surface view; inner layers composed of larger and more cubical cells with thinner walls. Abaxial hypodermal layers often interrupted by vascular buttresses, cells always smaller than outermost adaxial layer, most irregular in intercostal regions; substomatal chambers each surrounded by an annulus of 2—4 small cells. CHLOREN- cHyYMA including 2 or 3 adaxial layers of palisade cells, uppermost layer most anticlinally extended. Abaxial mesophyll divided into rectangular compartments between longitudinal and transverse veins, including rather loose, isodiametric but shortly lobed cells. Vetns (Figs. 3, 5) all attached to exch hypodermis by well-developed fibrous buttresses: adaxial but- tresses narrow except for massive buttresses above few main veins, com- posed of unlignified fibers with wide lumina; abaxial buttresses made up of narrow lignified fibers with narrow lumina. Outer parenchymatous bundle sheath 1-layered lateral to all veins, composed of cubical or slightly longitudinally extended cells. Extended protoxylem only present in larger 1960] TOMLINSON, ANATOMY OF PHENAKOSPERMUM 289 Fics. 1-5. ANATOMICAL DETAILS OF THE LEAF OF PHENAKOSPERMUM. l, Adaxial epidermis of lamina, surface view, < 300. 2, Abaxial epidermis of lamina, surface view, including part of a costal band, & 300. 3, Transverse section of lamina, including a main vein, * 54; sclerenchyma solid black, phloem stippled, xylem lined. 4, Longitudinal section of lamina including a transverse vein in transverse section, * 125. 5, Transverse section of lamina, including a small longitudinal vein, * 145; s.b. = silica bodies. 290 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI veins, vascular tissues rather inconspicuous in smaller veins. Probable free vein extensions, pectinating with longitudinal veins, visible as occa- sional fibrous strands below adaxial hypodermis near main veins. TRANS- VERSE COMMISSURES (Fig. 4) frequent, all buttressed to either abaxial hypodermis or epidermis by short, unlignified, thick-walled prosenchyma- tous elements or short fibers with wide lumina; adaxial sheath usually con- sisting of a single layer of colorless parenchyma in contact with the palisade; vascular tissues reduced, xylem often consisting of a single file of tracheal elements, phloem incompletely sheathed by lignified elements. ExPANsION CELLS represented by well-developed hypodermal layers above main veins (Fig. 3). Indistinct “pulvinar band” also present at junction of midrib and lamina, veins in this region with very tall and well-developed adaxial buttresses. Leaf axis (lamina midrib |Fig. 6], petiole [Fig. 7], leaf sheath). EPIDERMIS varying in structure in different parts: (i) Adaxial epidermis of midrib and upper part of leaf sheath: cells cubical or slightly longitu- dinally extended, outer walls distinctly thickened, anticlinal walls often slightly sinuous in midrib; rarely containing silica bodies. (ii) Adaxial epidermis of leaf base (Fig. 8): cells thin-walled, rectangular and con- siderably longitudinally extended, anticlinal walls not sinuous. (iii) Abaxial epidermis of midrib and leaf sheath together with whole of epi- dermis of terete part of petiole (Fig 12): cells more or less cubical, inner wall becoming markedly thickened and often enveloping a spherical silica body, especially in leaf sheath (Fig. 9). Sometimes slight differentiation between costal and intercostal regions. Sromata occasional on both surfaces, guard cells and lateral subsidiary cells like those in stomata of lamina, arrangement of other subsidiary cells varying in different parts of leaf axis: (i) Adaxial epidermis of midrib and upper part of leaf sheath: arrangement more or less as in lamina. (ii) Adaxial epidermis of leaf base (Fig. 8): terminal subsidiary cells short, thin-walled. (iii) Abaxial epidermis of leaf sheath and whole of epidermis of terete part of petiole: terminal subsidiary cells with uniformly thickened and conspicuously pitted walls, similar cells also usually adjacent to lateral subsidiary cells (Figs. 10, 12). Hypoprrmis usually conspicuously different from ground paren- chyma cells of leaf axis: (i) Abaxial hypodermis in all parts consisting of 2 to many layers of small, short cells with massively thickened, pitted lignified walls, interrupted by small, loose, lobed cells below stomata but these cells also becoming sclerotic in old leaves (Fig. 10). (i) Adaxial hypodermis in midrib and upper part of leaf sheath including at least 1 regular layer of stone cells either immediately below epidermis (leaf sheath) or separated from epidermis by a thin-walled layer (midrib). (iii) Adaxial hypodermis of leaf base always with thin-walled, markedly tangentially extended cells, CHLORENCHYMA present as an inconspicuous abaxial band. Arr LACUNAE in one to many series extending from leaf insertion to distal part of midrib (Fig. 6). Most conspicuous in petiole 1960] TOMLINSON, ANATOMY OF PHENAKOSPERMUM 291 (Fig. 7) as several arcs of lacunae abaxial to main vascular arc, separated from each other by vertical partitions; represented vy a single arc of air canals in leaf sheath; small system of air canals also adaxial to main vascular arc in petiole. Air lacunae segmented at regular intervals by thick TRANSVERSE SEPTA, each septum consisting of 2-3 layers of large, fairly compact, colorless parenchyma cells on either side of a central plate of 2 or 3 layers of small, stellate cells, the arms of which include wide inter- cellular spaces. Air lacunae in the midrib often containing a loose net- work of algiform cells. Abaxial rFIBROUS STRANDS abundant in all parts of leaf axis, together with the sclerotic abaxial hypodermis forming a pe- ripheral rigid mechanical zone. Fibrous bundles pectinating with abaxial vascular bundles, usually separated from epidermis by a sclerotic hypo- peeray s.b at : EF aa a) ee) ~(2| Fics. 6-12. ANATOMICAL DETAILS OF THE LEAF OF PHENAKOSPERMUM. 6, Transverse section of midrib, with only main arc of vascular bundles shown, stoma, * 430. , Transverse section of stoma from abaxial epidermis of lamina, X 430. 12, Epidermis of petiole in surface view, X 300. 292 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XLI dermis but in petiole often immediately adjacent to epidermis. VEINS arranged in 3 distinct systems: (i) Main V- or U-shaped arc of large vascular bundles mostly adaxial to main system of air lacunae (Figs. 6, 7). (ii) A very feeble system either of narrow adaxial bundles with re- duced, often obliquely or inversely orientated vascular tissues or of purely fibrous strands, most conspicuous in petiole. (iii) Abaxial system of numerous, mostly medium-sized vascular bundles below main arc, bundles uniformly scattered and not arranged in obvious arcs, frequent in longi- tudinal partitions between air lacunae, bundles in this position in midrib often inversely orientated. Large vascular bundles with normal Scitaminean construction (cf. Solereder and Meyer, 1930; Tomlinson, 1956); phloem parenchyma often thick-walled and lignified. TRANSVERSE COMMISSURES connecting longitudinal veins at irregular intervals, sometimes extending across transverse partitions of air lacunae. GROUND PARENCHYMA uniform, often including tannin cells. Stem. Narrow cortex abruptly delimited from central cylinder by a dense sclerotic zone of congested vascular bundles. EpmperMis as in abaxial part of leaf sheath with files of cubical cells, each with inner wall thickened and enclosing a spherical silica body. Sromara occasional, each surrounded by sclerotic subsidiary cells as in the leaf sheath. HypopERMIS usually including 2 or 3 (locally many more) markedly sclerotic layers. Cortex including 3 or 4 innermost irregular series of large vascular bundles together with smaller bundles in outer cortex. Each bundle with a well- developed fibrous sheath external to the phloem; xylem sheathed by parenchyma, usually including only one wide metaxylem vessel and with- out extended protoxylem. CENTRAL CYLINDER with a peripheral zone of congested vascular bundles separated from each other by narrow bands of sclerotic parenchyma. Vascular tissues of peripheral bundles reduced to a single xylem element and a narrow phloem strand and completely sheathed by a wide fibrous cylinder. Inner limit of peripheral sclerotic zone showing a sharp transition to central region of central cylinder with its irregularly scattered bundles. Largest vascular bundles (apparently re- cently entered leaf traces), each including many wide metaxylem vessels, conspicuously extended protoxylem and a small phloem strand; fibrous sheath narrow and inconspicuous. Other vascular bundles (traces well below their insertion into a leaf), each with a single wide metaxylem vessel, no extended protoxylem and a complete, fibrous sheath. GROUND PAREN- CHYMA of inner part of central cylinder spongy and enclosing large inter- cellular spaces, individual cells rounded but not lobed, including abundant starch. Stolons. Notable features include a wide cortex containing many scattered vascular bundles, the largest of which are either bicollateral or amphivasal, each surrounded by a complete fibrous sheath; a narrow cen- tral cylinder delimited from the cortex by a discontinuous endodermis of 1960] TOMLINSON, ANATOMY OF PHENAKOSPERMUM 293 cells with U-shaped thickenings; peripheral tissues of central cylinder immediately within endodermis consisting of a congested region of irregu- lar vascular bundles. Central vascular bundles less congested, usually collateral. Root. (Structure described possibly not typical since only underground roots from juvenile plants and aerial roots from base of mature stems available.) (i) Underground roots: piliferous layer sometimes persistent, but usually shrivelled in old roots. Exopermis of narrow compact thick- walled cells with truncate ends, often separated from piliferous layer by 2 or 3 thin-walled layers. Cells of inner corTEX very uniform; air lacunae absent. ENDOpDERMIS of mature roots with U-shaped wall thickenings; pericycle thin-walled, mostly 1-layered. STELE as in Ravenala and Strelitzia with wide vessels scattered throughout the central tissues to- gether with scattered phloem strands each containing a wide sieve tube. (ii) Aerial roots: normal polyarch monocotyledonous root structure. Silica, crystals, tannin, starch. Srzica certs. (i) “False stegmata”’ with thin walls present in discontinuous files, each cell including a stellate, druselike silica body (Fig. 5, s.b.); common in leaf and stem adjacent to vascular bundles; in lamina common adjacent to fibrous vascular buttresses and in cells of outer parenchymatous sheath; also common adjacent to transverse commissures. In leaf axis common adjacent to vascular bundles but absent from vicinity of abaxial fibrous strands; also frequent adjacent to transverse veins. (ii) Epidermal cells of stem, abaxial epidermis of leaf axis (Figs. 9, 10) and sometimes costal abaxial epidermal cells of lamina (Fig. 2, s.b.) developed as “true” stegmata, each with a spherical, spinulose silica body becoming enveloped by thickenings of inner tangen- tial wall of cell; most obvious in leaf base. Crystars. Calcium oxalate abundant as small rhombohedral crystals usually in clusters in ground parenchyma, common in assimilating cells of leaf, especially abundant in cells of transverse diaphragms of air lacunae. Raphide-sacs frequent in all parts but especially abundant in ground parenchyma of stem, common on surface of transverse diaphragms in air lacunae. TANNiN. Abundant in all parts of ground parenchyma in unmodified cells. STARCH. Abundant in ground parenchyma of stem as large, more or less ellipsoidal but not flattened or markedly eccentric grains; small grains often common in cells adjacent to vascular bundles of leaf axis. Vascular tissues. VESSELS present only in root and stem; elements in underground roots (see root above) of the order of 1500 » but up to 5000 y. long, 160-200 » wide with either simple or scalariform perforation plates with few thickening bars on slightly oblique end walls. Elements of aerial roots shorter, narrower, never with simple perforation plates. Elements in stem 2000-2500 p (or as short as 1250 ») long, 125-160 » wide, with oblique or very oblique scalariform perforation plates with many thick- 294 JOURNAL OF THE ARNOLD ARBORETUM [VOu. 201 ening bars. Vessels absent from stolons and leaf. SIEVE TUBES apparently with simple sieve plates in stem and leaf, DISCUSSION THE ANATOMY OF PHENAKOSPERMUM IN RELATION To Its SYSTEMATIC PosITION Phenakospermum in relation to the Musaceae as a whole. From the information recorded here and in the previous publication (Tomlinson, 1959) it is evident that Phenakospermum shares most of the features by which Strelitzia and Ravenala can be distinguished from other members of the Musaceae. The following combination of anatomical features is com- mon to these three genera: Terminal subsidiary cells short and distinct from other epidermal cells; other cells adjacent to stomata sometimes different from remaining epi- dermal cells; adaxial hypodermis multiseriate (up to. six layers deep); abaxial hypodermal cells arranged concentrically around substomatal chambers; longitudinal veins of lamina all buttressed to each hypodermis; all transverse septa of lamina including vascular tissues; petiole including several arcs of air canals; transverse diaphragms of air canals each several cells thick; chlorenchyma of leaf axis situated in an abaxial band; stem often an erect woody trunk without an endodermis; subterranean roots with an anomalous stele including uniformly scattered metaxylem vessels and phloem strands, the latter each including a single, wide sieve tube: internal silica cells thin-walled, each including a stellate, druselike silica body; silica cells often situated adjacent to transverse as well as longitu- dinal veins. On the basis of this close anatomical similarity it is evident that these three genera form a very distinct and natural subunit within the Musaceae. This unit corresponds to the Strelitziaceae of Nakai (1948) but not that of Hutchinson (1959) which also includes Heliconia, Phenakospermum, however, is of peculiar interest because it possesses a number of features which are not found in Ravenala and Strelitzia but which may be present in other members of the Musaceae. These features are indicated below together with the names of other genera of the Musa- ceae, in parentheses, which also possess them: Stomata not sunken (Heliconia, Musa, Orchidantha); anticlinal epi- dermal walls of lamina markedly sinuous (Heliconia); inner wall of abaxial epidermal cells of lamina thicker than outer; transverse septa of lamina attached to abaxial hypodermis by well-developed, mostly sclerotic buttresses (the corresponding buttresses in Ravenala and Strelitzia are never markedly sclerotic; in Musa the transverse septa only occasionally include sclerotic elements); Jateral parenchyma of veins of lamina always one-layered (Orchidantha) ; leaf axis including fibrous and vascular bun- dles adaxial to main vascular arc (Heliconia, Musa, Orchidantha) ; starch 1960] TOMLINSON, ANATOMY OF PHENAKOSPERMUM 295 grains neither as flattened as in Ravenala nor as spherical as in Strelitzia ; stegmata, i.e. silica cells with unequally thickened walls, common in abaxial epidermis of leaf (such cells have been recorded and illustrated for Ravenala by Solereder and Meyer but they have never been observed by me in either Strelitzia or Ravenala) « vessels in the root sometimes bearing simple perforation plates (Musa). Lane (1955) has emphasized that Phenakospermum is sufficiently dis- tinct from Ravenala to make its acceptance as a separate genus very de- sirable. The above information supports this contention, since the anatomi- cal differences between Strelitzia and Ravenala, both of which are accepted as good genera, are not greater than the structural differences between Phenakospermum on the one hand and Strelitzia and Ravenala on the other. Indeed, it is anatomically easier to recognize Phenakospermum than it is to distinguish Ravenala from Strelitzia. On the whole Phenako- spermum seems to be something of a connecting link between Ravenala and Strelitzia and the rest of the Musaceae. The relation between Phenakospermum, Strelitzia, and Ravenala. Lane has suggested that Phenakospermum has more affinity with Strelitzia than with Ravenala even though it was originally thought to be congeneric with Ravenala, Anatomical information which might throw light on the interrelationships between these three genera is presented below. Phenakospermum and Ravenala share the following features which are not found in Strelitzia: Wax on abaxial surface of the lamina tabular; abaxial hypodermis of leaf axis markedly sclerotic in most regions; sub- sidiary cells adjacent to stomata of abaxial surface of the leaf axis sclerotic, especially in Phenakospermum (occasionally so in Strelitzia according to Solereder and Meyer); periderm apparently never developed. In Strelitgia the wax on the abaxial surface of the lamina is rodlike, the abaxial hypo- dermis of the leaf axis is apparently never sclerotic, and root and stem frequently develop a specialized periderm. Phenakospermum and Strelitzia share the following features which are not developed by Ravenala: Abaxial hypodermis mostly one-layered; only two arcs of air canals in the petiole; air lacunae apparently absent from the underground roots; vessels present in the aerial stems. Ravenala on the other hand has a consistently two-layered abaxial hypodermis, several arcs of air canals in the petiole which may be a result of its great size, air lacunae in the cortex of the root and no vessels in the stem, The last observation, however, is open to doubt and needs confirma- ion. Even if these differences are constant, they are not sufficiently large to be of taxonomic significance, although it should be noted that the anatomical resemblances between Phenakospermum and Ravenala are to be found in features which are likely to be plastic and influenced considerably by environmental conditions and so may not be a reliable indication of phylo- genetic affinity. On the basis of this assumption it could be said that 296 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI Phenakospermum shows more qualitative anatomical resemblance to Strelitzia than to Ravenala. POSSIBLE PHYLOGENETIC IMPLICATIONS FROM THE ANATOMICAL OBSERVATIONS In some respects Phenakospermum forms a connecting link between Strelitzia and Ravenala on the one hand and the rest of the Musaceae on the other, since, as has been shown above, many of the features in which it differs fron Serelinnio and Ravenala are to be found in the remaining mem- bers of the Musaceae. In this respect it is particularly noteworthy that sinuous epidermal walls have previously only been recorded in Heliconia. This is of significance as it makes it easy to speculate on the possible evo- lution of the whole of the Musaceae from some ancestor with a Phenako- spermum-like habit. Ravenala and Strelitzia would be least modified descendants, Heliconia and particularly Orchidantha would be most modi- fied. It may also be significant that the anatomy of the stolons of Phenakospermum closely resembles that of the rhizomes of these herba- ceous types. Thus by elimination of the erect woody trunk and persistence of the stolons it would be possible to produce rhizomatous plants of the Heliconia type. The fleshy corm of Musa and Ensete could have evolved from a woody trunk by telescoping of its internodes. The derivation of the Musaceae from an ancestor with a caulescent habit is implied by Lane in his paper, and he also suggests how the inflorescences of members of the family may be various modifications of the type shown by Ravenala. Thus it is possible to envisage the ancestral type from which the Musaceae have evolved as possessing the growth habit of Phenakospermum with the lateral inflorescence of Ravenala. SUMMARY The anatomy of the vegetative parts of Phenakospermum guianense is described for the first time. The systematic implications of these observa- tions are that: (a) Phenakospermum is a valid genus; (b) together with Ravenala and Strelitzia it forms a natural unit within the Musaceae: (c) it may possibly be more closely allied to Strelitzia than to Ravenala. It is suggested that the Musaceae have evolved from an ancestor with the vegetative habit of Phenakospermum and the inflorescence of Ravenala. BIBLIOGRAPHY Hutcuinson, J. The families of flowering plants. Vol. II. Monocotyledons. 2nd ed. Oxford. 1959. Lane, I. E. Genera and generic relationships in Musaceae. Mitt. Bot. Staats- samm. Munchen 13: 114-141. 1955. Nakal, T. A new attempt to the classification of the Strelitziaceae. Bull. Tokyo Sci. Mus, 22: 19-24. 1948. 1960] TOMLINSON, ANATOMY OF PHENAKOSPERMUM 297 SoLEREDER, H., and F. J. Meyer. Musaceae. Jn Systematische Anatomie der Monokotsiedonen, Heft 6: 1-26. 1930. eee P. B. Studies in the systematic anatomy of the Zingiberaceae. r. Linn. Soc. Bot. 55: 547-592. 1956. ie anatomical approach to the classification of the Musaceae. /bid. 779- 809. 1959, DEPARTMENT OF BOTANY UNIVERSITY OF LEEDS 298 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI THE GENERA OF CONVOLVULACEAE IN THE SOUTHEASTERN UNITED STATES ! KENNETH A. WILSON CONVOLVULACEAE Jussieu (MorninG-GLoRY FAMILY) Annual or perennial herbs or woody plants, climbing or prostrate, rarely erect [or shrubby (more rarely trees) |] without tendrils, the stem usually twining and with milky sap. Leaves alternate, simple, without stipules, rarely reduced to scales (Cuscuta). Flowers solitary or in cymes sub- tended by two opposite or subopposite bracts. Sepals 5 [rarely 4] free (rarely united at the base), imbricate. Corolla sympetalous, entire, 5- lobed or 5-parted rrarely 4], induplicate-valvate or contorted in bud. Stamens 5 (rarely fewer), inserted on the corolla tube, alternate with the corolla lobes; pollen spherical and spinulose, or spherical and smooth, or ellipsoid with parallel folds. Gynoecium syncarpous, the style solitary and simple or forked, or styles 2 (rarely 3 or 4), the ovary superior, 2- or 1- locular or at times subdivided by false partitions into 4 locules, rarely 3- locular. Ovules erect, anatropous, sessile, 2 in each locule (rarely solitary). Fruit a capsule, 2- or 3-locular with 1 or 2 seeds in each locule, usually dehiscing by valves or circumscissile, irregularly dehiscent, or indehiscent. Embryo straight, cotyledons folded or spirally coiled, rarely small or absent A world-wide family, primarily of the tropics and subtropics, but ex- tending also into the temperate regions, and best represented in Asia and the Americas. Characterized by the presence of milky sap, bicollateral vascular bundles, plaited corollas, axile placentation, basal erect ovules, and folded cotyledons, the family consists of about 50 genera and 2000 species and is represented in our area by 10 genera. On the basis of the structure of the pollen grains Hallier divided the family into two groups, the Psiloconiae with smooth pollen and_ the Echinoconiae with spinulose pollen (rank not designated) and into nine *Prepared for a biologically oriented generic flora of the southeastern United States, a joint project of the Gray Herbarium and the Arnold Arboretum made possi- and includes North Carolina, Tennessee, Arkansas, and Louisiana. Material included in descriptions which is inapplicable to our species is placed in brackets. References which have not been checked are marked by an asterisk. 1960] WILSON, GENERA OF CONVOLVULACEAE 299 tribes. Van Ooststroom has recently divided the family into two sub- families (Cuscutoideae and Convolvuloideae), three tribes, and nine sub- tribes which are equivalent to Hallier’s tribes. Generic lines are difficult to draw in this family, and treatments vary with different authors depending upon the emphasis placed on the taxonomic characters used (e.g., pollen, bracts, corolla, style, stigmas, and fruit). Hallier’s studies on the members of this family form the basis for the sys- tem of classification that is most generally followed. This system has found support in the studies of Van Ooststroom and of Meeuse, based on the Malaysian and the South African species respectively. Other studies (e.g., Roberty), discount the value of Hallier’s conclusions but add little new information to our knowledge of the family. It is evident that the entire family is in need of intensive study, and that all characters must be thoroughly re-evaluated. Both flowering and fruiting material is necessary for the correct deter- mination of most of the genera of Convolvulaceae. It is in general not possible to determine species which are sterile. Collectors should therefore make every effort to collect material in both flowering and fruiting condi- tion, to record the shape of the corolla and its color; and to press opened corollas carefully between waxed paper or facial or other similar tissue. REFERENCES: Avrarp, H. A. The direction of twist of the corolla in the bud, ane twining of the stems in Convolvulaceae and Dioscoreaceae. Castanea 12: 88-94. 1947 BarLton, H. Convolvulacées. Hist. Pl. 10: 305-331. 1891. BentTuHAM, G., and J. D. Hooker. Convolvulaceae. Gen. Pl. 2: 865-881. 1876. Cuoisy, J. en DC. Prodr. 9: 323-462. 1845. HALLIER, H. Bausteine zu einer Monographie der Convolvulaceae. Bull. Herb. Boiss. 5: 366-387, 736-754, 996-1013, 1021-1052. 1897; 6: 714-724. 1898; 7: 408-418. 1899. ————. Versuch einer natiirlichen Gliederung der Convolvulaceen auf mor- BS eee und anatomischer Grundlage. Bot Jahrb. 16: 453-591. 1893. ———. Convolvulaceae africanae. Zbid. 18: 81-160. 1893. Zur Convolvulaceenflora Amerika’s. Jahrb. Hamburg. Wiss. Anst. 16 (Beih. 3): 19-56. 1899. House. H. D. Notes on Convolvulaceae. Muhlenbergia 5: 65-72. 1909. [Notes and new species in Convolvulus, Jacquemontia, Operculina, Ipomoea, and Rivea. Jounston, M. C. Convolvulaceae in southern coastal Texas. Tex. Jour. Sci. 11: 191- oe 1959.* Knicut, R. J. An examination of certain Sven ay trends in the Convol- vulgeeae: Diss. Abs. 19: 1175, 1176, 1958. Characters differentiating common morning-glories occurring in Vir- ginia. Va. Jour. Sci. 10:°63-69. 1959: [Species of Ipomoea and Convol- vulus. | Macsripe, J. F. Flora of Peru. Publ. Field Mus. Bot. 13:57) 1=536,, 1959. [Convolvulaceae, 455-536. | Meeuse, A. D. J. The South African Convolvulaceae. Bothalia 6: 641-792. 300 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI O’DonELL, C. A. Convolvuloideas de Uruguay. Lilloa pa 349-376. 1959. Convolvulaceas argentinas. /bid. 87-348. pls. ee Ooststroom, S. J. vAN. The Convolvulaceae of Malay I. The genera Cuscuta, Dichondra, Evolvulus, Bonamia, Neuropeltis oa Porana. Blumea 3: 62-94. 1938; II. The genera Jacquemontia, Aniseia, Convolvulus, Caly- stegia, Shutereia, Merremia, Operculina and Decalobanthus. Ibid. 267-371. 1939; III. The genus cnoes. Ibid. 481-582. 1940; IV. The genera Mina, Lepistemon, Stictocardia and Argyreia. Ibid. 5: 339-411. 1943; V. Key the genera. /bid. 689-691. 1945. VI. The genus Argyreia in the Philip. pine Islands. /bid. 6: 337-348. 1950. . Convolvulaceae. Jn D. G. G. J. vAN SteENIS. Flora Malesiana I. 4: 388-512. 1953. PARULEKAR, N. K. Stem anatomy of Convolvulaceae. Proc. Indian Sci. Congr. Assoc. 42(3, abs.): 229. 1955.* Peter, A. Convolvulaceae. Nat. Pflanzenfam. IV. 3a: 1-40, 375-377. 1897. Rao, K. V. R. Gametogenesis and embryogeny in five species of the Convol- . Jour. Indian Bot. Soc. 19: 53-69. 1940, [Species of Ipomoea, Evolvulus, Argyreia. | Roperty, G. Genera Convolvulacearum. Candollea 14: 11-65. 1952. [An un- satisfactory new system of classification of the family. ] SAYEEDUD-D1n, M. Observations on the anatomy of some of the Convolvulaceae. Proc. Indian Acad. Sci. B. 37: 106-109. 1953.* SHARMA, A. K., and A. K. Cuatreryi. A cytological investigation of some oa as an aid in understanding their lines of evolution. Phyton Buenos Aires 9: 243-257. 1957. VeRpcourT, B. Notes from the East African Herbarium — VI. Notes on Afri- can Convolvulaceae (Part 2). Kew Bull. 1958: 185-197. 1958. | Merremia, Stictocardia, Astripomoea.|; (Part 3). Ibid. 199-217. |Zpomoea, Tur- bina. | . Some notes on tropical African Convolvulaceae. Webbia 13: 321-330. 1958.* Wotcott, G. B. Chromosome numbers in the Convolvulaceae, Am. Nat. 71: 190-192, 1937. [Undocumented. | KEY TO THE GENERA OF CONVOLVULACEAE A. Leaves scale-like or absent: pet without chlorophyll (parasitic); corolla bearing scales at base of stamens ; embryo coiled, without cotyle oe AR hte fae evi bs ee aL aie aoe as ek oh ds 4, & Uae RaeY wd La xlel a x emia eed Cran A. Leaves well developed; plants green (autotrophic); corolla ae scales; embryo straight or merely curved; cotyledons generally folded. B. Styles 2, gynobasic, corolla deeply 5-lobed, ovary and capsule deeply 2- lobed, the carpels distinct or aed slightly nee creeping a aca at the nodes with minute flowers. .................... . Dichondra. B. Style 1 or 2, terminal; ovary er ls not 2-lobed, the 3 ieee C. Styles 2, free or fused near the bas D. Styles free, each 2-cleft, cen 4, linear-filiform; leaves small, plants never twining. .......................... 3. Evolvulus, D. Styles united near the base, stigmas 2, peltate. ...... 4. Bonamia. C. Style solitary. 1960] WILSON, GENERA OF CONVOLVULACEAE 301 E. Capsule valvate, irregularly dehiscent or the pericarp basally circumscissile. igmas 2. G. Stigmas elliptic or oblong, flattened. 5. Jacquemontia. G. Stigmas filiform or subulate. ........ 6. Convolvulus. F. Stigmas 1, capitate or 2- or 3-lobed. H. Pollen smooth; stigmas 2-lobed, globose; fruit a 4-valved capsule or the pericarp basally circumscissile. ........ 7. Merremia. H. Pollen spinulose; stigma 1, capitate or 2- or 3-lobed, lobose; fruit a 4—6-valved capsule. .... 8. Ipomoea. . Capsule indehiscent, pericarp leathery or woody. I. Ovary 2-locular, fruit woody, mostly 1-seeded; leaves rod Fo 6) cc) a ne) Oe een nem et. 9.. 9. Turbina. T. Ovary 4-locular, fruit leathery, mostly are leaves densely sericeous on the lower surface. .... me gyreia. es) 1. Cuscuta Linnaeus, Sp. Pl. 1: 124. 1753; Gen. Pl. ed. 5. 60. 1754. Parasitic yellowish, orange or reddish (rarely greenish) twining plants with filiform stems and alternate scale-like leaves; stems attached to the host by small haustoria. Flowers in cymose clusters, white. Calyx of 5 (rarely 4) united (rarely free) sepals. Corolla of 5 (rarely 4) petals, imbricate, with fimbriate scales at the base of the tube, as many as and below and opposite the stamens [or rarely lacking]. Pollen ellipsoid, smooth. Gynoecium of 2 (rarely 3) carpels; ovary 2(rarely 3)-locular, each locule with 2 basal, anatropous ovules; styles distinct or united. Fruit an indehiscent, irregularly dehiscent, or circumscissile capsule. Embryo spirally coiled, without cotyledons. Endosperm fleshy. Typr SPECIES: C. europaea L. (Derivation of name uncertain, perhaps from Arabic.) — DoppER (also LOVE VINE, STRANGLE WEED, DEVIL’S GUT, GOLD THREAD, PULL DOWN, HELLBINE, HAIRWEED, DEVIL’S HAIR, HAILWEED). A world-wide genus of over 150 species centered primarily in the Americas where they occur from southern Canada south to Chile and Argentina; about 32 native and 18 introduced species in the United States; represented in our area by at least 19 species in two of the three sub- genera. None of-the species seems to be limited to a single host although certain ones are serious agricultural pests, particularly of clover and alfalfa (C. Epithymum Murray, C. indecora Choisy, and C. Gronovii illd Subgenus GRAMMICA (Loureiro) Engelm. (flowers with distinct styles and globose, peltate or convoluted stigmas) includes most of the native American species and about sixteen of our species. Subgenus MonocyNa (Engelm.) Yuncker (flowers with united styles) is primarily Old World in its distribution and is represented in our area by one native (C. exaltata Engelm.) and one introduced species (C. japonica Choisy). The species of subg. Cuscuta (flowers with distinct styles and elongate cylindrical stig- 302 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI mas) are all native to the Old World, although several of them, including C. Epithymum and C. Epilinum Weihe, have been introduced into North America. Cuscuta Epithymum may perhaps occur in our area on legumi- nous crops. Cuscuta is at times treated as a separate family Cuscutaceae, primarily on the basis of its parasitic habit and imbricate corolla with scales. Embry- ological details, such as the absence of parietal cells in the nucellus, the per- sistence of one of the synergids, and the spirally coiled embryo without cotyledons and without tissue differentiation, have been cited in support of this separation. Both floral structure and embryology seem to be very similar to those of the Convolvulaceae, however. Such differences as do exist support the separate tribe Cuscuteae Benth. & Hook. or perhaps even the subfamily Cuscutoideae Peter. Since the generative cell divides before the pollen is shed, the mature pollen grain is three-celled. All species which have been studied have a monosporic embryo sac of the “Polygonum” type, with the exception of Cuscuta reflexa Roxb., which has a bisporic, “Allium” type embryo sac. The ovule has a single integument which is differentiated in the seed into an epidermis, two layers of palisade-like cells, and several layers of small, thin-walled, starch-filled cells. Keys have been prepared for the identi- fication of the seeds of various species, but the seeds require a considerable amount of sectioning in order to show the layers of the seed-coat. The seed can germinate as soon as it matures, or it may remain viable in the soil for five or more years. The seedling is leafless with a very small root which dries up as soon as the plant becomes established on a suitable host. In order to control and eradicate Cuscuta from agricultural fields it is recommended that it be prevented from seeding during its first season’s appearance. Diploid chromosome numbers of 14, 28, 32, 42, 56, and 60 have been found. Both self-fertile and self- sterile species are known, It is possible, however, that the self-sterility of C. suaveolens Ser. and C. subinclusa Durand & Hilgard is a result of abnormal meiosis, such as has been demonstrated in C. reflexa Roxb. (2n = 28, 32, 42), rather than of self- incompatibility. REFERENCES: (For a more extensive bibliography see DEAN 1936, 1937, and YUNCKER 1921, 1927 BENNETT, C. W. Studies of dodder transmission of plant viruses. Phytopathol- ogy 34: 905-932. 1944. BEzRUCHENKO, N. Z. Biology of dodders. (In Russian) Soviet. Bot. 15: 213- 217. 1947. Bunninc, E., and R. Kauttr. Uber den Chemotropismus der Keimlinge von Cusciin europaea. Biol. Centralb. 75: 356-359. 1956.* Dean, H. L. Dodder overwintering as haustorial tissues within Cuscuta- induced galls. Proc. Iowa Acad. Sci. 61: 99-106. 1954. . Cuscuta as a parasite on pteridophytes. Ann. Bot. 50: 823-825. 1936. 1960] WILSON, GENERA OF CONVOLVULACEAE 303 ————. An addition to bibliographies of the genus Cuscuta. Univ. Iowa Stud. Nat. Hist. 17: nae LOR AOS. FEDORTSCHUK, W. Embryologische Untersuchung von Cuscuta monogyna Vahl und Cuscuta Epithymum L. Planta 14: 94-111. 1931. Friyn, V. V. Vergleichende Embryologie und Karyologie einiger Cuscuta-Arten. (In Ukranian; German summary.) Zhur. Inst. Bot. URSR 12(20): 83- ie 99. 193 FOGELBERG, S. O. The cytology of Cuscuta. Bull. Torrey Bot. Club 65: 631- 645 FritscHe, E., M. BovurLLENNE-WALRAND, and R. BOovILLENNE. Quelques observations sur la biologie de Cuscuta europaea L. Bull. Sci. Acad. Roy. Belg. V. 44: 163-187. 1958. GAERTNER, E. E. Studies of seed germination, seed identification, and host relationships in dodders, Cuscuta spp. Mem. Cornell Univ. Agr. Exp. Sta. 294: 1-56. 1950. Lackey, C. F. Reaction of dodders to stems of other dodders and to their own stems. Phytopathology 36: 386-388. ; Ler, W. O., and F. L. Trmmons. Dodder and its control. U. S. Dep. Agr. Farmers’ Bull. 2117: 1-20. 1958 Loo, S. W. Cultivation of excised stem tips of dodder in vitro. Am. Jour. Bot. 33: 295-300. 1946. MacPuHerson, G. E. Comparison of development in dodder and morning glory. Bot. Gaz. 71: 392-398. 1921. [C. Gronovii and Convolvulus sepium.] NaraAyANA, H. S. Diffuse type of parasitism in Cuscuta hyalina Roth. Sci. Cult. 21: 447-450. 1956.* PazoureEK, J. Anatomical observations of the parasitism of Cuscuta. Preslia 30(2): 121-125. 1958 RaGHAVAN, R. S. Chromosome numbers in Indian medicinal plants. Proc. Indian Acad. Sci. B. 45: 294-298. 1957. [C. reflexa, ] SMITH, B. E. A taxonomic and morphological study of the genus Cuscuta, dod- ders, in North Carolina. Jour. Elisha Mitchell Sci. Soc. 50: 283-302. 1934. SovecgEs, R. A propos de l’embryogénie ue Cuscuta. Bull. Soc. Bot. Fr. 100: 28-34. 1953. ae hyalina Ruth, C. planiflora Tenore, C. reflexa Roxb.] STEVENS, O. A. Dodder (Cuscuta) on eee plants. N. Dak. Agr. Exp. Sta. Bimonthly Bull. 15: 80, 81. 1952.* Tract, B. A contribution to the morphology and embryology of Cuscuta hyalina Roth and C. planiflora Tenore. Phytomorphology 1: 9-21. 1951. TRONCHET, J. Action de l’acide 2.4-dichlorophénoxyacétique a 0,2% sur de jeunes plantules de Cuscuta gronovii Willd: élongation intense, géotropisme inversé et développement d’entre-noeuds. Compt. Rend. Acad. Sci. Paris 244: 1811-1813. 1957. Mise en évidence de l’hydrotropisme des plantules de Cuscuta gronovii Willd. Ibid. 245: 979-981. 1957. lations periodique de masse fraiche des plantules de Cuscuta pronoun Willd cultivées en conditions constantes éventualité d’un rythme endogéne. Bull. Soc. Fr. Physiol. Veg. 4(3): 88, 89. 1958.* Truscott, F. H. On the regeneration of new Bheots from isolated dodder auctor: Am. Jour. Bot. 45: 169-177. 1958. VALENTA, V. A new yellows virus causing flower proliferations in the dodder, Cuscuta eevee Yunck. Phytopath. Zeitschr. 33: 316-318. Yuncker, T. G. Revision of the North American and West Indian species of 304 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI Cuscuta. Ill. Biol. Monogr. 6(2/3): 1-141. pls. 1-13. 1921. [Includes bibliography. | . Additions to a ee of the genus Cuscuta. Proc. Indiana Acad. Sci. 36: 259-262. . The genus 72 Mem. Torrey Bot. Club 18: 113-331. 1932. —. Cuscuta americana in Florida. Torreya 35: 60. 1935. . Convolvulaceae Lindl. 1. Cuscuta (Tourn.) L. Jn Fl. Texas 3: 123- 150. South. Methodist Univ. ‘ . Nomenclatural changes in the genus Cuscuta, and notes on some Ameri- an species. Bull. Torrey Bot. Club 70: 61- 1943 . Cuscuta japonica Choisy, an Asiatic species new " America. Torreya 44: 34, 35. 1944. [Florida and Texas. | . Observations on the presence of stomata in some species of Cuscuta. Proc. Indiana Acad. Sci. 53: 100-104. 1944. OQ 2. Dichondra J. R. & G. Forster, Char. Gen. Pl. 39. pl. 20. 1776. Creeping, perennial, pubescent herbs with reniform to orbicular alter- nate leaves, the stems rooting at the nodes. Flowers small, inconspicuous, solitary in the axils of the leaves, the peduncle subtended by a minute bract. Calyx of 5 sepals, free or united only at the base. Corolla 5-parted, broadly campanulate to rotate. Styles 2, gynobasic, the stigmas capitate; ovary deeply 2-lobed, each lobe with 2 basal, anatropous ovules. Fruit a 2-lobed, indehiscent or irregularly dehiscent capsule, each lobe 1 (rarely 2)- seeded. Typr species: D. repens Forst. (Name from Greek, di, two, and chondros, a grain, in reference to the fruit.) A genus of perhaps five species, primarily of the Americas, but with one pantropical species (Dichondra repens Forst.) which is very variable, especially in the length and shape of the petals, the shape of the leaves, and in the pubescence. In our area the genus is represented by D. repens var. carolinensis (Michx.) Choisy which occurs on roadsides, in pinelands, hammocks, and as a weed in lawns from Florida to Virginia, west to Texas and Arkansas. Dichondra repens is also grown as a ground cover, often as a lawn plant. Although the genus is small and the species are rela- tively common, the group has not been well collected and is poorly under- stood. A thorough study is needed to determine the nature and variation of the species and to evaluate the present classification. The differences evident in the size of both plants and leaves of var. carolinensis may be in response to the habitat or the difference may be more fundamental. There seem to be no characters other than the gynobasic style and the two nearly separate carpels of Dichondra to support the segregate family Dichondraceae, and the genus otherwise agrees well with the genera of the Convolvulaceae. However, additional studies, both morphological and cytological, should provide evidence with respect to the relationships of this genus. Dichondra is most nearly related to Falkia L. f., a small genus of about four species of Africa, which is characterized by a 4-lobed ovary and a gamosepalous calyx. 1960] WILSON, GENERA OF CONVOLVULACEAE 305 REFERENCES: Drew, W. B. Dew drop “grass” as a lawn plant in central Missouri. Ecology 25: 246, 247. 1944. [Dichondra repens. ] House, H. D. A new species of Dichondra. Muhlenbergia 1: 130, 131, 1906. [D. occidentalis in California. ] Howe.L1, J. T. A ground cover. Leafl. West. Bot. 2: 174. 1939. [D. occi- dentalis. | Paropr, D. Nota sobre una nueva especies del género Dichondra. Ann. Soc. Cient. Arg. 13: 5-10. 1882. [D. villosa.] 3. Evolvulus Linnaeus, Sp. Pl. ed. 2.1: 391. 1762. Diffuse herbs or suffrutescent plants with prostrate or ascending branches, never twining, bearing entire leaves. Flowers sessile or pediceled, solitary in the axils of the leaves [or in terminal cymes]. Calyx of 5 free sepals. Corolla white, blue or purple, funnelform to rotate, the limb S-angled or distinctly 5-lobed. Stamens 5, inserted on the corolla tube, exserted or included. Pollen globular, smooth. Styles 2, free or united at the base, each 2-cleft, the stigmas linear-filiform; ovary 2-locular, each locule with 2 ovules. Fruit a 1-4-seeded capsule. Seeds smooth or minutely verrucose. ‘TYPE SPECIES: E. nummularius L. (Name from Latin evolvere, to unroll, in reference to the nontwining habit.) A primarily tropical genus of perhaps 100 species, all of the Americas, from southern United States south to Argentina, but with two (E. alsinoides L. and E. nummularius L.) extending into tropical areas in the Old World. The genus, characterized by its distinct, 2-cleft styles, has been divided into eight sections, primarily on the basis of the inflorescence,\ the length of the peduncles, and the habit of the plant. Five species occur in our area, all members of sect. ALSINOIDES Meissn. Evolvulus alsinoides L. is distributed throughout the tropical regions of the world and occurs in waste places and in hammocks in the Florida Keys. It is a polymorphic species in which 15 varieties have been de- scribed, although none seems to be very clear-cut; var. Grisebachianus Meissn. is reported from our area. Our other species are E. glaber Spreng., E. sericeus Sw., E. Grisebachii Peter (E. Wrightii House), E. Nuttallianus Roemer & Schultes (EZ. argenteus Pursh, E. pilosus Nutt.). Evolvulus macilentus Small, described from the lower Florida Keys, may represent a glabrous form of E. sericeus. REFERENCES: OostsTRoom, S. J. vAN. A monograph of the genus Evolvulus. Meded. Bot. Mus. Utrecht 14: 1-267. 1934, Perry, L. M. Evolvulus pilosus an invalid name. Rhodora 37: 63. 1935. [ = E. Nuttallianus Roemer & Schultes.] SAYEEDUD-Din, M. Some common Indian herbs with notes on their anatomical character. VIII. Evolvulus alsinoides Linn. Jour. Bombay Nat. Hist. Soc. 42: 816-818. pls. 1-3. 1941. 306 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLI VarADAN, K. S. S., T. S. VAIDYANATHAN, and M. V. R. Rao. Phytochemistry of Epolvulus alsinoides Linn. Indian Jour. Pharm. 20(4): 100-105. 1958.* 4. Bonamia DuPetit-Thouars, Hist. Vég. Isl. France Réunion, Mada- gascar 1: 33. pl. 8. 1804, nom. cons. Perennial, herbaceous [or woody], prostrate or twining vines [rarely erect undershrubs] with alternate, herbaceous or subcoriaceous entire leaves. Flowers axillary, solitary, or in cymes. Sepals 5, equal or sub- equal. Corolla 5-lobed, blue or white, campanulate or funnelform. Sta- mens 5, inserted on the corolla tube, included or slightly exserted; pollen smooth. Style deeply 2-cleft (or styles free, 2, rarely 1), stigmas peltate, ovary 2-locular, each locule with 2 ovules. Fruit a 2-, 4- or 8-valved capsule. (Including Breweria R. Br., Stylisma Raf.; not Bonamya Neck., nom. rejic.) Type species: B. madagascariensis Poir. (Named in honor of Francois Bonami, 1710-1786, French physician and botanist.) A genus of 40 or more species widely distributed in the tropics of the world, represented in our area by seven species. The genus includes species of very different habit but it may be distinguished by the bifid style and the capitate stigmas. Various attempts have been made to split the genus, and, as a result, our species, with the exception of Bonamia grandiflora (Gray) Hallier, have been placed in Stylisma Raf., primarily on the basis of their more slender stems and more herbaceous habit. This group does not seem to be clearly separable from Bonamia, although it may represent a natural group within the genus. The problem of the generic limits of Bonamia is further complicated by Seddera Hochst. of the Old World, and a detailed study of all these species with a view of clarifying the genera would be highly desirable. All of our species, except Bonamia Pickeringii (Torrey in M. A. Curtis) Gray, occur in Florida in sandy soil or in dry pinelands. Of these B. humistrata (Walt.) Gray extends into eastern Texas and north into southeastern Virginia, B. aquatica (Walt.) Gray (Stylisma trichosanthes of Small) north to Alabama and North Carolina, B. Michauxii * (Stylisma aquatica of Small) west to Texas and north to North Carolina. Bonamia villosa® and B. angustifolia * are both limited to Florida. Bonamia Pick- eringii is represented in the sandhills of Georgia and from Wilmington, North Carolina, by plants which have been interpreted as two different varieties. Additional variants of this species occur in the pine barrens of New Jersey, and from Texas north to Iowa and Hlinois. 2 Bonamia Michauxii (Fern. & Schubert) K. A. Wilson, comb. nov. Breweria Mey Fern. & Schub., Rhodora 51:37. 1949 3 Bonamia villosa (Nash) K. A. Wilko. comb. nov. Breweria villosa Nash, Bull. Torrey Bot. Club 22: 154. 1895. *Bonamia angustifolia (Nash) K. A. Wilson, comb. nov. Breweria angusti- folia Nash, Bull. Torrey Bot. Club 22: 155. 5. 1960] WILSON, GENERA OF CONVOLVULACEAE 307 REFERENCES: See family Petes HALLIER (1893, 16: 527, 528), Van Ooststroom (1938, 75-80; and 1 FERNALD, M. Tees ae B. G. SCHUBERT. Some identities in Breweria. Rhodora 51: 35-43. pls. 1121-1129. 1949. [Reprinted as Contr. Gray Herb. 169. 1949. | House, H. D. Studies in the North American Convolvulaceae. III. Calycobolus, Bonar and Stylisma. Bull. Torrey Bot. Club 34: 143-149. 1907. 5. Jacquemontia Choisy, Mém. Soc. Phys. Genéve 6: 476. 1833. Herbaceous or woody, prostrate or twining vines [rarely erect under- shrubs] with entire or lobed leaves. Flowers axillary in cymes or dense terminal clusters [rarely solitary], with or without an involucre. Sepals 5, equal or unequal. Corolla entire, 5-toothed or 5-lobed, funnelform or campanulate. Stamens 5, inserted on the corolla tube, included: pollen smooth. Style simple, included: stigmas 2, elliptic or Gane flattened: ovary 2-locular, each locule with 2 ovules. Bui a 4- or 8-valved capsule. (Including T cue Raf.) Type species: J. ferruginea (Steud.) Choisy. (Name in honor of Victor Jacquement, 1801- 1832, French botanical explorer. ) About 120 species in 4 sections, primarily of the American tropics and subtropics, but also with a few species in the Old World; represented in our area by five species in two sections. Sections ANOMALAE Meissn. (flowers 1-3, or in loose, many-flowered inflorescences) and Caprrutt- FLORAE Ooststr. (flowers in the axils of the upper leaves, aggregated into dense globose or ovoid terminal spikes) are apparently limited to the American tropics in their distribution; neither occurs in our area. Sect. CAPITATAE Meissn. (inflorescence many-flowered, flowers in dense terminal clusters usually surrounded by bracts) is represented in the southeastern United States by Jacquemontia tamnifolia (L.) Griseb. (Thyella tamnifolia (L.) Raf.) which occurs in fields and thickets, often as a weed in cultivated areas, from Florida north to southeastern Vir- ginia and west to Louisiana, Araneae and Texas. Our other species are in sect. CyMOSAE Meissn. Guilorescence 5—many-flowered loose dichasia), and in our area are found only in southern Florida. Jacquemontia jamai- censis (Jacq.) Hallier and J. pentantha (Jacq.) G. Don are known also in the West Indies, while J. Curtissii Peter ex Hallier and J. reclinata House are found only in southern peninsular Florida. REFERENCES: See Tay, references, House (1909), and Van Ooststroom (1939, p. 266; 1953 Fox, W. Be R. K. Goprrey, and H. L. BLromauist. Notes on distribution of North Carolina plgnts le Rhodora 54: 165-182. 1952. [Jacquemontia tamnifolia in North Carolina. ] SMALL, J. K. Jacquemontia reclinata. Addisonia 18: 35, 36. pl. 594. 1934. 308 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI 6. Convolvulus Linnaeus, Sp. Pl. 1: 153. 1753; Gen. Pl. ed. 5. 76. 1754. Herbaceous or more or less woody, perennial [or annual], prostrate, twining or erect plants with ovate to oblong leaves, cordate, sagittate or hastate at the base. Flowers axillary, solitary or in peduncled cymes. Calyx of 5 subequal sepals, the bracts minute, linear, and remote from the calyx, or persistent and enveloping the calyx. Corolla campanulate to funnel- form, white or pink. Stamens 5, included, unequal in length; pollen smooth, ellipsoid and tricolpate or globular and polyforate. Ovary 1-locu- lar, 2-locular or imperfectly 2-locular; stigmas 2, ovoid to ellipsoid, linear or filiform. Fruit a 4-valved capsule with 4 or fewer seeds. (Including Strophocaulos Small, Calystegia R. Br.) Typr species: C. arvensis L. (The name from the Latin convolvere, to entwine.) — BINDWEED. A genus of about 250 species in the temperate and tropical regions of both hemispheres, represented in our area by four or perhaps more species. The genus is interpreted here in the broad sense to include species fre- quently segregated as Calystegia R. Br., here maintained as a section. Section ConvoL_vuLus (Strophocaulos Small) is represented in our area by the single weedy species C. arvensis, field bindweed (2 = 50), native to Eurasia but now widespread in the temperate areas throughout the world and often becoming a serious pest. The species occurs in our area in fields, roadsides, and waste places from northern Florida, north and west throughout our area and beyond it. It is very variable in the shape of its leaves, and at least three of the forms have been named. Section Catystecia (R. Br.) Gray, which includes about 25 species of both hemispheres, is represented in our area by three or perhaps more species. Members of this section are distinguished by the two large bracts which subtend the calyx and envelop it, the generally larger flowers, the 1-locular or incompletely 2-locular ovary, the ovoid or ellipsoid stigmas, and the globose, polyforate pollen. Convolvulus sepium L. (2n = 22, 24), a very variable species distributed throughout most of the temperate areas of the world, is represented in the Southeast by at least three named varieties based primarily on the shape of the leaves. Convolvulus sepium var. sepium (including C. americanus (Sims) Greene), with broadly ovate leaves with angulate, truncate or rounded basal lobes and a U- or V-shaped leaf sinus, is the European form and is now distributed in eastern North America as far west as Missouri and Illinois, and also in New Mexico and Oregon. Varietas fraterniflorus Mack. & Bush (leaves with quadrangular sinuses) is reported from Pennsylvania to North Dakota, south to Vir- ginia, Kentucky, and Arkansas, growing on roadsides and in fields and wasteplaces. Varietas repens (L.) Gray (leaves lance-ovate to lanceolate, mostly longer than broad) occurs on shores, beaches, and dunes along the coast from Florida north to New Brunswick, west to Texas, and also north to Ohio, Indiana and Wisconsin. Convolvulus sericatus House, closely allied to Convolvulus sepium, was described from the mountains of Georgia and has also been reported from the mountains of North Caro- lina. 1960] WILSON, GENERA OF CONVOLVULACEAE 309 Convolvulus spithamaeus L. (2n = 22) (including C. Purshianus Wher- ry) occurs on sandy or rocky soil, shale barrens, road cuts, rocky woods and dunes from southeastern Canada south to Georgia and Alahara and west to Minnesota and Iowa. The variation of this species is complex and needs careful study on a population basis. Two to four varieties or subspecies have been recognized. Convolvulus pellitus Ledeb. forma anestius Fern. (Convolvulus japonicus Thunb. of Small), with double flowers and sterile, has been introduced as a cultivated plant and is now naturalized from Maine to Michigan, south to Virginia, Tennessee, and Missouri. REFERENCES: See family references, House (1909), KNIGHT (1959), VAN Ooststroom (1939, pp. 282-286; 1953). Brown, E. O. ‘Notes on some variations in field bindweed (Convolvulus arvensis L.). Iowa State Coll. Jour. Sci. 20: 269-276. : Catt, L. W., and R. E. Getty. ae eradication of bindweed. Kansas Agr. Exp. ce. Circ. 101: 1-18. 1923 Crarts, A. S., and P. B. KENNEDY. The physiology of Convolvulus arvensis (morning- ee or bindweed) in relation to its control by chemical sprays. Pl. Physiol. 5: 329-344. 0 FEHER, J. Kleistopetalie bei Calystegia sie und ee ee (Hun- garian; German summary.) Bot. Kéz FERNALD, M. L. Studies of eastern ree oe ee ndations in the order Tubiflorae. Rhodora 51: 70-85. 1949. [Notes on C. spithamaeus, C. sepium, and C. pellitus.} . Errors in citation in Agrostis and Convolvulus. Rhodora 51: 192, 193. 1949. [Correction in citation of C. sepium var. repens Frazier, J. C. Nature and rate of development of root ean of Convolvulus arvensis. Bot. Gaz. 104: 417-425. 1943. Amount, distribution and seasonal trend of certain organic reserves in the root system of field bindweed, Convolvulus arvensis L. Pl. Physiol. 18: 167-184. 1943. Food reserve depletion and synthesis in field bindweed, Convolvulus arvensis L., as related to 7-day and 14-day intervals of cultivation. /bid 315-323. Hamitton, R. D., C. J. Wuirrietp, and H. E. Rea. Soil treatments for field bindweed (Convolvulus arvensis) eradication in northwest Texas. Tex. Agr. Exp. Sta. Prog. Rep. 1304: 1-4. Control of eae infestation of bindweed (Convolvulus arvensis ) in northwest Texas. Tex. Agr. Exp. Sta. Prog. Rep. 1392: 1-5. House, H. D. A new southern Convolvulus. Torreya 6: 149, 150. 1906. [C. sericatus from Georgia. ] MackENZIE, K. K., and B. F. Busu. New plants from Missouri. Missouri Bot. Gard. INA Rep. 16: 102-108. 1905. [C. fraterniflorus.] MacPHerson, G. > Comparison of development in oe and morning glory. Bot. Gaz. 712 392-398. piss 25-27. 19201 Gerace PALMER, a C. Convolvulus Wallichiana at Swarthmore, Pee a Rhodora 47: 3 1945. PERSY, a ae observations sur le comportement du nucléole dans la 310 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI caryocinése somatique de Calystegia sepium R. Br. (Convolvulus sepium L.) et sur sa néoformation. Bull. Soc. Bot. Belg. 68: 222-233. 19 Pryor, M. R. Morning glory: a oe report. Calif. Weed Conf. Brak 8: 116-118. 1956.* [ Convolvulus arve Purer, E. A. Growth behavior in eh Be Soldanella L. Ecology 17: 541- 550. 1936 SOUEGES, R. Emb rogenie des Convolvulacées. Développement de l’embryon chez le Comeau arvensis L. Compt. Rend. Acad. Sci. Paris 205: 813- B15, Torrey, J. a " Endogenous bud and root formation by isolated roots of Con- volvulus grown in vitro. Pl. Physiol. 33: 258-263. 1958. Tryon, R. M., Jr. The varieties of Convolvulus spithamaeus and of C. sepium. Rhodora 41: 415-423. pls. 557, 558. 1939 Wuerry, E. T. Four shale-barren plants in Pennsylvania. Proc. Penn. Acad. Sci. 7: 160-164. 1933. [C. ines ee p. 163.] r dwarf bindweeds. Bartonia 28: 32, 33. 1957. [Convolvulus spithamoeus: recognizes four oh ane 7. Merremia Dennstaedt ex Hallier, Bot. Jahrb. 16: 581. 1893. Herbaceous or woody twining vines [or plants prostrate, rarely erect | with entire, dentate, or palmately lobed or compound leaves. Flowers axillary, solitary or in few- to many-flowered cymose [or variously branched] inflorescences. Calyx of 5 subequal sepals, coriaceous to herbaceous, accrescent in several species. Corolla campanulate or funnel- form, white or yellow to orange, slightly 5-lobed. Stamens 5, included, unequal in length, the anthers often contorted; pollen smooth. Ovary 2- or 4-locular, stigma 2-lobed, globose. Fruit a 1—4-seeded, 4-valved capsule or an irregularly dehiscent capsule with the pericarp circumscissile at the base. (Operculina of Small.) Type species: M. hederacea (Burm. f.) Hallier. (Named in honor of Blasius Merrem, 1761-1824, professor at Marburg.) About 80 species in five sections widely distributed in the tropics of both hemispheres. Characterized by its bilobed, globose stigmas and its 4-valved or irregularly dehiscent capsule, the genus is represented in our area by four species in two sections. Section XANTHIPS (flower bud obtuse or subacute; midpetaline bands indistinctly defined, never with dark lines) is known in our area by only Merremia umbellata (L.) Hallier (Jpomoea polyanthes Roem. & Schult.), a very variable species of the tropics of the world represented in the Americas by the yellow-flowered var. umbellata. Section STREPTANDRA (flower buds mostly acute; midpetaline bands in the dried state often with 5 dark lines) is represented in our flora by three species. Merremia tuberosa (L.) Rendle (Operculina tuberosa (L.) Meissn.), 2% = 30, a glabrous vine with yellow flowers and entire- margined leaf segments, is reported from pinelands, hammocks, and waste places in Florida and Texas (fide Small) and is known also from the West Indies, tropical America, tropical Africa, India, and Ceylon. This species has been placed in the genus Operculina S. Manso because the 1960] WILSON, GENERA OF CONVOLVULACEAE SUL pericarp of the irregularly dehiscent capsule separates from the receptacle, thereby forming an operculum. Van Ooststroom, however, has pointed out that this manner of dehiscence is clearly different from that of Operculina in which the pericarp is two layered and only the outer layer, not the entire pericarp, is circumscissile, while the inner layer remains attached. Mer- remia dissecta (Jacq.) Hallier (Operculina dissecta (Jacq.) House), 2n = 30, a pubescent vine with white flowers and coarsely toothed leaf segments which occurs on roadsides and in hammocks and waste grounds from Florida to Georgia and Texas, ranges from the southeastern United States south to Argentina. It is also cultivated in other tropical areas where it occa- sionally escapes. The capsule of M. dissecta is 4-valved and the pericarp does not loosen from the receptacle as it does in M. tuberosa. Merremia tridentata (L.) Hallier ssp. angustifolia (Jacq.) Ooststr. (Ipomoea angus- tifolia Jacq.), a common African species, now occurs in waste places on the Coastal Plain and in other scattered localities. REFERENCES: See family Coa MAcsRIDE (1959), VAN OoststROOM (1939, p. 292; 1953), VERDCOURT (195 Faucao, J. I. A. een ao estudo das espécies brasileiras do género Mer- remia Dennst. Rodriguesia 16-17: 105-125. 1954 House, H. D. Studies in the North American Gonvole ulaceae, IJ. The genus Operculina. Bull. Torrey Bot. Club 33: 495-503. 1906. O’DoneELL, C. A. Las especies argentinas del genero ‘“Merremia.” Lilloa 5: 1940. $564. Bie. . Revision de las especies americanas de “Merremia.” bid. 6: 467-554. pls. 1-9. 1941. 8. Ipomoea Linnaeus, Sp. Pl. 1: 159. 1753; Gen. Pl. ed. 5. 76. 1754. Herbaceous or shrubby, annual or perennial, twining, prostrate or erect plants with entire, lobed or divided leaves. Flowers axillary, solitary or in few- to many-flowered cymes. Bracts variable. Calyx of 5 sepals, herba- ceous or subcoriaceous. Corolla campanulate or funnelform, rarely salver- form, with distinct midpetaline bands. Stamens 5, alternate with the corolla lobes, inserted on the corolla tube, included or rarely exserted; pollen globular, spinulose. Ovary 2- or 4-locular, rarely 3-locular, with 2 anatro- pous ovules in each locule. Style solitary, the stigma entire, 2(or 3)-Idbed, globular, included or exserted. Fruit a 4—6-valved capsule with 4-6 or fewer seeds. (Including Batatas Choisy, Quamoclit Moench, Pharbitis Choisy, Exogonium Choisy, Calonyction Choisy.) Type species: J. Pes- tigridis L. (Name from Greek ips, a worm, and homoios, resembling, in reference to the twining habit.) — MorNING-GLory. A large genus of the tropics and subtropics with about 500 species and perhaps 30 or fewer in our area. The limits of the genus and its subdivision have been matters of considerable disagreement. Following Van Ooststroom, who adopted the system of Hallier with but little modification, the genus $12 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI is divided into eight sections on the basis of the habit, inflorescence, flower, and seed characters, and includes a number of groups sometimes recog- nized as separate genera. Several other subdivisions of the genus have been proposed (cf. Choisy, Bentham, Peter, House) but the system most widely accepted is that of Hallier, although often with some minor modifications. A truly satisfactory generic treatment, however, must await a thorough monographic revision. Section Ipomora (§ Pharbitis (Choisy) Griseb.) (mostly high-twining plants, hispid or lanate; flowers mostly showy; sepals herbaceous, oblong, lanceolate or linear; seeds glabrous, puberulent or shortly arachnoid) is represented in our area by about five species including /pomoea purpurea (L.) Roth (2m = 30), introduced from tropical America as an ornamental, and now a pernicious weed along roadsides, waste places and in cultivated grounds throughout our area and also northward and westward. /pomoea hederacea (L.) Jacq. (2 = 30), introduced from tropical America, is equally weedy on roadsides and waste places, and J. Nil (L.) Roth (2n = 30), a native of Africa has spread from cultivation into hammocks and cultivated grounds in Florida and Louisiana. /pomoea barbigera Sweet and J. congesta R. Br. (Pharbitis cathartica (Poir.) Choisy) also occur in our area. Section Batatas (Choisy) Hallier (flowers mostly small, axillary in umbellate cymes; sepals mostly subcoriaceous, oblong or lanceolate; corolla funnelform:; seeds glabrous) includes the widely cultivated J. Batatas (L.) Lam. (2n = 90), the sweet potato, with a large number of cultivars. The origin of the sweet potato has been the subject of con- siderable discussion and study. Present cytological evidence seems to in- dicate that it is an allopolyploid, resulting from the hybridization of a tetraploid and a diploid species. Ipomoea trichocarpa Ell. (J. trifida (HBK.) G. Don), /. tiliacea (Willd.) Choisy, and J. triloba L. are mem- bers of this section, and J. pandurata (L.) G. F. W. Meyer (2 = 30) and I. lacunosa L. may belong here. Section LetocaLyx Hallier (plants mostly glabrous, flowers solitary or in subumbellate dichasia. sepals mostly oblong or lanceolate, seeds mostly glabrous) is represented by Jpomoea Pes-caprae (L.) R. Br. in Tuckey ssp. brasiliensis (L.) Ooststr. (2n = 30) and J. stolonifera (Cyr.) J. F. Gmel., both of which occur on sandy beaches and coastal sand dunes from Florida to Texas and in tropical and subtropical countries of both hemi- spheres. Also to this section belong J. cairica (L.) Sweet (2n = 30), J. sagittata Lam. (2n = 30), and J. heptaphylla (Rottl. & Willd.) Voigt. Section CaLoNycTIon (Choisy) Griseb. (annual or perennial twiners, mostly glabrous; flowers axillary, solitary or in a cincinnus or a dichasial cyme, nocturnal; sepals herbaceous to membranaceous, glabrous, or some- times hirsute; corolla salverform, the tube long; stamens and style often exserted; ovary 2-locular or rarely 4-locular; capsule 4-valved; seeds gla- brous) is represented with us by /pomoea alba L. (Calonyction aculeatum (L.) House) in southern Florida where it grows in hammocks, often 1960] WILSON, GENERA OF CONVOLVULACEAE 313 luxuriously following fire. A native of tropical America, it has been widely cultivated and has now escaped in the tropics of both hemispheres. Section QuaAmMocLit (Moench.) Griseb. (annual or perennial twiners, mostly glabrous; flowers axillary in a cyme, rarely solitary; sepals herba- ceous to membranaceous, glabrous; corolla salverform, often bright red, rarely yellow or white; stamens and style exserted; ovary 4-locular; cap- sule 4-valved, seeds glabrous rarely puberulent) includes about 15 species of the Americas, three in our area. [pomoea Quamoclit L., with pinnately dissected leaves, a native of tropical America and widely grown as an ornamental, has spread from cultivation as far north as Virginia and Mis- souri. Ipomoea coccinea L., with cordate leaves, ranging from Georgia north to Rhode Island, Pennsylvania and Illinois and west to Kansas, Oklahoma and Arkansas, is limited to the United States in its distribution. This species has often been confused with J. hederifolia L., which ranges from Florida, Louisiana, Georgia, and Texas, south to northern Argentina (also in Malaysia and Africa), but may be distinguished by having the inner sepals up to 3 (rarely 4) mm. wide, the cordate leaves with the blade entire, dentate, or 3-5 lobed; fruiting pedicels always erect. (See O’Donell, 1959, pp. 45-51, for extensive synonymy. ) Species of this section are Section Er1osPpERMUM Hallier (very variable perennial plants; sepals mostly obtuse; seeds with long-bearded edges, or rarely the whole surface villose) is known in our area by three species. Ipomoea Tuba (Schlecht.) G. Don (Calonyction Tuba (Schlecht.) Colla = C. grandiflorum (Jacq.) Choisy) grows along beaches and in saline situations in coastal hammocks in southern Florida and is distributed throughout tropical America, east tropical Africa, and Asia to Polynesia. 7pomoea crassicaulis (Benth.) B. L. Robinson (J. fistulosa Mart. ex Choisy) occurs from Florida and the West Indies, Mexico and Central America southward to Brazil and Para- guay, and is cultivated in other tropical countries where it occasionally escapes. Ipomoea microdactyla Griseb. (Exogonium microdactylum (Griseb.) House) a woody vine with bright red flowers, obtuse sepals, and seeds with long hairs on the edges, occurs in southern Florida and the West Indies. A number of species of 7pomoea are popular in cultivation in the United States. Two of the best known are /. Nil Roth ‘Scarlett O’Hara’ and J. tricolor Cav. ‘Heavenly Blue.’ REFERENCES: Most of the extensive technical agricultural literature on /pomoea Batatas has, of necessity, been omitted. See family references, House (1909), VAN Oost- sTROOM (1940, pp. 481-582; 1953), Rao (1940), and VERpcouRT (1958). Autres, H. E. Ipomoea trichocarba Ell. and Ipomoea trifida G. Don. Jour. Elisha Mitchell Sci. . 75: 129, 1959. [The two are conspecific. | Axtarp, H. A. A new pe of the moonvine Calonyction aculeatum with divided corolla limb, and length-of-day behavior and flowering of the com- mon form. Jour. Wash. Acad. Sci. 35: 33-36. 1945. [Forma apopetalum.] Asusy, E., et al. Studies in the morphogenesis of leaves. I. New Phytol. 47: 314 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI 153-176. 1948; II. 7bid. 177-195. 1948; IV. Ibid. 49: 23-35. 1950; V. Ibid. 189-192; VI. Ibid. 375-387; XI. Ibid. 55: 91-110. 1956. [Mostly experiments on /. caerulea. | BAEHNI, C. /pomoea heptaphylla in Georgia and Mexico. Rhodora 38: 164. 1936. Britton, N. L. Exogonium microdactylum. Addisonia 1: 33, 34, pl. 17. 1916. BurKILL, I. H. Aji and Batata as group-names within the species /pomoea Batatas. Ceiba 4: 227-240. 1954. [History of discovery and diffusion of the sweet potato by the Spaniards and Portuguese. | Dormer, K. J., and J. A. BENTLEY. Some complex relationships between auxin content and leaf area in Jpomoea caerulea Koen. New Phytol. 51: 116-126. 1952 FeporTscHUK, W. Entwicklung und Bau des mannlichen Gametophyten bei den Arten der Convolvulaceen-Gattung Quamoclit. Planta 16: 554-574. 1932. [J. coccinea, I. hedertfolia, I. Quamocht. | FEHER, J. scr aoe bei Calyste gia ae und el piles (In Eaineaian Germany summary.) Bot. Kozlem. 27: GreuLacH, V. A. Photoperiodic after fects in ee on ie Oci,.43% 65-73. 1943. iva hederacea and 1. purpure Hacrwara, T. Genetic studies of flower- em in Japanese morning glories. (In Japanese.) Bot. Mag. Tokyo 44: 573-595. 1930. [Z. Nil.] Harter, L, L. Bud sports in sweet potatoes. Jour. Agr. Res. 33: 523-525. pl. 1. — and J. L. Wermer. A mga study of sweet-potato diseases and their control. U.S. Dep. Agr. Tech. Bull. 99: 1-117. pls. 1-26 HartTMAn, J. The non-flowering character of sweet potatoes of the Teves type. Pi, Phacicll 22: 322-324. 1947. Haywarp, H. E. The seediiny anatomy of Jpomoea Batatas. Bot. Gaz. 93: 400-420. pls. 6,7. 1932. HorneEL., J. How did the sweet potato reach Oceania? Jour. Linn. Soc. Bot. 53: 41-62. 1946. House, H. D. Studies in the North American cae The genus Exogonium. Bull. Torrey Bot. Club 24: 97-107. pls. 1, 2. 1908. . The North American species of the genus /pomoea. es N. Y. Acad, Sci. 18: 181-263. 1908. . Studies in the North pets Convolvulaceae — V. Quamoclit. Bull. Torrey Bot. Club 36: 595-603. 1909. INAMURA, S. I., and A. TAKImMoTo. aes responses in Japanese morning glory, Pharbitis Nil Chois., a sensitive short day plant. Bot. Mag. Tokyo 68: 235-241. 1955. (See also 68: 260-266. 1955; 69: 23-29, 289-297, 353- 358, 1956; 70: 53-57. 1957; 71: 254-260. 1958.) IsBELL, C. L. Regeneration in leaf cuttings of Jpomoea Batatas. Bot. Gaz. 91: 411-425. 1931. Keur, A. E., Y. C. Trxc, and J. C. Mriiter. The site of carotenoid and antho- cyanin synthesis in sweet potatoes. Proc. Am. Soc. Hort. Sci. 65: 396-398, 1955, Kenyan, F. M. G. A morphological and cytological study of Ipomoea trifida. Bull, Torrey Bot. Club 55: 499-512. 1929. Kino, J. R., and R. BAmrorp. The Sea number in 7pomoea and re- lated genera. Jour. Hered. 28: 279-282. 1937 1960] WILSON, GENERA OF CONVOLVULACEAE SHE) KnicuT, R. J. Morphological variation within a Clarke County Boa ae of Ipomoea hederacea (L.) Jacquin. (Abs.) Va. Jour. Sci. II. 7: 1956. . Hybridization in 7pomoea and its relatives. (Abs.) Va. nae Scale 8: 291, 292. de Kuse, G. Bud inhibition and correlative growth of ee, in sweet potato stem. Mem. Coll. Sci. Univ. Kyoto B. 21: 107-114. Lam, S. L., and H. B. CorpNEr. Flowering hormone in ee to blooming in sweetpotatoes, Science 121: 140, 141. 1955. Lauritzen, J. I. Some effects eee temperature on sweetpotatoes. Jour. Agr. Res. 42: 617-627. pl. 1 Factors affecting ee ‘and decay of sweet potatoes by certain storage rot fungi. Jour. Agr. Res. 50: 285-329. 1935. Leg, S. H., and C: H. SHEo. Developmental morphology of the tubers of sweet polio. (In Chinese; English summary.) Acta Bot. Sinica 5: 207-221. pl. 1, figs. 1-7. 1956. McCormick, F. A. Notes on the anatomy of the young tuber of /pomoea Batatas Lam. Bot. Gaz. 61: 388-398. 1916. MaAHESHWARI, P. The seed structure of /pomoea, a criticism. Sci. Cult. 9: 557. 1944.* Mepscer, O. P. The Man-of-the-earth or wild potato vine. Torreya 24: 39-42. 1924. [J. pandurata. | MIKELL, J. J. The influence of chemical! treatments on the growth and flowering of sweet potatoes. Proc. Am. Soc. Hort. Sci. 60: 295-298. 1952.* Miter, J. C. Inducing the sweet potato to bloom and set seed. Jour. Hered. 28: 347-349. 1937. . Further studies and technic used in sweet potato breeding in Louisiana. Jour. Hered. 30: 484-492. 1939. NAKAJIMA, G. Cytogenetical studies on the amphidiploid of Quamocht. (In Japanese; English summary.) Jap. Jour. Genetics 26: 203-209. 1951.* NAKAYAMA, S. Photoreversible control of flowering at the start of inductive dark period in Pharbitis Nil. Ecol. Rev. 14: 325, 326. 1958.* Nyoxu, E. The effect . aa on leaf shane in Ipomoea caerulea. New Phytol. 55: 213-228. The effect of ae iG and temperature on leaf shape in lpomors caerulea. Ibid. 56: 154-171. Nouara, S. Genetical oe on n Quam pone Jour. Coll. Agr. Univ. Tokyo 11: 21-44. pls. 3-7 O’DonELL, C. A. Nota ne ie fistulosa Martius ex Choisy. Bol. Soc. Argent. Bot. 4: 175, 176. 1952 . Las especies americanas de “Ipomoea” L. sect. “Quamoclit” (Moench) Griseb. Lilloa 29: 19-86. 1959. Rao, N. S. Chromosome studies in the genus /pomoea. Curr. Sci. Bangalore 16: 156. 1947. Rosa, J. T. Mutations in the sweet potato. Jour. Hered. 17: ae 168. 1926. SMALL, J. K. Ipomoea macrorhiza. Addisonia 13: 43, 44. pl. 4 1928, Stout, A. B. The flowers and seed of sweet potatoes. Jour. 4 - Bot. Gard. 25: 153-168. pl. 292. 1924. : eae Batatas. Addisonia 9: 35, 36. pl. 306. 1924. rther notes on the flowers and see of sweet potatoes. Jour, N. Y. Bot. Card. 27: 129-135. 1926. 316 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI TAKENAKA, Y. The genes of Japanese morning glory (Pharbitis Nil). (Abs. ) (In Japanese.) Jap. Jour. Genetics 33: 326. 1958. Taximoto, A., K. Ikepa, and S. I. Imamura. Uber den lichtunempfindlichen Teilprozess in der bliihinduzienrenden Dunkelperiode bei Pharbitis Nil. Bot. Mag. Tokyo 71: 317, 318. 1958. Txompson, H. C., and J. H. Beattie. Group classification and varietal de- scriptions of American varieties of sweet potatoes. U. S. Dept. Agr. Bull. 1021: 1-30. pls. 1-8. 1922. Tinc, Y. C., and A. E. Keur. Meiotic studies in the sweet potato (Ipomoea Batatas Lam.). Jour. Hered. 44: 207-211. 1953. , —— , and J. E. Mitrer. A cytological study of the sweet potato plant Ipomoea Batatas (L.) Lam. and its related species. Am. Nat. 91: 197-203. 1957. Verpcourt, B. Typification of the subdivisions of /pomoea L. (Convolvulaceae ) with particular regard to the East African species. Taxon 6: 150-152. 1957. : e names of the morning glories cultivated and naturalized in East Africa. Ibid. 231-233. [Discussion of the typification of 7. congesta, J. hederacea, I. purpurea, and I. Nil. | Warmke, H. E., and H. J. Cruzapo. The flowering and seed-setting of sweet potatoes in Puerto Rico. Science 109: 62, 63. 1949. 9. Turbina Rafinesque, FI. Tellur. 4: 81. 1836 [1838]. Herbaceous or more or less woody, trailing or climbing, perennial vines with entire, ovate leaves, cordate at the base. Flowers in axillary peduncled cymes borne on the terminal portion of the branches, Calyx of 5 unequal, ovate or elliptic, obtuse sepals with thin margins. Corolla campanulate, white, with distinct midpetaline bands. Stamens 5, included, unequal in length; pollen spinulose. Ovary 2-locular, the stigma biglobular, the style included, Fruit a 1-locular, usually 1-seeded, indehiscent, ovoid-oblong capsule with a thin, woody pericarp. (Legendrea Webb & Berth.) Type species: T. corymbosa (L.) Raf. (Name from Latin, turbinatus, turbi- nate, in reference to the shape of the capsule.) 4 genus of about ten or perhaps more species of the American tropics and Africa, represented in our area by a single species, Turbina corymbosa (Rivea corymbosa (L.) Hallier), which occurs throughout tropical South America, Central America, the West Indies, and Mexico, and is reported from southern peninsular Florida, the Florida Keys, and south Texas growing in hammocks and thickets. It has been introduced in various places in the Old World and in some areas has become naturalized. T. corymbosa was used by the Aztecs as a narcotic in divination, as a medi- cine, and as an ingredient in magical and analgesic ointments. Turbina may be distinguished by the indehiscent, ovoid-oblong, mostly 1-seeded, woody capsules, and by the ovate to oblong, obtuse, spreading sepals which are not accrescent or only slightly so. Rivea Choisy has been shown to include species with linear-oblong stigmas and indehiscent, dry or nearly dry fruit and to be restricted to tropical Asia. 1960] WILSON, GENERA OF CONVOLVULACEAE ply! REFERENCES: See family references, House (1909), MEEUSE (1957, pp. 773-784), VAN Oost- STROOM (1943, 354), and VeRDcouRT (1958). ScHuLtes, R. E. A contribution to our knowledge of Rivea corymbosa, the nar- cotic ololiuqui of the Aztecs. 45 pp. Botanical Museum, Harvard Univ., Cambridge, Mass. 1941. ; narcotic morning-glory (Una convolvulacea narcotica). Revista Soc. Cub. Bot. 10: 61-68, 101-107. 1953; 11: 14-23. 1954. [T. corymbosa. | TayLor, N. Come and expel green pain. Sei Month, 58: 176-184. 1944. [His- torical account of the narcotic uses of T. corymbosa and Lophophora williamsii. | 10. Argyreia Loureiro, Fl. Cochinchin. 1: 134. 1790. Perennial, woody, twining vines with ovate to orbicular leaves, cordate at the base and densely sericeous on the lower surface. Flowers in ped- uncled axillary cymes. Calyx of 5 equal or subequal, coriaceous sepals. Corolla funnelform, campanulate or tubular, rose-purple, with distinct midpetaline bands, Stamens 5, included, unequal in length; pollen spinu- lose. Ovary [2- or] 4-locular, the stigmas biglobular, the style included. Fruit thick-walled, leathery, 4-seeded or less, indehiscent. TYPE SPECIES: A. obtusifolia Lour. or A. acuta Lour. (Name from Greek argyreios, silvery, in reference to the pubescence on the lower surface of the leaves.) — WOOLLY MORNING-GLORY. A genus of about 90 species in tropical Asia, Malaysia, and Queensland. A single species of sect. PryxantHus G. Don, A. nervosa (Burm. f.) Bojer (A. speciosa (L. f.) Sweet), a native of India which is cultivated in tropical areas as an ornamental and occasionally escapes, has been re- ported in Florida from hammocks along the Everglades (Small). REFERENCES: See family references, Rao (1940) and VAN OoststRoom (1950, 1953). Hoocranp, R. D. The Convolvulaceae of Malaysia, VIII. The genus Argyreia in the Malay Peninsula. Blumea 7: 179-192. 1952. Srncu, P. Pharmacognosy of leaf of Argyreia speciosa Sweet. Jour. Sci. Indus. 16C(10): 204-206. 1957.* 318 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. XLI BIBLIOGRAPHIC DATA ON THE HILLCREST GARDENS BOOKS, 1911-1941 Ricuarp A. Howarp Tue HiLicrest GARDENS BOOKS constitute a most unusual series of publications, being at once documents on the history of the present Case Estates of the Arnold Arboretum, in Weston, Massachusetts, and a report of the horticultural and agricultural pursuits of Miss Marian Roby Case. In 1909, her mother, Laura Lucretia Williams Case, gave Miss Case a portion of the family estate in Weston. In succeeding years Miss Case bought adjacent pieces of land, including a house which became her home, until a total of 100 acres was acquired. Miss Case operated this land, known from 1910 to 1919 as the Hillcrest Farm and from 1920 to 1944 as the Hillcrest Gardens, as a truck farm, plant introduction station, and school of practical agriculture for boys. The history of the farm and its methods of operation will be described in an issue of Arnoldia to be pub- lished soon. Between the vears 1911 and 1941, Miss Case published the annual “green books” to record the development of the farm, the gardens and the school. From the beginning, the books consisted primarily of student reports made at the school’s Labor Day exercises which concluded the summer’s work. In addition, there were included in these books special papers by such horticulturists and botanists as John G. Jack, Elmer D. Merrill, Arthur Williams, E. H. Wilson, and John Wister. The student reports were generally of high quality and lasting value, covering such subjects as agricultural practices, weather data, hurricane damage, monthly flowering lists of herbaceous plants, lists of birds, wild flowers and native trees, trees and shrubs under cultivation in Weston, and records of plant introduction trials. These data have been of value in documenting the age of certain plants at the Case Estates and in determining the date of introduction and the persistence in New England of exotic plant intro- ductions. During the summer of 1910 Miss Case employed local boys to help with the farm work. The following year the Hillcrest Farm School was organ- ized and the first booklet describing the farm and its activities was pub- lished that September under the title “‘The Second Summer at Hillcrest Farm, Weston, by Marian Roby Case.” Similarly, the title of each suc- ceeding “green book” indicated the number of years the school had been functioning, but the complete title and/or the authorship changed from time to time, as follows: 1912-1917: The Summer at Hillcrest Farm, Weston, by the Boys for the Boys. 1960 | HOWARD, HILLCREST GARDENS BOOKS 319 POTS LOO ener By the Farmers. 1920-1923: Hillcrest Gardens, Weston, Mass., The Summer, by the 1924-19872 a By Marian Roby Case, F.H.R.S. and the Gardeners. POS Sige act Tres. By Marian Roby Case, F.R.H.S. and the Hillcrest Boys. 1939-1941: ...... By Marian Roby Case, F.R.H.S., Jack A. Williams and the Hillcrest Boys. The first booklet consisted of eleven pages with a green cover printed in black. Beginning in 1915, the cover printing was in gold, and from 1916 until 1941 the cover design, with one exception, consisted of a draw- ing by one of the boys or a visitor to Hillcrest. The sole exception, the issue of 1921, had pasted on the cover a photograph showing the display of produce exhibited by Hillcrest Gardens at the Massachusetts Horticul- tural Society’s fall show. Beginning with the issue of 1924 a group photo- graph was included in the booklet, and during the succeeding years there appeared photographs of Hillcrest exhibits or produce, student reunions, aerial photographs of the estate, hurricane damage and repair, or pic- tures of gardens seen by Miss Case in her travels. The cost of publishing these books was borne by Miss Case and she determined their distribution. She mentioned, or quoted from, letters of acknowledgment received from Herbert Hoover, then in charge of a gov- ernment program to increase food supplies; David Fairchild, Agricultural Explorer in Charge, Bureau of Plant Industry, U.S. Department of Agri- culture; A. Lawrence Lowell, President of Harvard College; and others. Although Miss Case planned to send books regularly to certain libraries — that of the Royal Horticultural Society, England; the City Library, Provi- dence, Rhode Island; the libraries of the Massachusetts Horticultural So- ciety, the Arnold Arboretum, the City of Boston, the Town of Weston, the Boston Athenaeum, the State Library; and the Jones Library, in Am- herst — not one of these has a complete file today. The “green books” remain an informal but highly informative account of horticulture, agriculture, and plant introduction in a period when pri- vate gardens flourished in the Boston suburbs. 320 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI STUDIES IN ARTOCARPUS AND ALLIED GENERA, V. A REVISION OF PARARTOCARPUS AND HULLETTIA FRANCES M. JARRETT Parartocarpus Baillon, Adansonia 11: 294. 1875; Benth. & Hook. f. Geos Pigs 375, 1880: Becc. For. Borneo, 632. 1902, Webbia 5: 559. 1923; Renner, Bot. Jahrb. 39: 361. 1907; Backer, Beknopte FI. Java. 6: 12. 1948: Jarrett, Jour. Arnold Arb. 40: 9, 11. fig. 2, a-f, fig. 3, g-t. 1959. Type species: Parartocarpus beccarianus Baillon (= P. venenosus (Zoll. & Mor.) Becc.). Gymnartocarpus Boerl. Ic. Bogor. 1: 73. t. 24, 25. 1897, Handl. Fl. Ned. Ind. 3: 335, . 1900; Koord. & Val. Bijdr. Boomsoort. Java 11: 28. 1906; Koord, Exkursionsfl. Java 2: 96, 1912, “Gymnoartocarpus.” TyPE SPECIES: Gymnartocarpus ee es & Mor.) Boerl. (= Parartocarpus veneno- sus (Zoll. & Mor.) Becc Medium to large trees. Leaves spirally arranged, simple, entire, penni- nerved, thinly to thickly coriaceous, glabrous to pubescent; epidermal gland-hairs superficial, long-stalked, heads cylindric, unicellular; spongy mesophyll compact, without resin-cells; juvenile leaves elongate. Stipules simple, axillary, nonamplexicaul, the apex entire or bifid, scar intrapetiolar. Inflorescences unisexual, rarely bisexual, capitate, head globose at an- thesis, becoming shallowly lobed in syncarp, pedunculate, solitary or paired in leaf-axils; stamens or ovaries sunken in numerous cavities closely set over the entire surface of the receptacle; receptacle armoured from numer- ous, closely set, indurated, spinous, conical, obtuse or truncate processes (perianth segments or interfloral bracts ?), those surrounding the open- ings connate basally in groups of 2—4 around each aperture, those cover- ing the intervening surface free; basal involucre of 3-4(—8) deltoid or ovate bracts present, sometimes obsolete. Azé anthesis: anthers or stigmas exserted between the connate processes; male head with 1-3 stamens in each cavity, the filaments free or united, anthers 2-3 mm. long, short- exserted; female head with the ovaries solitary in each cavity, unilocular, the style apical with a short-exserted, lanceolate or fimbriate stigma, the ovule subapical. Mature syncarp formed by the enlargement of the en- tire female head, with (1—)3 to many flowers forming fruit; mature ovary thick-walled, with a firm exocarp finally decaying and leaving the stony endocarp, the style lateral, the seed large, attached laterally, testa mem- branous with a thickened apical cap, the embryo curved, the cotyledons incumbent, unequal, the inner one smaller, the appressed faces at an angle of 90° to the median plane of the ovary, the large, straight radicle directed upwards at the hilum with the tip enclosed in a little remaining endo- sperm; germination epigeal. 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, V 32 DISTRIBUTION: peninsular Siam, Malaysia (except the Moluccas and the Lesser Sunda Islands), Solomon Islands. The genus Parartocarpus, in which two species are recognized in this revision, has capitate, armoured inflorescences which have frequently been confused with the inflorescences of Artocarpus. The characters distin- guishing the two genera have been discussed fully in the introductory paper of this series (Jour. Arnold Arb. 40: 1-29. 1950; keys to the genera on page 26) and they will therefore be recapitulated only briefly here. The inflorescences of Parartocarpus differ from those of Artucurpus in having a usually well-developed involucre of 3—4(—8) basal bracts and in having the stamens or ovaries enclosed in what are considered to be cavities of the receptacle (formed chiefly by intercalary growth in the walls separating them), rather than enclosed in perianths. The surface of the receptacle is covered by solid, indurated (sessile) processes which vary in shape from cylindric and truncate to spinous and which bear a superficial resemblance to the indurated, tubular, perforate, free perianth apices of Artocarpus sect. Duricarpus. Many of the processes are con- nate, either at the base only or for nearly their whole length, in groups of 2-4 around the openings of the receptacular cavities, with the stamens or stigmas exserted between them at anthesis. The processes covering the intervening surface are free but are otherwise identical in appear- ance to those surrounding the apertures. These processes may represent fertile and sterile perianths respectively, but, taking into account their similarity to the indurated heads of the interfloral bracts in some species of the related, though less reduced, African genus Treculia, it is possible that they are secondarily modified interfloral bracts. The leaves in Parartocarpus are spirally arranged as in Artocarpus subg. Artocarpus, but the stipules are nonamplexicaul and they are not paired as in Artocarpus, but are simple and axillary, each being formed from a pair of stipules fused along the intrapetiolar margins. The generico-specific description of Parartocarpus beccarianus was pub- lished by Baillon in 1875 and was based on a Beccari collection from Borneo bearing male inflorescences. These were rather poorly preserved and Baillon described them erroneously as being covered by stamens in- termixed with [stalked] bracts having obtuse, somewhat thickened heads. He stated that Parartocarpus differed from Artocarpus in having a basal involucre, in the lack of perianths, and in the nature of the stipules. (But he apparently compared the last only with the amplexicaul stipules of subg. Artocarpus, since he described them, in contrast, as lateral and nonamplexicaul which, if correct, would have corresponded with the con- dition present in subg. Pseudojaca. The genus was described again as Gymnartocarpus by Boerlage in 1897 with a single species, G. venenosa, based on Artocarpus venenosa Zoll. & Mor. from Java. The latter was published in 1845 and is the earliest account of a species of Parartocarpus. Boerlage described the stamens and ovaries in Gymnartocarpus as being enclosed in cavities of the re- o22 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI ceptacle, and he regarded the processes on the surface as representing bracteoles, not observing any fusion between them. He distinguished the genus from Artocarpus by the absence of perianths and the different form of the interfloral bracts. On the basis of Baillon’s description he also sep- arated it from Parartocarpus. He did not comment on the stipules or on the involucre, but the latter appears, from the plates, not to have been well developed in his material. Several further species that must be referred to Parartocarpus had mean- while been described under Artocarpus. Artocarpus tylophylla Miq. (1859) and A. cerifera Mig. (1867) were both described from Java and were reduced to Gymnartocarpus venenosa by Boerlage in 1897, Artocarpus riedelii Miq. (1867), from Celebes, and A. involucrata Schum. (1889), from New Guinea, were transferred to Parartocarpus in 1907 and 1900, respectively. In describing Artocarpus bracteata from Malacca and A. forbesti from Perak and Sumatra, King (1888, 1889) mistook the spinous processes on the syncarp for the perforate perianth apices of Artocarpus, and the discordant stipular characters were the cause of his rejection of Trécul’s subgenera in Artocarpus when revising the species of ‘British India” (i.e., India, Pakistan, Ceylon, Burma and Malaya), as explained in an earlier paper (Jour. Arnold Arb, 40: 123. 1959). King made no mention of Parartocarpus and Baillon’s description may not have been available to him. In 1902 Beccari, in an appendix to his Nelle Foreste di Borneo (re- printed in Webbia 5: 550-565. 1923), discussed the inflorescence struc- ture of Parartocarpus and listed seven species in the genus: P. beccarianus Baill., P. venenosus (Zoll. & Mor.) Bece. (Artocarpus venenosa Zoll. & Mor., Gymnartocarpus venenosa (Zoll. & Mor.) Boerl., Artocarpus for- besti King), P. bracteatus (King) Becc. (Artocarpus bracteata King), P. borneensis Becc., P. excelsa Becc. (also from Borneo), P. papuana Becc. and P. involucrata (K. Schum.) Schum. & Lauterb. He reduced Gymnar- tocarpus to Parartocarpus, removed King’s two wrongly placed species to the correct genus, and described three new species. He minimized the differences between Artocarpus and Parartocarpus, stating that the only distinguishing character was the presence of an involucre in the latter. He regarded the stamens and ovaries in Parartocarpus as being enclosed in perianths which were tubular and laterally fused with each other below, but free and 2—3-fid above. The “sterile” processes he thought were sterile perianths, but he also apparently regarded them as equivalent to the interfloral bracts of Artocarpus. In considering this structure as similar to that found in Artocarpus he failed to recognize significant differences that exist between the genera, regardless of the morphological interpre- tation of the inflorescences in Parartocarpus. In Artocarpus the 2-4-fid or -partite male perianths enclosing the stamens are always completely free from each other. The tubular and perforate female perianths are frequently connate only in a ‘peripheral layer forming an external wall to the syncarp, while they remain free in the proximal region containing the ovaries. Thus each ovary is separated from the next by two perianth 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, V a0 walls instead of the single (receptacular) wall found in Parartocarpus. . Smith, in describing Gymnartocarpus triandra in 1922 (trans- ferred by him to Parartocarpus in 1924) apparently took the same view of the inflorescence structure as Beccari. He was the first author to de- scribe the stipules correctly as connate and intrapetiolar. The two species recognized in this revision are Parartocarpus brac- teatus, showing little variation and confined to Western Malaysia, and the variable P. venenosus, the distribution of which corresponds to that of the genus, extending beyond Malaysia to peninsular Siam and the Solo- mon Islands. Four fairly distinct subspecies are here distinguished within the latter and the three that are additional to the type subspecies corre- spond to the three new species described by Beccari in 1902. For one of these (P. excelsus) King’s species, Artocarpus forbesti, is taken as the basionym for the subspecific epithet.' The remaining specific names men- tioned above, with the addition of Artocarpus woodu described by Mer- rill in 1908 (transferred to Gymnartocarpus in 1921 and to Parartocarpus in 1923), are distributed in synonymy among the various subspecies of Parartocarpus venenosus. A few comments on the inflorescences in Parartocarpus may be added here. The male and female heads are indistinguishable externally before anthesis, in contrast to Artocarpus in which they usually differ in shape. The fusion of some of the processes is only conspicuous in the female head, especially at maturity, although it may be established by cutting trans- verse sections through the bases of the processes in the male head. While the inflorescences are normally unisexual, heads are found in which male and female flowers occur on different areas of the receptacle, but only very rarely intermingled. The involucre, again, is normally well de- veloped, but in P. venenosus the bracts may be indistinct and merged into an annulus at the base of the inflorescence. Both Parartocarpus and Hullettia are unusual among the Moraceae in that the anthers are extrorse. Parartocarpus has copious latex which is poisonous (unlike that of Artocarpus) and is used as an arrow poison. The ripe fruits are, how- ever, frequently described as being edible, although the young seeds are stated to be poisonous. The bark on the trunk is very distinctive through the presence of abundant, large, pustular lenticels. KEY TO THE SPECIES OF PARARTOCARPUS aves having 11-15 pairs of lateral veins, rufous pubescent beneath with the see a distinctly prominent, 5—10 on eae side of midrib; processes on the syncarp spinous, the bases + bulbous, on inflorescences at anthesis narrowly spinous, c. 3 < 1 mm.; involucral bracts 5-10 mm. long. .... 1. P. bracteatus. Leaves having 6-15 pairs of lateral veins, thinly pubescent to glabrous beneath with the intercostals not or shallowly prominent, fewer; processes on the syn- 1Jt should be noted that in the introductory paper this was referred to as a distinct species within Parartocarpus; it is now considered that the taxon can be recognized only at the subspecific level. 324 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI carp truncate to spinous, the bases not bulbous, on inflorescences at anthesis truncate to acute, never narrowly spinous; involucral bracts to 5 mm. lo sb Politi Ath aie eh lees Seah Wise ied grades eediain ecards kan # pac na eet 2. P. venenosus. 1. Parartocarpus bracteatus (King) Becc. For. Borneo, 632. 1902, “bracteata”; Renner, Bot. Jahrb. 39: 362. 1907 Artocarpus bracteata King in Hook. f. Fl. Brit. Ind. 5: 540. 1888, et in Ann, Bot. Gard. Calcutta 2: 7. 1889, pro parte quoad ¢. 1B, syncarpium solum, et spec. Griffith 4663; Ridley Fl. Malay Penin. 3: 352. 1924. Syntypes, Malacca, Griffith 4663, Maingay 2411 (Kew Distrib. 1476) (cat); lecto- type, Griffith 4663 (CAL). Artocarpus rufescens auct. non Miq., Kurz, For. Fl. Burma 2: 431. 1877. Trees, height to 45 m., buttressed or not, bark grey, smooth, with very large lenticels. Twigs 5— 8 mm. thick, rugose, densely rufous pubescent. Stipules 5-10 mm. long, broadly lomeenlate: rufous pubescent. Leaves 10-28 6-14 cm., obovate-oblong, obtuse or short-acuminate, base rounded or shallowly cordate, margin entire; venation prominent be- neath; glabrous or nearly so above, except the short-pubescent main veins, rufous pubescent beneath; lateral veins 11-15 pairs, curved, basal 2 or 3 pairs slightly crowded; intercostals numerous, usually parallel; dark green, drying yellowish to purplish brown above, red-brown beneath; petiole 20-35 mm. long. Inflorescences solitary in leaf-axils. At anthesis: male head 25—35 mm. across, globose, echinate from closely set, rigid, spinous, often slightly curved processes c. 3 1 mm.; stamens 2 in each cavity, to 8 mm. long, filaments free or united at the base, anthers oblong, apiculate, 2.5—-3 mm. long; basal involucre of 3 ovate, concave, rufous-pubescent bracts, 5-10 x 5-10 mm.; peduncle 20-35 & 2-3 mm., rufous pubescent; female head echinate as in male head, many of the processes basally connate in groups of 2—4 with a lanceolate style exserted to 3 mm. between the free apices. Syncarp to 9 cm. across, subglobose, shallowly lobed, drying red-brown, echinate from closely set, rigid, spinous processes, many connate in groups of 2—4, those on the lobes + bulbous below, 6-8 * 2-5 mm., the rest slender; wall c. 2 mm. thick; “seeds” (pericarps with a thick, stony endocarp) numerous, ellipsoid, 15 & 12 mm.; core c. 40 mm. across; involucre as in male head; peduncle 40-55 > 5 mm., rufous pubescent. VERNACULAR NAME: ipoh, Malaya (cf. Antiaris toxicaria Lesch.). Uses: the latex is used as an arrow poison. DISTRIBUTION: in evergreen forest to 2000 ft.; Malaya, Sumatra, Banka, Borneo Malaya. SELANGoR. Bukit Cheraka For. Res., Walton KEP 28387 (xeEp, infl.); Kuala Lumpur, Ginting Simpah, KEP 64942, 71252 (KEP, 2); Sungei Lalang Kajang, Symington CF lage (SING). Necrt SEMBILAN. Senawang Reserve, Yakim CF 507 (x, KEP, 6). Matacca. Alvins 465 (sinc), Maingay 2411 (Kew Distrib. 1476). 10h 8 (GH, K, ¢); Bukit Kemuning, Derry 1022 (sinc); 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, V S25 between Ching and Roombiya, Griffith 4663, 1842 and 1845 (CAL, K, L, P, U, ¢, mixed, @). Penanc. Curtis, March 1893 (sinc, 6, %). StnGAPorE. Chanchu Kang, Ridley 4128 (CAL, sING); MacRitchie reservoir, Sinclair SFN 39426 (L, sING, 2); Mandai road, Nur, Oct. 1917 (stnc, 2); Serangoon road, Ridley 8408 (K, 2); Tanjong Kling, Corner, March 1938 (stnc). PuLau Tioman. Ayer Surin, Henderson SFN 21692 (BM, BO, K, SING, infl.). Sumatra. TAPANULI. Barus, Penkalan Tapus, bb 29549 (a, BO, L). EAST Coast. Huta Padang, near Kisarin, Krukoff 327 (Bo, SING). PALEMBANG. Ban- juasin, Bajunglintjir, N/FS T 761 (Bo, L, 2); Lematang Ilir, Gunong Megang, NIFS T 816 (so, L, 2); Lematang Ilir, Semangus, bb 32216 (Bo, 1, SING) ; Lematang Ilir, Tandjong, NJFS T 645 (xo, infl.); Rawas, Grashoff 1001 (Bo, L, 2). Banka. Blinju, Berkhout 151, Grashoff 90 (Bo). Borneo. EAST AND NORTHEAST BorNeEo. Balikpapan: Pemaluan, bb 24741, 24753 (A, BO, L); Sungei Warin region, Kostermans 4303 (K, L, 9). Berouw: Inaran, bb 12173 (Bo). E. Kutei: Loa Djanan, w. of Samarinda, Kostermans 6550 (BO, K, L, ¢, mixed). W. Kutei: Mendom, Sungei Alan-Klindjang, bb 29257 (A, BO, L). BritisH NortH Borneo. Bukit Garam, near Kinabatangan river, Wood A 4659 (A, KEP, L, SING, infl.). In publishing the description of Artocarpus bracteata, King cited two collections, Griffith Kew Distrib. 4663 and Maingay Kew Distrib. 1476, both from Malacca. While the first of these collections is Parartocarpus bracteatus as here understood, the second is a mixture of this species and Artocarpus rigidus ssp. rigidus, owing to the allocation of the same Kew Distribution number to two different collections. At Kew there are under the number 1476 a sheet of Parartocarpus bracteatus bearing Maingay’s own number 2411 and a sheet of Artocarpus rigidus with his number 2414. Duplicates distributed under the number 1476 consist at Calcutta and Leiden of A. rigidus and at the Gray Herbarium of Parartocarpus brac- teatus. Although King at some period annotated the Kew specimens cor- rectly, he appears to have drawn up his description from the material at Calcutta, namely the Griffith specimen consisting of detached leaves, a small portion of twig, and a mature syncarp of Parartocarpus bracteatus, and the Maingay specimen which was a sterile shoot of Artocarpus rigi- dus. The vegetative characters are a mixture of the two species, and the twigs are described as annulate (as in Artocarpus subg. Artocarpus). The plate is also a mixture, since comparison with these specimens shows that the artist drew the syncarp of Griffith 4663 as if attached to the shoot of Maingay 1476. The confusion in the type material, the description, and the plate has not been noted previously, and King’s specific epithet has always been applied to this species of Parartocarpus. Kurz incorrectly identified this species with Artocarpus rufescens Miq. (= A. dadahk Miq.) and included it under that name in his Forest Flora of British Burma (1877), presumably because on Griffith 4663 at Cal- cutta the provenance is given as Burma. This is certainly an error, since no other collections have been seen from farther north than Penang and the Kew specimen has notes by Griffith stating that it was collected in Malacca. King corrected Kurz’s identification in 1888 and gave the dis- tribution as Malacca only. 326 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI The collection made by Curtis in Penang which is cited above resembles Parartocarpus bracteatus in the shape and venation of the leaves, but is anomalous in that the shoot is subglabrous except for the youngest parts, and the processes on both male and female inflorescences are much shorter than is usual in this species. In the male head they are shortly conical and in the syncarp they are obtuse with a small, acute umbo. In the absence of other, similar collections this specimen is not treated as distinct, but it would be of interest to have further material from Penang. 2. Parartocarpus venenosus (Zoll. & Mor.) Becc. For. Borneo, 632. Trees, height to 35 m., not or scarcely buttressed, bark yellow to erey- brown, with numerous large lenticels. Twigs 3-7 mm. thick, rugose, appressed pubescent to puberulent, glabrescent or not. Stipules to 3 mm. long, deltoid, appressed pubescent. Leaves 4.5-23(-30) > 3-10 cm., obovate-oblong to obovate- or oblanceolate-elliptic, obtuse to short-acumi- nate, base rounded or cuneate, varying auriculate, margin entire; main veins prominent beneath, reticulum not or shallowly so; glabrous or nearly so above, thinly pubescent to glabrous beneath; lateral veins 6—15 pairs, curved; intercostals few, not parallel; dark green above, paler beneath, drying grey-green to purplish brown above, yellow-green to red-brown beneath; petiole 15-45 mm. long. Inflorescences solitary in leaf-axils. At anthesis: male head 15-30 mm. across, globose, covered by closely set, indurated, conical, obtuse, um- bonate or truncate processes, 1-1.5 %& 1-1.5 mm.; stamens 1-3 in each cavity, to 4.5 mm. long, filaments united for half to nearly all their length, anthers oblong, obtuse or apiculate, 2-2.5 mm. long; basal involucre of 3—4(—8) deltoid, short-pubescent to glabrous bracts, to 5 <5 iim,* peduncle 12-60 % 2-3 mm., short-pubescent to glabrous, often somewhat thickened below the involucre, sometimes markedly so, the involucre re- duced to an annulus; female head, surface as male head, many of the processes basally connate (often inconspicuously so) in groups of 2—4 with a lanceolate or fimbriate style exserted to 0.5—1.5 mm. between the free apices. Syncarp to 18 cm. across, subglobose, shallowly, sometimes markedly, lobed, brown, covered by closely set, indurated, spinous, um- bonate, obtuse or truncate processes, many often clearly connate in groups of 2-4, those on the lobes 1-10 & 2-5 mm., the rest smaller; wall 3-5 mm. thick; “seeds” (pericarps with a thick, stony endocarp) (1—)3 to many, ellipsoid, 15-30 « 13-20 mm.;: core c. 30 mm. across; involucre as in male head; peduncle 30-100 x c. 5 mm., short-pubescent to glabrous. Uses: the latex is used as arrow poison; the ripe fruit is said to be edible, but unripe seeds are poisonous. DISTRIBUTION: peninsular Siam, Malaya, Sumatra, Simalur, Enggano, Riouw-Lingga Archipelago, Borneo, Java, Philippine Islands, Celebes, New Guinea, Bismarck Archipelago, Aru Islands, Solomon Islands. 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, V S20 Fic. 20. Distribution of the species of PARARTOCARPUS and HULLETTIA, Ts Parartocarpus bracteatus; 2, P. venenosus, a, ssp. venenosus, b, ssp. papuanus, c, ssp. borneensis, d, ssp. forbesit; 3, Hullettia dumosa; 4, H. griffithiana., Parartocarpus venenosus has the widest range of any species within the group of genera under study and extends over the whole of Malaysia and the Solomon Islands, though it is apparently absent from the Moluccas and the Lesser Sunda Islands. Within this area, it shows considerable variation and four subspecies are recognized here with a geographical and ecological basis. specimens bearing syncarps having a knobbly surface, with processes of intermediate shape, are found which are referable on their vegetative characters to both ssp. venenosus and ssp. forbesii. T he vegetative dis- tinctions between the four taxa lie in slight differences in indumentum, leaf shape and size, and number and prominence of the lateral veins, as may be seen from the key to the subspecies. The peduncles in ssp. forbes are also rather shorter than in the other subspecies. These taxa are not absolutely separable, but most of the collections seen can be assigned to one or another without difficulty. In view of their fairly clear geographical and ecological separation, recognition at the subspecific level would seem to be justified. The distribution of ssp. forbesii is of particular interest in that it appears to be almost completely confined to peat-swamp forest. Very few other species of this group of genera are found in this habitat, apparently because they require fairly good drainage. 328 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI KEY TO THE SUBSPECIES OF PARARTOCARPUS VENENOSUS 1. Peduncles eangue: the male 12-35 mm., the female 30-40 mm. long; syn- carp processes umbonate to spinous; leaves usually to less than 12 cm. long, thickly coriaceous, glabrous, obovate-elliptic, base cuneate, with 6-10 Dale lateral WEIS. 55 sick ba 4p key we ooo a a ok Herken 5 ssp. forbesit. — Peduncles ssbrous to pubescent, the male 25-60 mm., the female 35-100 mm. long; syncarp processes truncate to spinous; leaves larger, or obovate- oblong and usually puberulent. 2. Leaves usually less than 12 cm. long, lateral veins 8-11 pairs, distinctly prominent beneath, reticulum shallowly sO; syncarp processes spinous. de Nie ER Ree oO 4 BOE RET ea Kaede oe ek ones ssp. borneensis. 2. Leaves usually more than 12 cm. long, lateral veins 6-15 pairs, less prominent, reticulum not or scarcely so; syncarp processes truncate or 3. Shoot appressed-pubescent to subglabrous; leaves obovate- -oblong to obovate- pee — cuneate or rounded, lateral 8-15 pairs, usually drying red-brown. ........................000.... SSP. VENENOSUS. Shoot Pea or nearly so; leaves obovate-elliptic, apex often rather broadly rounded, lateral veins 6-10 pairs, usually drying greyish to VEMOWISH BLOODS 6 44 and. de tab SiS evade oc-¥bdc been ssp. papuanus. ies) ssp. venenosus Artocarpus venenosa Zoll. & Mor. Natuur- en Geneesk. Arch. Neerl.-Ind. 2: 213. 1845, Flora 30: 471. 1847; Mig. in Zoll. Syst. Verz. Ind. Archip. 2: 89, 95. 1854: Mig. FI. Ind. 1(2): 289. 1859. Holotype, Java, Zollinger ord (2): oe (BM, BO, Artocarpus venenosa Zoll. var. " tylophylla Mig. in Zoll. Syst. Verz. Ind. Archip. 2: 89, 95. 1854. Holotype, Java, Zolliger 2983 (Pp); isotypes (BM, L, P, U). Artocarpus tylophylla Miq. Fl. Ind. Bat. 1(2): 289. 1859. Artocarpus callophylla Zoll. & Mor. in Teysm. & Binnend. Cat. Hort. Bog. 85. 1866, nomen nudum Artocarpus cerifera Mig. Ann. Mus. Lugd.-Bat. 3: 212. 1867. Holotype, Java, Blume 2145 (1). Parartocarpus beccarianus Baillon, Adansonia 11: 294, 875; Becc. For. Borneo, 632. 1902; Renner, Bot. Jahrb. 39: 363. 1907. cao Sarawak, Beccari PB 2557 (P): isotypes (A, K). Gymnartocarpus venenosa Boerl. Ic. — 1: 73. t. 24, 25. 1897: Koord. & Val. Bijdr. Boomsoort. Java 11: 28. Radermachia cerifera Blume ex an Ic. Bogor. 1: 73. 1897, pro syn. Gym- nartocarpus venenosa. Parartocarpus venenosus (Zoll. & Mor.) Becc. For. Borneo, 632. 1902, “venenosa’; Renner, Bot. Jahrb. 39: 362. 1907; Backer, Beknopte FI, Java 6: 12. 1948. Twigs, lower surface of leaves, peduncles and involucral bracts nee pubescent or puberulent, glabrescent or not. Leaves to 12-23 3(-30 S-10 cm., obovate-oblong to obovate-elliptic, varying to oe oblong or oblanceolate- elliptic, base rounded or cuneate; main veins promi- 1960 | JARRETT, ARTOCARPUS AND ALLIED GENERA, V 329 iy Pe Fic. 21. Leaves and syncarp surface in Parartocarpus. a—c, Leaves (X 12): a, P. bracteatus; b, c, P. venenosus, b, ssp. venenosus, c, ssp. forbesi. d-h, Surface of the mature syncarp (approx. X 1): d, P. bracteatus; e-h, P. veneno- Sus, @, SSp. venenosus, f, ssp. papuanus, g, h, ssp. forbes. nent beneath, reticulum not or scarcely so; lateral veins 8-15 pairs. Male head, peduncle c. 30-60 mm. long. Syncarp with obtuse, or truncate and often depressed processes; peduncle c. 40-100 mm. long. VERNACULAR NAMES: bulu ongko (Sundanese), purut (Javanese), Java. DISTRIBUTION: in evergreen and mixed forest to 3000 ft., tolerating a short dry season; peninsular Siam, northern Malaya, Sumatra, Simalur, Enggano, Java. Peninsular Siam. Islands off e. coast: Kaw Pa-ngan, Put 1256 (sm, 6); Kaw Samui, Put 844 (Bm, 2); Kaw Tao, Kerr 12771 (pm, ¢, 2). Malaya. Prov. WELLESLEY. Bukit Juru For. Res., KEP 9831 (KEP), 9843 (KEP, 2), Durant 330 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI KEP 9840 (xep). LANGKAWI IsLAnps, Langkawi, Gunong Raya, Abdullah KEP 42014 (KEP). Sumatra. TAPANULI. Sibolga: Aek Labuan Talang, bb 19303 ooh Kuala Badung, bb 19335 (A, BO). InprAGIRI, Danau Mengkuang, bb 27558 (A, BO, L). StmaLur. Achmad 155 (Bo, L, 6), 920 (Bo, K, L, 2); ree Tapah, Defajan, Achmad 1604 (Bo, L). EnccaNno. Bua Bua, Littjeharms 4258 (A, BO, K, L, P), 4464 (A, BO, K, L, P, infl.). Borneo. Sarawak. Beccari PB 2557 (A, XK, Pp, S) Java. Blume 2145 (L, 3); between Mt. Smeru and Mt. Kelut, Blume HB 7314 (Bo). West Java. Bantam: Gunong Karang, we: Koorde ers 8686 (Bo, K, L), i (Bo, L); Menes, Backer 7063 (xo, 9); Tjimara Udjung, Gunong Rompang, Koorders 8685 (BO, K , L). Batavia: yee Salak, Koorders 24444 (Bo), Eh (BO, er Tjiampea, Koorders 30362, 30363 (Bo). Buitenzorg: Djasinga. De Voogd, April 1941 (Bo, kK, L); Handjere, NIFS Ja 6816 (k, L); Natuur Monument, ee Iwul, NIFS Ja 1961 (Bo); Pasir Pogor, ae a van den Brink 8017 (Bo). Preanger: Palabuanratu, Koorders 5246 (Bo, L 8684 (BO, K, L, P, 2), 8688 (Bo, infl.), 12357 (Bo, L, 2), 42907 (Bo, 2). Tjilumpang, near ‘Tiidadap, Winckel 1912 (Bo). CENTRAL Java. Banjumas: Pringombo, Bandjarnegara, Koorders 8695 (Bo, K, L, @). Pekalongan: Batu near Pie sn. (BO). East Java. Besuki: Banjuwangi, Koorders 8028 (BO), 0 (Bo, L, P), 8681, 8682 (Bo, L), 8683 (BO, K, L, P); Tjuramanis, caernes 21066, 21406 (Bo), 21407 (Bo, kK, L, infl.), 22238 (L), 38336 (Bo, é). Madiun: Gunong Wilis, Ngebel, Koorders 38756 (Bo, K, L, P, 4, 2). Pasuran Tangkil-Zuidergebergte, Koorders 22441 (Bo, L), 23891 (Bo, K, L, P, SING, v). Probolinggo: Malang, Zollinger 2371, Oct. 1844 (BM, Bo, L, P, @). NUSA KAMBANGAN. Koorders 8694, 20225 (Bo, L), 24126 (A, BO, K, L, 2), 24627, ae 24728, 24729 (BO, L), 27036 (Bo, L, 6). Cultivated. Java. Bogor: Hort. , Teysmann, 1860 (1), Zollinger 2983 on LP, U8 i: Hort, Bot, cult a VII B 1a, Forman 60, Sutrisno 78 (KX, ; The type subspecies of Parartocarpus venenosus shows some variation in the outline of the leaves and rather more in the shape of the processes on the syncarp. The leaves are, however, characteristically obovate-oblong with a somewhat acute apex, and persistently puberulent beneath, at least on the main veins. The leaves are usually larger and more thinly coriaceous than in either ssp. borneensis or ssp. forbes. The venation is less i nent than in the former and the syncarp processes are never spinou The type collection of Parartocarpus beccarianus, Beccari PB 3557, has small, thinly coriaceous, narrowly obovate-oblong leaves (to 12 & 4 cm.), which are puberulent beneath, but lack a prominent reticulum. The male inflorescences have processes with obtuse apices and the syncarp (presumably at Florence) is described by Beccari in 1902 as having the surface tessellate from the depressed pyramidal apices of the “polygonal scales” or perianth apices. Except in the rather small size of the leaves, the collection agrees with P. venenosus, sensu stricto, rather than with ssp. borneensis or ssp. forbesit. ssp. papuanus (Becc.) Jarrett, stat. nov. Artocarpus riedelii Miq. Ann. Mus. Lugd.-Bat. 3: 213. 1867. Syntypes, Celebes, 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, V oa Manado, peeee HB 5778 (uw), De Vriese s.n. (L); lectotype, Teysmann HB 5778 Artocarpus eae i K. Schum. in Schum. & Hollrung, Fl. Kais. Wilhelms Land, 39. 1889. Holotype, northeast New Guinea, Hollrung 522 (8); iso- type e (B 0). Parartocarpus involucrata (K. Schum.) Warb. ex Schum. & Lauterb. Fi. eutsch. Schutzgeb. Siidsee, 267. 1900; Becc. For. Borneo, 632. 1902; Renner, Bot. Jahrb. 39: 363. 1907; Lauterb. ibid. 62: 303. 1928; Diels, ibid. 67: 174. 1935; White, Jour. Arnold Arb. 31: 82. 1950. Parartocarpus papuana Becc. For. Borneo, 633. 1902, non S. Moore, 1923. Holotype, New Guinea, Vogelkop, Beccari PP 719 (¥1); isotype (FTI). Parartocarpus riedelit Warb. ex Renner, Bot. Jahrb. 39: 362. 1907. Artocarpus woodii Merr. Philip. Jour. Sci. Bot. 3: 221. 1908; Elmer, Leafl. Philip. Bot. 2: 623. 1909. Holotype, Mindanao, Merrill 5259 (PNH, destroyed); isotypes (L, P); lectotype (P). Gymnartocarpus woodii (Merr.) Merr. Philip. Jour. Sci. 18: 52. 1921; Brown, Bull. Bur. For. Philip. 22(2): 270. fig. 19. 1921 Parartocarpus woodiit (Merr.) Merr. Enum. Philip. Pl. 2: 39. 1923 Artocarpus venenosa Zoll & Mor., Schum. Notizbl. Bot. Gart. Berlin 1: 48. 1895. Twigs, lower surface of leaves, peduncles and involucral bracts ap- pressed pubescent or puberulent, soon glabrescent except often the in- volucre. Leaves to 12-23 & 5-10 cm., obovate-elliptic, varying to oblance- olate-elliptic or obovate-oblong, base cuneate, varying to rounded or auri- culate; main veins prominent beneath, reticulum not or scarcely so; lateral veins 6-10 pairs. Male head, peduncle c. 25-35 mm. long. Syncarp with truncate, often depressed processes; peduncle 35—70 mm. long. DISTRIBUTION: in evergreen forest to 2000 ft.; Philippine Islands, Celebes, New Guinea, Bismarck Archipelago, Aru Islands, Solomon Islands. Philippines. yienee Cagayan: Klemme FB 6650 (us, 2). Isabela: San Marino, Ramos & Edano BS 47015 (sm, sinc, @), 47176 (SING). Quezon: Baler, ee e. of Castillo river, Quisumbing PNH 2518 (A, PNH, Q). Bataan: Alvarez FB 12942 (p, us, 2), Curran FB 17584 (BM, BO, L, US, ¢). Rizal: Maneja FB 23961 (pv, us, 2); Orind, Loher 6946 (x, infl.). Laguna: Tamesis FB 11993 (L, Pp, us, 2); Calauan, McGregor moe 12392 (BM, P, 2); Mt. pea Forestry School FB 20140 (BM, L, P, é), Ramos oe (u, US, 2), Villamil FB 20394 (us, 2), Whitford BB Gor (BM, 2); ene Los Banos, Elmer 18291 (A, BM, K, L, P, 6); San Antonio, ee BS 15051 (e). Camarines: Alvarez FB 21453 (BM, P, US, 2), 23748 (A, BO, 2), Hsia FB 21107 (us, infl.); Camarines Sur, Ahern 41 (Bo, us, infl.), Alambra FB 28088 (a, Pp, 2). Albay: Rapu-Rapu Island, Vidal 3837 (A, K, 2). Sorsogon: Irosin, Mt. Bulusan, Elmer 14650 (A, BM, K, L, 2). Samar. Sherfesee et al. FB 21084 (us, 2). Leyte. Dagami, Ramos BS 15184 (K, us, 6, 2). Mun- DANAO. Surigao: Ramos & Pascasio BS 34684 (a, P, 2). Iinauat Island: Ramos & Pascasio BS 32540 (a, 2). Bucas Grande Island: Merrill 5259, Oct. 1906 (L, Pp), Ramos & Pascasio BS 35056 (Bo, L, infl.), 35092 (a, K, US, 2). Celebes. NorTH PENINSULA. Manado, DeVriese (1); Manado, Ratahan, ee HB 5778 (go, L, U); Minahassa, Koorders 19050 (Bo, 6), 19315 (Bo, L). CENTRAL CELEBES. Malili: Kawata, N/JFS Cel./II-409 no. 10 (Bo, K, S32 JOURNAL OF THE ARNOLD ARBORETUM [Vou Seer L, 3); Lawoli, 66 23251 (Bo); Usu, NIFS Cel./II-324 (a, L, SING), 409 no. 409 (Bo, L, 6); Mantano Meer, Kjellberg 2805 (so, 2). Masamba, Mina, bb 24504 (BO, L, SING). New Guinea. VoceLKop. Bomberai: Fakfak, Lundquist 266 (L); Rauna, bb 22536 (Bo). Inanwatan, bb 32657 (Bo, L). Manokwari: Andai, Beccari PP 719, 1872 (F1, 6, 2); Momi, bb 33492 ae Fe Oransbari, Brouwer BW 2584 (1); Warnapi, Kostermans 408 (BO, K, L). Sorong: Warsamson, 25 km. e. of Sorong, Schram BW 2923, 5949 (L). DurcH ete NEw GuINEA, Hollandia: Versteeg BW 4825 (1, 6, 2); Holtekang, Schram BW 1508 (L); Idenburg river, a hard Camp, Brass 13547 (a, infl.), Brass & Versteegh 13547A (a, 2); Tam Schram BW 2723 (x, L, 2); mouth of Tami river, ial BW 2678 (xk, L, infl.), 2811 (XK, L, infl.), 2812 (1, infl.), Versteegh BW 3805 (L). DutcH Sout New Guinea. Merauke, Bot river, halfway between eae and Lake Wam, Van Royen 4737 (kK, L). Papua. Central Division: Koitaki, Carr 12623 (A, BM, K, SING, 6, 2). Northern Division: Hydrographers Range foothills, Hoogland 3855 (A, BM, L, Western Division: Lake Daviumbu, Middle Fly river, Brass 7476 (A, L, 6). MANDATED TERRITORY OF NEW Gums NEA. Madang Dis- trict: Constantinhafen. Hollrung 522, Feb. 1887 (B, Bo, 6); Gogol valley, hills ne. of Mawan village. Hoogland 4891 (kK, 2); Ramu valley, c. 5 mi. se. of Faita airstrip, Saunders 274, 539 (1). Morobe District: near Finschhafen, Hellwig 286 (sinc); Umboi Island, White NGF 9646 (x, 2). Sepik District: Leder- mann, 1912-13 (stnc). New Britain. West Nanakai, Gorea village, near Cape oskins, Floyd 6448 (kK, ¢). ScHouTeN IsLanps. Biak, bb 30684 (Bo, kK), 30830 (BO, L, SING). JAPEN. bb 30281 (Bo, L, SING); Mentebu, bb 30228 (a, BO, L, SING, @); Rendawaja, Malinka BW 7014 (1). Satawatr. Kaloal, Koster BW 4243 (1). Aru Istanps. Kobroor: Dosinamalu, bb 25319 (a, BO, L, SING, infl.), Buwalda 5096 (A, K, L, SING, infl.). Trangan: Lutor, Beccari, June 1873 (I, 6); Ngaibor, bb 25453 (a, BO, sING, 2), Buwalda 5420 (A, K, L, PNH). Solomon Islands. BoUGAINVILLE. Buin, Marmaromino, Kajewski 2199 (A, K, infl.). TREASURY ISLAND. Guppy 55 (x, infl.). SANTA IsaBEL. Kalina bay, Walker BSIP 238A (A, kK, ae GUADALCANAL. nee river, Kambou river, Walker BSIP 238 (A, kK, Although ssp. papuanus is widely distributed, it is rather constant in its characters. It is nearest to ssp. venenosus but is distinguished by the con- sistently subglabrous mature shoot and by the rather few lateral veins and broadly rounded apex of the leaves, which usually dry yellowish or greyish green. The syncarp is apparently always tessellate. ssp. borneensis (Becc.) Jarrett, stat. nov. Parartocarpus borneensis Becc. For. Borneo, 634. 1902; Renner, Bot. Jahrb. 39: 362. 1907. Holotype, Sarawak, Beconn PB 2005 (¥1); isotypes (FI, K). Twigs, lower surface of leaves, peduncles and involucral bracts ap- pressed pubescent when young, usually persistently puberulent. Leaves to 8.5-12 & 5-7 cm., obovate-oblong, base rounded or broadly cuneate; main veins distinctly prominent beneath, reticulum shallowly so; lateral veins 8-11 pairs. Male head, peduncle 25-35 mm. long. Syncarp with spinous processes; peduncle 40-80 mm. long. 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, V RS DISTRIBUTION: in evergreen forest on sand or loam to 200 (—1200) ft. in low undulating or hilly country; Borneo. Borneo. SARAWAK. Mt. Mattang, Vallombrosa, Beccari PB 2005, June 1866 (K, FI, 29); Sungei Semenggoh For. Res., Wyatt-Smith KEP 79306 (x, L, @). Brunei. Andulau For. Res., Ashton BRUN 585 (K, KEP, L, 2), Smythies et al. SAN 17500 (K, KEP, L, 6), Wyatt-Smith KEP 80076 (KEP, 2). EAST AND NORTHEAST BorNEO. Central Kutei: Belajan river region, Kostermans 10258 (Kk, @). W. Kutei: Longbleh, bb 16047, 16054 (A, Bo, L); Mujup, bb 16782 (a, Bo, L, infl.); near Tandjong Isui, Endert 1898 (xk, L). Tidung: Birik, bb 17930 a BO, L, SING, 6). BritisH NorTH Borneo. Kabili-Sepilok For. Res., Enggoh KEP 48970 os 6). Laspuan. Motley 254 (x, infl.). This subspecies is readily distinguished by the consistently spinous syn- carp and the small, obovate-oblong, rather thickly coriaceous, usually thinly pubescent leaves with rather prominent venation. ssp. forbesii (King) Jarrett, Jour. Arnold Arb. 41: 137. 1960. Artocarpus forbesiti King in Hook. f. Fl. Brit. Ind. 5: 539. 1888; King, Ann. Bot. Gard. Calcutta 2: 7. ¢. 1A. 1889; Ridley, Fl. Malay Penin. 3: 352. 1924: Moore, Jour. Bot. 63, Suppl. 112. 1925. Syntypes, Malaya, King 10829 (CAL, not seen; duplicates examined, K, P), Sumatra, Forbes 3080 (CAL, not seen; aunlioniee examined, BM, L). Parartocarpus errese Becc. For. Borneo, 634. 1902; Renner, Bot. Jahrb. 39: 363. 1907. Holotype, Sarawak, Beccari PB 673 (¥F1); isotypes (A, BM, FI, Kee ieee triandra J. J. Smith, Bull. Jard. Bot. Buitenzorg III. 4: 233. t. 6-8. 1922. Syntypes, era Beguin 316, 536, 583, Grashoff 798 (Bo); lectotype Beguin 583 (BO Parartocarpus triandra J. J. Smith, Bull. eee us Buitenzorg III. 6: 80. 1924; Browne, For. Trees aes 35 Twigs, and involucral bracts appressed pubescent or puberulent, soon glabrescent. Leaves to 8-13 & 4-6 cm., obovate-elliptic, base cuneate; main veins prominent beneath, reticulum not or scarcely so; lateral veins 6-10 pairs. Male head, peduncle 12-35 mm. long. Syncarp with spinous or umbonate processes; peduncle 30-40 mm. long. VERNACULAR NAME: tenggajun (Malay), Sumatra, Borneo. DISTRIBUTION: in evergreen forest to 2500 ft., usually in low-lying peat-swamp forest; western and southern Malaya, eastern and southern Sumatra, Riouw-Lingga Archipelago, Borneo. Malaya. Perak. Ulu Bubong, King 10829, Aug. 1886 (K, P, 2). Jou 8th mile Kota Tinggi-Mawai road, Corner, Feb. 1935 (SING); Gunong See Corner, Jan. 1937 (SING); Pengkalan Raja, Ngadiman SFN 36661, 36682 (SING, 9); Sungei Kayu, Mawai-Jemalaung road, Kiah SFN 32185 (x, SING, @). PENANG. Ayer Hitam For. Res., Strugnell KEP 49702 (KEP, 2). SINGAPORE. Jurong, Corner SFN 21845 (A, BM, BO, K, SING, 8, 2), 28147 (K, sING); 15th mile Jurong, Corner SFN 26194 (a, K, SING, infl.). Sumatra. East Coast. Asahan, Masihi For. Res., Krukoff 4124 (A, Bo, L, 334 JOURNAL OF THE ARNOLD ARBORETUM [VOL XLI SING, @); Benkalis, Sengoro, Beguin 536 (Bo); Benkalis, Sungei ae Beeuin 583 (Bo, L, 6, 2); Benkalis, Tamansari, Beguin 316 (Bo, L ; Batu, Sungei Palas, bb 10642 (po); P. Mendal, Kelumang, bb 124 72 ee SING, infl.): P. Tebing Tinggi, bb 12926 (xo, infl.). aa ee Belimbing, bb 28537 (a, Bo, L); P. Gelang, bb 29150 (aA, Bo, L). PALEMBANG. Banjuasin and Kubustreken, Grashoff 798 (Bo, L, 4); Muara Menai. River Rawas, Forbes 3080, 1880 (BM, L, 2). Riouw-Lincca ArcHrp. Karimon: Rutan, Simulur, bb 6303 (so). P. Sinkep: Manggu, Ketjil, bd 5366 (so, infl.). Borneo. SARAWAK. Binatang, Surong, Daro For. Res., Tahir 9715 (kK, L infl.); near Kuching, Beccari PB 673, Nov. 1865 (A, BM, FI, K, 6). BRUNEI. Sebatu-Arur Mangan watershed, Astian BRUN 348 (K, KEP, 2), WEsT BorNEO. Kubu, Baru, Telok Meranti, bb 8041 (Bo); Sambas, Paloh, bb 13891 (Bo, 2); Simpang, Dienu, bb 12699 (Bo). SOUTH AND SOUTHEAST Borneo, Lower Dajak, Danau Rawah, bb 13483 (Bo); Sampit, Sungei Kereng Bindjai, Sabangau, bd 7941 (Bo). British NortH Borneo. Tambunan, Wvyatt-Smith KEP 80436 (K, KEP, infl.). P. NUNUKAN. Kostermans 8664 (K, L, 2); S. Simengkadu, Meijer B 2372 (K, KEP, @). This subspecies may be distinguished from the type and from ssp. borneensis by the obovate leaves, the consistently subglabrous adult shoot, and the shorter peduncles. The syncarp processes are rather variable in shape, overlapping those of both these entities. However, as noted above, the ecological preferences of ssp. forbesii seem to be quite distinct from those of the other subspecies. From ssp. papuanus, ssp. forbesii may be distinguished by the smaller, more thickly coriaceous leaves drying red- brown, and by the svncarp processes. Artocarpus forbesii was reduced by Beccari to Parartocarpus venenosus in 1902, and King’s epithet has not been taken up previously in the latter genus, most specimens having been determined as P. triandra J. J. Smith SPECIES EXCLUDENDAE PARARTOCARPUS PAPUANA S. Moore, Jour. Bot. 61, Suppl. 52. 1923, non Becc., 2 = Prainea papuana Becc. For. Borneo, 635. 1902. PARARTOCARPUS sp., Benth. & Hook. f. Gen. Pl. 3: 375. 1880 (Beccari PB 667) = Prainea frutescens Becc. For. Borneo, 635. 1902. Hullettia King in Hook. f. Fl. Brit. Ind. 5: 547. 1888; King, Ann. Bot. Gard. Calcutta 5(2): 163. ¢. 197. 1896; Engler & Prantl, Nat. Pflanzenfam. Nachtrage II-IV. 122. 1897; Boerl. Handl. Fl. Ned. Ind..35. 338, 272. 1900; ora 7 Malay Penin. 3: 358. 1924; Jarrett, Toa Arnold Arb. 9, 11. fig. 2, g-k, he 3, joel. 1959. LECTOTYPE SPECIES: ieee Ba eek (Karz) King Kurzia King in Hook. f. Fl. Brit. Ind. 5: 479. 1888, in as Shrubs to small trees. Leaves spirally arranged, simple, entire, penni- nerved, thinly to thickly coriaceous, lower epidermis with numerous cells having strongly thickened, pitted walls and bearing hairs or not, the lower 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, V 355 surface hence hispid-pubescent, scabrid or smooth, minutely punctate be- neath in sicco from the whitish stomata; epidermal gland-hairs superficial, long-stalked, heads globose, unicellular; spongy mesophyll compact, with- out resin-cells; juvenile leaves elongate. Stipules paired, linear, lateral, scars small, round. Inflorescences unisexual, capitate, head pulvinate or obconical at an- thesis, becoming subglobose in syncarp, pedunculate, solitary or paired in leaf-axils; stamens or ovaries sunken in cavities closely set on the upper surface of the receptacle; receptacle naked (perianths and inter- floral bracts lacking), fleshy; involucre of 3-6 obtuse to lanceolate, fleshy, flattened bracts present. At anthesis: anthers or stigmas exserted through perforations in the upper surface of the receptacle; involucre marginal; male head with up to 30 flowers; stamens paired in each cavity with the filaments partially or almost completely united, anthers 0.7-1 mm. long, long-exserted; female head with up to 6 flowers, ovaries solitary in each cavity, unilocular, the style apical with a short-exserted, capitate stigma, the ovule apical. Mature syncarp formed by the enlargement of the entire female head, with 2-6 flowers forming fruit and completely filling the receptacle; involucre equatorial or becoming sub-basal through the greater expansion of the upper surface of the receptacle; mature ovary pergamentaceous, scar left by the style apical, the seed large, attached apically, testa membranous except the thickened apical cap, endosperm none, embryo straight, orientation longitudinal, cotyledons equal, appressed faces at an angle of 0-90° to median plane of ovary, radicle and plumule small, basal, DISTRIBUTION: southern Tenasserim and peninsular Siam, Malaya, Sumatra (Indragiri). The genus Hullettia was described by King in 1888 with the two species recognized in the present revision, namely, a species from Tenasserim based on the wrongly assigned Dorstenia griffithiana Kurz, and a new species from Malaya, Hullettia dumosa King. The new genus appeared in the key to the Urticaceae in the Flora of British India as “Kurzia,” but, finding that this name was preoccupied [by the algal genus Kurzia Mar- tius, Flora 53: 417. 1870], King changed it in the systematic treatment to Hullettia, after R. W. Hullett, a schoolmaster in Singapore who served on the Gardens Committee there. In 1896 King published a somewhat more detailed account, with a plate of H. dumosa, in “A Century of New and Rare Indian Plants.” As with Prainea, which he described in the same works, King wrongly stated that the ovule was basal and erect and hence that the position of Hudlettia was in the tribe Conocephaleae (= subfamily Conocephaloideae). It was rejected from this group by Renner in his study of the leaf anatomy of the Artocarpoideae and Conocephaloideae, but he omitted the genus as being of doubtful affinity (Bot. Jahrb. 39: 419, 1907). Hullettia has otherwise been mentioned only in the Nachtrage to the Natiirlichen Pflanzenfamilien (1897), in the floras of Boerlage (1900) and Ridley (1924), and in Burkill’s Dictionary (1935). 336 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI The morphology of the inflorescences has been discussed in the intro- ductory paper of this series (Jour. Arnold Arb. 40: 1-29. 1959). The attachment of the ovule has been found to be apical in the two species both at anthesis and at maturity, which indicates (with the erect stamens) that Hullettia is related to the genera at present placed in the Artocar- poideae. The capitate inflorescences resemble those of Parartocarpus in being clearly involucrate (with fleshy marginal to sub-basal bracts) and in having the stamens and ovaries apparently enclosed in cavities of the receptacle rather than in perianths. Externally, however, the inflorescences of the two genera have a very different aspect, since those of Parartocarpus are armoured from indurated processes, whereas those of Hullettia have a smooth, fleshy, pubescent surface, with perforations leading to the recep- tacular cavities. Perianths and interfloral bracts appear to be entirely absent in the latter genus. The mature syncarp of Hudllettia differs internally from that of Pararto- carpus in having pergamentaceous, not indurated, pericarps, and straight, rather folded embryos. However, the testa has a eis developed, thickened cap occupying an apical position in the pericarp, as in Parartocarpus. The tips of the cotyledons lie in contact with this cap, which may be compared with the similar thickening of the testa in Prainea. The latter, however, is basal (corresponding with the sub-basal attachment of the mature seed) and the orientation of the embryo relative to the pericarp is the reverse of that found in Hullettia, the radicle being apical. It should be noted that this view of the inflorescence structure of Hul- lettia is in complete accord with the account of ‘“Dorstenia sp.” given by Griffith (published in 1854), on which Dorstenia griffithiana Kurz (= Hullettia griffithiana (Kurz) King) was largely based. Griffith apparently dissected fresh inflorescences and the details given by him of the pericarps and seeds support the present description, which is based on an examina- tion of dried syncarps (boiled up in water) of both species. King, on the other hand, regarded the stamens and ovaries as being enclosed in tubular perianths, which were connate with each other and the receptacle. How- ever, as in Parartocarpus, no sign of fusion between adjacent perianths has been found in the walls separating the receptacular cavities. In view of the other characters indicating an affinity between the genera, it seems more probable that the inflorescence structure of Hullettia is homologous with that of Parartocarpus. In the latter the walls have been shown to be largely, if not wholly, of intercalary origin. No material has been available for the study of development in Hudllettia and it is possible that vestiges of perianths, indistinguishable at later stages, are incorporated in the surface of the receptacle. The leaves of Parartocarpus and Hullettia agree in having long-stalked, superficial gland-hairs with unicellular heads. As was noted in the intro- ductory paper, Hullettia is distinguished by the numerous enlarged cells with strongly thickened, pitted walls in the lower epidermis. However, it should be observed that the minutely punctate appearance of the under 1960] JARRETT, ARTOCARPUS AND ALLIED GENERA, V Sy Fic. 22. The syncarp in Sa ee a, rae head; b, pericarp, side view; c, pericarp cut in median ne d embryo moved; d, embryo from 1), p ” same view; e, embryo seen from ‘x (all approx. a surface of the leaves when dry is due, not to these cells, as was stated there, but to the whitish guard cells of the stomata. The distribution of Hudlettia as a genus is very limited. Hullettia grif- fithiana is known from southern Tenasserim and peninsular Siam, while H. dumosa occurs quite widely in Malaya and has recently been collected in central Sumatra. Both species may well be more common than is indi- cated by the number of collections, since they are small trees or shrubs which are probably relatively inconspicuous except at the time of fruiting. The fleshy, orange syncarps presumably attract arboreal mammals or birds in the same way as the fruits of those species of Artocarpus that have fleshy syncarps. The two species in the genus are readily separated on inflorescence characters, especially those of the mature syncarp. The vege- tative characters show some overlapping, but there is usually no difficulty in assigning sterile collections to one or other of the species. KEY TO THE SPECIES OF HULLETTIA Male head c. 10 mm. across, peduncle to 20 mm.; syncarp with obtuse to obso- lete involucral bracts, peduncle to 45(-55) mm.; leaves smooth to scabrid beneath, base cuneate to rounded, petiole to 50 mm. .......... 1. H. dumosa. Male head to 7 mm. across, peduncle 25-45 mm.; syncarp with lanceolate involucral bracts, peduncle c. 75 mm.; leaves hispid-pubescent to scabrid beneath, base narrowly and abruptly rounded or auriculate, petiole to 18 mm. ........ 2 1. Hullettia dumosa King in Hook. f. Fl. Brit. Ind. 5: 547. 1888; King, Ann. Bot. Gard. Calcutta 5(2): 163. #. 197. 1896; Ridley, Fl. Malay Penin. 3: 358. 1924; Burkill, Dict. 1202. 1935. Syntypes, Perak, King 3959, Scortechini s.n. (CAL, not seen; duplicates examined, xk). Shrubs or small trees, height to 10 m. Twigs 3-6 mm. thick, acutely rugose, thinly pubescent to subglabrous, hairs pale yellow, subappressed. 338 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI Stipules 3-10(-15) & c. 1 mm., puberulent. Leaves 10-40 & 4-12 cm., elliptic to oblong- or obovate-elliptic, rarely oblanceolate-elliptic, with an acumen to 2.5 cm. long, base cuneate to rounded, margin entire; main veins prominent beneath, reticulum less so; smooth on both surfaces or scabrid beneath; lateral veins 8-12(-15) pairs, curved; intercostals few, often one median between and parallel to the lateral veins, mid-green, drying greyish to yellowish green, paler beneath and minutely punctate. venation straw-coloured; petiole 10-35(—50) mm. long. Inflorescences solitary or paired in leaf-axils. At anthesis: male head c. 10 mm. across, obconical or pulvinate, short-pubescent, with c. 4 marginal, subacute, fleshy, involucral bracts, c. 3 * 5 mm.; stamens 3 mm. long, paired, filaments united to just below anthers, anther-cells oblong, 1 mm. long; peduncle 10-20 x 1 mm., short- pubescent, often enlarging gradually into receptacle: female head with clavate stigmas exserted 0.5 mm. through perforations in the surface. Syncarp to 5.5 cm. across, depressed globose, orange with orange-yellow flesh, drying olive-brown, with surface smooth, velutinous, with c. 4 equatorial to sub-basal, obtuse, fleshy, involucral bracts to 3 & 12 mm., often nearly obsolete; wall c. 2 mm, thick; “seeds” (pergamentaceous pericarps) 2-5, aaa 22. 4 1A i, . peduncle 15- 45(—55) & 1.5 mm., short- spibescen DIsTRIBUTION: in forest to 6700 ft.; Malaya, Sumatra (Indragiri). Malaya. PERAK. Scortechini 656 («), [Scortechini | sm. (K, 2); Batu Kuran, Curtis, 1892 (stnc); Bujong malacca, Curtis, 1892 (si ie Ridley 9617 (SING, 60% * 84% > 64% Paspalum orbiculare i a. 25. = 29 Sacctolepis contracta ig = 4 277 = 28 Chrysopogon aciculatus 9 < 20 2 oe Paspalum conjugatum Go 202 0 = |] Ferns Nephrolepis exaltata oO =: 0 Ca 0 Stenoloma chinensis Ay 0 0) Pteridium aquilinum Oe eae i yee Shrubs Stachytarpheta cayennensis 15 LS 24. = 26 Psidium ee Ort 4 = 4 Lantana cama a ns 8 =) 7 Melastoma Sn aoe Ae) Se and Psidium cattleianum Cheey p> Ohl On. a e Elephantopus mollis Sen eo 2 = 4 Centella asiatica Sipe ore ee) Oi: “eae 28 Cassia leschenaultiana 1 , decrease; <, increase; =, no change. examination except as a casual occasional plant on disturbed areas. The increase in Elephantopus is not impressive, partly because of the biennial habit of the plant and partly because the change was more pronounced in the more moist Setaria—Nephrolepis habitat type. Not all the plants appeared to be of biennial habit since some had flowered at the time of examination. The increase of Cassia following fire is large and involves both numbers of plants and size of individuals. It is possible that a slight seasonal effect tends to reinforce this difference. Cupkea was only mildly stimulated by the burning. Chrysopogon has been observed to thrive on the exposed and eroded south slopes and was predictably improved in its competitive position by fire. Passiflora foetida increased slightly in frequency in both associations. 402 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLI Species showing a strong tendency to decrease in cover or frequency following fire in the two associations studied were Lantana camara, Steno- loma chinensis, Melastoma malabathricum (Fic. 6), and Paspalum con- jugatum. To a lesser degree, Spathoglottis plicata showed a decrease in frequency while Setaria geniculata declined in cover only. Stenoloma was completely obliterated by the fire in places where it had been re- corded a year before, while the woody Lantana and Melastoma were markedly reduced in cover but observation revealed that they were by no means killed. Species that were little affected by burning include guava, which was found only in the Setaria—Nephrolepis community and which was ob- served elsewhere to be severely damaged by fire, Centella asiatica, Saccio- lepis contracta, Paspalum orbiculare, and Stachytarpheta cayennensis which increased much less than expected. Following the simulated strip-mining and planting of the stripped alea with forage, field and fruit crops, a number of species new to the area were observed. An inventory of plants new to this area and appearing in the disturbed areas was taken and these are designated as “invaders” in the APPENDIX. Since all the fruit and ornamental plants were brought in from elsewhere, some as potted plants, undoubtedly some of the in- vaders came in with the plants. Others such as Solanum nigrum (popolo) may have already been in the soil in dormant condition awaiting proper conditions before breaking their dormancy and making their appearance. Stripping, followed by field preparation, apparently supplied the necessary conditions for the germination of long dormant seeds. SUMMARY AND CONCLUSIONS The evidence from the present study points to conclusions similar to those from the Jamaican study of Howard and Proctor (1957) i that none of the species studied shows a definite affinity for ecologic niches characterized by high aluminum content of the soil. The species present are nearly all those adapted to a generally moist environment with warm equable temperatures, acid, infertile soils, and a shallow root zone. Since the aluminum content of the soil increases with depth while the rooting of the dominant plants is unusually shallow, it must be inferred that they are not dependent on high levels of aluminum for survival, but rather are restricted in their root development to the organic and more fertile upper horizons. In addition, the major changes in the vegetation pattern apparently are not governed by alterations in the soil profile, but rather are related to slope, exposure and effective rain- fall. The high degree of disturbance of the natural pattern through fire, grazing and manipulation of the cover has further removed the expression of the natural vegetation determinants. Despite the fact that of aluminum (Dicranopteris linearis, Paspalum orbiculare, Melastoma malabathricum and sometimes Nephrolepis exaltata) and are growing on 1960] MOOMAW & TAKAHASHI, VEGETATION ON GIBBSITE 403 soils classed as aluminous, it can only be said that they are tolerant of the condition and are capable of accumulating the aluminum ion. The most important single factor in the present distribution of species in the area is the locus and date of introduction of each of them and their competitive relationship both to the endemic species which they have replaced and to each other. A tabulation of the species present reveals only 5% to be endemics, while 30% are indigenous species of fairly wide Island distribution (see species list), and the remainder are introductions. The indigenous and endemic species occur more frequently in the gulch bottoms and protected areas than on the exposed and disturbed slopes and ridges, clearly indicating their remnant nature. Other vegetation anomalies, such as the presence of Scaevola frutescens (beach naupaka) and other invaders, can be attributed to the importation of sand, coral and cinders at various times for road and building purposes. Few of the species in the initial stages of succession on the Jamaican bauxites are the same as those observed on the Kauai site, but the later, more persistent weedy plants include Lantana, Stachytarpheta, and Sida which were present in Jamaica also. Several of the grasses used for re- vegetation in Jamaica are the same ones favored in the Wailua Reserve, such as Digitaria decumbens (Pangola grass), Melinis minutiflora (mo- lasses grass), and Panicum purpurascens (Para grass). Panicum maxi- mum (Guinea grass) was useful in the Jamaican study, while the site on Kauai was considered too wet for the successful establishment of any large-scale planting of Guinea grass. The overburden removed from the mining sites was observed to be unusually fertile in Jamaica, as it was in Hawaii, partly owing to the high content of organic matter and partly owing to the content of weed seeds, roots and living plant parts. Studies with the major plantation and pasture crops in Hawaii have shown that successful production can be attained with applications of fertilizer in fairly large amounts. Six to eight hundred pounds per acre of complete fertilizer, with or without lime, will give good results with Digitaria decumbens and Desmodium intortum on either the mined sub- strate or the returned topsoil and produce a rapid vegetative cover. Sugar cane and pineapple can be grown at plantation levels of production or above, if well fertilized. REFERENCES CITED BRAUN-BLANQUET, J. 1932. The study of plant communities. McGraw-Hill Book Co. Brown, D. 1954. Methods of surveying and measuring vegetation. Comm. Bur. Past. and Field Crops Bull. 42. CHENERY, E. M. 1948. Aluminium in the plant world. Kew Bull. 1948: 163- 135 . 1949. Aluminium in the plant world. Jbid. 1949: 433-466. Cine, M. G., et al. 1955. Soil survey of the Territory of Hawaii. U.S. D. A. Soil Survey Series 1939, No. 25. 404 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI Crocker, R. 1952. Soil genesis and pedogenic factors. Quart. Rev. Biol. 27(2)2 139-168. DEGENER, O. 1932—date. Flora Hawaiiensis 1-5. Published by the author. Howarp, R. A., and Proctor, G. 1957. The vegetation on bauxitic soils in Jamaica. Jour. Arnold Arb. 38: 1-41, 151-169. MacDonaLp, G. A., Davis, D. A. and Cox, D. C. 1954. Kauai, an ancient Hawaiian volcano. Volcano Letter No. 526: 1-3. Monk, E. C. J., and Van Baren, F. A. 1954. Tropical soils. Interscience Pub- I Moomaw, J. C. NakKAMuRA, M. and SHERMAN, - D. 1959. Aluminum in some Hawaiian nlants, Pacific Science 13: 335-34 NEAL, Marie C. 1948. In gardens of Hawaii. - P. Bishop Museum Spec. Ripperton, J. C., and Hosaka, E. Y. 1942. Vegetation zones of Hawai. Hawaii Agr. Exp. Sta. Bull. 89. SHERMAN, G. 1958. Gibbsite-rich soils of the Hawaiian Islands. Hawaii Agr. Exp. Sta. Bull. 116. ———. and Kaneutro, Y. 1948. The chemical composition of Hawaiian forest floors. Hawaii Agr. Exp. Sta. Bien. Rep. 1946-48 WEATHER BuREAU. 1959. Hawaii, annual summary, 1958. Climatological data 54(13): 154-160. Wess, L. J. 1954. Aluminium accumulation in the Australian-New Guinea flora. Austr. Jour. Bot. 2: 176-196. EXPLANATION OF PLATES PLATE. 1 Fic.. 1, “The SR nna association of the ridge-top, Wailua Game Refuge, Kauai. Sporobolus capensis is seen along the roadway. Shrubs in the background include Eugenia a Pandanus odoratissimus, Psidium guajava, and Melastoma malabathricum, an aluminum accumulating species. Fic. 2. The Setaria-Nephrolepis association on a north-facing slope. The “Ama’uma’u” tree fern, Sadleria cyatheoides, is seen in the middleground. Setaria and Pas- palum are the dominant grasses PLATE II Fic. 3. Dicranopteris linearis, known to be an aluminum accumulating fern, occurs primarily on the lower slopes and valley bottoms. Fic. 4. The vegetation of the protected valleys is dominated by Psidium guajava and Pandanus odora- tissimus. The understory contains high percentages of ferns (Dryopteris) and Oplismenus hirtellus, the shade-tolerant “basket grass.” Most of the indigenous species were found in this association. PLATE III Fic. 5. A plot in the Setaria—Paspalum association following fire. A young orchid plant, Spathoglottis plicata, occurs on the left side of the plot, while the broad leaves are those of Elephantopus mollis. The stoloniferous grass on the bare soil is Chrysopogon aciculatus. Fic. 6. Regeneration of Melastoma malabathricum six months following fire. Top growth is completely killed back. Note that the Setaria has already produced mature inflorescences PLATE I Jour. ARNOLD ARB. VoL. XLI AKAHASHI, VEGETATION ON GIBBSITIC SOILS Moomaw & T Jour. ARNOLD Ars. VoL. XLI PuaTe II Moomaw & TAKAHASHI, VEGETATION ON GIBBSITIC SOILS Jour. ARNOLD Ars. VoL. XLI Pirate III Moomaw & TAKAHASHI, VEGETATION ON GIBBSITIC SOILS 408 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI APPENDIX. Species List with Occurrence and Origin Indicated. Symbols: A, abundant; C, common; U, uncommon; rare; I, indigenous; E, endemic; *, known in Hawaii at that date; P, ee (usually by Board of Agriculture and Forestry); +, invader of stockpiled topsoil or of exposed subsoil. OCCURRENCE ORIGIN INDIGENOUS SPECIES OR RIDGE Stope VALLEY ENDEMIC INTRODUCED FERNS AND FERN ALLIES Asplenium insiticium Athyrium microphyllum Blechnum occidentale Cibotium chamissot Dicranopteris linearis 1923 Trop. Am.* Ore Elaphoglossum reticulatum Lycopodium cernuum U Microlepia setosa Microsorium scolopendria N ao biserrata var. U furca N Ne exaltata C A Nephrolepis hirsutula R Phlebodium aureum Pleuropeltis thunbergiana Psilotum nudum Pteridium aquilinum var. decompositum Sadleria cyatheoides Stenoloma chinensis C I ee eee es es I. Pantropic I Trop. Am. Gq 64 See Groat eaecae == —_ vu C4 bed GYMNOSPERMAE Araucariaceae Araucaria excelsa U P, Norfolk Island MONOCOTYLEDONEAE Araceae Colocasia esculenta U ut Commelinaceae Commelina diffusa CC 1888, Pantropic * Cyperaceae Cladium meyenit Cyperus cylindrostachyus Cyperus kyllingia f. humilis Cyperus rotundus Fimbristylis diphylla U U Widespread aq 1898, Madagascar * 1900, Pantropic * 1850 * ae 1960] MOOMAW & TAKAHASHI, VEGETATION ON GIBBSITE Species List (Continued) 409 OCCURRENCE ORIGIN INDIGENOUS SPECIES OR RipcE SLopeE VALLEY ENDEMIC INTRODUCED cence lavarum R Scirpus va U 1888, N. Am.* Dioscoreaceae Ee ee bulbifera R I Gra Ventas affinis C C 1912, Trop. Am. Chloris inflata U + 1906, Trop. Am.* Chrysopogon aciculatus U Cc iT Coix lacryma-jobi U 1888, Indonesia * Cynodon dactylon U + 1835, Old World * Digitaria henryi 16) 1932, Formosa Digitaria violascens U + 1917, Trop. Asia Eleusine indica U + 1840, India * Oplismenus hirtellus C 1841, Trop. Am. Panicum purpurascens U 1902, Africa Paspalum conjugatum C C Cc 1840, me Guiana * Paspalum orbiculare A A U 1888 Asi Paspalum urvillet U 190625, i Sacciolepis contracta C A + 1906, Indonesia Setaria geniculata A A 1851, Euro Setaria verticellata U + 1860, saa a nal olus capensis U 1903, Africa Liliacea Gee terminalis U I Orchidaceae pinnae plicata CG e ? Asia—Malaya Pandanac F Ste arborea U E Pandanus odoratissimus G C I Pandanus variegatus U P Taccaceae Tacca leontopetaloides U I Zingiberaceae Zingiber zerumbet U U I India an cena Amaranthacea ean hybridus U I + Pantropic Amaranthus spinosus U + 1900, Trop. Am. Amaranthus viridus U + Trop. Am. Anacardia ae eee U + 1917, S. Am. Caryophyllaceae Drymaria cordata U 1900, Asia—Malaya 410 JOURNAL OF THE ARNOLD ARBORETUM Species List (Continued ) [ VOL. XLI OCCURRENCE ORIGIN INDIGENOUS SPECIES OR RipGE SitopE VALLEY ENDEMIC INTRODUCED Casuarinaceae Casuarina equisetifolia U P 1895, Malaya Convolvulaceae mea pes-caprae R IT + Compositae Bidens pilosa U + 1864, Trop. Am.* Eclipta prostra U Widespread Elephantopus mals A A : + 1931, Trop. Am. Emilia soncht c + Trop. Asia Erechtites valerianifolia U + 1870, N. Am Erigeron canadensis C R + ?N. Am. Pluchea odorata U 1931, 5. Am.* Sonchus oleraceus U ? Europe Vernonia cinerea U + Trop. Afr., Asia Euphorbiac yar ae U I Euphorbia hirta U + 1888, Pantropic * Euphorbia hypericifolia U + Pantropic Euphorbia thymifolia U + Widespread Goodeniaceae Scaevola frutescens var. U U I sericea Scaevola gaudichaudiana U il I Lauraceae Cinnamomum cam phora R ? China Leguminosae Abizaia moluccana U 1917, Malaya Cajanus cajan U P 1909, India Canavalia sericea U I Cassia leschenaultiana c Cc 1888, India * Crotalaria incana U + Trop. : Desmodium canum U P 1916, Trop. Am.* Desmodium uncinatum R rop. Am. Mimosa pudica (e Cc 1800, Trop. Am.* Vigna marina U Pantropic Lobeliaceae Lobelia sp R E Lythraceae Cuphea carthagenensis @ + 1900, Trop. Am. Malvaceae Hibiscus tiliaceus U U (cS PP Sida acuta U is Melasto ee malabathricum C C 1916, India 1960] MOOMAW & TAKAHASHI, VEGETATION ON GIBBSITE 411 Species List (Continued) OCCURRENCE ORIGIN INDIGENOUS SPECIES OR RipGE SLOPE VALLEY ENDEMIC INTRODUCED Myrtaceae Eugenia cumini U C 1866, Trop. Afr. Eugenia malaccensis U Metrosideros collina ymorpha U c E Psidium cattleianum C U iE Psidium cattleianum f. U 1888, Trop. Am.* lucidum Psidium guajava C ce A 1800, S. Am.* see he tomentosa Cc 6 1920, Asia Onagrac Jussiaea suffruticosa var. U 1888 Pantropic * ligustrifolia Oxalidaceae Oxalis eae U 1888, N. Am.* Passiflorac ee Phe U U U 1880, S. Am. Passiflora foetida var. U U 1888, Trop. Am. gossypifolia Phytolaccaceae Phy ge acinosa U + Asia Piperace See memobranacea var. R E waimeana Portulacacea Portulaca oleracea U + 1888, Pantropic * Primulaceae Anagallis arvensis 6) + Europe Proteaceae Grevillea robusta U U Australia Macadamia ternifolia U P 1890, Australia Rubiaceae Richardia scabra R 1888 Trop. Am.* Solanaceae Solanum nigrum R Po +t Sterculiaceae Waltheria americana U 1819, Trop. Am.* Umbelliferae Centella asiatica C C 1888, Asia * Verbenaceae antana camara A U 1858, Trop. Am. Stachytarpheta cayennensis A A Trop. Am. 412 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLI NOMENCLATURAL CHANGES IN DAPHNOPSIS (THYMELAEACEAE) Lorin I. NEVLING, JR. Two NOMENCLATURAL ERRORS have come to light since the publica- tion of a revision of the genus Daphnopsis (Nevling, 1959). These in- volve (1) the illegitimacy of the subspecific name Daphnopsis americana ssp. tinifolia (Sw.) Nevl. and (2) the combination Daphnopsis anomala (HBK.) Nevl. which should have been ascribed to Domke. The subspecific combination Daphnopsis americana ssp. tinifolia (Sw.) Nevl. (1959, p. 313) is an illegitimate name and must be replaced. Daph- nopsis americana, originally described by Miller (1768) as Laurus amert- cana, was based upon a Houston specimen and Houston’s “Manuscript Cata- logue of the Plants which he had observed growing in the islands of Jamaica and Cuba; as also at Campeachy and La Vera Cruz, in the years 1728, 1729 and 1732” (introductory comments). In the subsequent edition of the Dictionary (Miller, 1797), Laurus americana was treated as a synonym of Daphne tinifolia Swartz, and Miller states that it was discovered at La Vera Cruz but is also a native of Jamaica. Swartz’s name Daphne tinifolia (1788, p. 63) is possibly based on Miller’s earlier epithet as evidenced by the inclusion of Laurus americana as a synonym. Swartz’s publication was concerned with new species of the Antillean flora and gives the habitat only as Jamaica. There is a real question as to whether or not Swartz was aware of the occurrence of the species in Mexico, a fact which he could not know without seeing the specimen of Houston. As a result of his citation of Jamaica only, the workers postdating Swartz, having benefit of information concerning the type and type locality contained in the 9th edition of Miller’s Dic- tionary, assumed that a second species quite distinct from the Mexican D. americana existed, even if not realized by Swartz, and treated it as such. Three combinations were made, some of them several times, involving the specific epithet ¢imifolia, and all were based on the concept that it was a distinct species restricted to various islands of the Antilles. The subspecies under consideration is restricted to Cuba, Jamaica and Hispaniola and I proposed the subspecific combination, Daphnopsis americana ssp. tinifolia (Sw.) Nevl. indicating the basionym of Swartz. However, the combination is illegitimate because it is based on Swartz’s illegitimate name (Article 64. International Code of Botanical Nomen- clature. 1956). Thus a new subspecific epithet must be chosen to replace ssp. tinifolia. It seems proper to choose Meissner’s indefinite infraspecific epithet cumingiit (D. tinifolia 8. cumingii Meissn. [1857, p. 523]) which is based upon a staminate specimen (Cuming 56) now on deposit in the Naturhistorisches Museum, Wien, and place it at subspecific rank. In 1960 | NOMENCLATURAL CHANGES IN DAPHNOPSIS 413 addition, the names based on the Swartz basionym are transferred to ssp. americana. Accordingly, the nomenclature is emended to read: Daphnopsis americana ssp. americana ee americana (Mill.) J. R. Johnston, Proc. Boston Soc. Nat. Hist. 1909; Contr. Gray Herb. ser. 2. 37: 242. June 1909. ee americana Mill. Dict. ed. 8. Laurus no. 10. 1768. (T.: Houston s.n.) Daphne tinifolia Sw. Prodr. Veg. Ind. Occ. 63. 1788 eae obovata Humb. & Bonpl. ex Wikstr. Diss. Daphne ed. 2. 40. 1820. : Humboldt & Bonpland 7549.) Daphne bonplandiana Kunth, Syn. Pl. 1: 447. 1822. Daphne lagetto Bonpl. ex Kunth, ibid. 1822, pro syn. Nordmannia tinifolia (Sw.) Fisch. aa Ge Meyer, AnneocloNatesers2. 20: 49. February 1843; Bull. Acad. St. Pétersb. Classe Physico-Math. 1: 355. June 1843. Hargasseria mexicana Schiede & Deppe ex C. A. Meyer, Ann. Sci. Nat. ser. . 20: February 1843 (based on Daphne bonplandiana sensu Cham. & Schlechtd. non Kunth); Bull. Acad. St. Pétersb. Classe Physico-Math. 1: 356. June 1843. Hargasseria tinifolia (Sw.) Endl. Gen. Suppl. 4°: 68. 184 Hargasseria schiedeana Endl. ibid. 1847 (based on Denies bonplandiana Kunth). Daphnopsis bonplandii (Kunth) Meissn. DC. Prodr. 14: 521. 1857. Daphnopsis tinifolia (Sw.) Meissn. ibid. 523. : Daphnopsis lindenti Meissn. ibid. 1857. (T.: Linden 95 (?).) Daphnopsis bonplandiana (Kunth) Standl. Contr. U.S. Natl. Herb. 23: 1013. Daphnopsis americana ssp. cumingii (Meissn.) Nevl., comb. & stat. nov. Daphnopsis tinifolia 8. cumingii Meissn. DC. Prod. 14: 523. 1857. (T.: Cuming 56 (2).) Daphnopsis americana ssp. tinifolia (Sw.) Nevl. Ann. Missouri Bot. Gard. 46: 313. An additional change involves the combination Daphnopsis anpmala (HBK.) Nevl. (1959, p. 323). This combination was previously made by Domke and should be credited to him. Daphnopsis anomala (HBK.) Domke, Bibliotheca Bot. 111: 57. 1934, in discussion. LITERATURE CITED MEISSNER, K. F. Thymelaeaceae. DC. Prodr. - 493-605. 1857. Miter, P. The Gardners Dictionary, . So oe The Gardners Dictionary, ed. i NEVLING, L. I., Jr. A revision of the ae Daphnopsis. Ann. Missouri Bot. Gard. 46: 257- 358. 1959. Swartz, O. Prodr. Veg. Ind. Occ. 1788. 414 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI SEEDLING LEAVES IN PALMS AND THEIR MORPHOLOGICAL SIGNIFICANCE P. B. TOMLINSON IN MOST PLANTS the first plumular leaves, or leaves which are pro- duced in succession to the cotyledons, differ from normal foliage leaves in shape and size. They are referred to as “juvenile” leaves (Primar- blatter of German authors) in contrast to the “adult” leaves on the older parts of the plant. Normally where adult leaves of a plant are compound or much divided the juvenile leaves are much simpler in outline. Succeeding leaves (Folgeblatter of German authors) are increasingly com- plex and a gradual transition from juvenile to adult foliage takes place. Exceptions to this generalization are known, as, for example, in certain members of the Quiinaceae (Foster, 1951) and Ficus alba (Corner, 1951; p. 681) in which adult leaves are less complex than seedling leaves. The adult leaves of most palms are large and compound, so it is not unexpected to find that their first foliage leaves are small and usu- ally simple. Although there is considerable variety in the shape of the first foliage leaf, it is very constant for each species and so is of con- siderable diagnostic importance. A practical result of this, of value to nurserymen, is that it is possible to identify, or at least find evidence of misidentification, in palms at a very early stage of growth. Un- fortunately, however, records of the diagnostic characters of seedling palm leaves are not often included in taxonomic writings but they are available for Howeia (Cook, 1926), Roystonea (Cook, 1935), and Veitchia (Moore, 1957). The present paper describes the main shapes exhibited by juvenile palm leaves. A second major observation is that the series of increasingly complex leaf types which occur in the transition between the first foliage leaf and later adult forms is constant and characteristic for each species. Six main series are described here, and by comparing them it is possible to arrive at certain interesting conclusions about evolutionary trends in juvenile palm foliage. Significant correlations are also found between the type of juvenile foliage and the adult foliage in palms and it has been found possible to explain these correlations on an evolutionary basis. There is an extensive literature dealing with the morphology and anatomy of the palm seedling (Gatin, 1912; Boyd, 1932), but, except in the writings of Micheels, Pfitzer and Drude, little attention has been paid to later stages of growth in which juvenile foliage is exhibited. Micheels (1889) and Pfitzer (1885) both emphasize the diagnostic value of seedling leaves in palms. Micheels (1889) and Drude (1889) both describe some of the series of leaves developed by young palms. 1960] TOMLINSON, SEEDLING LEAVES IN PALMS 415 Germination of palm seeds and a new term. A brief résumé of the morphology of the palm seedling as a whole is necessary for an under- standing of the observations recorded below. Excellent accounts of this subject have been written by Gatin (1906, 1912). Germination of palm seeds is hypogeal. The cotyledon is never expanded as a green, aerial photosynthetic organ because its apex remains embedded in the endosperm of the seed and becomes modified into an absorptive organ or haustorium. This converts the reserve food material of the endosperm into a soluble form. The food material is then transmitted to the rest of the embryo via a tubular “middle piece’ which represents the petiole and sheath of the cotyledon. In many palms the middle piece elongates considerably and buries the seedling some distance below the seed. In other palms, in which the cotyledon does not elongate, the seedling develops next to the seed, as, for example, in the coconut. In examining palm seedlings it must be realized that the first green leaves are not immediate post-cotyledonary leaves because one or more of the first plumular leaves are bladeless and appear as sheathing scale- leaves. Their number seems to be fixed in each species of palm. They apparently have a protective function since they envelop the subsequent bladed leaves and are therefore mechanically useful in permitting buried shoots to break through to the soil surface without damage to the en- closed foliage leaves. The present account deals largely with the morphology of the first foliage leaves. In this account it is proposed to apply the term eophyll (Greek eos-early; phyllon-leaf) to the first few leaves with a green expanded lamina developed by the seedling. This term seems necessary in order to identify these organs which are only one of a series of leaf forms to which the term “juvenile” can be applied. In the present paper “eophyll” is used with reference to palms, but it may have a wider appli- cation. OBSERVATIONS The first eophyll of palm seedlings. In most palms the first eophyll is simple, although it may be either entire or bifid, i.e., with a deeply emarginate apex. Within these two main types there is a ton- siderable range of form, although for a given species the shape and size is very constant. Other diagnostic eophyll features are the type and distribution of armature and indumentum. As a consequence, it is often possible to identify at least the genus to which a palm belongs from its first eophyll. In all, three main categories of eophyll may be distinguished. ENTIRE EOPHYLLS. These are usually linear or linear-lanceolate (Fics. la, 2a). Sometimes their apices are truncate as in Corypha, Licuala (Fic. 3) and Livistona. In the caryotoid palms Arenga, Didymosperma and Wallichia the first eophyll is flabellate (Fic. 4), the margin and apex being irregularly and distantly toothed. 416 JOURNAL OF THE ARNOLD ARBORETUM [VOL Er BIFID EOPHYLLS. The depth of the apical incision varies considerably. In species of Drymophloeus and Iriartea the apical incision is so incon- spicuous that the leaflet appears to be entire at a casual glance. Normally, however, the segmentation is so complete that the leaf is evidently com- posed of two opposite, terminal leaflets (Fics. 7b, 8a). The apex of each segment is either acute or, as in Hyospathe and many of the Ptychosper- mate palms, truncate (Fic. 10). In Aiphanes the margin is irregularly toothed (Fic. 9 COMPOUND EOPHYLLS. Many palms have a first eophyll which is always compound. This is palmate in fan-palms such as Borassodendron, Latanta (Fic. 5) and Lodoicea but pinnate in feather-palms such as Hedyscepe, Metroxylon, Nephrosperma, Phytelephas and Raphia. In several genera of feather-palms certain species have simple first eophylls whilst those of other species are compound, as in Calamus, Chamaedorea, Euterpe, and Howeia. The significance of compound first eophylls is discussed later. When the distribution of types of first eophyll throughout the whole family Palmae is considered, certain significant correlations between the shape of this organ and the morphology of the adult foliage become ap- parent Comparisons between the first eophyll and adult foliage in palms. The following arrangement of the tribes corresponds to that given by Drude (1889). A. INDUPLICATE PALMS (segments V-shaped in section). Phoeniceae: Adult leaves imparipinnate, terminal leaflet always distinct; first eophyll entire, lanceolate. Sabaleae: Adult leaves palmate or costapalmate; first eophyll entire, lanceolate, apex truncate in Corypha, Licuala and Livistona. Borasseae: Adult leaves palmate or costapalmate; first eophyll entire, lanceolate but with a truncate apex in Borassus and Hyphaene,; digitately compound in Borassodendron, Latania and Lodoicea. Caryoteae excluding Carvyvota: Adult leaves imparipinnate, terminal leaflet often inconspicuous; first eophyll entire, somewhat flabellate. Caryota: Adult leaves bipinnate; first eophyll bifid. B. REDUPLICATE PALMS (segments A-shaped in section). Lepidocaryineae: Adult leaves mostly paripinnate, pair of terminal leaflets often obscure; costapalmate in Lepidocaryum and Mauritia; first eophyll usually bifid, but pinnately compound in Metroxylon, Raphia and species of Calamus and Daemonorops. Areceae and Geonomeae: Adult leaves mostly paripinnate, terminal pair of leaflets usually conspicuous and equal, rarely adult leaves persistently bifid as in species of Geonoma, Hvospathe and Reinhardtia; first eophyll mostly bifid, but pinnately compound in species of Acanthophoenix, Euterpe, Hedyscepe, Howeia and Nephrosperma, and entire, lanceolate in Roystonea and Stevensonia. 1960 | TOMLINSON, SEEDLING LEAVES IN PALMS 417 Iriarteae: Adult leaves paripinnate, terminal pair of leaflets usually conspicuous and equal, apices of segments usually irregularly toothed; first eophyll bifid, almost entire in Jriartea because of a very shallow apical incision; entire and lanceolate in the anomalous genus Ceroxylon. Morenieae: Adult leaves usually paripinnate, adult leaves persistently bifid in some Chamaedorea species; first eophyll usually bifid, rarely pinnately compound in some species of Chamaedorea; entire and lanceolate in the anomalous genus Pseudophoenix. Attaleae and Elaeideae: Adult leaves irregularly pinnate, terminal leaflets arranged irregularly, often obscure; first eophyll entire, lanceolate, but bifid in Cocos nucifera. Bactrideae: Adult leaves paripinnate, terminal leaflets conspicuous and equal; rarely adult leaves persistently bifid in species of Astrocaryum and Bactris; first eophyll always bifid. Phytelephas and Nypa: Adult leaves pinnate; first eophyll in Phytele- phas pinnately compound. The morphology of the seedlings in these two genera is not well understood and they are not considered further in this account. It may be noted that the first leaves of basal suckers of caespitose palms are usually simple, i.e., they revert to the juvenile condition. Transitions to the adult foliage. In most palm seedlings, after one or more leaves of the same shape as the first eophyll have been pro- duced, a long series of transitional leaves of increasing complexity is developed until the type of foliage characteristic of the adult plant appears. Although each species of palm produces its own characteristic series of transitional leaves, six main classes can be recognized. These are de- scribed below in relation to both the first eophyll and the adult type of eaf. A. INDUPLICATE PALMS, Crass 1. First eophyll entire; adult foliage leaves imparipinnate (Fics. la—c). The eophylls (Fic. 1a) are succeeded by imparipinnate leaves of increasing size in which the odd terminal leaflet resembles the whole lamina of the first eophyll (Fics. 1b, c). This type occurs in Phoenix and in.all the Caryotoid palms except Caryota. In Phoenix the odd terminal leaflet can always be recognized in undamaged adult leaves but often in the Caryoteae the situation is less regular so that the imparipinnate condition is only clear in small transitional leaves. Crass 2. First eophyll entire; adult foliage leaves palmate or costapal- mate (Fics. 2a, b). This class is confined to the fan-leaved tribes Sabaleae and Borasseae. The first, linear foliage leaves (Fic. 2a) are followed by broader leaves in which the lamina is incompletely split into a number of narrow segments (Fic. 2b). Later leaves are broader still and with more segments as the adult foliage is gradually approached. Normally the segmentation of these transitional leaves is not symmetrical, although 418 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI rarely an odd terminal segment can be recognized, as in Coccothrinax, Li- vistona and Trachycarpus. In Cryosophila and Sabai the transitional leaves are often bifid as a result of a deep median split. This split persists in mature leaves of Cryosophila, the blades of which consequently have equally segmented halves. In all fan palms it is assumed that the terminal leaflet has been displaced. Ib 6 FI 1-6. Lamina of eophylls in induplicate-leaved palms. 1, Pho pumila: a, first eophyll. b, c, succeeding transitional eophylls. 2, W eis filifera: a, first eophyll. b, transitional juvenile leaf. 3. Licuala peltata, first eophyll. 4, Arenga pinnata, first eophyll. 5, Latanza sp.. first eophyll. 6, Caryota urens, first eophyl Caryota. First eophyil bifid; adult foliage leaves bipinnate. This type ,is anomalous. In the adult foliage the primary rachis ends in a pair of leaflets comparable to the bifid lamina of the first eophyll (Fic. 6). The transition series between juvenile and adult foliage is long and complex. Some of the intermediate stages in which the leaves are once-pinnate with basal leaflets tending to become secondarily segmented resemble the adult foliage of species of Didymosperma and Wallichia. 1960] TOMLINSON, SEEDLING LEAVES IN PALMS 419 7a | 7b 7c Fics. 7-11. Lamina of eophylls in reduplicate-leaved palms. 7, Roystonea regia: a, ‘first eophyll, b, c, succeeding transitional juvenile leaves. 8, Chrysal- docarpus lutescens: a, first eophyll, b, transitional juvenile leaf. 9, Azphanes sp., first eophyll. 10, Ptychosperma macarthurii, eophyll. 11, Cocos nucifera, eophyll with basal perforati ions. 420 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI B. REDUPLICATE PALMS. Crass 3. First eophyll entire; adult foliage leaves paripinnate (Fics. 7a—c). The first lanceolate eophylls (Fic. 7a) are succeeded by bifid eophylls (Fic. 7b) and then by paripinnate compound leaves of succes- sively larger size (Fic. 7c). This transition series is the longest exhibited by reduplicate-leaved palms since it includes all possible types of first eophyll which they are capable of producing. It is known only in Roy- stonea and Stevensonia. Crass 4. Firsi eophyll bifid; adult foliage leaves paripinnate (Fics. 8a, b). The bifid eophylls (Fic. 8a) are eventually succeeded by pinnate leaves bearing a pair of terminal leaflets which resemble the whole lamina of the first eophyll (Fic. 8b). This last may be striking in palms in which the terminal leaflets are broader than the remaining leaflets. Sometimes, however, the arrangement of distal leaflets in the adult foliage is not regular and the adult foliage leaf is then not obviously paripinnate. This series corresponds to the previous one but for the omission of the initial lanceolate eophyll. It is found in the majority of palms belonging to the tribes Areceae, Bactrideae, Iriarteae, Lepidocaryineae and Morenieae together with Cocos of the Attaleae. Crass 5. First eophyll entire; adult foliage irregularly pinnate. In this class there is great variation in the shape of transitional leaves within a single species and even within a single individual. It is impossible to describe the juvenile leaves as either paripinnate or imparipinnate since they are not segmented regularly. Often the later eophylls have incom- plete splits so that the blade is irregularly fenestrate. Other examples are common in which one half of the blade is entire whilst the other is partly or wholly segmented. The same irregularity and obscurity characterizes the terminal segments of adult leaves because the most distal leaflets are often filamentous and difficult to distinguish from the filament which terminates the rachis. This class characterizes most members of the Elaeideae and Attaleae and probably also occurs in Ceroxylon and Pseudophoenix. Cocos, together with other rare exceptions from the above mentioned tribes (e.g., Attalea allenii), has uniform paripinnate leaves and exhibits the series of Class 4. C. MIxED REDUPLICATE AND INDUPLICATE PALMS. Crass 6. First eophyll compound; adult foliage leaves either pinnate or palmate. The occasional and mostly unrelated genera in which the first eophyll is compound have a transitional series which includes no new leaf forms, there being merely an increase in size until the adult type of foliage is produced. DISCUSSION A fairly constant correlation between the shape of the first eophyll and the morphology of the adult foliage is apparent in the information presented above. It is that the induplicate (V-folded) palms normally 1960 | TOMLINSON, SEEDLING LEAVES IN PALMS 421 have an entire, lanceolate first eophyll and that the adult leaf is impari- pinnate. On the other hand the reduplicate (A-folded) palms normally have a bifid first eophyll and an adult leaf which is paripinnate, where- ever this condition can be recognized with certainty. It is suggested below how this correlation could arise and also how to account for apparent exceptions. From the above it might also appear that there is a primary connection between fan-leaves and an entire first eophyll. However, it is generally regarded (see Eames, 1953) that the palmate leaf is merely a pinnate leaf with a condensed rachis or, more precisely, is equivalent to a pinnate leaf in which the rachis fails partly or wholly to elongate. It so happens that this trend has occurred mainly in the induplicate-leaved palms so that the two major tribes of this group, the Sabaleae and Borasseae, are entirely palmate. On the other hand only Lepidocaryum and Mauritia of the reduplicate-leaved group have developed palmate leaves. Although palmate leaves of the borassoid and sabaloid tribes are thus morphologically equivalent to imparipinnate leaves, the development of an odd terminal leaflet, even in juvenile foliage or in costapalmate leaves which have a short but distinct rachis, is infrequent. In these palms it is assumed that the terminal leaflet has been displaced. Before these correlations can be satisfactorily explained it is first neces- sary to account for the variation in eophyll morphology. Phylogenetic trends in eophyll shape. The considerable range of eophyll shape and type in transitional juvenile foliage in palm seedlings (see Fic. 12) seems to result from certain evolutionary trends. It is assumed that a long and gradual transition between the first eophyll and the adult foliage leaf, including an extensive series of intermediate forms, is a primitive one, and that in more recently evolved palms one or more of the members of this series is omitted during seedling development. Essentially the same suggestion has been made by Dufour (1910) for some members of the Ranunculaceae in a paper describing juvenile foliage in Anemone and Ranunculus which was brought to my attention during the preparation of this article. The situation is somewhat complicated in palms because two main series of leaves are present: those with re- duplicate and those with induplicate vernation. Parallel evolutionary trends have occurred independently within each series, and most possible stages of eophyll elimination occur. REDUPLICATE PALMS. The longest series of different eophyll forms in this group are those exhibited by Roystonea and Stevensonia which form Class 3 (Fic. 12, lower left). This is assumed to be the primitive condi- tion. The large Class 4 to which most reduplicate palms belong has been derived by omission of the initial entire eophyll (Fic. 12, lower middle left). In Class 6 are found those species in which both simple types of eophyll are omitted so that the first foliage leaf is pinnately compound (Fic. 12, upper middle left). This represents the most advanced condi- 422 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI l2 REDUPLICATE INDUPLICATE Adult Juvenile Juvenile Adult : v ) | c —> , Fic. 12. Diagrammatic representation of blade shape in adult and juvenile palm leaves. Further explanation in text. Solid arrows indicate ontogenetic series. Broken arrows indicate presumed phylogenetic series. Numbers refer to classes described in text. tion. It is evidently a derived and not a primitive condition since it has appeared quite independently in unrelated genera. The cocoid palms of Class 5 (Fic. 12, upper left) are anomalous. The first eophyll is entire and certainly represents a primitive condition but the subsequent irregularly pinnate transitional leaves which are character- istic of this class indicate a special evolutionary trend. A possible explana- tion of this trend is given below. INDUPLICATE PALMS. Class 1, exemplified by Phoenix represents the primitive condition in this group (Fic. 12, lower right). Since only two types of eophyll are present, only one derivative type is possible. This 1960] TOMLINSON, SEEDLING LEAVES IN PALMS 423 is found in the three borassoid genera in which the initial entire eophyll is omitted so that the first eophyll is palmately compound (Fic. 12, upper middle right). None of the sabaloid palms show this advanced condition. One other possibility has never been realized because none of the pinnate induplicate palms possess a pinnately compound first eophyll. Caryota is obviously a derived type although it is difficult to account for its bifid first eophyll, a type of leaf not otherwise found in the indupli- cate palms (Fic. 12c, upper right). This genus merits a detailed develop- mental study. Permanent juvenile foliage. In the evolutionary trend described above, certain types of juvenile foliage are omitted from the ontogenetic series. The converse trend, in which the adult type of foliage is never produced, is common, usually being found in palms with a reduced habit in which the stems are short and the internodes narrow. In these palms the juvenile foliage persists throughout the life of the plant (Fic. 12j, upper middle left), or at the most compound leaves with very few segments are produced. This feature has had a polyphyletic origin since it is exhibited by several unrelated genera. Amongst the fan palms some species of Licuala and Teysmannia have an undivided, orbicular lamina, or at the most segmentation consists of shallow marginal incisions. In the feather palms more obvious juvenile foliage is retained in several species of the genera Astrocaryum, Chamae- dorea and Geonoma. In these species the simple foliage leaves resemble the first eophylls of related species with reduplicately pinnate adult foliage. In Bactris militaris, B. wendlandiana, and Hyospathe concinna the leaves are always simple, unlike the pinnate adult leaves of the remaining species in these genera. Normally these persistent juvenile leaves are small, al- though in Bactris militaris the undivided leaves may be ten feet long. The genus Reinhardtia is of special interest. Reinhardtia elegans has pinnate leaves and is considered by Moore (1957) to be the most primitive member of the genus. Other species are smaller and have fewer leaflets. Reinhardtia latisecta and varieties of R. gracilis have either simple leaves or at the most leaves with two or three segments. An additional peculiarity in these last two species is the presence of small perforations at the base of the lamina, close to its insertion on the rachis. These perforations are evidently incomplete splits. Somewhat homologous ‘‘fenestrate’’ leaves occur among the juvenile foliage of many cocoid palms and they are par- ticularly striking in seedling coconuts (Fic, 11). Evidently it is common for Reinhardtia to have persistent juvenile foliage. The general conclusion is that where simple leaves characterize the adult foliage of palms, they represent a derived and not a primitive condi- tion. Possible origin of the palm leaf. On the basis of his studies on the development of the palm leaf, Eames (1953) makes the following state- 424 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI ment: “The compound leaf of the palm has been derived from the simple leaf by the dissection of the blade into leaflets attached to a rachis which represents the midvein of the ancestral leaf. The number and form of the leaflets are controlled by the venation pattern of the ancestral simple leaf.” It is doubtful, as has just been shown, that any of the simple leaves in existing palms represent this ancestral type. The only primitive simple leaf in palms is the lanceolate first eophyll of Phoenix and Roystonea. However, it would be unwise to compare this juvenile leaf with archaic adult forms, although it is very probable that the ancestors of existing palms had a first eophyll essentially the same as that in Phoenix and Rovstonea. On the other hand, it is possible to construct mentally an ancestral palm leaf from which all existing leaf types can be derived by the splitting proc- esses described by Eames. This archetype leaf would have a regularly plicate, entire lamina, with numerous lateral major veins in the form of ribs occupying the dorsal and ventral crests of the folds. Each vein would have a somewhat sigmoid course, its end in the leaf margin either passing to the apex or fusing with the ends of lower veins to form a marginal com- missure. The thickened midrib would extend to the apex of the leaf and there become an abaxial rib (Fic. 13a). This is essentially the type of leaf envisaged by Eames. The nearest approach to it in living palms is found in such genera as Manicaria, Mascarena, Stevensonia, and Vershaffeltia in which the congested leaflets often are persistently coherent along the margins. From a distance leaves of these genera appear to be entire. It is appropriate to mention at this point that the leaf of a fossil plant (Sanmiguelia lewisii Brown) has recently been described and interpreted as that of a primitive palm (Brown, 1956). Whilst this leaf does not bear all the essential features listed above, the chief difference being the absence of a midrib, it could still serve as a fundamental type from which modern compound palm leaves have evolved by the methods suggested below. The age of this plant is also noteworthy since, if it is interpreted correctly, it would be the remains of the earliest known flowering plant. The splitting which produces individual segments may occur in three possible ways. First, the leaf may split along the ventral (abaxial) crests of the folds (Fic. 13b). This would produce reduplicate segments. At the leaf apex there could be either two equivalent splits, one on each side of the rachis, or a single split could bisect the rachis. In both examples a pair of equal terminal leaflets would result, but in the former type of splitting the free end of the rachis would persist as a whip-like filament, a feature not uncommon in paripinnate palms but particularly prominent in Cocos (Venkatanaryana, 1957). This situation accounts satisfactorily for the paripinnate, reduplicate leaf but it will only arise if there is a regular arrangement of ribs at the leaf apex. In contrast the cocoid palms of class 5 apparently owe their irregularly pinnate leaf to the fact that the ribs, at least at the leaf apex, are not equally spaced and that splitting is not regular. The segments here are reduplicate and the leaf is essentially 1960 | TOMLINSON, SEEDLING LEAVES IN PALMS 425 paripinnate as is revealed by the few rare examples in which the arrange- ment of distal leaflets is regular. The second possibility is that the splits which separate adjacent seg- ments take place along the dorsal (adaxial) crests (Fic. 13c). This pro- duces induplicate leaf segments and an odd terminal leaflet so that the leaf is automatically imparipinnate. This is the type of leaf found in Phoenix and the Caryoteae. The palmate leaf of the Borasseae and Sabaleae is a similar leaf with a partly or wholly condensed rachis in which it is assumed that the terminal leaflet is normally displaced and is no longer recognisable. I3a AA AAAS NI INS Fic. 13. Suggested origin of reduplicate and induplicate leaves from a hypo- thetical unsegmented ancestral leaf. a, Diagrammatic transverse section through apex of ancestral leaf with plicate but eeneiea lamina. b, Origin of re- duplicate segments by cleavage along ventrai crests. c, Origin of induplicate segments by splitting along dorsal crests. Rachis crosshatched, terminal seg- ments lined. Splits are assumed to eliminate the ribs, otherwise the ribs them- selves could be divided and the segments would then have marginal half-ribs. The third possibility is for splits to occur without reference to the folds. This condition is known only in the small fan palm Rhapis which as a consequence has unequal and irregular segments (Naumann, 1887). Although this theory accounts for the correlation between leaf shape and vernation it is an over-simplification of the situation. Eames, in con- firmation of the observations of several early workers has clearly shown that the plication of palm leaves is a result, not of simple folding, but of a cleavage and invagination of a solid tissue. This process is completed before the leaf primordium is more than a few millimeters high and results in a plicate blade, the margin of which is con- 426 JOURNAL OF THE ARNOLD ARBORETUM [VOLS XLT nected by a continuous strip of tissue. The splitting that results in the separation of adjacent segments occurs later and, according to Naumann, may take place in a variety of ways. However, this complex developmental process does not invalidate the above argument, since it is possible that the primitive palm leaf developed an entire, plicate lamina by the same initial cleavage process. On the other hand, the view of Arber (1922) that the palm leaf is a modified phyllode may mean that at no stage in its phyletic history was there an entire unsegmented blade. The diagram constructed by Eames (1953, Fic. 11), based on Roystonea, showing the origin of equally-spaced folds in the leaf primordia will only result in mature leaves in which the segments are equal, uniformly spaced and either opposite or regularly alternate on the two sides of the rachis. This regularity is relatively uncommon in palms. Many palm leaves may be described as irregularly pinnate (since the leaflets are not evenly spaced) or unequally segmented (since the leaflets are not all of the same width) and combinations of these two types are common (Tomlinson, 1961). In irregularly pinnate leaves the initial clefts must be irregularly spaced in the primordium, The most difficult situation to account for is one in which the pinnae are fasciculate, individual clusters being separated by long, naked portions of the rachis, and in which clusters of leaflets are subopposite or even alternate. It is easier to account for unequally seg- mented leaves which evidently arise because splits separating individual segments occur at irregular intervals and along only a few furrows. Even so, the enormous range in the morphology of individual leaflets has still to be accounted for and only detailed studies of the development of the more outstanding leaf forms will reveal the mechanism of their origin. It is clear, on the other hand, that the fundamental process described by Eames takes place in the early stages of development of all adult palm leaves. The palms are undoubtedly an ancient group with a long phylo- genetic history so it would not be unexpected to find minor divergence from the fundamental mechanism of leaf development inherited from primitive forms. Much of the controversy which exists in the literature dealing with the development of the leaf in palms is likely to be a consequence of this diversity. Morphogenetic considerations. From the account given by Eames it is apparent that three essential processes are responsible for the development of the palm leaf:— (i) a cleavage and folding of solid tissue during very early stages of leaf ontogeny, (ii) elongation of the rachis to a greater or lesser extent during the later stages, (iil) a splitting whereby adjacent leaf segments become free. The vast range of leaf forms in palms is a result of variation in the intensity of these processes. Consequently, if any one of them could be influenced experimentally it should be possible to modify the final leaf shape. Leaves vary considerably in the extent to which external influences modify their ultimate shape (Ashby, 1948) but the susceptibility of the palm leaf to 1960 | TOMLINSON, SEEDLING LEAVES IN PALMS 427 changes in environmental conditions has never been explored. There are obvious technical difficulties involved in using palms as experimental objects. However, this account has shown that the essential morphology of the palm leaf is established very early during the production of juvenile foliage and therefore the possibility of carrying out experiments on palm seedlings grown in greenhouses in a cool temperate climate is not excluded. One suggested experiment would be to establish whether failure of the rachis to elongate in typical palmate leaves is the result of an inhibition process which could be overcome by application of growth substances and, if so, whether it would be possible to recreate a presumed ancestral leaf. SUMMARY In the development of the palm seedling there is a gradual transition from small, usually simple leaves to the large compound leaves of the adult foliage. It is suggested that in the evolution of palms, one or more of the simple types of leaf which were present in the transitional series in the primitive palm and which are still extant in Phoenix and Roystonea, have been omitted and so the different types of transition series in existing palms produced. The most highly evolved seedlings have a compound first foliage leaf. A reverse trend is one in which the juvenile foliage persists and is present throughout the life of the palm. There is found to be a significant correlation between (a) the shape of the first foliage leaf (first eophyll), (b) the series of transition forms between juvenile and adult foliage, (c) the morphology of the adult leaf, (d) the type of folding in the individual leaf segments. Palmate leaves are not essentially different from pinnate leaves. Two main groups exist in palms, the induplicate-leaved palms being imparipinnate, the reduplicate palms being paripinnate. It is suggested that these two forms are a result of a simple difference in the way in which a hypothetical entire, plicately- folded prototype palm leaf has been segmented. The great variety in the morphology of adult palm leaves can be ac- counted for on the basis of variation in three fundamental processes which occur during their development. It is possible that these processes can be influenced artificially. The present article is partly the outcome of several long discussions with Dr. H. E. Moore, of the Bailey Hortorium, Cornell University. I should like to thank him for clarifying many details and for correcting me on many points. BIBLIOGRAPHY ARBER, AGNES. On the development and morphology of the leaves of palms. Proc. Roy. Soc. London. B. 93: 249-261. 1922. Asusy, E. Studies on the morphogenesis of leaves. I. An essay on leaf shape. New Phytol. 47: 153-176. 428 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI Boyp, Lucy. Monocotylous seedlings. Trans. Bot. Soc. Edinburgh 31: 1-224. 1932 Brown, R. W. Palmlike plants from the Dolores Formation (Triassic), S. W. Colorado. U.S. Geol. Surv. Prof. Paper 274H: 205-209. 1956 O. F. A new genus of palms based on Kentia forsteriana. Jour. Wash. Acad. Sci. 16: 392-397, 1926. Nege ee of royal on mcience 81;-590,-.1935, Corner, 1 an el ayside trees of Malaya. ed. 2. vol. 1. 1951. Drupe, O. ee In Engler and Pent ‘Die naturlichen Pflanzenfamilien II. 3: 1-93. 1889 Durour, L. Etudes des feuilles primordiales de quelques plantes. Rev. Gén. Bot. 22: 367-384. Eames, A. J. Neglected morphology of the palm leaf. Phytomorphology 3: 172-189. 1953 Foster, A. 8. Heterophylly ae foliar venation in Lacunaria. Bull. Torrey Bot. Club 78: 382-4 195 GATING. (Ge Le a a et chimiques sur la germination des palmiers. Ann. Sc. Nat. Bot. IX, 3: 191-314. Les palmiers: histoires naturelles et horticoles des different genres. Paris. 1912 MIcHEELs, H. Récherches sur les jeune palmiers. Mém. Cour. Acad. R. Belg. 55: 1-126. 1889. Moore, H. E. Synopses of various care of Arecoideae. 0g 21. Veitchia; Art. 23. Reinhardtia. Gent. Herb. 481-536, 541-576. NaAuMANN, A. Beitrage zur aL OTT der Ralmenniiien: Flora 70: 193-253. 1887. Pritzer, E. Ueber Fruchte, Keimung und Jugendzustande einiger Palmen. Bee Deutsch. Bot. Ges. 3: 32-52 TomMLinson, P. B. Anatomy of the Monocotledons: Vol. 2. Palmae. 1961 (In press). VENKATANARYANA, G. On certain aspects of the ae of the leaf of Cocos nucifera L. Phytomorphology 7: 297-305. 1957 DEPARTMENT OF BOTANY UNIVERSITY OF LEEDS 1960] THE DIRECTOR’S REPORT 429 THE DIRECTOR’S REPORT Tue ARNOLD ARBORETUM DuRING THE FISCAL YEAR ENDED JUNE 30, 1960 The Staff: It is with regret that the sudden death of Dr. Ivan M. Johnston on May 31, 1960, is recorded. Dr. Johnston, Associate Professor of Botany and a member of the staff of the Arnold Arboretum since 1931, was known in the botanical world for his monographic studies on the Boraginaceae and for his floristic work in the deserts of the United States, Mexico and South America. A biographical report of his life and professional career, as well as a bibliography of his published works, will appear in the next number of the Journal of the Arnold Arboretum. Dr. Karl Sax, Professor of Botany, retired on August 30, 1959 and has served as Visiting Professor of Botany at the Gibbs Research Laboratory, Yale University, for the remainder of the year. Three staff resignations also were accepted. Dr. Frances M. Jarrett, Botanist, resigned October 31st to accept a post at the Royal Botanic Gardens, Kew, England. Mrs. Claude Weber, Botanist, resigned Septem- ber 1st to enter the graduate school of Radcliffe College. Mrs. Mary Lehmer, Business Secretary, resigned June 15th and will live in Albu~ querque, New Mexico. Each contributed devoted and valuable services to the Arnold Arboretum and each carries with her our appreciation. Two new scientists were appointed to the staff. Dr. Lorin I. Nevling was appointed Assistant Curator, beginning October 15. Dr. Nevling, who received his Ph.D. degree from Washington University in St. Louis, has been working on the Thymelaeaceae of the New World. Dr. Joab L. Thomas, a graduate of Harvard University, was appointed Cytotaxonomist as of July 1. Dr. Thomas, whose doctoral thesis dealt with the Cyrillaceae, will work primarily on the cytology and genetics of the cultivated plants in the living collections of the Arboretum. During the year Mr. Henry Draper was appointed Superintendent of the Case Estates at Weston, Miss Ann Waterman was appointed jointly with the Gray Herbarium as Assistant Librarian, and Miss Ann Close became Business Secretary. The Ninth International Botanical Congress was held in Montreal, Canada, between August 19th and 29th, 1959. Approximately three thousand botanists, representing many fields of research, plus wives and children attended these meetings. We were pleased that approximately 100 foreign scientists visited the combined herbaria and the living collec- tions before, during and after the meetings. Eleven members of the Arbo- retum staff attended the pre-congress nomenclatural meetings. During the congress Dr. Howard presented an invitational paper titled ‘““The Vascular 430 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, xEr Anatomy of the Petiole as a Taxonomic Character” at a symposium on plant anatomy. Dr. Hu reported on “The Taxonomy, Geography and Economic Importance of the Genus Fraxinus in China” at a session on plant taxonomy. Dr. Wood was appointed to the Subcommittee for Family Names of the Committee for Spermatophytes. Dr. Howard was elected President and Dr. Wagenknecht, Secretary of the International Associa- tion of Botanical Gardens, both to serve for five years. The annual meeting of the American Association of Botanical Gardens and Arboreta, held in conjunction with that of the American Horticultural Council at Rochester, New York, in October, was attended by Drs. Howard, Wagenknecht and Wyman. Following this meeting Dr. Wyman was ap- pointed chairman of the committee on Plant Registration for the Associa- tion. The Arboretum staff was represented at the annual meeting of the Plant Propagators Society by Mr. Fordham. The staff again made the necessary arrangements for the annual meeting of the New England section of the American Society of Horticultural Sciences which was held in Cambridge. Five staff members attended these meetings with three presenting technical papers. Dr. Howard was a participant in the Tropical Botany Conference sponsored jointly by the National Academy of Science and the National Research Council and held at the Fairchild Tropical Garden, May 5-7, 1960. During the year the combination of the American Horticultural Society and the American Horticultural Council was effected. Dr. Wyman served for five years as Secretary of the latter organization which maintained its business office at the Arnold Arboretum. Following the union of the two groups as the American Horticultural Society, Dr. Wyman resigned as Secretary of the Council and was then elected First Vice-President of the new society, the headquarters of which will be in Washington, D. wards of recognition came to Drs. Wyman and Sax during the year. A special award to a professional horticulturist for outstanding accom- plishments and contributions was given to Dr. Wyman by the Horticultural Society of New York. Dr. Sax received the Jackson Dawson Medal from the Massachusetts Horticultural Society and was named ‘Horticulturist of the Year” at the University of Massachusetts students’ horticultural show. Horticulture: The winter of 1959-60 was unusually mild with only a single heavy snowfall. These conditions allowed for nearly continuous out-of-door work in the cleaning of many shrub plantings and in the pruning and removal of dangerous trees. Much of this work was necessitated by the extreme weather conditions of the previous year. The effects of the severe winter of 1958-59 became more evident in many plants as the season progressed and were again visible in the spring of 1960. The most conspicuous damage occurred on some, but not all, plants of the hardy strain of Albizzia juli- brissin var. rosea. 1960 | THE DIRECTOR’S REPORT 431 The Arnold Arboretum in June, 1960. ABove: Bussey Brook as it flows through the Rhododendron collection. BeLtow: A view across the lilac collec- tion toward the torch azaleas in the oak woods. The spring season of 1960 appeared to be a short one. It was not possible to complete all the transplanting from the nursery in Weston to the permanent plantings in Jamaica Plain, but 214 species, varieties and 432 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI cultivars were moved. Some of these were used to augment existing group plantings and others replaced weak or poorly grown plants. Many were new, representing additions to the permanent plantings. Several plantings received special attention during the year. The collec- tion of tree peonies, moved the previous year, is being pampered until it is well established in its new area. Approximately 1700 bulbs of Narcissus cultivars were added to the naturalized area being created along Bussey Brook. The mass planting of torch azaleas along the South Street bank was pruned, cleared and fertilized. The Taxus and Juniperus collections which had become crowded during recent years were given special atten- tion, several of the large yews and junipers were moved, and some duplicate plants taken out. New specimens have been added to the main collection and the general appearance is much improved with favorable growth con- ditions again present in the area. The city of Boston through its Department of Parks and Recreation continued the annual program of improvement of the roads, paths and benches which is its responsibility. A new surface was placed on the full extent of the road between South Street and Bussey Street in the area of the Rhododendron and conifer collections. The meadow road from the Arborway passing in front of the administration building to the maple collection was also resurfaced. These two areas not only represented the most serious deterioration of the roadways, outside of the Peters Hill area, but were also the sections most heavily travelled by visitors and by Arboretum equipment. It is hoped that additional areas may be resurfaced as city funds permit. The Department of Parks and Recreation also com- pleted the last portion of the work of clearing the cobblestone gutters within the Arboretum. The area of Peters Hill is now clear and the whole road- way much improved. The dirt which had accumulated in the gutters had become overgrown and was salvaged for fill. Chemical sprays are now used to keep the cobblestones free of weeds. During the year Mr. Williams, the superintendent, worked out a simple system of spraying with ‘“Dowpon” to edge sharply and effectively the many plantings and beds in the shrub collection. Experiments continue on chemical methods of weed control as a means of improving the appear- ance of the grounds and of reducing the amount of hand labor required. This experimental program involves the type of chemical, the rate and frequency of application, The results are measured in cost of application, the control obtained and the effects on different woody plants. In previous reports the use of large quantities of organic material as mulch on the Arboretum collections has been indicated. At various times the Arboretum has had available at low cost or simply for the hauling adequate supplies of spent hops, sawdust, wood chips, buckwheat hulls, or coffee grounds. The mulch used most recently has been cocoa shells for which there had been no commercial demand. The cocoa-shell mulch was used experimentally at first but with notable success. The mulch is fire proof, retains moisture, remains in place, i.e., does not float or blow away, 1960 | THE DIRECTOR’S REPORT 433 has good fertilizing value which is released slowly, and is attractive in appearance. Through staff comments to classes, publications, and the questions of many visitors a horticultural demand for this product has now been created so that gratis supplies are no longer available to us in the quantities needed. Thus, during the year it has been necessary to locate new sources of mulching material; it now appears that animal manures from local horse stables and the city zoo will be available for our needs. In an attempt to have within the collections representatives of genera, species, varieties and cultivars offered by the nursery trade within our area, an attempt was made to locate sources for 480 advertised taxa not currently in our living collections. Many of these taxa have been tried on previous occasions and have not proved hardy; others represent new selections or new cultivars. It is desirable, however, to have information on the growth and habits of such plants, even though previous experience has shown that many may not be botanically distinguishable from already known plants or that the commercial source is unwise in offering for sale a plant which is not hardy or desirable for the Boston area. The small plants which have been obtained will be held at Weston for growth and initial hardiness tests. One of the striking features of the Arnold Arboretum living collections is the completeness and accuracy of the labeling and mapping systems. Both of these are maintained only by constant effort. Many staff members codperate in the continuous work of checking identifications, to insure the accuracy of the names. During the past winter a complete inventory of the plants in the living collections at Jamaica Plain and Weston was made and the nomenclature checked against Rehder’s Bibliography, as well as more recent monographs. The work of verifying plants in the living collec- tions from which type specimens were collected by Sargent, Rehder, and others has been continued. A special label has been prepared to identify such plants properly. During the fall the Arboretum was approved as the National Registration authority for cultivars of woody ornamental plants not now handled by special societies. This delegation is from the American Association of Botanical Gardens and Arboreta and the American Horticultural Society and is for a two-year trial period. In this work the Arboretum staff cooperates with any individual or organization wishing to register specific groups of plants. To the present, about a dozen individuals and organiza- tions have expressed an interest in or a willingness to carry on such regis- tration activity. For groups not delegated to others, the members of the Arboretum staff, in an effort codrdinated by Dr. Wagenknecht, will register cultivar names and compile registration lists as interest is expressed. A set of registration regulations and forms was prepared and is distributed on request. The return of these forms, plus a small registration fee, a repre- sentative herbarium specimen, and a living plant establishes the registration of a name. If a check of the literature shows that the cultivar name has not been used previously and the name is in accordance with the Inter- 434 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI national Code of Nomenclature for Cultivated Plants, the sender is so notified. Most of the plants received for registration during the first few months of this work had been given original and valid names. A few names were not acceptable, being either pre-occupied or incorrectly formed according to the Code. In each instance the sender agreed to alter the cultivar name in accordance with suggestions made. The Arboretum has no responsibility for the distinctness of the cultivar registered and legally none for the identification, although both points are of immediate concern to the staff. Where necessary, additional correspondence often has clarified the relationships of cultivars, producing information of lasting value which will be kept with the type specimen. The living plants will be kept in a special area of the Arboretum for reference and testing. If it is obvious that the species is not hardy in this area most of the registered plant mate- rial is sent to a codperating arboretum in a more appropriate environment. The presence of authentic living and herbarium material of cultivars will facilitate the work of the staff in horticultural plant taxonomy. The end of the two-year trial period will coincide approximately with the 1962 International Horticultural Congress when steps may be taken to coordi- nate registration activities for all ornamental woody plants at the inter- national level. In the 1959 session of the Massachusetts Legislature a bill was passed to obtain from the Bussey Institution several acres of land, including that on which the Arnold Arboretum maintains its greenhouses, nursery area, hedge collection, and collection of bonsai. Although the bill was incorrectly filed, a new bill will be prepared. For this reason the staff has spent con- siderable time during this fiscal year drawing plans for new greenhouses and the necessary accessory buildings. A site was selected on land owned by Harvard University for the Arnold Arboretum along Centre Street and a program of fertilization and soil preparation is under way in this area. The new greenhouse plans will improve research facilities by providing experimental cold chambers, laboratory space and modern head-house equipment. A small classroom for graduate students and adult classes is also contemplated. The overall increase in greenhouse space will be approximately thirty-three per cent while the nursery area will be increased threefold. The President and Fellows of Harvard College, as trustees of the Arnold Arboretum, approved the plans in principle and it is expected that an architect will be selected during the summer with construction possible during the next year. During the year the staff of the plant propagation department made a special study of the plants represented in the living collections by single specimens and those which were not in the best of condition. The unique taxa were selected for immediate propagation. Published information is lacking for the vegetative propagation of many of these and an experi- mental approach is necessary. In all, 257 taxa are to be considered. Initial propagation efforts have not been successful with all of these and this program will continue. Many of the plants represented by single specimens 1960 | THE DIRECTOR’S REPORT 435 in the collections are desirable ornamental shrubs and it is the difficulty of propagation that presumably has limited both the number of specimens which we have and their availability in trade sources. The routine work of filling requests for plant materials and of handling shipments of seeds and plants received may occupy a large percentage of the department’s work time. During the past year 245 shipments were made by air mail or air freight, parcel post and express within the United States and to twelve countries beyond. A regular exchange is maintained with most of the institutions represented. Thus, during the year, the Arboretum received 235 shipments of living material of 693 species and varieties from the same or other sources in the United States and eleven other countries. A total of fifty-two shipments of seeds representing 186 taxa was received and processed from seventeen countries, in addition to the United States. Included were many oaks and related plants and several species of blueberries from higher altitudes of southeastern Asia. A collection of seeds of Acacia drummondii, said to be one of the hardiest and most attractive of the acacias, was received from New South Wales. The seeds germinated well and seedlings have been distributed to loca- tions in the South and Far West. During the past season trees of the hardy strain of Cedrus libani fruited heavily. Although these plants have grown in the Arboretum for many years this clone is relatively rare in cultivation. It was possible to distribute the seed in quantity to twenty-eight locations in fourteen of the colder northern states, with shipments also going to Canada and north- ern Europe for trial. A replicate experiment with these seeds has shown that two months of stratification at 41° F. allowed 80% germination within four days. The seedlings are extremely susceptible to damping-off organisms, but with the rapid germination obtained it was possible to separate seedlings quickly and to minimize the spread of these diseases. dawn redwood, Metasequoia glyptostroboides, is an Arboretum introduction which has become a popular tree for general planting. While none of the plants grown from seed introduced in 1948 has yet produced staminate inflorescences, many reports of ovulate cones have been re- ceived. With no fertile seeds available the species must be reproduced by cuttings. Work during the past year revealed that cuttings taken on August Sth rooted earlier and in far better percentages than those taken on June 5th. Winter hardwood cuttings of the same plant rooted poorly. In an experiment to determine the viability of scions in storage, crab- apple scions which had been collected in December 1958 and kept ina polyethylene bag with a small amount of moistened sphagnum moss at a temperature of 41° F. were finally grafted onto appropriate stock in March 1960. Although a small number of scions was involved, the via- bility and successful union of most of these suggest that many materials of this kind can be stored under appropriate conditions and retained from one grafting season to the next. The use of sulphuric acid baths to break seed coats and dormancy is 436 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI well known. Seeds of Koelreuteria paniculata germinated in six days following such treatment or could be stored (as in this instance) in a refrigerator for sixty days before planting. In all cases there was no evidence of the reported secondary dormancy. Seeds of Cytisus nigricans produced general germination in ten days following treatment in sulfuric acid for one hour. It is worth reporting that a flat of seeds of this species which had received the normal hot water treatment the previous year had failed to germinate. When these seeds were recovered and treated with sulfuric acid they, too, germinated in ten days. Case Estates: The house at 101 Wellesley Street in Weston adjacent to the ground- cover demonstration area was made available to the Arboretum during the summer of 1959. This house, one of the oldest in Weston, was given to the Arboretum by Miss Marian Case but has been occupied by Miss Case’s former gardener for a number of years. The house was recondi- tioned and a new heating system was installed. It is now occupied by the superintendent at the Case Estates. The New England Daffodil Society expressed an interest in obtaining space for a test garden for Narcissus varieties at the Case Estates and beds were prepared for trial plots of these bulbs. The Case Estates have in various areas many naturalized bulbs which are not only colorful but draw considerable attention in the spring. A trial garden will be of value to the staff in relation to its teaching programs and to its work in cultivar registration. The New England Daffodil Society has agreed to plant and maintain the beds. Work has continued in a general program of cleaning plantings and rejuvenating some of the older trees. Thirty old apple trees were removed during the year to reduce maintenance costs of this area and to allow for an expansion of other shrub and tree collections. Large amounts of manure were obtained without charge and many of the temporary nursery beds were fertilized following the removal of small plants. Again this year the surplus plants were offered to the Department of Buildings and Grounds in accordance with the terms of the Arboretum trust. Approximately 430 plants of 144 taxa were offered and most of these were utilized on the Harvard campus and in conjunction with the Harvard building program. Education: None of the Arboretum staff was scheduled to offer formal courses at Harvard University during the past year. Several staff members assisted in the guidance of graduate students at Harvard and Radcliffe. A series of luncheon seminars for graduate students and staff was continued in which various staff members and students presented papers which were followed by discussion. The popular classes in horticulture and botany, conducted by the staff, 1960] THE DIRECTOR’S REPORT 437 were eleven in number. The field classes at Jamaica Plain and Weston continued to draw the largest numbers. Plant propagation was again enrolled to the limit of the facilities. More technical courses on plant geography, plant ecology, the gymnosperms, economic botany and a course in French on the contributions of French horticulture to American gardens drew qualified students. A course in economic botany which visited restaurants offering foreign foods proved to be a novel and appeal- ing method of teaching. After an enjoyable meal, the class studied and discussed the plants and plant parts used for food. This course has received considerable newspaper publicity, one report even returning from Caracas, Venezuela. Another expression of the educational activity of the Arboretum staff is found in the outside lectures which they are requested to give. Such lectures are offered at various levels. On occasion the Arboretum staff has through lectures or guided tours dealt with students from the elementary grades of the Weston schools to the graduate students of many colleges. Many requests for speakers are received from Garden Clubs and similar groups interested in horticulture. Such lectures can be both entertaining and instructive and the staff speakers attempt to make them so. Again this year approximately fifty groups numbering at least twenty-five persons were given conducted tours of the living collections in Jamaica Plain or Weston. Dr. Wood represented the Arboretum on the program of the annual Williamsburg Garden Symposium, speaking on “Birds, Bats, Bees and Botany: the Story of Pollination.” Dr. Howard was a speaker on the campuses of Duquesne and Pittsburgh universities and the State University Teachers College, Oneonta, New York, under the sponsorship of the American Institute of Biological Sciences. He was the American Association for the Advancement of Sciences speaker at Franklin and Marshall College in Lancaster, Pennsylvania, and was the principal speaker for the Pennsylvania Catholic Round Table of Science at its annual meeting in Pittsburgh. Dr. Wyman was the principal speaker at the annual meetings of the Colorado Nurserymen and the Arborist Association, the Michigan Nurserymen’s Association, and the Pennsyl- vania Nurserymen’s Association. He was also the horticultural speaker at the University of Maine’s Farm and Home Week and took part in the nurseryman’s short course offered by the Waltham Field Station. Exhibits and Displays: The annual Spring Flower Show of the Massachusetts Horticultural Society was held at a new location, Wonderland Park in Revere, Massachu- setts. As display space was limited, the Arboretum staff was asked to display again a pruning exhibit which was used five years ago. This exhibit, demonstrating the methods and equipment of pruning shrubs and trees, also contained examples of plants showing both correct and improper pruning. While the exhibit lacked the color of forced flowers it was awarded a gold medal. Interest in the exhibit was such that it was 438 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI again set up in the lecture hall in Jamaica Plain for the use of visitors and classes during the spring. An open house was held at the Case Estates on May Ist to allow inspection of the nursery area, ground-cover and small-tree demonstra- tion plots and other plantings. During the day staff members were on the grounds to answer questions. A rainy day reduced the crowds which eh ne created a parking and traffic problem in the vicinity. As a result of such demonstrations the number of visitors and the number of tours in Weston increase each year. A portion of the Arnold Arboretum exhibit on the aaa of the United States exhibition at the Floriade, in Rotterdam, the Netherlands, June, 1960. From March to September, 1960, an International Horticultural Ex- position, the Floriade, is being held in Rotterdam, the Netherlands. The American exhibit is under the sponsorship of the Foreign Agricultural Service of the Department of Agriculture and the American Horticultural Arboretum. Dr. Wyman prepared a list which, with few exceptions, consists of plants available in European nurseries. a few plants not avail- able in Europe were secured from our own nurseries or from commercial sources and shipped to Holland. The Arboretum plants in the American exhibit comprised part of a private arboretum as well as established landscape plantings around a typical American home. The selection 1960 | THE DIRECTOR’S REPORT 439 of sixty-four taxa was made to afford some bloom or point of attraction throughout the six month period of the Floriade. Elsewhere in the ex- hibit, the American Association of Botanic Gardens and Arboreta pre- pared a display of Kodachrome transparencies, photographs, and technical data, illustrating and describing American arboreta and botanic gardens. Dr. Howard served on the organizing committee, and the staff supplied pictures and details concerning the Arnold Arboretum. The lecture hall in Jamaica Plain was again the location for temporary displays of flowering plants for classes and various special meetings. The 24th Inter-scholastic Judging and Identification Contest in Floriculture and Ornamental Horticulture was held there in April. Nearly 100 students took part in the team and individual competitions. Special displays of Arboretum plants were also prepared both at Jamaica Plain and at the Harvard Club in Boston for the annual meeting of the Associated Harvard Clubs. A special display of ornamental crab apples was prepared and exhibited at the Fall Show of the Massachusetts Horticultural Society. Comparative Morphology: Professor I. W. Bailey, Professor of Plant Anatomy, Emeritus, con- tinued to serve as curator of the wood collection of the Arnold Arboretum. In addition, Professor Bailey is continuing his investigations on the com- parative anatomy of the Cactaceae under a three-year grant from the National Science Foundation. Thus far, he has devoted his attention largely to the leaf-bearing genera Pereskia, Pereskiopsis and Quiabentia. If these putatively primitive representatives of the Cactaceae are to be utilized as an initial fundamental basis in studying salient trends of phylogenetic specialization within the family, and in obtaining evidence of valid taxonomic significance, it is essential that both the exomorphic and the endomorphic structures of all organs and parts of the three genera be thoroughly investigated. Dr. Uttam Prakash, of the Birbal Sahni Institute of Palaeobotany, Lucknow, India, has been a visiting scholar during the year and has utilized the wood and slide collections of the Arboretum in connection with his investigations of the fossil floras of central India. Library: The efforts of the library staff were again concentrated on classifying and cataloguing the extensive reprint and pamphlet collection. This collection, formerly alphabetized by author as a separate unit, has now been distributed among the various library categories (e.g., monographs, floras, forestry, etc.). Three hundred fifty-six pamphlets were added to the collection, making the total number 17,340 on June 30, 1960. T'wo hundred sixty-eight books were added to the bound volumes bringing the total of 50,783 volumes. Fourteen hundred cards were added to the main catalogue, an indication of the extensive system of cross referencing of pamphlets and volumes now being developed. 440 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLI Requests for the interlibrary loan of books remained high and 177 volumes were sent on loan from the combined libraries. An equal number of requests was handled by supplying contura copies, microfilms or photostats. The cost of such reproduction is charged to the person or institution making the request whenever possible. Filling requests in this manner requires more effort on the part of our own staff but reduces the wear on the books and allows them to remain available to our own staff. The Gray Herbarium Index to American Plants, maintained in Jamaica Plain, received 3000 new entry cards during the year. The Linnaean Society of London made available 825 microfiche re- productions of the Linnaean Herbarium. A set of these cards was pur- chased jointly with the Gray Herbarium and will be available in the library. Botanical Congress, which was attended by Arboretum staff members, photo- graphed at McGill University, Montreal, Canada, on August 18, 1959. Herbarium: During the year 8,766 specimens were mounted and added to the herbarium, bringing the total collection to 714,136 specimens on July 1, The specimens added were in part received during the year and in part collections on hand from previous years that required study or preparation before mounting. During the past year 11,767 specimens 1960 | THE DIRECTOR’S REPORT 441 were received: 10,585 in exchange, 1074 by gift, 54 for special identifica- tion, and a similar number by special subsidy. The largest numbers of specimens came from institutions or collectors in Asia with 2931, 2523 and 1458 specimens from Eastern Asia, Western Malaysia and Papuasia respectively. During the year the Arboretum staff sent 1,664 specimens in exchange to other American institutions and 6,277 to institutions in other countries. The staff filled ninety-eight requests for loans which comprised 10,903 specimens from the combined Arnold Arboretum and Gray herbaria borrowed by fifty-three different institutions, thirty-six in the United States and seventeen foreign countries from Canada to Argentina and from England to Japan. For their own work or that of students, the staff of the combined herbaria borrowed sixty loans of 5290 specimens from thirty-five institutions, sixteen in the United States and nineteen abroad. The outgoing loans averaged 111 specimens, while the incoming loans averaged 88 specimens. These figures clearly indicate the relative wealth of the herbarium collections in the Harvard herbaria and the desire of other botanists to consult this material. The research of the herbarium staff, in part supported by special gifts and grants, involves floristic, monographic and anatomical studies. Work is in progress on the floras of the Lesser Antilles, the southeastern United States and several areas of Asia. Monographic studies of genera in Chloranthaceae, Polygonaceae, Theaceae, Thymelaeaceae and Urticaceae are progressing. Three staff members are using detailed anatomical studies for taxonomic purposes in broad surveys of plant structures and in specific scientific problems. The results of this work will be reflected in future papers similar to those cited in the bibliography of the staff. Travel and Exploration: During July 1959, Dr. Howard made a final trip to Hawaii in order to record growth data on some experimental projects established on the island of Kauai. On this trip he completed a report on the use and re- habilitation of agricultural lands following strip mining for aluminum ores. The report was submitted to the Commissioner of Public Lands and was used for reference in preparing mining-rehabilitation legislation for the state of Hawaii. During two weeks of December and January, Dr. Howard and Dr. Wagenknecht collected in Florida, Jamaica and the Cayman Islands. The primary purpose of this trip was to observe agri- cultural and forest plantings on mined-out bauxite lands in Jamaica. The opportunity was taken to collect specimens of cultivated ornamental shrubs and trees in all three areas. The Arboretum contributed to the support of the expedition of Drs. Ernst and Lucy Abbe, of the University of Minnesota, to Thailand, Cam- bodia, the Malay Peninsula, and Borneo. The purpose of this expedition was a study of the Asiatic oaks and their relatives. Dr. Abbe sent acorns and other fruits from the mountain areas for trial at the Arboretum 442 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XLI where some species from the area are now under cultivation. Approxi- mately a thousand collection-numbers of herbarium specimens were gathered by this expedition. Gifts and Grants: During the spring the Arboretum was notified that it had been desig- nated as one of three beneficiaries of a perpetual charitable trust fund established under the will of Mrs. Martha Dana Mercer, of Doylestown, Pennsylvania, a long-time friend of the Arboretum, who died February 21, 1960. The fund is being handled by the Old Colony Trust Company of Boston with the income to be paid to the Arnold Arboretum. The use of this generous gift is unrestricted, and the income when available will be applied to improvement of the physical facilities and for increases in the staff of the Arboretum. Again this spring the “Friends of the Arnold Arboretum” were generous in their response to the annual appeal for gifts to support the horticultural activities of the Arboretum. Even though such gifts are, with few excep- tions, unrestricted they have been used for the most part in the areas of plant propagation and the care of the living collections. Special gifts or grants from individuals and foundations were received to support the research of various staff members. A gift from Mr. George R. Cooley will support the collaborators of Dr. Wood on the flora of the southeastern United States. In addition, Dr. Reed C. Rollins, Director of the Gray Herbarium, and Dr. Wood received a three year grant from the National Science Foundation for the preparation of a biologically oriented generic flora of the southeastern United States. Dr. Perry re- ceived a one year grant from the Public Health Service for investigations on the native medical uses of plants of southeastern Asia. This grant, which may be renewed twice, is utilizing the extensive herbarium collec- tions of the Arnold Arboretum from that area. Grants from the National Science Foundation to Dr. Bailey and Dr. Howard for their research have been continued. Publications: Under the editorship of Dr. Wyman, Volume 20 of Arnoldia was issued during the year. As a policy the articles in Arnoldia are of timely horti- cultural interest. The issues appear at irregular intervals with twelve numbers comprising each volume. The Journal of the Arnold Arboretum, by contrast, is issued quarterly on a regular basis with technical articles largely representing the research of the staff members. Volume 40 of the Journal was issued during 1959, Dr. Wyman advised and codperated with the Electric Council of New England in the publication of a twenty-four-page booklet, “Trees in Your Community —- A Handbook of Selected Trees for New England.” The recommendations were often based on growth habits of trees observed in our own small-tree demonstration plot in Weston and on the Arboretum 1960] THE DIRECTOR’S REPORT 443 grounds in Jamaica Plain. The booklet is being distributed by New England wire-using utilities to home owners interested in the subject. Orders for 156,000 before printing indicate the need and popularity of such handbooks. Bibliography of the Published Writings of the Staff and Students July 1, 1959 — June 30, 1960 BarLey, IrviNc WipMer. Some useful techniques in the study and interpreta- tion of pollen morphology. Jour. Arnold Arb. 41: 141-148. 1960. Howarpb, RicHArD ALDEN. An additional note on Tillandsia paniculata, Brome- liad Soc. Bull. 9: 54-56. 1959. _ The Director’s Report. The Arnold Arboretum during the fiscal year ended June 30, 1959. Jour. Arnold Arb. 40: 420-436. 9 _ Studies in the genus Coccoloba, VII. A synopsis and key to the species in Mexico and Central America. Jour. Arnold Arb. 40: 205-220. 1959. _ Studies in the genus Coccoloba, VIII. Nomenclatural changes. Jour. Arnold Arb. 41: 40-46. 1960 _ Studies in the genus Coccoloba, IX. A critique of the South American species. Jour Arnold Arb. 41: 213-229. 0. Hv, Suru-yinc. Chinese hollies. Tsing Hua Jour. Chinese Stud. Spec. No. Nat. Sci. 1: 150-184. 1959 _ A monograph of the genus Paulownia. Quart. Jour. Taiwan 12: 1-54. 1959. _ A revision of the genus Clethra in China. Jour. Arnold Arb. 41: 164- 190. 1960. Jounston, IvAN M. Some noteworthy American Borages (Studies in the Boraginaceae, XXX). Wrightia 2: 13-22. Kopuskt, CLARENCE E. A revised key to the Chinese species of Jasminum. Jour. Arnold Arb. 40: 385-390. 1959. NEVLING, Lorin I., Jn. Araliaceae. /”: Woodson, R. E. e¢ al. Flora of Panama. Ann. Missouri Bot. Gard. 46: 223-242. 1959. _ A revision of the genus Daphnopsis. Ann. Missouri Bot. Gard. 46: 257-358. 1959. Sax, Kary. The cytogenetics of facultative apomixis in Malus species. Jour. Arnold Arb. 40: 289-297. 1959. Meiosis in interspecific pine hybrids. Forest Sci. 6: 135-138. 1960. Standing room only. (Revised paperback edition.) Boston. 1960. SCHWARTEN, LAzELLA. Index to American botanical literature. Bull. Torrey Bot. Club 86: 147-157, 206-217, 270-282, 353-309, 1959: 87: 73-83, 161-171, 229-245. 1960. Tuomas, Joas. A monographic study of the Cyrillaceae. Contr. Gray Herb. 186: 1-114. 1960. WAGENKNECHT, BurpeTTEe L. Christmas plants around the world. Arnoldia 19: 50-77) pl. 10-15, 1959) _ Note on Elaeagnus commutata. Rhodora 62: 29. 1960. - Revision of Heterotheca, section Heterotheca (Compositae). Rhodora 62: 61-76, 97-107. 1960. WateRMAN, ANN H. The mints (family Labiatae) of Michigan. Mich. State Univ. Biol. Ser. 1: 271-302. 1960 444 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI WEBER, CLAUDE. Botanical motifs of the coats of arms on stamps. Topical Times 11: 14-19. 1959, National flowers and trees on stamps. Topical Times 10: 200-202. 9 ——.. ms on postage stamps. Principes 4: 9-16. 1960. WILsSon, KENNETH A. The genera of Hydrophyllaceae and Polemoniaceae in the southeastern United States. Jour. Arnold Arb. 41: 197-212. 1960. The genera of the Arales in the southeastern United States. Jour. Arnold Arb. 41: 47-72. 1960. The leptosporangium of the eae a fern Anarthropteris dic- tyopteris. Contr. Gray Herb. 187: 53- : e sporangia of three Sree ne of Polypodium. Am. Fern Jour. 49: 147-151. 1959. (with C. E. Woop, Jr.). The genera of Oleaceae in the southeastern United States. Jour. Arnold Arb. 40: 369-384. 1959. Woop, Carroty E., Jr. The genera of Sarraceniaceae and Droseraceae in the southeastern United States. Jour. Arnold Arb. 41: 149-163. 1960. ‘ e genera of Theaceae in the southeastern United States. Jour. Arnold Arb. 40: 413-419. 1959. (with R. B. CHANNELL). The genera of the Ebenales in the south- eastern United States. Jour. Arnold Arb. 41: 1-35. ———. The genera of Plumbaginaceae in the southeastern United States. Jour. Arnold Arb. 40: 391-397. 1959, . The genera of the Primulales of the southeastern United States. Jour. Arnold Arb. 40: 268-288. 1959. (with K. A. Witson). The genera of Oleaceae in the southeastern United States. Jour. Arnold Arb. 40: 369-384. 1959, Wyman, Donatp. The dwarf evergreens. Horticulture 37: 472, 473, 485, 486. 1959. w hawthorns selected from numerous species. Am. Nurseryman 110(9). iG 11, 98-101. 1959, . Firs eontabute stately specimens to northern gardens. Am. Nursery- man 110(11): 12, 13, 70-72, 1959. Hydrangeas include old favorites in limited list of garden species. Am. Nurseryman 110(3): 14, 15, 53-55. 1959. . Magnolias hardy in the Arnold Arboretum. Arnoldia 20: 17-28. 1960. . Ma forms and uses rank maples high among shade trees. Am. Nurseryman 110(7): 10, 11, 88-94. =o Maples liven the autumn scene. Horticulture 37: 512, 513, 541. 1959. . Plant introduction by the Arnold Arboretum, Bull. Univ. Wash. Arb. 22: 74-76, 100. 1959. . Plants of possible merit? Arnoldia 20: 9-16. 1960 . Shrub honeysuckles with pink to red flowers. Aumolaia 20: 29-32. 1960. Sorting the woody ornamentals — Dogwoods yield varied foliage, flowers, fruits. Am. Nurseryman 111(9): 10, 11, 105-111. 60 Sorting the woody ornamentals — Few spireas popular despite range in height and blooming time. Am. Nurseryman 111(11): 10, 11, 71-76. 1960. Sorting the woody ornamentals — Magnolias provide flowering trees for almost all sections. Am. Nurseryman 111(7): 12, 13, 69-80. 1960, 1960] THE DIRECTOR’S REPORT 445 Sorting the woody PE a ee eae popular for spring white. Am. Nurseryman 111(5): 15, 68-73. . Sorting the woody cae a second ion at the popular elm. Am. Nurseryman 111(1): 12, 13, 78-82. 1960. . Sorting the woody ornamentals — The variable Euonymus. Am. Nur- seryman 111(3): 16-18, 93-96. 1960. —. These make spring focal points. Horticulture 38: 207. 1960. . Viburnums. Arnoldia 19: 47-56. 1959. . Viburnums valued for varied species useful in gardens. Am. Nursery- man 110(5): 14, 15, 61-71. ; . Weigela list reflects changing styles. Am. Nurseryman 110(1): 12, 13, 30-33. 1959. . Why not plant many viburnums? Horticulture 37: 34, 35, 41. 1960. RicHarp A. Howarp, Director 446 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI Staff of the Arnold Arboretum 1959-1960 RicHAarp ALDEN Howarp, Ph.D., Arnold Professor of Botany, Professor of Dendrology, and Director. IrvING WipMER BalLey, S.D., Professor of Plant Anatomy, Emeritus. Jos—EpH Horace FAuLL, Ph.D., Professor of Forest Pathology, Emeritus. Kari SAx, S.D., Professor of Botany, Emeritus. MicHaEL ANTHONY CaANoso, M.S., Curatorial Assistant.* ANN CtosE, A.B., Business Secretary.** Henry Draper, Superintendent, Case Estates. ALFRED JAMES ForpuHaM, Propagator. HeEMAN ARTHUR Howarp, Assistant Horticulturist. SHIU-YING Hu, Ph.D., Botanist. FRANCES Mary JARRETT, Ph.D., Botanist. Ivan Murray Jounston, Ph.D., Associate Professor of Botany. CLARENCE EMMEREN Kopsusk1I, Ph.D., Curator.* MARGARET CATHERINE LEFAvouR, Herbarium Secretary. Mary SCANLAN LEHMER, Ed.M., Business Secretary.*** SusAN DreLtano McKetvey, A.B., Research Associate. Lorin I. Neviine, Jr., Ph.D., Assistant Curator. Lity May Perry, Ph.D., Botanist. LAZELLA SCHWARTEN, Librarian.* JoasB Lancston THomas, Ph.D., Cytotaxonomist. BuRDETTE LEwis WAGENKNECHT, Ph.D., Horticultural Taxonomist. ANN H. Waterman, M.S., Assistant Librarian.* CLAUDE WEBER, Botanist, Special Project.* KENNETH ALLEN WILson, Ph.D., Botanist, Southeastern Flora Project.* ROBERT GEROW WILLIAMS, B.S., Superintendent. CARROLL EMory Woop, Jr., Ph.D., Associate Curator and Editor. DonaLp Wyman, Ph.D., Horticulturist. * Appointed jointly with the Gray Herbarium ** Appointed June 15, 1960 #4" Resigned June 15, 1960 INDEX Achras, 13 Artocarpus brevisericea, 130 Acorus, 50-51 —callophylla, 328 Albizzia moluccana, 410 —canarana, 137 Aleurites canes 410 —cannoni, 137 Alsodeia japurana, 244 —cerifera, 137, 328 Amaranthus fy Rec 409 —cumingiana, 84, 95 Bee 40 — -— stenophylla, 100 —virl 09 — dadah, 83, eee arvensis, 411 — — pubes ,9 Anamomis, 276 = dasyphyl, o 103 — dicrana, 276 — — flava, —simpsonii, 276 Sea Anatomy of Phenakospermum (Musa- an. macrophylla, 139 ceae), The, 287 — eberhardtii, 12 Rane Seat 103 —— poilanei, 128 Anubing, 85 —elongata, 137 Aporosa nitida, 138 —erythrocarpa, 92, 103 Araceae, 47-50 — exsculp 37 am. Aroideae, 58 — ficifolia, —subfam. Calloideae, 51 — finlaysoniana, 139 —subfam. Colocasioideae, 56 — foeniformis, —subfam. Philodendroideae, 55 — forbesii, 137, 333 — subfam. Eee. 61 — fretessii, 82, age —subfam. Pothoi , 50 — frutescens, 1 Arales in the qn United States, — fulvicortex, oY 116-117 The Genera of the, 4 — glaucescens, 118 Araucaria excelsa, 408 — —tephrophylla, 119 Argyreia, 317 —glaucus, 80, 118-121 Arisaema, 58-61 — — villosiusculus, 119 rrow-arum, 56 gomeziana, Artocarpus and Allied Genera, Studies in, —— griffithii, 1 IV. A Revision of Artocarpus Sub- — gomezianus, a 88-91, 128 genus Pseudojaca, 73, 111. V. A Re- —-—gomezianus, 82, 88-9 vision of Parartocarpus and Hullettia, _— — zeylanicus, 90-91 320 — griffithii, 128 Artocarpus sect. Glandulifolium, 134 — humilis, 126 —sect. Pseudojaca, 73, 83 ate 80, 130, 132-133 —ser. Clavati, 130 — inaequa 11 —ser. Peltati, —inconstantissima, 92 —subg. Pseudojaca, 73-83 —involucrata, 137, 331 cuminatissima, 8 Seas , Ae — africana, 136 — lacuc —altissimus, 80, 134, 135-136 apnea Te 81, 83, 90, 92, 111-114 — angustifolia, 1 wae Say, 88, 9 — antiarifolia, 100 —malayan 2 — bengalensis, 111 — lamellosa, fen 123, 139 — bicolor, 133 — lanceolata, 102, 123, 128 —pbifaria, 139 — leytensis, — biformis, 118 —limpato, 13 — blumei sarawakenss, 139 —lingnanensis, 124 borneen —longifolius, 81, 83-84 eee eee pe 324 — madagascariensis, 139 448 Artocarpus masticata, 88 scare tenuis, 81, 107- 108 — riedelii, 138, 330 — rotunda, — rotundifolia, — rbrovenius, of —rufescens, 91, 32 —sampor, 124 102, 114-115 wn re) iat) jo] Q oO =} a _ ue oo — styracifolius, 80, 133-134 — subrotundifolius, 81, 108-109 1 — — papillosus, 99-100 Parish | Je inocampus. oF 83, 102-103 — yunnanensis, — zollingeriana, ae Arum Family, 47 Asplenium insiticium, 408 Athyrium mic — 408 Axonopus affinis Comparative Anatomy a I. Folia BAILEY, the eat bearing Cactaceae, JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI Vasculature of Pereskia, Pereskiopsis and Quiabentia, 341 BaiLey, I Some Useful Techniques in the udy and in ae of Pollen nee 14 Barhal, 1 Batatas, a Betoh, 12 Bibliographic Data on the aa Gar- dens Books, 1911-1941 ers cel of the ean aa Writings he Staff and Students July 1, 1959- 4 Bindw 08 Biccant aa 408 Bog-bugles, 15 Bonami pine. 153 Cactaceae, Comparative Anatomy of the Leaf- hearin ng, I. Foliar Vasculature of Pereskia, Pereskiopsis and Quiabentia, 341 Caj ajanus cajan, 410 3 Campderia gracilis, 241 Canavalia sericea, 41 Carex sp., a 8 Carlomohr 6 Cassia eechenw ea 410 Casuarina equisetifola, 410 Catch-fly, Cay cha i Centella asiatica, 411 CHANNELL, B. and C. E. Woon, Jr. e the Ebenales in the Southeastern United States na, A Revision of the Gents Clethra Chloris inflata, 4 Chrysophyllum, Chrysopogon aciculatus, 409 Chytraculia, 2 Chytralia, 274 1960] INDEX Cibotium chamissoi, 408 Coccoloba argentinensis, 218 Cinnamomum camphora, 410 —ascendens, 218-219 Cladium meyenii, 408 — barbadensis, 249 Clethra in China, A Revision of the —barbeyana, 219, 232 Genus, 164; History, 165; Taxonomic — billbergii, 219, 358 Characters 167; Ec alosical Notes and eae a 360 Geographical Distibition 174; Tax- —bolivarana, 2 0 Apel — bracteolosa, a 360 Clethra, 177-1 — brasiliensis, 22 Clethra, 180 —candolleana, 221-222, 225 —sect. Cuellaria, 187 — caracasana, —sect. Euclethra, 180 —_—— 22 —sect. Monostigma, 179 — caribaea, 40, 227 —ser. Alnifoliae, 184 — carinata, —ser. Delavayanae, 180 —caurana, 222, 236 — ser. Esquirolianae, 183 — cerifera, er. Faberianae, 186 — chacoensis, 223, 381 —alnifolia, 177, 180, 184 —charitostachya, 223 — barbinervis, 184-1 —conduplicata, 225 — bodinieri, 179-180 —confusa, 223 — — bodinieri, 179 —cordata, 225-226 —— latifolia, 179 — a ey —— parvifolia, 180 — cordifolia, 226, 245 — brammeriana, 188 — coriacea, 3! — canescens, 185-186 Corona 40-41, 226-227 — cavaleriel, 183 —corrientina, 227, 366 — delavayi, 180, 181- oe — crescentiifolia, nw 217, 227-228 — delavayi, 181-18 —— obtusata, 44, 217 — — glabra, — ere: 228 ——] — cujabensis, 228-2 — esquirolii, eh 184 —cyclophylla, 222, 229 — faberi, 186-187 — cylindrostachya, 229, 239 — fargesii, 185-186 darienensis, — ferruginea, 187 — declinata, 223, 231 —kaipoensis, 187, 188 —— major, — kwangsiensis, ta 189 —— minor, 223 —liangii, 186 vello a, 231 — lineata, —densifrons, 232-233 ene bracteats 183 —dioica, 40, 227, 233 —monostachya, 180-181 — diversifolia, — pinfaensis, 187 — douradensis, 45, 232, 251 polyneura, 189 — dugandiana, 2 — sinica, —erecta, 234, 367, 379 — sleumeriana, 185 —ernstil, 228, 23 est 189-190 Phssewe 234-235 ra, 367 See Sane in the Genus, VIII. —excoriata, 235 Nomenclatural Changes, 40. IX. A — fagifolia, Ae 236 Critique of the South ererican Srecies — fallax, 236 213, 231, 357 — fasciculata, 44, te 236-237 Coccoloba aaa 215-216 — fastigiata, 237-23 —acuminata, 216-2 —— glabrata, 238, ar —-— glabra, 21 — ferruginea, 2 — ubescens, 216 — filipes, 238-239 —alagoensis, 217 — firma, —alJnifolia, 2 — gar 17 — Mercere 44-45, 217-218 — glaziovii, ae 241 450 JOURNAL OF THE ARNOLD ARBORETUM Coccoloba goudotiana, 237, 241, 357 — guaranitica, 243 — — opaca, — guianensis, 45, 243, 250 —gymnorrhachis, 243-244 —- cae 248-249 — longependula, 249, 383 — longiochreata, 228, 249 s>anICRO puntata, 234, 252 tachya, 219 — — ovalifolia, "358 — mollis, 253-255, 372 — novogranatensis, 40, 227, 257 —nutans, 257-258 —nymphaeifolia, 258, 367 5 | ° o a) 3 R 2 Rees _ 360 — orbicular ee nek ee Coccoloba ovata, 361-365 —— hostanakt, 234 n 0 Bel oes 231, 372 —ramosissima, 372-373 —recurva, 45, 373 —riedelii, 373, 374 — rigida, 373-374 —sp., 380-381 — spruceana, 382-383 [ VOL. XLI 1960] Coccoloba squamosa, 383 — sticticaulis, 383-384 — striata, 384-385 — stricta, 385 — strobilulifera, 185, 385 — sublobata, 240, 385 — trianael, —trinitatis, 45, 251 —uvifera, 387-388 — vellosiana, 44, 217, 388 —venosa, 235, 257, 388 — warmingil, — williamsil, — zernyi, 390 Coccolobis padifolia, 249, 365 Coix lacryma-jobi, 409 i 8 389 248, 389 341; Species with Dominantly Pinnate Pseud Venation, 343; Pre- liminary Observations upon the Micro- scopic Structure of Veins, Veinlets and eaf Traces, 343; Discussion, 346 Conanthus, 204 Convolvulaceae in the Southeastern Unit- ed States, The Genera of, 298 298-301 Crotalaria incana, 4 Cuellaria, 187 Cuphea carthagenensis, 410 Cuscuta, 301-304 Cynodon dactylon, 409 Cyperus cylindrostachyus, 408 —kyllingia humilis, 408 —rotundus, 408 Dacrydium taxoides, 36 iene aa aaa 413 — laget ne a INDEX 451 Daphne tinifolia, 413 Daphnopsis (Thymelaeaceae), Nomencla- tural Changes in, 412 Dheu, 112 Dionaea, 158-160 Dieseores bulbifera, 409 Diospyros, 17, 22 Dipholis, 6-7 Biphziucs sandwichianum, Director’s Report oa 408 , rnold Arbore- tum ee the a ee 441; Gifts and Grants, Publications, 442; Bibliography of ae Published Wiitines of the Staff and Students July 1, 1959— June 30, 1960, ee Dodder oa Cope 338 Downy-myrtle, 278 Drosera, 157, 160-163 Droseraceae in the So Sees utheastern Unite Aa Genera of ae eee 1 Droweactae 156-157 ria cordata, 409 Doe dentata, 408 — gongylodes, 408 Duckweed, Duckweed Family, 63 Dumb-watches, 153 452 Ebenaceae, 17-18 Ebenales in the Southeastern United States, The Genera of the, 1 Eclipta a aaa Emilia sonchifolia, 410 Enkianthus (Ericaceae), Polyploidy in, 191 Enkianthus campanulatus, 193 — albiflorus, 193 —perulatus, 193 — subsessilis, 193 Ensete, 296 Erechtites valerianifolia, 410 Erigeron canadensis, 410 aceon. 311 Ficus cannoni, 137 —exsculpta, 1 ee 92 mpan Fibro ee 408 Fly-tra See a ae 409 Frog-bonnets, 153 Galax, Genera of Convolvulaceae in ie South- eastern United States, The, 2 Genera of the Ebenales in the South- eastern United States, The, Genera of Hydrophyllaceae or Pole- moniaceae in a Southeastern Date States, The, Genera of a in the Southeastern United States, The, 2 Genera of Sarraceniaceae and Drosera- ceae in the Southeastern United States, The, 152 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLI Gibbsitic Soils in Hawaii, Vegetation on, 391 Gold thread, 301 A Taxonomic Revision a Solera XII. Section Micro- carpus, 36 Greek. elrea: 210 277 Gymnartocarpus, 320 — triandra, 333 — venenosa, 320, 328 — woodii, 331 Hailweed, 301 Ww _— Halesia, 26- Hargasseria mexicana, 413 413 112 Hawaii, Weeetatlon on Gibbsitic Soils in, 391 Hedyosmum Rg gees 221 Heliconia, Sot Hellbine Hibiscus tice 410 Hillcrest Gardens Books, 1911-1941, Bib- liographic Data on the, 318 Hill- eg or Hopea, = welt 137, 138 p A. Bibliographic Data on the Hillcrest Gardens Books, 1911- 1941, 318 Howarp, Ricuarp A. The Director’s Re- por 29 Howarp, RicHarp A. Studies in the Genus Coccoloba, VIII. Nomencla- es, 40; IX. A Critique of the South American Species, 213, 231, 357 Hu, SHiu-yinc. A aa of the Genus Clethra in China, Hullettia, Studies in ks and Al- lied Genera, V. A Revision of Pararto- 0 car and, Hullettia, 334-337 — dumosa, 337-338 1960] Hullettia griffithiana, ae 338-340 Hung se: anaes 12 Hydrolea, Sine and Polemoniaceae in the Southeastern United States, The Genera of, 19 Hydrophyllaceae, 197-198 Hydrophyllum, 199-200 Indian turnip, 58 Ipoh, Ipomoea, 311-316 — pes-caprae, 410 Ipomopsis, 211-212 Ironwood, 8 Jack-in-the- See 58 pus and Allied Genera, IV. A Revi- sion of Artocarpus Subgenus ee jaca, 73, 111; V. ae eee of Parar carpus and Et lettia, 320 Jussiaea Sees lizustrifolia, 411 Kanagona, 90 Kelembi, 10 Kingdonia uniflora, 265 Klutum a Kurzia, Kwai nae 125, 132 Lakuch, 112 Lantana camara, 411 413 Leaf Venation and Pubescence in Lemnaceae, 63-66 Lobelia sp., 410 Love vine, 301 Lowi, [ener 11 Lychnis, 207 eee cernuum, 408 Ma hat, 125 Maba, 17, 18 Macadamia ternifolia, 411 Macrocalyx, 2 INDEX 453 Manilkara, 2, 13-17 Marilaunidium, 204 Marion Roby Case, 318 Mastic, Mastichodendron, 5-6 M i, 104, 107 Melnctoma ye a 410 Merremia, 310-31 ee ee polymorpha, 411 — spur eee setosa, 408 eee scolopendria, 408 Mimosa pudica Mimuso 4 1 Minuartia aretioides, 262 26 S oD ad ° Q endron, 26 Monographic Study of the West Indian Bee of Phyllanthus, Supplement to Mess J. C. and TAKAHASHI. Veretiaod on Gibbsitic ne in Hawaii, 391 Morning-glory, 311 Morning-glory Family, 298 Morus altissima, 135 Mosiera Muhlenbeckia eee 233, 377 —tamnifolia laxiflora, 255 Musidendron, Feil Myauklok Mo rete Myrtaceae in the ee United States, The Genera of, Myrtaceae, 270-272 —subfam. Leptospermoideae, 273 Ean hiay Myrtoideae, 273 Myrtle-of-the- eer 274 Myrtus, 276-277 Nakedwood, pe Nama, 204, Nemophila, Nephrolepis eae furcans, 408 0 NEVLING, Lorin I, Jr. Nomenclatural 454 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. XLI Changes in Daphnopsis (Thymelae- Phenakospermum (Musaceae), The Anat- omy of, 287 aceae), Nomenclatural Changes in Daphnopsis Plenakornen rmum guianense, 287 Thymelaeaceae), Phlebodium aureum Nomenclatural Changes, Studies in the Phlox, 207-21 Genus Coccoloba, VIII, 40 Phlox Family, 2 Nordmannia tinifolia, 413 Phyllanthus, ae eee to a Nyctelea, 20 graphic Study of the West Indian Species of, 27 Operculina, 3 Phyllanthus sect. ee 279 Oplismenus | ae 409 — sect. Botryanthus Orchidantha, 294, 296 — sect. Conam i, 281 Orontium, 54-55 — sect. Elutanthos, 281 Otamb, — sect. Epistylium, 281 Oxalis corniculatus, 411 — sect. Floribundi, 279 — sect. Kirganelia, 279 Palms and Their Morphological Signifi- | — sect. Nothoclema, 281 cance, Seedling Leaves in, 414 — sect. Phyllanthus, 280 Pandanus odoratissimus, 40 —sect. Xylophylla, 282 — variegatus, — subg. Kirganelia, 279 Panicum purpurascens, 409 — subsect. Niruri, 280 Parartocarpus and Hullettia, Studies in oe Swartzient: 280 Artocarpus and Allied Genera, V. A —ama Revision of, 320 ee 285 Parartocarpus, 320-324 — arbuscula, 285 — beccarianus, oe 328 — axillaris, 282 — borneensis, 33 — caribaeus, 280 — bracteatus, oe 324-326 — casticum, 279 — excelsa, 333 — cauliflorus, 282 — involucrata, 331 a 282 —pa eve oo. 334 — debilis — en — ine 285 , 334 — exiImius, 283-286 — a4 333 — fraternus, 280 — venenosus, 137, 138, 320, 326-334 — mimicus, 280 — — borneensis, 332-333 —niruri, 280 — — forbesii, 137, 333-334 — nutans, 281 — — papuanus, 330-332 —-— nutans, 281 — — venenosus, 328-330 — reticulatus, 279 — woodii, 331 — stipulatus, 280 Paspalum conjugatum, 409 — tenellus, 279 — orbiculare, 409 Phytolacca acinosa, 411 —urvillei, 409 Pistia, 61-63 Passiflora edulis, 411 Pistiaceae, 47 — foetida encase: 411 Pitcher- plant pi 152 Peltandra, 55-56 Pitcher p Peperomia membranacea waimeana, Pleuropeltis thunbersian, 408 Pereskia, Pereskiopsis and Qu ahem Pluchea odorata, Comparative Anatomy of the Leaf- Podocarpus, A a nomic apes of, inane Cactaceae, I. Foliar Vascula- XII. Section Wiicroanea S, ture 1 Podocarpus ustus, 36-3 Pereskiopsis and Quiabentia, Compara- Polemoniaceae in ‘the Southeastern United ive omy of the Leaf-bearing States, The Genera of Hydrophyllaceae rman I. Foliar fi ees of and, 197 Pereskia, 341 a 05-207 Persimmon, 18 —su . Cobaeoideae, 206 Phacelia, 202-204 i Polemonioideae, 206 Pharbitis, 311 — tribe Bonplandieae, 206 1960] Polemoniaceae tribe pea 206 ribe ea 206 Sees 206, 207 Polemonium, 206, 210-211 Pollen MeaEneS Some Useful Tech- niques in the Study and Interpretation of, 141 Polygonum arborescens, 43, 44, 217 23 — declinatum Phe eeger 43- 44, 388 — uvifer Polyplody in Enkianthus (Ericaceae), pater es oleracea, 411 Pouteria, 11-1 Prainea frutescens, 137, 334 38 Ww 08 Pteridium aquilinum decompositum, 408 Pull down, 301 Quamoclit, 3 Quiabentia, ae Anatomy of the Leaf- pone Cactaceae, I. Foliar Vas- culature of Pereskia, Peresiapsis and, 341 Quiina pteridophylla, 266 Rademachia rotunda, 139 —subg. E — eximia, — glabra, 268 — goyenli, 268 — monroi, 268 INDEX Raoulia parkii, 268 — petriensis, 268 Ravenala, 287, 295, 296 Revision of the Genus Clethra in China, A, 164 Rhodomyrtus, 278 — tomentosa, 411 Rhynchospora lavarum, 409 Richardia scabra, 411 Rinorea ee 244 Robertia ees zernyl, 390 Sacciolepis contracta, 409 Sadleria aaa 408 — subfam. ‘Mimusopoidene, 13 — subfam. Sideroxyloideae, 5 Sapote Family, 2 Droseraceae in the Southeastern United States, The Genera of, 152 Sarraceniaceae, 152-153 Satin-leaf, 10 Sax, Hairy J. ge in Enkianthus (Ericaceae), Scaevola ratesen sericea, 410 — gaudichaudiana, 41 Schinus rbinthifaiv, 409 ] and Their Significance, 414; Ob- servations, 415; Discussion, 420; Sum- ry, Selanking, 12 Setaria geniculata, 409 Silverbell-tree, ae Skunk-cabbag Sloetia ay aa), 139 T. ae Venation and Pubescence in the Genus Raoulia 59 Techniques in the Study and Interpretation of Pollen Mor- 456 JOURNAL OF THE ARNOLD ARBORETUM phology, 141; a and Serene Occ Monoethanolamine, 145; Discussion, 146 ae oleraceus, 4 South American Species: Studies in the Genus Coccoloba, IX. A Critique of the, 213, 231, 357 Southeastern United States, The Genera of the Arales in the, 4 Southeastern United ee ies Genera of Convolvulaceae in the ssa crn United aah ae Genera Ebenales in the, 1 Brees United States, The Genera of Hydrophyllaceae and Polemoniaceae in the, 197 Southeastern United States, The Genera of Myrtaceae in the, 27 Southeastern United States, The Genera f Sarraceniaceae and Droseraceae in the, Spathiflo 47 Eerie pre plicata, 409 of the Arnold anne ore 1960, pry: Stenoloma chinensis, 408 Stopper, 274, 2 Storax, 24 Strangle weed, 301 Strelitzia, 287, 295, 296 Strophocaulos, 308 Studies in Artocarpus and Allied Genera, I A Revision of Artocarpus Sub- genus Pseudojaca, 73, 1 e- vision of Parartosarpus and Hullettia, i Nomenclatural erent : Critique of the South American Sere 216, 931,3 Stylisma, 306 Styracaceae, 22-24 Styrax, 2 Styrax cos 22 Sun 160 ew, Sundew Family, 156 Supplement to a Monographic Study of the West Indian Species of Phyllanthus, Sweet-flag, 50 Sweetleaf Family, 31 Symplocaceae, 31-32 Symplocarpus, 51-54 Symplocos, 32-35 Tacca sp aaa 409 Taew a ae sone and J Moomaw Vegetation on ene Soiled in Hawaii, 39 aaa 93, 126, 128 Taro, 57 Taxonomic Revision of Podocarpus, A, XII. Section Microcarpus, 36 Tomiinson, P. B. The Anatomy of Phenakospermum (Musaceae), 287 Tomuinson, P. B. Seedling Leaves in ae Sash Their Morphological Sig- nificance, 414 Tr cilia sticana 136 Trum ae ae oe on Supt ~ in Hawaii, Geolog Th Vig Viticella, 200 Waltheria americana, 411 ba, 90 Water-lettuce, 62 279 Species of Phyllanthus, Supplement to a Monographic Study of the, ae Wild-oliv WILSON, ae A. The Genera of the Arales in the Southeastern United es, 47 Witson, KennetH A. The Genera of Convolvulaceae in the Southeastern United States, 298 [| VOL. XLI 1960 | INDEX 457 Witson, KennetH A. The Genera of Sarraceniaceae and eee in the Hydrophyllaceae and Polemoniaceae in Southeastern United a the Southeastern United States, 197 be C. E., Jr. and R. ave NELL. Witson, KrennetH A. The Genera of Genera of the Ebenaes in the Myrtaceae in the Southeastern United sao United States States, 270 Woolly morning-glory, 317 Wolffia, 71-72 Wolffiella, 70-71 Zingiber zerumbet, 409 Woop, CarroLt E., Jr. The Genera of