JOURNAL OF THE ARNOLD ARBORETUM HARVARD UNIVERSITY B. G. SCHUBERT EDITOR L. I. NEVLING, JR. C. E. WOOD, JR. LAZELLA SCHWARTEN CIRCULATION VOLUME XLV CAMBRIDGE, MASSACHUSETTS 1954 Reprinted with the permission of the Arnold Arboretum of Harvard University KRAUS REPRINT CORPORATION rk 1968 . 1 (pp . 2 (pp. . 3 (pp . 4 (pp a) DATES OF ISSUE . 1-160) issued January 15, 1964. 161-284) issued April 15, 1964. . 285-400) issued July 15, 1964. . 401-505) issued October 15, 1964. Printed in U.S.A. TABLE OF CONTENTS Tue GeNneERA OF BERBERIDACEAE, LARDIZABALACEAE, AND MENI- SPERMACEAE IN THE SOUTHEASTERN UNITED States. Wallace R. Ernst . A Proposep EXPLANATION FOR THE ORIGIN OF cea INDUCED DieLoi Mutants 1n SorGHUM. Mary E. Sanders and Clifford J. Franzke Strupigs IN ALYSSUM: NEAR EASTERN REPRESENTATIVES AND THEIR Auuiss, I. T. R. Dudley Tue Rvupiaceous GENUS MusSAENDA: THE SPECIES OF THE PHILIPPINE IsLANDS. Don M. A. Jayaweera CoMPARATIVE ANATOMY OF THE LEAF-BEARING CACTACEAE, XI. Tue XYLEM OF PERESKIOPSIS AND QUIABENTIA. I. W. Bailey TyPIFICATION IN Dirca. Lorin I. Nevling, Jr. Tuer GeNuS CHAENOMELES (ROSACEAE). Claude Weber ............ THE GENERA OF CELASTRALES IN THE SOUTHEASTERN icon States. George K. Brizicky Tue GENERA OF LYTHRACEAE IN THE SOUTHEASTERN UNITED Srates. Shirley A. Graham PoLYEMBRYONY IN EvonyMmus (CELASTRACEAE). George K. Bri- C2) 5] 4) ee ea eEO PTET YON ry on are re ree eae Conn MorpPHouocy IN PINUS SABINIANA. James R. Griffin Tuer ELAEAGNACEAE IN THE SOUTHEASTERN UNITED STATES. see. BB TEIN. scacecscncepe nese ne eee ot Senne veer Notes on ROSACEAE IN THE LESSER ANTILLES. Richard A. Howard Tue GENERA OF RHIZOPHORACEAE AND COMBRETACEAE IN THE SouTHEASTERN Unirep States. Shirley A. Graham ............... Ture Genus CHAENOMELES (RosacKEAE) (Concluded). Claude THE GENERA OF CISTACEAE IN THE SOUTHEASTERN UNITED STATES. George K. Brizicky Synopsis OF THE GeNus AtyssuM. T. R. Dudley .. COMPARATIVE ANATOMY OF THE LEAF-BEARING Cie XIL PRELIMINARY OBSERVATIONS UPON THE STRUCTURE OF THE EpIDeRMIs, STOMATA, AND CuticLE. I. W. Bailey .................. Synopsis oF THE Genus AuRINIA IN TurKEY. T. R. Dudley ........ A RE-EVALUATION OF THE GENUS AMBROSIA (CompPosITAE). Willard W. Payne ..... a Tue GENERA OF RHAMNACEAE IN THE SOUTHEASTERN UNITED States. George K. Brizicky ...... A New Diospyros FROM THE MISANTLA REGION IN Mexico. Ar- HNO. GO? PONDS cca eee eit eetann 36 o7 A Furtuer Note oN CEANOTHUS HERBACEUS VERSUS C. OVATUS. George K. Brizicky THE Drirector’s REPORT BIBLIOGRAPHY OF THE PUBLISHED WRITINGS OF THE STAFF AND SrupENTs, JuLy 1, 1963—JuNE 30, 1964 STAFF OF THE ARNOLD ARBORETUM, 1963-1964 InDEx TO VOLUME XLV JOURNAL OF THE ARNOLD ARBORETUM VoL. XLV JANUARY 1964 NUMBER 1 THE GENERA OF BERBERIDACEAE, LARDIZABALACEAE, AND MENISPERMACEAE IN THE SOUTHEASTERN UNITED STATES * WALLACE R. ERNST BERBERIDACEAE A. L. de Jussieu, Gen. Pl. 286. 1789, “Berberides,’ nom. cons. (BARBERRY FAMILY) Perennial, rhizomatous herbs or shrubs [to small trees], often with perulate vegetative buds; plants glabrous or hairs unicellular [or uni- seriate, sometimes glandular], occasionally glaucous. Leaves mostly peti- olate, exstipulate (or substipulate), cauline and alternate (in Berberis some reduced to spines, the others fasciculate), or + opposite (Podophyl- lum), or radical (Jeffersonia), simple or variously divided, pinnately or ternately compound, palmately (or pinnately) veined; stomata ranuncula- ceous. Inflorescences 1 to many flowered, terminal or axillary, usually bracteate, scapiform, racemose, Ccymose, umbelliform (or paniculiform). Flowers regular, hypogynous, often calyculate, bisexual. Perianth usually 3(—5)-merous, aposepalous and apopetalous, usually multiseriate (some- n tinued through volum e area covered, as in earlier treatments, 1s bound by and includes North Carolina, Tennessee, Arkansas, and Louisiana. The rip- tions apply primarily to the plan f this area with supplementary information in Helpful information or assistance of various kinds was given by H. E. Ahles, G. K. Brizicky, G. R. Cooley, W. H. Duncan, Shiu-ying Hu, Jean H. Langenheim, R. Ornduff, Lily M. Perry, D. G. Rhodes, Bernice G. Schubert, and Lazella Schwarten. Mrs. Gordon W. Dillon prepared the final typescript. The illustrations, the con- tinuing work of Dorothy H. Marsh, were drawn from fresh or preserved materials hannell, and Louis Bass. A number of the drawings of aly- ’ m those of Cocculus, were made under that of C. E. Wood, Jr., some distributional data, made numerous suggestions, and devoted much time to the ; leti manuscript in various stages of completion. 2 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV times partly acyclic). Sepals often petaloid, usually 6 (4 to many) in 2 (1-4) series. Petals usually 6 (—9) [4] in 2 (3) series, sometimes glandular or compere Ruetniie usually as many as and opposite the petals (or stamens twice many or more), equal, appearing as 1 or 2 series; filaments eerie andes anthers 2-locular at anthesis, i pollen sacs lateral to + introrse [or extrorse?], longitudinally dehiscent, sometimes forming narrow valves hinged laterally or at the apex; pollen mostly 3-colpate (spiny in Diphylleia). Gynoecium 1-carpellate; stigma various; style present or absent; ovary superior, 1-locular, usually + gibbous; placenta solitary, parietal or subbasal; ovules anatropous, 2- integumented, mostly + ascending, the micropyle down, the raphe usually above. Fruit few (1) to many seeded, fleshy or dry, nonfollicular, in- dehiscent or irregularly ruptured (or transversely [obliquely] loculicidal). Seeds oblong, straight to slightly curved (or + spherical or disciform- subhemispheric), sometimes arillate; endosperm present; embryo + straight, usually small (elongate in Berberis). (Including Podophyllaceae DC. and Nandina Thunb.; excluding Glaucidium Sieb. & Zucc. and Hydrastis Ellis.) Typr GENUS: Berberis L. About 10 or 12 genera, variously estimated to include 300 to perhaps over 600 species (Berberis, ca. 500 species, Mahonia, ca. 100; cf. Ahrendt), mostly in the Northern Hemisphere. The herbaceous species probably amount to fewer than 50. One species in each of five genera is indigenous in the eastern United States; three other genera are indigenous in the western United States. With the exception of Berberis, which is repre- sented in both Eurasia and South America, the American genera lack close counterparts in Europe but exhibit strong affinities with vicarious taxa in eastern Asia (Gray), where they also are distributed in deciduous woodlands. At least one species of Berberis has been naturalized in our area. One or two other taxa are to be looked for as possible escapes from cultivation. The conformation of the leaves and, especially, the venation patterns are diverse, but at the generic level more or less palmate venation is frequent, even in the leaflets of pinnately divided leaves. The most con- spicuous morphological tendencies uniting Berberidaceae include the calyculate flowers; the multiseriate, deciduous, more or less imbricate perianth; the stamens opposite the petals; and the peculiar, valvelike dehiscence of the anthers of many taxa. The solitary, unicarpellate, non- follicular gynoecium is a unifying characteristic. The petals in some respects resemble staminodia but are situated in the position of the corolla and are often glandular or nectariferous. The tendency for a multiseriate perianth with more than one series of sepals and of petals places the stamens opposite the petals and the petals opposite the sepals. Some- times, however, the insertion appears to be slightly acyclic or spiral (see Schmidt) or the stamens may seem to be uniseriate. The opposition of stamens and petals as it occurs in Berberidaceae and Menispermaceae contrasts with the theoretically dissimilar situation in Primulaceae, 1964 | ERNST, GENERA OF BERBERIDACEAE 3 Rhamnaceae, etc. (Bentham; see also developmental stages in Payer). In Berberis, Mahonia, and Ranzania, the stamens are tactile, bending inward when touched (Chaveaud, Dop, Kumazawa, and Miller). Ini- tially the anthers are 4-locular and more or less introrse or simply lateral. Later, their orientation may be obscured by somewhat unequal develop- ment or by the various modes of dehiscence of the anther wall. In all cases, the anthers are longitudinally dehiscent and more or less 2-locular at anthesis. All of the tissue on the outer side of a locule may function as a single vertical, abaxially hinged valve, as in Podophyllum, or as two opposing, laterally hinged valves, as in Nandina. In Ranzania all of the outer wall of tissue forms an apically hinged valve, while in the remaining genera only a portion of the wall forms a similar but narrower valve (see Kumazawa). The monomeric nature of the gynoecium (see Eckardt) has been chal- lenged (Chapman, Eames), but the arguments against the older interpre- tation do not seem always to be convincing. In some clearly teratological examples (Podophyllum), more than one ovary of similar structure is formed in the same flower. Although the basic morphology of the gynoe- cium appears more or less uniform, some of the characteristics of matura- tion become important taxonomically. The limits of the higher systematic units of the family are unclear when based exclusively on either floral or vegetative morphology. Subfamily Berberidoideae (including Mahonia and Nandina) is woody, and the rachis of the imparipinnate (rarely trifoliolate) leaves is either jointed and somewhat swollen at the insertion of the leaflets (or secondary rachises) or, in Berberis, the simple leaves are articulated, usually near the base. Subfamily Podophylloideae is herbaceous, usually with scarlous ground-level bracts and with scattered vascular bundles in the aerial stems (Agardh, Brown, Harvey-Gibson & Horsman, Metcalf & Chalk). The leaves are simple (sometimes two-parted) and palmately veined or are bi- or trifoliolate or ternately decompound (pinnately compound in Bongardia); the rachis of the leaves is not jointed as in Berberidoideae. Although the venation sometimes is obscure and may vary in a single species and among the leaflets on a single leaf, the tendency for palmate or ternate venation is more conspicuous in Podophylloideae than in Ber- beridoideae. Close morphological relationships among some taxa are clearly evident, but the overlapping of major characteristics seems to pre- vent the separation of the genera into tribes. For points of view and arrangements of the genera see Heintze, Janchen, Kumazawa, Miyaji, and Tischler. About 40 named alkaloids have been identified in about eight genera of Berberidaceae (see Willaman & Schubert); of these, about half (20) are reported only for Berberidaceae. Others are reported also for Menisperma- ceae (9), Ranunculaceae (7), Rutaceae (4), Leguminosae (3); An- nonaceae, Monimiaceae, and Papaveraceae (2 each); Aristolochiaceae, Compositae, Lauraceae, Magnoliaceae, and Solanaceae (1 each). Phylo- genetic interest naturally focuses on the examples where chemical as well 4 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV as morphological similarities are noted. Hydrastis and Glaucidium are vegetatively similar to some Berberidaceae, but on the basis of floral structure are better referred to Ranunculaceae. Other superficial similari- ties recall some Papaveraceae and Fumariaceae. The insertion of perianth parts and stamens in Berberidaceae suggests alliance with Menispermaceae and with Lardizabalaceae where the stamens also tend to be opposite the petals. Although phylogenetic relationships among these families are im- plied, the degree of the relationships is difficult to estimate. In the above instances, it is the solitary, monomeric nature of the gynoecium that clearly distinguishes Berberidaceae. The fruits are neither follicles (in the sense that they do not open longitudinally through the placenta), achenes, nor drupes. The range of chromosome numbers in the family seems fairly well ex- plored at the generic level, with reports of 2n = 10, 12, 14, 16, 28, and 56. The useful properties of some taxa are discussed by Baillon and by Gray; see also genera in The Merck Index, ed. 7., 1960, for medicinal and poisonous properties. In some species, the roots or rhizomes and foliage are believed poisonous but the ripe fruits are considered edible (Fernald & Kinsey). No part of the plants should be considered entirely safe for eating, since the family is rich in alkaloids and contains some potentially poisonous compounds. Many taxa are desirable garden plants (cf. Berberis). Some of the exceptional characteristics of taxa not covered in the family description are as follows: in Achlys DC.., inflorescence spicate, flowers in tight clusters of two or three, perianth absent, stamens several, fruits one- seeded; in Epimedium L., petals long-saccate; in Leontice L., stipules leafy; in Leontice and Bongardia C. A. Mey., fruits + inflated, scarious, and gaping; in Ranzania T. Ito, pollen 6—12-rugate; in Vancouveria C. Morr. & Decne., filaments + connivent. REFERENCES: Under Menispermaceae see BENTHAM. AcARDH, J. G. Theoria systematis plantarum, etc. xcvi + 404 pp., index, ex- plicatio iconum, pls. 1-28. Lund. 1858. | Lardizabaleae, Nandineae, 71, pl. 5; Podophylleae, 74, 75, pl. 5; Berberideae, 138, 139, pl. 5; Menispermaceae, 241-243 pl BarLton, H. Monographie des Menispermacées et des Berbéridacées. Hist. Pl. 3: 1-76. 1871. [Includes Lardizabalaceae and some Flacourtiaceae. See English translation by M. M. Hartoc, Nat. Hist. Pl. 3: 1—75. 1874. ] Berbéridacées. /n: Traité du développement de la fleur et du fruit. Adansonia 12: 351-354. 1879, BECKER, H. F. Two new species of Mahonia from the Grant-Horse Prairie Basin in southwestern Montana. Bull. Torrey Bot. Club 89: 114-117. 1962. | Fossil spp. | Brown, R. Appendix No. V. Pages 420-485, in J. K. Tuckey, Narrative of an expedition to explore the River Zaire, etc. Ixxxii + 498 pp., 14 pls. London. 1818. [Commentary on Berberidaceae, footnote, 441, 442. See also Misc. Bot. Works Robert Brown (J. J. BENNETT, ed.) 1: 124. 1866.] 1964] ERNST, GENERA OF BERBERIDACEAE 5 CaLLoni, S. Contribuzione allo studio del genere Achlys nelle Berberidacee. Malpighia 2: 25-34. pls. 8, 9. 1888-1889. CuapmMaAN, M. Carpel anatomy of the Berberidaceae. Am. Jour. Bot. 23: 340-348. 1936. [Thinks gynoecium derived originally from 3 spirally ar- ranged carpels and in living representatives to be equivalent to 3 fused carpels in one phylad but to only 2 fused carpels in the other. | CiteRNE, P.-E. Berbéridées et Erythrospermées. Thése. 161 pp., errata, pls. 1-8. Paris. 1892. Dorm_r, K. J. The acacian type of vascular system and some of its derivatives. _ New Phytol. 53: 301-311. 1954. [Menispermaceae, Lardizabalaceae, Berberidaceae. | Eckarpt, T. Untersuchungen iiber Morphologie, Entwicklungsgeschichte und systematische Bedeutung des pseudomonomeren Gynoeceums. Nova Acta Acad. Leop.-Carol. II. 5(26): 1-112. pls. 1-25. 1937. [Berberidaceae, 95 ff. | ———. Das pseudomonomere Gynoeceum. Chron. Bot. 4: 206-208. 1938. EIcHier, A. W. Berberidaceae. Bliithendiagramme 2: 134-138. 1878. [Meni- spermaceae, 138-143; Lardizabalaceae, 143, 144. Feppe, F. Pflanzengeographische Verbreitung der Gattung Mahonia. Jahresb. Schles. Ges. Vaterl. Cult. Zool.-Bot. 77: 8-17. 1899. [Reprinted as “Ueber pflanzengeographische . . .” Versuch einer Monographie der Gattung Mahonia. Bot. Jahrb. 31: 30-133. 1901. [Morphology, anatomy, taxonomy, 37 spp. | FERNALD, M. L., & A. C. Kinsey. Edible wild plants of eastern North America. xiv + 452 pp. Cornwall, New York. 1943. [Podophyllum peltatum, 45, 206, 207; Berberis vulgaris, B. canadensis, 208; Akebia quinata, 209. See also revised ed., 1958. Gray, A. Diagnostic characters of new species of phaenogamous plants .. . Mem. Am. Acad. Arts Sci. II. 6: 377-449. 1859. [Floristic relationships between North America and Asia; see review in Proc. Am. Acad. Arts Sci. 4: 131-135. 1859; see also L1.] _ Berberidaceae. Jn: Gray & Watson, Syn. Fl. N. Am. 1(1): 66-72. 1895. Haturer, H. L’origine et le systéme phylétique des angiospermes exposés a aide de leur arbre généalogique. Arch. Néerl. Sci. Nat. III. B. 1: 146- 234. 1912 Harvey-Grpson, R. J., & E. Horsman. Contributions towards a knowledge of the anatomy of the lower dicotyledons. II. The anatomy of the stem of the Berberidaceae. Trans. Roy. Soc. Edinb. 52: 501-515. 1 pl. 1919. HeintzeE, A. Cormofyternas fylogeni. 170 pp. Lund. 1927. [Berberidaceae, 101, 102 Hime Baur, W. Die Berberidaceen und ihre Stellung im System. Denkschr. Akad. Wiss. Wien Math. Naturw. 89: 733-796. pls. 1-4. 1914. [ Bibliog- raphy; references to Lardizabalaceae, Menispermaceae. | Hu, S. Y. Medicinal plants of Chengtu herb shops. Jour. W. China Border Res. Soc. B. 15: 95-177. map. 1945. [Includes Akebia, Epimedium, Ma- honia, Nandina, Podophyllum. | Ito, T. Berberidearum Japoniae conspectus. (In Latin.) Jour. Linn. Soc. Bot. 22: 423-437. pl. 21. 1887 JANCHEN, E. Die systematische Gliederung der Ranunculaceen und Berber- 6 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV idaceen. Denkschr. Akad. Wiss. Wien Math. Naturw. 108(4): 1-82. 1949. [ Biochemistry, infrafamilial synonymy, bibliography. | Kitamura, T., & M. Sucamoto. Studies on the alkaloids of berberidaceous plants. XXXI. (In Japanese; English summary.) Jour. Pharm. Soc. Japan 81(2): 254-261. 1961.* Kumazawa, M. Morphology and biology of Glaucidium palmatum Sieb. et Zucc. with notes on affinities to the allied genera Hydrastis, Podophyllum and Diphylleia. Jour. Fac. Sci. Univ. Tokyo Bot. 2: 345-380. 1930. [Places Glaucidium and Hydrastis in Podophyllaceae, 379. ] . Structure and affinities of Glaucidium and its allied genera. (In Japanese.) Bot. Mag. Tokyo 44: 479-490. 1930. . Pollen grain morphology in Ranunculaceae, Lardizabalaceae and Ber- beridaceae. Jap. Jour. Bot. 8: 19-46. pls. 2-6. 1936. [ Ranunculaceae, 200 spp., 33 genera; Lardizabalaceae, 4 spp., 2 genera; Berberidaceae, 19 spp., 11 genera. | anzania japonica (Berberidac.). Its morphology, biology and sys- tematic affinities. bid. 9: 55-70. 1937. [Stamens irritable. Important analysis of staminal dehiscence of family, 57-60. See Contents (Abstracts), pp. ix, x, at beginning of volume for other papers. | . Comparative studies on the vernation in the Ranunculaceae and Ber- beridaceae. (In Japanese.) Jour. Jap. Bot. 13: 573-586. 1937 - On the morphology and anatomy of Achlys japonica Maxim. (In Japanese; English summary.) Bot. Mag. Tokyo 51: 660-668. 1937. . Systematic and phylogenetic consideration of the Ranunculaceae and Berberidaceae. /bid. 52: 9-15. 1938, [ Bibliography. | On the ovular structure in the Ranunculaceae and Berberidaceae. Jour. Jap. Bot. 14: 10-25. 1938. [Ranunculaceae, 29 genera; Berberidaceae, 10 genera. | Kurita, M. Karyotype studies in Berberidaceae. I. Mem. Ehime Univ. Sect. II. Nat. Sci. Biol. 2(3): 247-252. 1956.* Mavritzon, J. Zur Embryologie der Berberidaceen. Acta Horti Gothob. 11 1-18. 1936. [Seven genera; embryo sac “normal” except for Caulophyllum, which has modified Peperomia type. | Miyaji, Y. Beitrige zur Chromosomenphylogenie der Berberidaceen. Planta 11: 650-659. 1930. [Includes Hydrastis and Glaucidium; bibliography. | Payer, J. Traité d’organogénie comparée de la fleur. 2 vols. Paris. 1857. | Berberidaceae, 237-240, pl. 52: Menispermaceae, 241-244, pl. 53.| PRANTL, K. Berberidaceae. Nat. Pflanzenfam. III. 2: 70-77. 1891. [Sectional names. | RupbENKO, F. E. Spermatogenesis and fertilization in Mahonia aguifolium [sic | L. (In Russian.) Bull. Mosk. Obshch. Isp. Pri. Biol. II. 66: 133-137. 1961. 1964 | ERNST, GENERA OF BERBERIDACEAE 7 SAUNDERS, E. R. On neve polymorphism. I. Ann. Bot. 39: 123-167. 1925. [ Berberidaceae, 131-133.] Illustrations of carpel a aha II. New Phytol. 27: 175-192. 1928. [Berberidaceae, 175-183 ScuMmiptT, E. Untersuchungen iiber Berberidaceen. Beih. Bot. Centralbl. 45: 329-396. 1928. [Morphology; finds flowers acyclic, hemicyclic, or sec- ondarily verticillate, 393. | STEARN, W. T. Epimedium and Vancouveria (Berberidaceae), a monograph. Jour. Linn. Soc. Bot. 51: 409-535. pls. 24-31. 1938. [Discussion of early ana 425. Taxepa, H. On the genus Achlys. Bot. Mag. Tokyo 29: 169-185. pl. 7. 1915. | Morphological and systematic study. TISCHLER, G. Die Berberidaceen und Podophyllaceen. Bot. Jahrb. 31: 596-727. 1902. [Hybrids between Berberis and Mahonia, 647; phylogeny, 719; classified bibliography. | Tomita. M.. & H. Isnt. Studies on the alkaloids of berberidaceous plants. XII. (In Pipanese: English summary.) Jour. Pharm. Soc. Japan 77: 114-116. 1957.* [Epimedium. | Torrey, J., & A. Gray. aibesincess Fl. N. Am. 1: 49-54. 1838. TUuRRILL, W. B. te japonica. Bot. Mag. 166: pl. 76. 1949. WILLAMAN, J. J., . Scuusert. Alkaloid-bearing plants and their con- tained es a e Dep. Agr. Tech. Bull. 1234: 1-287. KerEy TO THE GENERA OF BERBERIDACEAE General — plants woody or herbaceous perennials ; leaves exstipulate, mostly alternate; flowers often calyculate, bisexual, hypogynous, regular ; erianth sil. of more than 2 series, + cyclic, mostly 3-merous, fugacious ; petals mostly as many (or twice as many) as sepals and opposite the stamens (or stamens twice as many as petals); anthers longitudinally dehiscent, often forming 2 apically hinged valves; gynoecium of a solitary 1 -locular carpel; placenta parietal or subbasal ; fruit mostly fleshy — loculicidal or ruptured by the young seeds), usually with more than 1 se A. Plants woody; leaves variously articulated above the base. Leaves pinnatel ecompound; inflorescences paniculiform; petals whitish, without distinct glands; anthers pose dehiscent. Nandina B. Leaves simple, usually fascicled in the axil of a simple or branched cauline spine; petals yellowish, usually with a pair of adaxial glands; anthers opening by 2 apically hinged valves. |. 2. Berberis. Plants herbaceous, rhizomatous; leaves not articula ted C. Flowers solitary on scapes; leaves all radical, 2-foliolate; petals white. usually twice as many as oan fruit transversely loculicidal; seeds with a laciniate aril. 5. Jeffersonia. C. Flowers 1 or more, borne on leaf- bearing stems; petals usually as many as sepals. Leaves ternately (or apparently pinnately) compound, the lpanets ed and . globular, mostly Pa on stout funiculi, soon bursting the WHELs ow, oc’ aca g Sd RE, Oe RE ee bs 6. Paulista. 8 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLV D. Leaves simple, palmately veined and lobed; petals white (or pink- ish), rounded and without glands; seeds several in a fleshy fruit. E. Flowers solitary, usually subtended by 2 + opposite (rarely alternate), several-lobed leaves; stamens usually twice as many as petals; anthers longitudinally dehiscent. .. 3. Podophyllum. E. Flowers in umbelliform cymes, usually subtended by 2 distinctly 2-cleft, alternate leaves; stamens as many as petals; anthers with 2 apically hinged valves. ................ 4. Diphylleia. Subfam. BERBERIDOIDEAE 1. Nandina Thunberg, Nov. Gen. Pl. 1: 14. 1781. Evergreen, glabrous shrubs with upright branches, to 2 m. tall. Leaves odd-pinnately decompound, the rachises swollen and articulated both at the insertion of the secondary branches and of the numerous pinnately veined, often reddish, + sessile, ovate or lanceolate leaflets; petioles persistent, enlarged basally, substipulate and + clasping. Inflorescences terminal (and axillary), many flowered, paniculiform, determinate, brac- teate, the rachises + scabrous. Perianth multiseriate, mostly 3-merous, the segments coriaceous to scarious and smaller below, + petaloid and larger above; the lowest 3 segments usually most durable and regularly or irregularly inserted; receptacle elongated. Sepals in about 4 series. Petals in 2 or 3 series, whitish and without distinct glands, + merging with the sepals. Stamens + in 1 series; anthers elongate, subsessile, slightly exceeded by the broad connective, longitudinally dehiscent; pollen 3-colpate. Stigma + 3-lobed; style persistent; placenta parietal; ovules 2 (3 or 4). Fruit globular, coriaceous, usually red and 2-seeded, indehis- cent. Seeds + horizontally oriented, + disciform, concave-convex with hollow faces adjacent. Type species: N. domestica Thunb. (Name de- rived from a portion of the Chinese for “plant from the south.’’) A single species, Nandina domestica, heavenly bamboo, 2n = 20, with many horticultural forms, native to China and perhaps Japan, occurring to an altitude of about 1200 m. An attractive shrub, it is commonly cultivated in the Southeast, where it freely reproduces by seeds in some gardens. It has been found growing without cultivation in Durham County, North Carolina, and should be looked for in other areas as a garden escape. Although usually classified with Podophylloideae, Nandina seems to have no immediate relationship to those herbaceous genera and is here placed with the Berberidoideae because of its woody habit and articulated leaves. The decompound leaves, the graduated multiseriate calyx that more or less merges with the corolla, the anthers without apically hinged valves, the lower chromosome number, and the peculiar seeds distinguish Nandina from the Berberis-Mahonia complex. The genus sometimes has been considered a separate family, Nandinaceae, but this seems unneces- sry. Furthermore, a legitimate family name is not available, since Nan- 1964] ERNST, GENERA OF BERBERIDACEAE 9 dinaceae Horaninow (1834) included the type of Berberidaceae, and Nandinaceae Nakai is a later homonym. REFERENCES: Under family references see Hu Duranp, J. Nandina domestica Thunb. Revue Hort. 28: 340, 341. 7 pl. 1923. [Supplementary notes by “F. L.” pipe DesLonccHamps, J. L. A. Nandine domestique. Nandina domes- tica. Herb. Gén. Amateur 4: pl. 281. 1820. SHEN, os Phylogeny and wood anatomy of Nandina, Taiwania 5: 85-91. 1954. [Illustrated; thinks Nandina only remotely related to Mahonia and Berberis. | Sims, J. Nandina spite lge Bot. Mag. 28: pl. 1109. 1808. ToMIrTA, M., & H. Isur. Studies on the alkaloids of berberidaceous plants. XXI. (In Carer English summary.) Jour. Pharm. Soc. Japan 79: 1092, 1093. 1959.* 2. Berberis Linnaeus, Sp. Pl. 1: 330. 1753; Gen. Pl. ed. 5. 153. 1754. Evergreen or deciduous, usually spiny shrubs to 2.5 [6] m. tall, with arching branches and yellow inner bark and wood. Leaves simple, + spatulate, articulated usually near the base, usually + fasciculate on 1 or more short shoots in the axil of a stout, simple or branched spine [or spine rarely foliaceous or absent]; venation + pinnately [or palmately| re- ticulate. Inflorescences axillary, few (1) to many flowered, racemose to subumbellate, terminating short, bracteate shoots. Flowers calyculate; perianth + yellowish [to reddish orange]. Sepals usually 6, 2[1-4]- seriate, + petaloid. Petals mostly 6, 2-seriate, usually each with a pair of adaxial glandular regions. Stanere 6; filaments tactile [sometimes + toothed]; anthers with 2 valves hinged at the apex; pollen 3-colpate. Stigma orbiculate, + umbilicate; style + absent |or present] ; placenta subbasal; ovules 2-9, erect. Fruit + oblong, indehiscent, fleshy. Seeds 1—9[-15], without arils: embryo long. (Excluding Mahonia Nutt., nom. cons.) Lectotype species: B. vulgaris L.; see Britton & Brown, Illus. Bl Noe U8. ed. 2. 2% 1274, .49ks, (Name Medieval Latin, apparently from berbérys, presumably the Arabic name of the fruit.) — BaRBERRY, BERBERRY. Perhaps 175 to nearly 500 species, mostly in Asia and South America, poorly represented in North Africa, Europe, and North America. Infra- generic relationships are poorly understood and seem complicated by numerous, presumably interspecific, hybrids. Twenty-one sections were recognized by Schneider, 32 sections by Ahrendt, and 15 series by Rehder. In our area, one species is indigenous, one is naturalized, and one or two others may have escaped from cultivation. In Berberis canadensis Mill. (§ Sinenses Schneid.; ser. Sinenses Rehd. ; § Canadenses Ahrendt), American or Allegheny baiteney y, 2n = 28, oc- curring in the mountains of West Virginia and Virginia, southward to 10 JOURNAL OF THE ARNOLD ARBORETUM | VOL. XLV Georgia, and sporadic westward to Missouri (absent from Canada), the leaves tend to be remotely dentate, the petals are apically notched, and the few-flowered racemes often are subumbelliform. In the southwestern United States B. Fendleri Gray is somewhat similar. Berberis Rehde- riana Schneid., of horticulture, is said to be a hybrid between the two. Berberis Thunbergii DC. (§ Sinenses Schneid.; ser. Sinenses Rehd.; § Tschonoskyanae Schneid. [{cf. Ahrendt]), Japanese barberry, 2n = 28, naturalized in our area and in much of the eastern United States, is dis- tinguished by its more or less entire leaves, its solitary flowers (or few- flowered umbels), and its immunity to wheat rust. The European B. vulgaris L. (§ Berberis; § Vulgares Schneid.; ser. Vulgares Rehd.), 2n = 28, common barberry, resembling B. canadensis but tending to have denticulate leaves, entire petals, and many-flowered racemes, should be looked for in our area, since it is widely naturalized in the Northeastern and the Midwestern States. Berberis ottawaensis Schneid. is said to be a hybrid between B. Thunbergii and B. vulgaris, and Berberis declinata Schrad. one between B. canadensis and B. vulgaris. The cauline spines of Berberis, no doubt the equivalent of leaves, are nearly ubiquitous (sometimes absent), usually obvious, stout, nonarticu- lated, and variable within the species and also on some plants. In some instances, a fully expanded but nonarticulated leaf is formed, rather than a spine, and in others the spines are soft and soon deteriorate. In any event, the spines, which subtend lateral branches, are distinct in ap- pearance from the scales usually associated with vegetative buds and from the comparable scales also found in some species of Mahonia. The simple leaf of Berberis seems like a reduction of the more complicated leaves of Mahonia, yet the seedling leaves of Mahonia probably also are simple. In some taxa of Berberis the articulation of the leaves is either subbasal, or midway on the narrowed, petiole-like base of the blade, or distal on a distinct petiole at the base of the expanded blade; and in one (or more) South American species the apparently simple leaves have double articula- tions, one subbasal and one distal on the petiole. The separate generic status of Mahonia (Berberis subg. Mahonia Gray and subg. Trilicina Gray), which seems to differ consistently from Berberis only in its compound leaves (see Ahrendt), is likely to remain debatable. Mahonia, however, does not have cauline spines, and the distributional pattern is somewhat more restricted than in Berberis. There can be no doubt that Berberis and Mahonia have much in common, including similar and somewhat unusual pollen, chromosome number, tactile stamens, and susceptibility to wheat rust. The question is whether it is better to treat these two groups of species as separate genera or as an inclusive genus, Berberis, with two subgenera. Close affinities between particular species of Mahonia and of Berberis, which might suggest normal recent pathways of genetic exchange, have not been pointed out, unless this is implied by the several taxa described under < Mahoberberis Schneid. (see Ahrendt, ensen, Levan, Melander & Eade, Vaarama, Wyman). These presumed intergeneric hybrids. with dimorphic leaves, would seem to bridge the gap 1964 | ERNST, GENERA OF BERBERIDACEAE 11 between Mahonia and Berberis, but thus far, only hybrids involving M. aquifolium (Pursh) Nutt. are available. The leaves of the elongated shoots of the hybrids tend to be simple and more or less sessile with coarsely spiny margins and are weakly articulated above the insertion on the stem; these leaves seem to be homologous with the prevalent spines of Berberis. The leaves of the shorter shoots tend to be trifoliolate, with the rachis articulated at the insertion of the leaflets, the margins of which are less coarsely dentate. The hybrid plants seem to bloom reluctantly and the flowers appear to be sterile. Information regarding the synthesis of hybrids under controlled conditions does not seem to be available. Chromosome numbers of 2” = 28 and 56 are reported for Berberis and 2n = 28 for both Mahonia and « Mahoberberis. Most species are orna- mental, but their cultivation is prohibited unless they are known to be either immune or highly resistant to black stem-rust of wheat (see Fulling). The fruits often are considered edible (Fernald & Kinsey) but should be sampled with caution, since potentially toxic substances occur in the family. In the Southeast, B. Thunbergii, B. Julianae Schneid., and Mahonia Bealei (Fort.) Carr. are commonly cultivated. REFERENCES: Under family references see FERNALD & KINSEY and GRay. Aszott, R. M. S. Developmental anatomy of the tracheary system in Berberis Thunbergii DC., with emphasis on the differences between protoxylem and metaxylem, primary xylem and secondary xylem. Diss. Abs. 20: 3482. 960.* AurenptT, L. W. A. Berberis and Mahonia; a taxonomic revision. Jour. Linn. Soc. Bot. 57: 1-410. 1961. [Morphological synopsis of taxa, 2-21; generic status of Mahonia, 296, 297.| AmEs, L. M. Barberries i immune or aa sears to black stem-rust of cereals. Arnold Arb. Bull. Pop. Inf. IV. 5: 1937. ANDERSON, E. The analysis of ee pee as illustrated by Berberis X gladwynensis. Ann. Missouri Bot. Gard. 40: 73-78. 1953. [Extrapolates origin of hybrid sp. | ARISUMI, T. Some Seer objectives 2 the improvement of Ma/fonia and Berberis. Proc. Pl. Propagators Soc. 8: 43-46. 1958.* Bary, [A.] pE. Neue Untersuchungen he die Uredineen, insbesondere die Aecidium Berberidis. Monatsber. Akad. Wiss. Berlin 1865: 15-49. 1 pl. 1866. [Classical paper on wheat-rust/barberry Ne ape Boynton, K. R. Berberis Thunbergi. Addisonia 10: 59. pl. 350. 1926. CHATTERJEE, R. The Rasanjana of the Hindus. Lloydia 12: 178- 182. 1949. [Medicinal uses of Berberis.] . BANERJEE. Plant alkaloids. V. Indian Chem. Soc. Jour. 30: 705- 707. 1953.* [| Berberis.] CHAvuvEAuD, G. Sur une nouvelle interprétation des mouvements provoqués dans les étamines de Berberis. Bull. Soc. Bot. Fr. 53: 694-698. 1907. [See Dop. DerRMEN, H. A study of chromosome number in two genera of iene Mahonia and Berberis. Jour. Arnold Arb. 12: 281-287. 19 12 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Dop, P. Recherches physiologiques sur le mouvement des étamines des Ber- béridées. Bull. Soc. Bot. Fr. 53: 554-572. 1906. [ Bibliography; see CHAUVEAUD. | Fuuine, E. H. Plant life and the law of man. IV. Barberry, currant and goose- berry, and cedar control. Bot. Rev. 9: 483-592. 1943. [Berberis, 485- 512; classified bibliography, 574-581. — litigation, eradication, quaran- tine of alternate hosts of wheat rust. | GRIFFEN, M. H. The chromosome cm of Berberis. Trans. Roy. Soc. S. Afr. 24: 203-206. 1937. Harrincton, H. D. Abnormal pistils in Berberis repens. (Abs.) Jour. Colo.- Wyo. Acad. Sci. 4(8): 36. 1956. [Mahonia.] Hooker, J. D. Berberis Thunbergii. Bot. Mag. 108: pl. 6646. JENSEN, H. Zwei neue Mahoberberis-Hybriden. Deutsch. fei A 300, 301, 310. 1950. [X M. aquisargentia, X M. aquicandidula. Jos, M. M. Los Berberis de la region de Nahuel Huapi. Revista Mus. La Plata Bot. II. 5: 21-72. 1942. [Systematic study; illustrated; bibliopraphy: 15 spp. See also ibid. 8: 169-178. 1953. Kern, F. D. Observations on the dissemination of the barberry. Ecology 2: ' 211-214. 1921. [B. vulgaris.] LEINFELLNER, W. Zur morphologie des gynézeums von Berberis. Osterr, Bot. Zeitschr. 103: 600-612. 1957. [Das normale Berberis-Gynézeum ist mithin als echt einkarpellig anzusprechen,” 610. | LevAN, A. On the chromosomes of a new a riee sacri hybrid. Hereditas 30: 401-404. 1944. [M. aquifolium * B. Sargen Levine, M. N., & R. U. Cotter. Susceptibility and ene of Berberis and related genera to Puccinia graminis. U. S. Dep. Agr. Tech. Bull. 300 1-26. 1932.* [See also L. Lino, Phytopathology 35: 417-420. 1945.] Li, H. L. The cultivated Mahonias. Morris Arb. Bull. 14: 43-50. 1963. [Includes spp. cultivated in the Southeast. | MaTHENY, G. E., R. S. Mutiin, & R. L. SHAVER. Studies of the germination, grow wth, and ‘precaination of seeds, pia and root fragments of Berberis canadensis Mill. Va. Jour. Sci. 1: 295, 940. MELANDER, L. W., & G. W. EapE. X art pe: Miethkeana, a new hybrid. Natl. Hort. Mag. 33: 257-260. 1954. [‘“M. aquifolium >< B. ‘Renton’ Hort.? MULLER, oF pe fertilisation : flowers. (Translated by D. W. THompson.) xii + 6 p. London. 1883. [Berberidaceae, 90-93. ] SAUNDERS, : _ Notes on some ete in the second generation of Berberis hybrids. Mem. Hort. Soc. N. Y. 1: 167, 168. 1904.* SCHNEIDER, C. K. Die Gattung Berberis (Rhea. Bull. Herb. Boiss. II. > 33-48; 133-148; 391-403 ; 449-464; 655-670; 800-812; 813-831. 1904— SHorT, G. R. A. Berberis aristata D.C. and other species of Berberis. A com- parative study of the structure of the stems. Pharm. Jour. IV. 63: 189-195. * Stmonian, V. H., & H. W. YounckKEN. gi ues iaatee study of a Berberis. Jour. Am. Pharm. Assoc. Sci. Ed. 42: 111-116. STAKMAN, E. C., M. N. Levine, R. U. Cotter, & : HINEs. Senn a bar- berry to the origin and persistence of physiologic forms of Puccinia graminis. Jour. Agr. Res. 48: 953-969. 1934.* THompson, N. E., & W. W. Rossins. Methods of eradicating the common 1964] ERNST, GENERA OF BERBERIDACEAE 13 barberry (Berberis vulgaris L.) U.S. Dep. Agr. Bull. 1451: 1-46. pls. 1-13. TomirTA, M., T. H. YANnc, & S. T. Lu. Studies on the alkaloids of berberidaceous plants. "XXIV-XXVL (In Japanese; English summary.) . Soc. Japan 80: 845-851. 1960.* [Berberis. See also ibid. 1302- ~1306.*] U.S. Dep. Acr. Bur. ENTOMOL. PL. QUARANT. Black stem rust quarantine and amended regulations. Notice of Quarant. 38. 7 pp. mimeogr. 1951. [See U.S. Dep. Agr. Res. Serv. Pl. Pest Control Div., Administrative instructions designating rust-resistant barberry, Mahoberberis and Mahonia plants. PI. Pest Control 557. Sth revision, 1959, for a list of resistant spp. ] Uster1, A. Das Geschlecht der Berberitzen. Mitt. Deutsch. Dendr. Ges. 8: 77-94. 1899. |[Berberis; bibliography; key to 100 spp. ] VaAARAMA, A. Contributions to the cytology of the genus Berberis. Hereditas 33: 422-424. 1947. [Somatic chromosomes of * Mahoberberis sp.] WILKINSON, R. E. Berberis Thunbergii, a host of cucumber mosaic virus (Mar- mor cucumeris). (Abs.) Phytopathology 43: 489. 1953. Wyman, D. Two new Mahoberberis hybrids. Arnoldia 18: 9-12. 1958. . Barberries [Berberis|. Ibid. 22: 9-16. 1962. Cae spp., vars., hybrids. ] Subfam. PODOPHYLLOIDEAE Lindl. 3. Podophyllum Linnaeus, Sp. Pl. 1: 505. 1753; Gen. Pl. ed. 5. 223. Herbaceous, rhizomatous perennials to about 50 cm. tall. Solitary leaves + orbicular, palmately veined, 5—9-lobed, + centrally peltate, long- petiolate; cauline leaves usually 2 (1-3), + marginally peltate, 3—7-lobed, short-petiolate, usually subopposite (or alternate), subtending a single flower (rarely 2); juvenile leaves + quadrate. Hairs unicellular. In- florescence terminal, of a single, pedunculate, calyculate, + 3-merous flower. Sepals 6, 2-seriate. Petals 6 or 9, in 2 or 3 series, white (pink). Stamens 12-18 (or more), usually twice as many as petals [or 6 and opposite the petals]; anthers longitudinally dehiscent. Gynoecium nor- mally unicarpellate and solitary (or sometimes of 2-8 free ovaries) ; stigma convoluted; style indefinite or absent; placenta parietal; ovules numerous. Fruit + ovate, to 5 cm. long, yellow to reddish orange or purplish, indehiscent, filled with the many seeds and the pulpy placenta, the latter forming + discreet arils, each inclosing a seed. Seedling cotyle- dons connate at the base. (Excluding Dysosma Woodson.) LrctotTyPE SPECIES: P. peltatum L.; typified by the removal of P. diphyllum L. to Jeffersonia Bart.; see Persoon, Syn. Pl. 1: 418. 1805. (Name from Greek, podos, foot, and acl leaf; contraction of Anapodophyllum Tourn.) — MAY-APPLE, MAND One species in the woodlands and meadows of eastern North America and one (or more) species in Asia. Podophyllum peltatum, 2n = 12, occurs from Florida to Texas, northward to Minnesota, Ontario, ad Quebec. The early spring plants are conspicuous, with the terminal 14 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV solitary flower bud characteristically placed between more or less paired, reflexed, furled leaves. The arrangement of the leaves and the position of the flower are variable (Foerste, Halsted, Harris, Holm, Wadmond); in exceptional instances the flowering stems lack leaves. The flowers are unusual in having twice as many stamens as petals; occasionally, the flowers have extra or irregularly inserted petals or two or more separate gynoecia (Clute). Eames insists that the normal fruits are the equivalent of three carpels and are only falsely monomeric (see also Chapman). The fruits usually are yellow but sometimes also orange-tinged or reddish; plants with reddish fruits seem to have pinkish petals (Raymond, Steyer- mark). The rhizomes, which sometimes cause dermatitis when handled, and other parts of the plants, including the green fruits, are bitter and may contain dangerous amounts of ‘“‘podophyllin,’ a mixture of potentially poisonous compounds. The ripe fruits, however, are considered edible and sometimes are preserved (Fernald & Kinsey, Muenscher, and Pammel). In Asia, Podophyllum emodi Wall. ex Hook. & Thoms., 2 = 12, with pink (or white) petals, as many stamens as petals, and reddish fruits, and occurring to an altitude of 4500 m., has the characteristic vernal appearance of the American species, but the cauline leaves (occasionally there are three) are more often alternate and usually are distinctly ternately parted. The cotyledons of Podophyllum are broadly oval or ovate and fused into an elongate, tubular, petiole-like base. During the first growing- season the epicotyl is undeveloped or retarded so that the cotyledons are the only photosynthetic organs. Later, the plumule bursts through the base of the cotyledonary tube, producing first a few bracts and then a peltate leaf with a long petiole (Holm, Dickson, Lubbock). The con- formation of the flowering stems in some respects recalls superficially some Ranunculaceae (Glaucidium, Hydrastis), some Papaveraceae (Hylomecon, Stylophorum); in our flora, Podophyllum vegetatively resembles Diphyl- leia. The relationship of Podophyillum to the one or two species of Dysosma Woodson, of Asia, is uncertain (see Kumazawa). In D. pleiantha (Hance) Woodson (Podophyllum pleianthum Hance), the leaves are of a distinc- tive texture and usually with distinctively toothed margin; the several pendent, usually dark flowers are arranged in an extra-axillary umbel; and the rhizome seems to resemble that of Diphylleia. In another taxon of Dysosma, the leaves appear to be more or less square and centrally peltate. REFERENCES: Under family references see AGARDH, FERNALD & Kinsey, Gray, Hu, KuMma- zAWA, and LEwIs BarTEK, J., et al. Isolation of some components of resina podophylli (Podophyl- lum peltatum L.) and comments on their structure. (In Czech.) Chem. Listy 49: 1550-1560. 1955.* BASTIN, . S. Structure of Podophyllum. Am. Jour. Pharm. 66: 417-424. 1894. [Anatomy. ] ay I. I. Medicinal plants of ao Pakistan: Podophyllum emodi L. Pakistan Jour. Sci. 8: 230-233. 1964 | ERNST. GENERA OF BERBERIDACEAE 15 Ciark, L. The embryogeny of Podophyllum peltatum. Minn. Stud. Pl. Sci. 1: 111-138. 1923. Ciute, W. N. A may-apple with multiple fruits. Am. Bot. 21: 92, 93. 1915. [P. peltatum with 2, 3, and 5 fruits from a single flower Dar.LINcTON, C. D. The analysis of chromosome cravenents I. Podophyllum versipelle. Cytologia 7: 242-247. 1936. [Dysosma. | Dickson, A. On the pranaey of Podophyllum emodi. Trans. Bot. Soc. Edinb. 16: 129, 130. pl. 9. 1886 ForersteE, A. F. The may sot Bull. Torrey Bot. Club 11: 62-64. 1884. [Stem morphology and variation. HAtstep. B. D. Pistillodia of Podophyllum stamen. Bull. Torrey Bot. Club 21: 269. 1894. Harris, J. A. The nb of Podophyllum. Bot. Gaz. 47: 438-444. 1909. [Variation in leaf number. | HimM_EL, W. J. A ech to the biophysics of Podophyllum petioles. Bull. Torrey Bot. ror 54: 419-451. pls. 30, 31. 1927. [Effect of pressure and tension on growth. | Hom, T. Podophyilun ites Bot. Gaz. 27: 419-433. 1899. [Analysis of seedling and stem; examples of similar ore bibliography. . Medicinal plants of North America. 7. eer peltatum L. Merck's Rep. 16: 250-252. 1907.* [See Bot. Jahresb. 36(1): 438. 1910.] Hooker. J. D. Podophyllum pleianthum. Bot. Mag. 116: pl. 7089. 1890. | Dysosma. | Hutcurnson, J. Podophyllum versipelle. Bot. Mag. 133: pl. 8154. 1907. | Dysosma. | KAUFMANN, B. P. Chromosome structure se its relation to the chromosome cycle. II. Podophyllum peltatum. Am. Jour. Bot. 13: 355-363. 1926 Kvuester. H. L. A chemical study of the ee and roots of Podophyllum peltatum L. Jour. Am. Pharm. Assoc. 15: 259-263. 1926.* Kumazawa, M. Podophyllum pleianthum Hance. Bot. Mag. Tokyo 50: 268- 276. 1936. | Morphology; comments on ieee Kuznetsova. G. A.. E. A. SELIVANOvA-Goropkova, A. S. SAMOKHVALOVA, & P. A. IaKtmov. The study of the may apple oa ieiee peltatum L.) cultivated in the Leningrad region. (In Russian.) Bot. 33 1340. 1959. [Chemical content. medical application. | ee R. pe. Remarque au sujet de quelques processus chromosomiques s les noyaux diploidiques du Podophyllum peltatum L. Compt. Rend. ree Sci. Paris 172: 1066-1068. 1921. MartTIN. F. W. Variation ae morphology of Podophyllum peltatum. Diss. Abs. 19(3): 424. mi 1958 MELLANOFF. I. & H. L. “SCHAEFFER. A study of resins of Podophyllum peltatum a ‘Am. Jour. Pharm. 99: 323-330. 1927 Mvut.inc. G. N.. & G. B. Witson. The chromosomes oF Podophyllum doo Rhodora 63: 267-275. 1961. [Microsporogenesis and mitosis in tapetu Newman. L. J. Chromosomal aberrations in Podophyllum peltatum. Evolution 13: 276-279. 1959. [Inversion. fragments. translocation OverToN, J. B. The organization of the nuclei in the root fs of Podophyllum peltatum. Trans. Wis. Acad. Sci. Arts Lett. 20: 275-322. pl. 7. 1922. [ Mitosis. See also A. RicHarps, Univ. Kan. Sci. Bull. 5: 87-93. pls. 15, 16. 16 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Porter, T. C. Variation in Podophyllum peltatum, Linn. Bot. Gaz. 2: 117, 118. 1877. [Variation in leaf number. Prain, D. Podophyllum emodi, var. chinense. Bot. Mag. 146: pl. 8850. 1920. Raymonp, M. A red-fruited form of Podophyllum peltatum. Rhodora 50: 18. 1948. [Forma Deamii. | SawYER, M. L. Carpeloid stamens of Podophyllum peltatum. Bot. Gaz. 82: 329-332. 1926. [Teratological examples. Scott, W. R. M., & E. J. Perry. Correlation of variation in resin content of Podophyllum with certain habitats. Rep. Mich. Acad. Sci. 21: 225-231. 1920 SELIVANOVA-GoroDKovA, E. A. On Podophyllum peltatum L. (In Russian.) Acta Inst. Bot. Acad. Sci. URSS. 6. Introd. Pl. 6: 262-297. 1958.* Sims, J. Podophyllum peltatum. Bot. Mag. 43: pl. 1819. 1816. Smirnova, E. S. In regard to a study of the genus Podophyllum L. in connection with the problem of relationship of monocotyledons and dicotyledons. (In Russian.) Dokl. Mosk. Sel’skokhoz. Akad. Timiriazeva 46: 217-226. 1959.* STEYERMARK, J. A. Color-forms of the may-apple. Rhodora 54: 131-134. 1952. SuLLIVAN, B. J., & H. I. WecHSLER. The cytological effects of podophyllin. Science 105: 433. 1947. [Compared with colchicine; affects spindle forma- tion, disperses chromosomes; perhaps useful cytological tool. | TayLor, J. H. Transition of meiosis to mitosis and visa versa in cultures of excised anthers of Podophyllum peltatum. Genetics 35: 136, 137. 1950. Wapmonp, S. C. Leaf retardation in Podophyllum peltatum. Asa Gray Bull. 6: 66, 67. 1898. [Aphyllous examples. ] WALLIS, T. E., & S. GotpBerc. The pian of Podophyllum peltatum. Quart. Jour. Pharm, 10: 40-51. 1937 The histology of Indian te Ibid. 311-318.* WARTBURG, A. von, E. ANGLIKER, & J. RENz. Lignanglucoside aus Podophyllum ee m L. 7. Mitteilung iiber mitosehemmende Naturstoffe. (English mmary.) Helvet. Chim. Acta 40: 1331-1357. 1957.* ase: E. G. The structures of the rhizome and root of Podophyllum peltatum L. Pharm. Arch. 3: 1-8. pls. 1, 2. 1900.* Woopson, R. E. A new genus of Berberidaceae. Ann. Missouri Bot. Gard. 15: 335-340. pl. 46. 1928. [Dysosma; synonymy of Podophyllum spp.; cf. KUMAZAWA. | 4. Diphylleia Michaux, Fl. Bor.-Am. 1: 203. pls. 19, 20. 1803. Perennial herbs, to 1 m. tall, from a characteristically jointed rhizome. Solitary leaves reticulately palmately veined, reniform-orbiculate, 2-cleft, to 70 cm. broad, + centrally peltate, long-petiolate; cauline leaves similar, usually 2, marginally peltate, distinctly alternate. Hairs unicellular. In- florescence long-peduncled, many flowered, cymose, umbelliform. Flowers 3-merous, noncalyculate. Sepals 6, 2-seriate. Petals 6, 2-seriate, white. Stamens 6; anthers dehiscent by 2 apically hinged valves; pollen con- spicuously spiny, ? irregularly aperturate. Style indistinct; placenta parietal; ovules 5 or 6, inserted near the base of the placenta. Fruit + gibbous, indehiscent, to 1 cm. broad, blue and glaucous. Seeds 2—4, oblong, slightly curved. Type species: D. cymosa Michx. (Name from Greek, dis, double or twice, and pAyllon, leaf.) — UMBRELLA-LEAF. 1964 | ERNST, GENERA OF BERBERIDACEAE 17 One species along mountain streams in our area from about 1000 to 1700 m. altitude and one or two species in Asia. Diphylleia cymosa, 2n = 12, is distributed in the higher altitudes of northern Georgia, eastern Tennessee, western North Carolina, and adjacent Virginia. The leaves and the ah as are sparsely pubescent. In the Asiatic D. Grayi F. S. Schmidt, = 12, the margins of the leaves usually are somewhat less incised, a ae pedicels are especially pubescent. The latter taxon, sometimes given varietal status under D. cymosa, although that combina- tion seems not to have been formalized, is considered by Li to be a separate species distributed in Japan and Sakhalin, while the Chinese plants, occurring to about 3700 m. altitude in Hupeh, Szechuan, and Yunnan, are referred to D. sinensis Li. In the Japanese material, the uppermost leaf tends to be more or less sessile. Diphylleia has been attributed, without the citation of specimens, to the Amur region of Manchuria-Siberia (see Kumazawa, Li). The distribution is given in the Flora of the USSR, however, as Japan and Sakhalin only. Diphylleia resembles Podophyllum and Dysosma but is distinguished by the two cleft leaves; the long-peduncled, umbelliform inflorescence; the white petals; the anthers with two uplifting valves; and the spiny pollen. REFERENCES: Under family references see Gray and KuMAz Li, H. L. Notes on the Asiatic flora. Jour. ree Arb. 28: 442-444. 1947. [| Diphylleia. | Lioyp, J. U., & C. G. Lioyvp. Diphylleia cymosa. Drugs Med. N. Am. 2: 120, 121. 1887. [Medicinal and historical discussion. ] MuraAKAMI, T., & A. MatsusHIMA. Studies on the constituents of Japanese Podophyllaceae plants. I. (In Japanese; nce summary.) Jour. Pharm. Soc. Japan 81: 1596-1600. 1961.* [D yt. | Sms, J. Diphylleia cymosa. Bot. Mag. 39: - “T666. 1814. 5. Jeffersonia Barton, Trans. Am. Philos. Soc. 3: 342. 1 pl. 1793. Low, acaulescent, glabrous herbs to about 50 cm. tall, from short rhizomes with fibrous, matted roots. Leaves few, radical, subtended by bracts, long-petiolate, the blade deeply 2-parted or 2-foliolate, the parts + palmately veined |or leaves undivided and + reniform]. Inflorescence scapiform with a solitary flower (rarely subtended by a linear bract). Perianth 4(3—5)-merous, 3-seriate, often + acyclic. Sepals 4 (3-5), pet- aloid, 1-seriate. Petals 8 (9), + 2-seriate, white [or lavender]. Stamens 8; anthers dehiscent by 2 narrow valves hinged at the apex. Stigma + 2-lobed; style + distinct; placenta parietal, the ovules many, indefinite. Fruit + clavate, + tuberculate, incompletely transversely [or obliquely | loculicidally dehiscent near the summit, the margins of the stoma be- coming extended and + revolute. Seeds many, oblong; aril laciniate, attached at the upper side of the hilum. Type species: J. binata Bar- ton = J. diphylla (L.) Pers. (Named in honor of Thomas Jefferson, 1743- 1826, Secretary of State under George Washington and third President of 18 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLV the United States, without “reference to his political character or to his reputation for general science, and for literature,’ but for his knowledge of natural history, especially botany and zoology.) — TWINLEAF, RHEU- MATISM ROOT. One species in eastern Asia and one in eastern North America, usually in rich deciduous woods. Jeffersonia diphylla, 2n = 12, occurs from Ala- bama and Tennessee northward to Iowa, Ontario, New York, and Virginia, often in calcareous soil. Frequently there is a minute apiculation at the terminus of the petiole between the two leaflets. The dehiscence of the odd fruits is more or less horizontal. The Asiatic J. dubia (Maxim.) Benth. & Hook. ex Baker & Moore (Plagiorhegma dubium Maxim.), 2n = 12, has undivided, slightly apiculate, cordate-reniform leaves, laven- der petals, and somewhat obliquely dehiscent fruits. The unusual features of Jeffersonia are the acaulescent habit, flowers with half as many sepals as either petals or stamens, and the definite loculicidal dehiscence of the fruits. Sometimes the insertion of the perianth segments is not quite regular. Both species superficially resemble Sangui- naria canadensis L., of the Papaveraceae. The asymmetrical leaflets of J. diphylla recall the lateral leaflets of Achlys. REFERENCES: Under family references see GRAY and Lewis Atry-SHAW, H. K. Jeffersonia dubia. Bot. Mag. 164: pl. 9681. 8. ANDREWS, F. M. hear of the embryo-sac of Jeffersonia diphylla. Bot. Gaz. 20: 423, 424. pl. 2 BARNHART, J. H. eae dnds diphylla. Addisonia 5: 31, 32. pl. 176. 1920. GRAENICHER, S. Some notes on the pollination of flowers. Bull. Wis. Nat. Hist. Soc. II. 4: 12-21. 1906. |Jeffersonia, 12-14; visited by bees. 9 spp.. and flies, 1 sp.] Hoim. T. Medicinal plants of North America: 3. Jeffersonia diphylla (L.) Pers. Merck’s Rep. 16: 125-127. 1907.* [See Bot. Jahresb. 36(1): 437. 1910. | HvuTcHINsOoNn, J. Jeffersonia ane Plagiorhegma. Kew Bull. 1920: 242-245. 1920. [Contrasts generic chara Sims. J. Jeffersonia diphylla. es Mag. 37: pl. 1513. 1812. 6. Caulophyllum Michaux, Fl. Bor.-Am. 1: 204. pl. 27. 1803. Herbaceous, glabrous, rhizomatous perennials. Leaves usually 2 (3), cauline, alternate, the lower 1 (2) ternately (to biternately) or apparently pinnately compound but + sessile and easily mistaken for 3 leaves (i.e., the common petiole mostly obscure or absent) ; leaflets 9 or more, variable in shape, the terminal ones + obovate, ternately lobed or divided, and + palmately veined; the lateral leaflets sometimes lanceolate and pinnately veined; the uppermost leaf + similar but smaller. Inflorescence terminal (often meh a smaller one axillary), bracteate, several flowered, race- mose or paniculiform. Flowers yellowish green to purplish, Ee 3-merous, Sepals 6, 2-seriate, petaloid. Petals 6, + uniseriate, small, + 1964] ERNST, GENERA OF BERBERIDACEAE 19 truncated, glandular. Stamens 6; anthers dehiscent by 2 uplifting valves. tigma minute, + introrsely decurrent; style subulate; placenta subbasal; ovules usually 2, on stout funiculi. Fruit soon ruptured by the enlarging seeds, small, thin walled, evanescent. Seeds usually paired, about 1 cm. broad, + spherical, fleshy, glaucous and blue (easily mistaken for fruits), elevated on conspicuous stout funiculi, and somewhat persistent; endosperm + spherical, hard, inconspicuously umbilicate ventrally; embrvo vertical, central, cylindrical, with radical below. Seedling cotyledons hypogeal. Type species: C. thalictroides (L.) Michx. (Leontice thalictroides L.). (Name from Greek, caulos, stem, and pAvilon, leaf, in recognition of the peculiar continuity of the stem and leaf.) — BLUE COHOSH, PAPOOSE-ROOT. One species distributed in eastern North America and another in eastern Asia. Caulophyllum thalictroides (Leontice subg. Caulophyllum Gray; $ Caulophyllum Prantl), 2n = 16, occurs in moist deciduous woodlands from New Brunswick and Nova Scotia to Manitoba, and southward, to an altitude of over 1000 m., in the mountains of North and South Carolina, Tennessee, Alabama, and Georgia. The rhizomes contain the alkaloid methylcytisine and perhaps glucosides. The herbage also is bitter, and children should be warned against eating the prominent berry-like seeds. Handling of the plants may cause dermatitis or irritation of mucous sur- faces (Muenscher). The roasted endosperm has been suggested as a possible coffee substitute (Gray). The Asiatic Caulophyllum robustum Maxim., 2n = 16, usually seeming to be more vigorous or larger than our species, occurs to an altitude of about 3500 m. The mostly narrower leaflets more often are pinnately veined. The unusual leaves, the peculiarly colored flowers, the truncated and glandular petals, and the unique maturation of the seeds are the salient characteristics of Caulophyllum. The leaflets and flowers somewhat recall those of Ranzania japonica (T. Ito) T. Ito, 2m = 14, of eastern Asia. Some floral similarities have been noted between Caulophyllum and the Eurasian Leontice L., having ternately decompound leaves, and Bongardia (Leontice § Bongardia Prantl), having imparipinnate leaves. The seeds in these two, however, do not exceed the finally gaping and scarious ovaries. REFERENCES: Under family references see GrAy, Lewis, and MAUvRITZON. Butters, F. K. The seeds and seedling of Caulophyllum thalictroides. Minn. Bot. Stud. 4: 11-32. pls. 4-10. 1909. [Cotyledons hypogeal; primary root persistent. | Hoim, T. Medicinal plants of North America: 2. Caulophyllum thalictroides (L.) Michx. Merck’s Rep. 16: 94-96. 1907.* [See Bot. Jahresb. 36(1): 437. 1910. HUTCHINSON, Ne Caulophyllum thalictroides. Gard. Chron. III. 67: 63. 1920. | A ‘“‘gymnospermous”’ dicotyledon.. 20 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Luovp, J. U., & C. G. Luoyp. Caulophyllum thalictroides. Drugs Med. N. Am. 2: 141-162. pls. 39, 40. 1887. [Historical and medicinal descriptions. | Rospertson, C. Flowers and insects. XVII. Bot. Gaz. 22: 154-165. 1896. [Caulophyllum, 154; proterogynous, outcrossed. Insect visitors: Hy- menoptera, 9 genera; Diptera, 8; Coleoptera, 2. LARDIZABALACEAE Decaisne, Arch. Mus. soos —_ Paris 1: 185. 1839, ‘“Lardizabaleae,”’ nom. (LARDIZABALA aa About five genera in Asia and two in western South America, probabl totaling less than 50 species. One species of Akebia, a garden escape, is established in several of the eastern United States. Among the more common characteristics of the family are the scandent [rarely arborescent], woody habit and perulate buds; the multiple articu- lation of the alternate, usually palmately [rarely pinnately] compound leaves; the regular, cyclic, trimerous structure of the functionally uni- sexual [rarely bisexual], hypogynous flowers; and aposepaly, apetaly [or petals reduced], and apocarpy. All or half of the [sometimes monadel- phous] stamens are opposite the perianth segments. Although not well exemplified by Akebia, the flowers of Lardizabalaceae tend to have two series of sepals and two of petals, so that a sepal subtends a petal, which, in turn, subtends a stamen or staminodium. This pattern also is charac- teristic of Menispermaceae and Berberidaceae, but contrasts with, for example, Papaveraceae and Fumariaceae, which have only a single series of sepals and conspicuous petals that are not so regularly opposed to as many stamens. Lardizabalaceae are distinguished from Berberidaceae in having unisexual flowers, extrorse anthers, and gynoecia of three or more free carpels. From Menispermaceae they are distinguished in being pri- marily monoecious [rarely dioecious or bisexual], in having more than a single seed per carpel, and in lacking a bony endocarp. Especially signifi- cant in Lardizabalaceae is the submarginal or laminal distribution of the ovules along the ovary wall rather than the restriction of the ovules to a single parietal or subbasal placenta. The fruits of Lardizabalaceae usually are fleshy and more or less follicular [or sometimes baccate| and are considered edible (see Fernald & Kinsey). The plants usually are glabrous [rarely with simple hairs]. Exceptional taxa are Decaisnea Hook. f. & Thoms., a hardy, upright shrub or small tree with pinnately compound leaves and bisexual flowers; and Lardizabala Ruiz & Pavon, nomenclatural type of the family, which is dioecious. The reported chromosome numbers are 2n = 28, 30, and 32. Only a few species are cultivated for ornament. In general, the family seems allied with Berberidaceae, Menispermaceae, and Sargentodoxaceae Stapf ex Hutchinson. REFERENCES: Under Berberidaceae see BAILLON, EICHLER, cea KUMAZAWA (1936), and Li; under Menispermaceae see Hooker & THOMSO 1964] ERNST, GENERA OF BERBERIDACEAE 21 BAILEY, I. W., & B. G. L. Swamy. The sg an carpel of Sart and its initial trends of specialization. Am. Jour. Bot. 38: 373-379. DECAISNE, J. Mémoire sur la famille des Perea ient Arch. Mus. a Nat. Paris 1: 143-213. pls. 10-13. 9. . Enumeratio Lardizabalearum. (In Latin.) Ann. Sci. Nat. Bot. II. 12: 99-108. 1839. GAGNEPAIN, F. Revision des Tae asiatiques de l’herbier du Muséum. Bull. Mus. Hist. Nat. Paris 14: 64-7 08: HemsLey, W. B. Asiatic ee Kew Bull. 1908: 459-461. 1908. [ Notes on Holboellia, Parvatia, Stauntonia. HENDERSON, E. M. The stem structure of Sargentodoxa cuneata, Rehd. et Wils. Trans. Proc. Bot. Soc. Edinb. 29: 57-62. 1924. [Comparison with genera of Lardizabalaceae. | HeralL, J., & R. BLotTirereE. Note sur les affinités des Lardizabalées. Bull. Soc. Bot. Fr. 33: 521-524. 1886. Hooker, W. J. Lardizabala biternata. Bot. Mag. 76: pl. 4501. 1850. [Stipules foliaceous. | PRANTL, K. Lardizabalaceae. Nat. Pflanzenfam. III. 2: 67—70. 1888. REAvuBOURG, G. Etude organographique et anatomique de la famille des Lardiza- balées. Thése. Univ. Paris, Ecole Supér. Pharm. 127 + 3 pp [Akebia, 31-52. ] Stapr, O. Sargentodoxa cuneata. Bot. Mag. 151: pls. 9111, 9112. 1926. [Six pages of discussion of family Teen Swamy, B. G. L. Some observations on the embryology of Decaisnea insignis Hook. et Thoms. tec Natl. Inst. Sci. India 19(2): 307-310. 1953.* 1. Akebia Decaisne, Arch. Mus. Hist. Nat. Paris 1: 195. 1839. Twining, glabrous, deciduous or evergreen, woody vines. Leaves alter- nate or + fascicled on short shoots, exstipulate, palmately compound, about 5[3]-foliolate; petioles long, articulated near the base and at the insertion of the petiolules, the latter articulated at the base of the + elliptical or obovate, + emarginate, usually minutely apiculate leaflets; blades + pinnately veined but often with 3 prominent veins from below. Plants monoecious; inflorescences axillary, unisexual or bisexual, bracteate, pedunculate, racemose, or + subumbelliform. Flowers pedicellate, regular, aposepalous, 3-merous, the sepals usually 3 (6), + petaloid, brownish or purplish, apetalous, functionally unisexual, hypogynous, apocarpous. 4 flowers several to many, distal on the raceme (sometimes a few also below) ; stamens usually 6, free, equal; filaments short; anthers elongate, extrorse, longitudinally dehiscent; pollen mostly 3-colpate; gynoecium rudimentary. @ flowers larger, solitary or few; stamens rudimentary; gynoecium of 3-12 free carpels; stigmas + broad, truncated; style ob- scure; ovary unilocular; placentae 2, laminal; ovules many, + ortho- tropous. Fruits follicular, many seeded, dehiscent along the suture, placenta + pulpy. Seeds + arillate?, endosperm present; embryo small. (Rajania Houtt. non L.) Lectotype species: A. quinata (Houtt.) Decne.; see A. Rehder, Bibl. Cult. Trees Shrubs 166. 1949. (Name de- rived from the Japanese word for the plant.) 22 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV About four species native to eastern Asia. In our area Akebia quinata, 2n = 32, occasionally grown for ornament, apparently is established and growing without cultivation in Madison County, North Carolina. Several authors warn that, if extensively naturalized, this species could become a nuisance by forming dense mats over other vegetation. The number of leaflets sometimes is variable, the terminal one, with slightly longer petiolule, usually somewhat larger. The clusters of slightly fragrant flowers are striking in appearance with their unusual color and velvety texture. The color has been likened to that of raw liver, sometimes appear- ing maroon; or the sepals of the staminate flowers appear rosy purple and those of the carpellate flowers purplish brown (Anderson). The unusual oblong fruits, to 8 cm. long, seldom seen in cultivation unless the plants are cross-pollinated, are glaucous, purple-violet. When the fruits ripen and open, the seeds are displayed in an elongated, gelatinous pulp derived from the inner portion of the ovary wall. The fruits and roots sometimes are used medicinally in China (Hu). Another species, A. trifoliata (Thunb.) Koidz., 2m = 32, usually with 3-foliolate leaves and paler sepals and fruits, sometimes also is cultivated. Akebia pentaphyilla (Mak.) Mak. is said to be a hybrid between this and A. quinata. REFERENCES: Under Berberidaceae see AGARDH, DORMER, FERNALD & KINSEy, and Hu. AHLEs, H. E., & A. E. Raprorp. Species new to the flora of North Carolina. Jour. Elisha Mitchell Sci. Soc. 75: 140-147. 1959. [A. quinata, 142.] ANDERSON, E. The genus Akebia. Arnold Arb. Bull. Pop. Inf. IV. 2: 17-20. 1934. [Illustrations of flowering and fruiting A. quinata. | & A. REHDER. New hybrids from the Arnold Arboretum. Jour. Arnold Arb. 16: 358-363. 1935. [A. & pentaphylla.] Hooker, J. D. Akebia lobata. Bot. Mag. 122: pl. 7485. 1896. [= A. trifoliata. | Hooker, W. J. Akebia quinata. Bot. Mag. 81: pl. 4864. 1855. LAvALLEE, A. Akebia quinata. Hort. Fr. 1869: 103-106. pl. 4. 1869. Li, H. L. Akebia as a weed in the Philadelphia area. Morris Arb. Bull. 5: 58. 1954 Sarto, K. Studies on the induction of polyploid flower plants and their utiliza- tion. XI. On the autotetraploid plant of fiveleaf Akebia. (In Japanese; English summary.) Jour. Hort. Assoc. Japan 26(1): 43, 44. 1957.* [Col- chicine treatment of A. quinata. | SARGENT, C. S. The fruit of Akebia paene gas Forest 4: 136, 137. 1891. [Illustrated: some plants fruitful, other Sawapa, T. Whence is or[i]ginated ed ge a conditional synonym of Akebia quinata. (In Japanese.) Jour. Jap. Bot. 5: 153-157. 1928. SHIMIzU, T. Taxonomic study of the genus Akebia, with special reference to a new species from Taiwan. Quart. Jour. Taiwan Mus. 14: 195-202. 1961. VESLER, J. Zur Entwicklungsgeschichte von Akebia quinata. Diss. Bonn. 1913.* ViLtMorIN, M. L. pe. Deux Lardizabalées a fruits comestibles. Bull. Soc. Acchim. Fr. 62: 89-93. 1915. [Akebia quinata, Decaisnea. | 1964 | ERNST, GENERA OF BERBERIDACEAE 23 MENISPERMACEAE A. L. de Jussieu, Gen. Pl. 284. 1789, ““Menisperma,”’ nom. cons. (MooNSEED FAMILY) Perennial herbaceous vines, woody at least at base [sometimes from tubers, rarely erect shrubs or small trees], sometimes with colorless, bitter juice. Leaves variable, alternate, simple [rarely 3-foliolate], usually exstipulate, the blades mostly palmately [to pinnately] veined, the margin entire to palmately lobed, sometimes peltate, deciduous or evergreen; the petiole sometimes swollen distally and/or basally. Hairs 1- or 2-celled (or multicellular-uniseriate) [sometimes glandular]. Plants mostly dioe- cious (? sometimes polygamous or flowers tending to be bisexual). Inflores- cences variable, commonly supra-axillary, bracteate, usually of many [rarely solitary] flowers, racemose to paniculiform or umbelliform, + determinate, sometimes concentrated at ends of branches. Flowers mostly small, usually unisexual, usually + regular (irregular), hypogynous, usually 3(2 or 1)[4 or 6]-merous, greenish, yellowish to whitish, some- times + calyculate, the perianth cyclic (to + acyclic), usually deciduous, aposepalous, apopetalous (sympetalous or apetalous). Sepals 4—9 (or 1) [to many], often in 3-merous cycles, + 2(3){or more]-seriate. Petals 6-8 (or 1 or absent), mostly + 2-seriate. Stamens usually 6 (3, 4) to many [40] in ¢ flowers, often opposite and as many as the petals (or indefinite): filaments free {monadelphous], sometimes embraced by the petals, or anthers sessile on an androphore; anthers 4(? 2)-locular, be- coming + 1(2)-locular at anthesis; pollen mostly prolate (oblate) and 3-colpate (colporate) ; staminodia often opposite the petals in @ flowers. Gynoecium of 3-6 (or 1) [-32]| free [to ?partially connivent]| carpels, usually in 1 series, often on a gynophore; stigmas various, subulate to laciniate; styles usually short; ovary unilocular, + gibbous; ovules 1 (or 2), usually inserted + medially on the solitary parietal placenta, 2- integumented, + anatropous, descending, the micropyle up (apotropous) ; gynoecium mostly absent (or rudimentary) in ¢ flowers. Fruits drupes, 1—5 per flower, + fleshy and indehiscent; endocarp bony, characteristically curved and sculptured, usually laterally compressed; seeds solitary; endo- sperm present [sometimes + ruminated or absent|; embryo + curved, ie pressed and fleshy (or foliaceous and laterally displaced) [sometimes un- equal]. Type GENUS: Menispermum Perhaps 80 genera and 370 species of tropical and subtropical regions. Four genera and five species are indigenous to the continental United States: four species, representing two tribes, occur in our area. Among the characteristics uniting most Menispermaceae are the peren- nial. mostly woody, dioecious, scandent habit; the simple leaves of variable conformation, the blade often with palmate venation and the petiole sometimes swollen distally and sometimes also basally; the perianth tend- 24 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV ing to have two series of sepals (sometimes also calyculate) and two series of petals (these often reduced or thickened, sometimes involute or absent) tending to be shorter than the sepals and with each opposite a stamen or staminodium; the apocarpous gynoecium, mostly of more than a solitary carpel, the ovules (usually solitary) medially attached with the micropyle up; and the fleshy, indehiscent fruits with bony endocarp and solitary, usually strongly curved, folded, or somewhat coiled seeds. The vascular bundles of young stems are separated by broad medullary rays. The wood is distinctive, with secondary thickening from fascicular or extrafascicular cambia, sometimes forming concentric or eccentric rings of bundles, resulting in many cambia and alternating layers of xylem and phloem (see Metcalf & Chalk and Esau). Within a taxon, the leaves often vary from entire to lobed, nonpeltate to peltate. The matching of the dioecious materials and the determination of sterile specimens often is difficult. Correlations between taxa resembling one another in the ana- tomical structure of the swollen portion of the petiole (which differs from the remainder of the petiole) and taxa resembling one another in fruit and seed characters have not been found (Rudolph, fide Metcalf & Chalk). Most of the hairs that seem to be unicellular probably are bicellular, the basal cell inconspicuous. The lack of tubers in our species should be verified. The (anatropous) ovules at first seem more or less straight and vertical, the micropyle up. During the maturation of the endocarp or stone, the placental region seems to buckle inward towards the seed, leaving a notch or hollow exteriorly and forming inwardly an intrusion, or condyle (Miers), of various proportions around which the seed and the embryo are curved. The condyle in tribe Menispermeae is seen superficially as a depression (nearly orbicular in Menispermum, excentric in Cocculus, or narrowly obovate in Cissampelos) on the broad faces of the laterally compressed endocarps bordered adaxially by a notch or indentation on the margin of the endocarp. Internally the condyle may be hollow, perforated, or re- duced to a narrow, partial partition of the seed. Although the shape of the fruits may be nearly spherical, abaxial deformation of the endocarp often is accompanied by extreme abaxial enlargement of the pericarp, so that the final position of the style (and the apex of the seed) is subbasal in tribe Menispermeae. Abaxially, a median-longitudinal line or crest is characteristic of many endocarps; adaxially the crest often is obscured by the condyle. Insight into the form of the closed condvles of tribe Menispermeae may be provided by the broad, open concavity of the endocarp of Calycocarpum (tribe Tinospo- reae) with its conspicuous median adaxial crest. Menispermaceae, distinguished by habit, dioecism, position of micropyle, drupaceous fruits, and curved seeds, resemble Berberidaceae and Lardiza- balaceae in the insertion of perianth (petals often reduced) and androe- cium, and in the apocarpous gynoecia. Flowers of Berberidaceae, how- ever, are bisexual, and the solitary carpels usually develop more than one seed, attached to a single parietal or subbasal placenta, apparently with 1964] ERNST, GENERA OF BERBERIDACEAE 25 the micropyles down. In Lardizabalaceae, the plants usually are monoe- cious (otherwise dioecious), the flowers unisexual, the ovules submarginal in more than one vertical row, and the leaflets of the compound leaves articulated with the rachis. The family customarily has been divided largely on characteristics of the fruits (endocarp) and seeds (endosperm and embryo). Originally six (later seven) tribes were recognized by Miers, five tribes by Hooker & Thomson, four series by Baillon, four tribes by Prantl, and, most recently, eight tribes by Diels. In each case, the genera were rearranged to some extent. In Diels’s treatment, the three largest tribes (comprising over 80% of the species and of the genera) are Triclisieae Diels (14 genera, 96 species), lacking endosperm; Tinosporeae Hook. & Thoms. (41 gene 78 species), having endosperm and thin, laterally displaced foliaceus cotyledons; and Menispermeae (Cocculeae Hook. & Thoms.) (16 genera, 140 species), having endosperm and appressed, nonfoliaceous cotyledons. Menispermaceae in our area are but the fringes of a large, complex family. When classified according to conformation of endocarps and embryos, Cocculus (q.v.), Menispermum, and Cissampelos (all of tribe Menispermeae), seem relatively similar, while Calycocarpum (tribe Tino- sporeae), seems remote. A classification based upon floral morphology, however, would lessen the apparent isolation of Calycocarpum and em- phasize the distinctness of Cissampelos (q.v.), with its strongly dimorphic flowers. About 90 named alkaloids have been identified in about 29 genera of Menispermaceae (see Willaman & Schubert); of these, about three-fourths (ca. 68) are reported for Menispermaceae only. Others occur also in Berberidaceae (9), Fumariaceae (6), Rutaceae (6), Ranunculaceae (5), Lauraceae (5), Monimiaceae (3); Papaveraceae, Magnoliaceae, Annona- ceae, and Buxaceae (2 each); and Aristolochiaceae and Hernandiaceae (1 eac The poisonous and medicinal properties of Menispermaceae are dis- cussed by Baillon. Commercial curare (Intocostrin), for medicinal use, is prepared from a species of Chondodendron (Merck Index, ed. 7. 1960). Portions of some taxa are used for aboriginal poisons and sometimes, with potentially dangerous consequences, for fish poisons, pesticides, or bitter flavorings. Only a few species of the family are cultivated for ornament. Chromosome numbers of 2” = 18, 20, 24, 26, 38, 48, 50, 52, 54, and 78 have been reported from among about ten genera. REFERENCES: Under Berberidaceae see AGARDH, BAILLON, og EICHLER, HIMMELBAUR, Li, Luspock, Payer, and WILLAMAN & SCHUBERT. BENTHAM, G. Notes on Menispermaceae. sats Linn. Soc. Bot. 5(Suppl. 2): 45-52. 1861. [General comments; floral morphology. BENTLEY, R., & H. TrrmMen. Med. Pl. 1: pls. 11-15. 1876. [Chondodendron, Tinospora, oe Anamirta, Cissampelos. | BLotTtTiERE, R. Etude anatomique de la famille des Ménispermées. Thesis. Ecole Supér. oe Paris. 76 pp., pls. 1, 2. 1886. 26 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV CoLeBrookE, H. T. On the Indian species of Menispermum. Trans. Linn. Soc. : 44-68. pl. 6. 1822. [Coscinium, Anamirta, Tiliacora, Cocculus.| CzapekK, F. Die Bewegungsmechanik der Blattgelenke der Menispermaceen. Ber. Deutsch. Bot. Ges. 27: 404-407. 1909. [Tinomiscium, Anamirta. | Diets, L. Menispermaceae. Pflanzenreich IV. 94(Heft 46): 1-345. DipasuPIL, P. C. M. Comparative anatomy wa five species of the family Meni- spermaceae. Diss. Abs. 17(1): 21. 1957. DREUILHE, A. Les Ménispermees et leurs ae 45 pp. bibliography. Mont- EIcHLer, A. W. Menispermaceae. /n: Martius, Fl. Bras. 13(1): 161-226. pls, 36-51. 1864. [Eleven genera . Versuch einer Charakteristik der natiirlichen Pflanzenfamilie Menisper- maceae. Denkschr. Bayer. Bot. Ges. Regensburg 5: 1-40. 1 pl. 1864. Ernst, A. A new case of parthenogenesis in the vegetable kingdom. Nature 34: 549-552. 1886. [Disciphania Ernstit. | ForMAN, L. L. The Menispermaceae of Malaysia: I. Kew Bull. 11: 41-69. 1956; II. Ibid. 12: 447-459. 1958; III. /bid. 14: 68-78. 1960. |For IV, see Cocculus.] Gray, A. Menispermaceae. Gen. Pl. U. S. 1: 69-76. pls. 28-30. 1848. [Coc- culus ; a Calycocarpum. | pingiee Syn. Fl. N. Am. 1(1): 64-66. 1895. Hooker, J. D., & T. THomson. Menispermaceae. Jn: FI. Indica 1: 167-206. 1855. [Observational commentaries; controversy with Miers. See also Berberidaceae (including Lardizabalaceae), 211-252. ] HUNKIARBEYENDIAN, R. Des produits fournis 4 la matiére médicale par la famille des Ménispermées. These. Univ. Paris, Ecole Supér. Pharm. 80 pp. 1 pl., map. 1887 Krarrt, K. Systematisch-anatomische Untersuchung der Blattstruktur bei den Menispermaceen. Diss. Erlangen. 1907. Krukorr, B. A., & H. N. MotpenkeE. Studies of American Menispermaceae, with special reference to species used in preparation of arrow-poisons. Brittonia 3: 1-74. 1938. . en ee ee notes on American Menispermaceae — V. Bull. Torrey Bot. Club 78: 258-265. 1951. [See bibliography for intervening papers. ] Kunitomo, J. I. Studies on the alkaloids of menispermaceous plants. CLXXXII-— CLXXXIV. (In Japanese; English summary.) Jour. Pharm. Soc. Japan 81: 1253-1266. 1961.* Bs ee also various authors in earlier volumes for previous papers in this se Mauevu, J. Recherches ni sur les Ménispermacées. Jour. Bot. Morot 16: 369-378. 1902. _ Sur les organes sécréteurs des Ménispermacées. Bull. Soc. Bot. Fr. 53: 651-663. 1906. Maurin, E. H. A. Essai sur a ae des Ménispermées. Thése. Fac. Méd. Strasbourg. 43 pp. 7 pl. . Miers, J. A few remarks on . een Ann. Mag. Nat. Hist. II. 7: 33-45. 1851. [Early designation of tribes, some names now invalid. | n the Menispermaceae. /bid. IIT. 13: 1-15, 122-135 | Calycocarpum |, 315- 323. 486-491: 14: 49-53, 97-103, 252-261, 363-374 | Menispermum |. 1864: 17: 128-138 [Cissampelos|, 265-270; 18: 12-22. 1866; 19: 19-29 [ Cocculus |, 84-95, 187-197, 319-330; 20: 11-20, 167-175, 260-266 | con- cluded]. 1867 1964 | ERNST, GENERA OF BERBERIDACEAE 27 ———. Menispermaceae. Contr. Bot. 3: 1-402. pls. 88-154. 1871. [Complete monograph, a reprinting, with additions and corrections of the 18 pioneer sey cited above. See discussion of Calycocarpum, 24 ff.; Cissampelos, 7 ff.; Cocculus, 249 ff.; Menispermum, 111 ff. ee FE. Contribuzione allo studio anatomico del caule delle Menispermacee. Mem. Accad. Sci. Ist. Bologna V. 10: 647-656. 1904. [Menispermum canadense, Cocculus, Cissampelos. | PRANTL, K. Menispermaceae. Nat. Pflanzenfam. III. 2: 78-91. RADLKOFER, L. Ueber das anomale Wachsthum des Stammes bei Menispermeen. Flora 41: 193-206. 1858. RupotpH, K. Zur Kenntnis des anatomischen Baues der Blattgelenke bei den Menispermaceen. Ber. Deutsch. Bot. Ges. 27: 411-421. 1909. [Twelve genera. SANTos, J. K. Stem and leaf structure of Tinospora Rumphii Boerlage and Tinospora reticulata Miers. Philip. Jour. Sci. 35: 187-208. pls. 1-7. 1928. . Anomalous stem structure in Archangelisia flava and Anamirta cocculus from the Philippines. /bid. 44: 385-407. pls. 1-8. 1931. TrouPIN, G. Menispermaceae. Jn: W. B. Turritt & E. MiLne-Repueap, FI. Trop. E. Afr. 32 pp. London. 1956. [Ten genera. ] KEY TO THE GENERA OF MENISPERMACEAE General characters (see Cissampelos): dioecious vines, woody at base; leaves exstipulate, alternate, petiolate, palmately veined, entire to lobed; flowers uni- sexual, hypogynous, usually regular (irregular), often calyculate; sepals 2(or more)-seriate (or 1); petals reduced, + 2-seriate (or 1, sympetalous, or absent) ; stamens free, often opposite and embraced by the petals (or anthers sessile on an androphore) ; iid longitudinally (or horizontally) dehiscent; gynoecium apocarpous, of (1) 3 arpels ; ovules usually solitary (or 2), the micropyle u fruits drupaceous, ie palais: bony, characteristically shaped, often peseneny 1-seeded; embryo usually curved or folde A. Plant pee only @ flow nth of 1 sepal oe 1 petal; gynoecium of 1 pubescent carpel; ee unbranched, several in the axil of a broad bract; fruit red. pubescent, the style subbasal; stone orbicular, a ee rani sculptured (subtropical Florida). .................. . Cissampelos. Perianth of 6 or more segments; “eynoecium of 3 or 6 Poe carpels; peduncles branched, minutely bract ate. C. Petals apparently absent; mae glabrate, usually each subtending two short, unequal staminodia; carpels 3, stigmas irregularly la- ciniate, extrorsely radiate; gynophore inconspicuous; fruits black, the style apical; stone + smooth, broadly cup eae i ae POORER, is.2.54 Su. TS he Calycocarpum. Cc. Petals present, shorter than the sepals; fruits os pers pene stone sculptured, laterally compressed. Carpels 3, stigmas extrorsely dilated and convolute; gynophore mostly longer than broad; sepals glabrate; fruits glaucous, blue- black, sessile on the 2-3 mm. long gynophore; stone crescent SHAQECE 5 55.50 Soe ete a Ret acs 3. Menispermum. Carpels 6, stigmas subulate; gynophore broader than long; a o 28 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLV sepals coarsely pubescent; fruits red, short-stipitate on an in- conspicuous fei stone appearing coiled. .. 2. Cocculus. A. Plant bearing only ¢ flow E. Anthers 4, sessile on a ee androphore, the dehiscence horizontal; sepals 4, coarsely pubescent; corolla sympetalous. .... 4. Ciss sampelos. _ Anthers 6 to many, filaments free; sepals more than F. Stamens 12 or more; sepals glabrous or glabrate G. Anthers included, introrse, oblong, 2- aie at oiaae the dehiscence vertical. ...................... 1. Calycocarpum. G. Anthers exserted, + apical, 4-lobed, erect, + ocslar at anthesis, the dehiscence + confluent apically. 3. Menispermum. F. Stamens 6, anthers included, + apical, 4-lobed, nodding inward, + 1-locular at anthesis, dehiscence + confluent apically; each filament embraced by a petal; sepals coarsely pubescent. .... 2. Cocculus. ica Tribe TrnosporEAE Hook. f. & Thoms. 1. Calycocarpum Nuttall ex Gray, Gen. Pl. U.S. 1: 75. pl. 30. 1848. Leaves thin, deciduous, not peltate, usually 5(7)-veined, 3—5—7-lobed; petiole often longer than blade. Hairs mostly multicellular- uniseriate. Flowers regular, + calyculate with 1-3 small sepal-like bracts. Sepals 6, + petaloid, 2-seriate, + equal, glabrate. ¢ flowers: petals absent; stamens about 12, arrangement + indefinite, not enfolded by perianth segments ; filaments free; anthers included, + introrse, attached subapi- cally, 2-locular at anthesis, longitudinally dehiscent; pollen prolate, 3- colpate (colporate). @ flowers: petals apparently absent; staminodia 12, narrow, short, thick, unequal, 2 opposed above each sepal (the lower staminodia probably equivalent to petals); gynophore inconspicuous ; carpels 3 (4); stigma extrorsely reflexed, irregularly laciniate, conspicu- ous; style short; ovules apparently solitary. Fruits 1-3 per flower, black, + ovoid, ca. 25 mm. long, the style apical, the gynophore inconspicuous ; endocarp large (to ca. 20 mm. long, 15 mm. wide), ovoid-cup-shaped, + smooth abaxially, broadly concave adaxially and toothed, + mucronate apically; embryo broadly curved, the cotyledons foliaceous, laterally dis- placed. Type species: C. Lyonii (Pursh) Gray. (Name from Greek, calyx, a cup, and carpos, fruit.) — CupSEED, LYONIA-VINE. Monotypic, distributed in rich or alluvial soils, in thickets, woods, or river banks, from northwestern Florida and western Georgia, northward to Tennessee, Kentucky, and southern Illinois, and westward to eastern Kansas (?), eastern Oklahoma, and Texas. The thin, nonpeltate leaves with petioles at least as long as the blades are notable. The blades, to ca. 25 cm. long and wide and usually with five or more veins, are deeply three or more lobed to near the center of the blade with the lobes narrowed basally. The flowers of both sexes seem to be apetalous. In the staminate, the included, introrse, finally two- locular anthers are distinctive. In the carpellate, two reduced staminodia are opposed above each sepal; the lower series of staminodia, apparently 1964] ERNST, GENERA OF BERBERIDACEAE 29 is Le Calycocarpum. a—m, C. Lyonii: a, portion of flowering staminate plant, 0 4 carpellate flower, showing sepals, staminodia, and carpels, pe ae rpel, Re side, showing irregularly cae stigma, X 16; i, carpel, lateral a ‘sho wing eae and short gynophore; j, mature drupe, showing terminal stigma-style, - k, semidiagrammatic cross section of fruit, showing foliaceous cotyledons of e mbryo surrounded by endosperm (white), eee en poe! aes Mi ee and ey fruit wall, x 2; 1, endocarp, adaxial side, m2 30 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV missed by Gray, were called petals by Miers and Diels. In texture and appearance, these organs resemble staminodia more than petals, but are in the appropriate position for petals. Unlike other Menispermaceae in our area, the style remains apical in the fruit, and the ovoid endocarps are smooth and not laterally compressed. While the embryo is not so tightly curved as in tribe Menispermeae, it is broadly so, conforming to the wide, open adaxial depression of the condyle. REFERENCES: Under family references see Gray and MIERs. 2. Cocculus A. P. de Candolle, Syst. Nat. 1: 515. 1817 (“1818”), nom. cons. Leaves + coriaceous, the blades mostly longer than the petioles, non- eltate, prominently 3-veined (often with a finer submarginal vein at either side), deltoid to ovate, sometimes 3- or 5-lobed, often apically mucronulate, basally oblique to cordate. Hairs 1- or 2-celled. Flowers + regular, calyculate with 1-3 small sepal-like bracts. Sepals 6, + pubes- cent, 2-seriate, the inner 3 larger, the margins + fimbriate to erose. Petals 6, + 2-seriate, shorter than the inner sepals, + fleshy [bifid], each embracing a stamen or staminodium. ¢ flowers: stamens 6, free, included, usually nodding inward; anthers 4-lobed, appearing terminal, + 1-locular at anthesis, dehiscence + apically confluent; pollen prolate, 3-colpate (?colporate). ? flowers: staminodia 6, + linear; gynophore distinct, broader than long; carpels 6, in 1 series; stigma + subulate, terete, turned outward, grooved along the upper surface; style short, Fruits 3—5 per flower, red [purple], glabrous, short-stipitate (—1 mm.), the style subbasal, the gynophore inconspicuous (less than 1 mm. long); endocarp sculptured, appearing coiled, laterally compressed, shallowly notched adaxially, the radicular end slightly projecting; embryo + coiled, the cotyledons appressed, narrow (but wider than the radicle). LECTOTYPE SPECIES: C. hirsutus (L.) Diels (Menispermum hirsutum L.), typ. cons.; see Rickett & Stafleu, Taxon 8: 271. 1959, and Int. Code Bot. Nomencl. 1961: 259. (Name from Greek, diminutive of coccos, berry; originally applied to “Cocculus indicus” |[Menispermum Cocculus L. = Anamirta Cocculus (L.) Wight & Arn.]) — SNAILSEED, MOONSEED. About 11 species, mostly of warmer regions, eight in the Old World, two in North America, and one in Hawaii; the conserved type species is native to Africa. Cocculus carolinus (L.) DC. (Menispermum carolinum L., Cebatha carolina (L.) Britt., Epibaterium carolinum (L.) Britt.), coralbeads, Carolina moonseed, 2m = 78, is distributed in woods, thickets, and fields from Florida to Texas northward to Kansas, Indiana, Virginia, and North Carolina, and in Mexico, from Coahuila to Tamaulipas. The nonpeltate leaves with three prominent veins, the blades to ca. 14 cm. long and wide; the pubescent sepals; the included stamens, each embraced by a petal; 1964 | ERNST, GENERA OF BERBERIDACEAE 31 Fic. 2. Cocculus. , staminate flower, saan inner eae two series of ant are Sees ne short outer sepals not visible, 15; c, inner sepal, clasping petal, and stamen, showing dehiscence of anther, * 15; d, carpellate nee outer sepal in fore- ground removed to show inner sepals, two series of petals, staminodia, and carpels, X 15; e, inner sepal, clasping petal, and psec re ™ 15: F, carpellate flower, with perianth re staminodia removed to show receptacle, gynophore, and gynoecium of gibbo eer ate with ee subulate stigmas, * 15; g, diagrammatic vertical pens of carpel, showing attachment of ovule, the micropyle up, X 10; h, diagrammatic se Cae section of fruit, showing stipitate base, subbasal stigma-style, coiled embryo surrounded by endosperm (white), bony endocarp (dark), and fleshy fruit wall, 4; i, subhelical sculptured endocarp, showing central depression of condy le, the micropyle to the lower left, * 3. the short, broad gynophore; and the glabrous, short-stipitate, red fruits with somewhat coiled endocarp are distinctive. The closest affinities may lie with some portion of the complex involving C. trilobus (Thunb.) DC. of eastern Asia, although in the latter the petals are ligulate and bifurcate and the fruits dark purple. West of our area, Cocculus diversifolius DC., having leaves with short 32 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV petioles and often narrow, oblong blades with revolute margins, and dark purple fruits, occurs in Texas and Arizona, and southward, in Mexico, to Oaxaca While the taxa of Cocculus primarily are scandent (resembling Meni- spermum), C. laurifolius DC., 2n = 26, of Southeast Asia, sometimes cultivated in the warmer parts of the United States, is an upright shrub or small tree with glossy leaves. In Socotra, C. Balfourii Schweinf., also shrubby, is conspicuous for its extra- or supra-axillary, flattened, some- times branched cladodes on which are borne the flowers and fruits. Coc- culus and Menispermum were placed in subtribe Menisperminae (Coccu- linae Hook. & Thoms.) by Diels, but Miers( see p. 114. 1871) separated them, placing Cocculus in his tribe ‘‘Platygoneae.” Chromosome numbers of 2m = 18, 20, 22, 26, 38, 50, 52, and 78 are reported. REFERENCES: Under family references see COLEBROOKE, Gray, Miers, and Morini. Forman, L. L. The Menispermaceae of Malaysia: IV. Cocculus A. P. de Can- dolle. Kew Bull. 15: 479-487. 1962. Lewis, W. H., H. L. Striptinc, & R. G. Ross. Chromosome numbers for some angiosperms of the southern United States and Mexico. Rhodora 64: 147- 161. 1962. [C. carolinus, n = 39. PRAIN, D. Cocculus trilobus. Bot. Mag. 1514: pl. 8489. 1913. Rao, K. V. J.. & L. R. Row. Chemical ee of Cocculus hirsutus DC. Jour. Sci. Indus. Res. 20B(3): 125, 126. 1961. 3. Menispermum Linnaeus, Sp. Pl. 1: 340. 1753; Gen. Pl. ed. 5. 158. 1754. Leaves + coriaceous (or thin), usually + peltate (nonpeltate); blade usually 5(or more)-veined, cordate, or deltoid to reniform in outline, shal- lowly to deeply + 5—7(3-9)-lobed or pointed (or unlobed), often apically mucronulate, the petioles at least as long as the blades. Hairs 1- or 2- celled. Flowers + regular and calyculate with 1 or 2 small sepal-like bracts; insertion of perianth and stamens or staminodia sometimes + flowers: stamens (9) 12-24, exserted, filaments erect, sometimes + em- braced by the petals; anthers 4-lobed, appearing terminal, + 1-locular at anthesis, the dehiscence + apically confluent; pollen prolate, 3-colpate. 9 flowers: petals not embracing the staminodia, spatulate, somewhat larger than in 6 flowers; staminodia (4) 6-12, + filiform, variously inserted or 1-seriate and opposite the petals, the rudimentary anthers introrse; gynophore distinct, longer than broad; carpels 3; stigmas ex- trorsely dilated, + convolute; style short, reflexed. Fruits 1 or 2 per flower, black, sida nonstipitate, the style subbasal, the gynophore 2—3 mm. long; endocarp sculptured, broadly crescent shaped, laterally compressed, conspicuously notched adaxially; embryo broadly crescent 1964] ERNST, GENERA OF BERBERIDACEAE 33 shaped, the cotyledons appressed, narrow, no wider than the radicle. LECTOTYPE SPECIES: M. canadense L.; see Britton & Brown, Illus. Fl. No. U.S. ed. 2. 2: 131. 1913. (Name from Greek, mene, moon, and sperma, seed.) — MOONSEED. One species of eastern Asia and one of thickets, woods, stream banks, and hedgerows in eastern North America. Many species originally de- scribed in Menispermum now are placed in other genera. Menispermum canadense L., 2n = 52, is distributed from southeastern Manitoba to Quebec, southward to western North Carolina, northern Georgia, northwestern Florida, ?Alabama, Tennessee, Arkansas, and Okla- homa. The leaf blades tend to be peltate, rather thin, mostly with five or more veins, and the petioles at least as long as the blades. Among other distinguishing characteristics are the glabrate sepals, the exserted stamens, the relatively long gynophore, and the nonstipitate, black fruits. Plants referred to M. mexicana Rose, occurring disjunctly in Nuevo Leon, Mexico, with leaves varying from peltate to nonpeltate, probably are conspecific with M. canadense. Our species somewhat resembles Cocculus carolinus vegetatively, but the leaves of the latter are nonpeltate and tend to have only three prominent veins. In eastern China and Japan, M. dahuricum DC., 2n = 52-54, resembles our species, but the leaves seem more strongly peltate. The fruits, looking like grapes, are bitter and probably poisonous. The taxonomy of Menispermum has been complicated by lack of agree- ment on the proximity of Cocculus and Cissampelos, but is simplified to the extent that the name of the tribe and subtribe must be formed on the root Menisperm-. Miers, placing Cocculus in another tribe, divided the Menispermeae (‘“Leptogoneae”) into subtribes Menisperminae and Cis- sampelineae (see Cissampelos). Tribe Cocculeae of Hooker & Thomson and of Diels included both Menispermum and Cocculus. In addition to the differences in floral structure, the stones and embryos of Menispermum are broadly crescent shaped, while those of Cocculus appear somewhat coiled. REFERENCES: Under family references see Gray, Miers, and Morinr; under Berberidaceae see AGARDH and PAYER. Hotm, T. Medicinal plants of North America. 78. Menispermum canadense L. Merck’s Rep. 22: 281-284. 1913.* [See Bot. Centralbl. 126: 60, 61. 1914.] Myers, L. Tyloses in Menispermum. Bot. Gaz. 78: 453-457. pls. 11, 12. 1924. [M. canadense. Sims, J. Menispermum canadense. Bot. Mag. 44: pl. 1910. 1817.* 4. Cissampelos Linnaeus, Sp. Pl. 2: 1031. 1753; Gen. Fi. ed, 35, 455. 1754 Leaves + coriaceous, + broadly reniform (to ovate or suborbicular), often apically emarginate, mucronulate [sometimes peltate], usually with or more veins. Hairs 1- or 2-celled. Flowers minute, symmetrically 34 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV dimorphic; perianth pubescent [glabrous], not calyculate. ¢ flowers: regular; sepals 4, valvate; corolla sympetalous [? or of 2—4 free petals]; anthers about 4 |? to 8 or more], extrorse, included, uniseriate, sessile on a short, central, columnar androphore, + 1-locular at anthesis, the dehiscence horizontal; pollen + oblate, 3-colpate. ¢@ flowers: irregular; sepal 1; petal 1 [sometimes divided], opposite the sepal; staminodia ab- sent; gynoecium of a solitary carpel; stigma 3-parted; style short, erect; ovary pubescent [glabrous]; ovules 2, the placenta toward the perianth. Fruit red, pubescent [glabrous], nonstipitate, the style subbasal; endocarp sculptured, + orbicular, laterally compressed, minutely notched adaxially ; embryo + conduplicate, the cotyledons appressed, narrow, no wider than the radicle. Lectotype species: C. Pareira L.; see Britton & Millspaugh, Bahama Fl. 142. 1920. (Name from Greek, cissos, ivy |Hedera|, and am pelos, grape | Vitis |.) Number of species uncertain, perhaps 20 to nearly 70, widely distributed in warmer regions. Pantropical and perhaps the most widely distributed taxon of Meni- spermaceae (cf. Diels), Cissampelos Pareira, 2n = 24, is known in our area from Dade County, Florida. The plants are reported to be high climbing. The leaf blades seem to be nonpeltate, somewhat reniform, with five or more veins, and the petioles at least as long as the blades; in other areas the blades sometimes are ovate and peltate. The carpellate flowers are borne in the axils of usually broad, relatively large bracts resembling small, thin, leaves on peculiar inflorescence branches. Several flowers, each on a separate, stout, upright, unbranched peduncle, these apparently arranged in two parallel ranks, are clustered in the axil of each bract. The bracts are variable in size but often equal or exceed the flowers and, later, the pubescent, red fruits; bractlets seem to be absent. The stones are nearly orbicular, with only a minute adaxial indentation; ex- ternally, the condyle appears obovate, and the embryo is strongly folded or horseshoe shaped. The staminate flowers are borne in very finely branched inflorescences with a few bracts and bractlets of various sizes; many of the filiform pedicels are ebracteate. The number of locules in the anthers before anthesis is not apparent. Floral dimorphism is especially conspicuous in Cissampelos, with the regular, sympetalous staminate flowers having a symmetrical androphore, and the irregular carpellate flowers composed of a solitary sepal, petal, and carpel. The general shape and sculpturing of the endocarps seem relatively similar to those of Cocculus, and especially Menispermum, but the floral morphology is distinctive; consequently, the systematic position of Cissampelos seems somewhat unclear. Miers placed Cissampelos (along with four other genera) in the subtribe Cissampelineae (‘‘Cis- sampelideae”’), one of his two subtribes under Menispermeae (‘‘Lepto- goneae’’), approximating the later treatment of Diels who, however, divided the Menispermeae (‘‘Cocculeae’’) into three subtribes. Hooker & Thomson placed Cissam pelos in their tribe Cissampelideae. 1964 | ERNST, GENERA OF BERBERIDACEAE 35 REFERENCES: Under a references see BENTLEY & TRIMEN, Miers, and Mori Houm, T. Cissampelos Pareira L. Merck’s Rep. 27: 60, 61. 1918.* See Bot. Jahresb. 50: 594. 1932.] MoeELter, J. Beitrage zur vergleichenden Anatomie des Holzes. Denkschr Akad. Wiss. Wien Math. Naturw. 36: 297-426. pls. 1-6. 1876. [C. Pareira, 364, 365.] Moxerjt, B., & R. Buanpart. Cissampelos Pareira L., source of a new curari- form drug, Planta Med. 7: 250-259. 1959.* tT. Hizarre, A. DE. Cissampelos ovalifolia, C. ebracteata. Plantes usuelles des Brasiliens. pls. 34, 35. 1825. [Separately paged text with each plate. De- velopment of fruit, 3, 4, with pl. 35. For explanation of date, see Jour Bot. 42: 86. 1904.] NM ARNOLD ARBORETUM Present address: AND DEPARTMENT OF BOTANY, GRAY HERBARIUM, SMITHSONIAN INSTITUTION, HARVARD UNIVERSITY WASHINGTON, D.C. 36 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV A PROPOSED EXPLANATION FOR THE ORIGIN OF COLCHICINE-INDUCED DIPLOID MUTANTS IN SORGHUM ! Mary E. SANDERS ? AND CLIFFORD J. FRANZKE COLCHICINE, AN ALKALOID extracted from plants of Colchicum species, especially C. autumnale L., has long been used for medicinal purposes. In tissues of animals which have been exposed to the drug, mitoses are arrested with the result that some cells are produced with a doubled number of chromosomes (Eigsti & Dustin, 1955, Chap. 1). Botanists realized that such tetraploid cells in plant tissues might give rise to tetra- ploid shoots and plants, and this proved to be the case (Blakeslee, 1937; Nebel & Ruttle, 1938). In many plant species, colchicine treatment has been used sincecshilly to produce polyploids for cytogenetic research and for breeding programs (Eigsti & Dustin, 1955, Chap. 11 In the breeding program at South Dakota State College, colchicine treatment of Sorghum was found to produce, in addition to polyploids, diploid mutant plants changed for many characters. When the mutants were selfed, some produced uniform progenies and continued to breed true in succeeding generations; others agrees progenies which segre- gated for many characters (Franzke & Ross, 7 ‘Experimental 3’ (Fic. 1), which has ze rise to many colchicine- induced diploid mutants, has been studied more extensively than other lines. It was developed from crosses involving three sorghum lines: ‘Da Milo’, a grain sorghum (Sorghum vulgare Pers. var. subglabrescens (Steud.) A. F. Hill); ‘Black Amber Cane’, a sorgo (S. vulgare var. sac- charatum (L.) Boerl.); and Sudan Grass, a grass sorghum (S. sudanense (Piper) Stapf). Two crosses, ‘Day Milo’, with ‘Black Amber Cane’ and ‘Day Milo’ with Sudan Grass, were made in 1932. An early dwarf grain type was obtained from each of these crosses by repeated selfing and selection. These two types were crossed together in 1939, and from the progeny of this cross, ‘Experimental 3’ was developed by continued selfing and selection. It had been through eight generations of selfing and was true-breeding when the first mutants were obtained in 1948. After more than 20 generations of selfing, it is still apparently the same true-breeding e authors are sincerely grateful to Dr. Richard A. Howard, Director of the Arnold Arboretum, for making possible the continuation of the work, and to members of the Arboretum staff for all their assistance. We also thank those who gave their time to review manuscript: Dr. G. Lefevre, is Dr. P. C. Mangelsdorf, Dr. M. R. Morris, Mr. J. Munoz, Dr. L. I. Nevling, Jr.. Mr. P. Nordquist, Dr. R. C. Rollins, Dr. chubert, Dr. O. T. Solbrig and Dr. O. J. Webster. Their many valuable sippestions. are areatly appreciated. e research on which the paper is based was supported in part by the National Science Foundation and the National Institutes of Health. “Research Associate, Arnold Arboretum. 1964] SANDERS & FRANZKE, MUTANTS IN SORGHUM 37 line and continues to give rise to colchicine-induced diploid mutants some of which resemble one of the original parents. For example, there are mutants with slender stems, narrow leaves and open panicles (Class I, Franzke & Sanders, in press) similar to Sudan Grass. Except where other- wise specified, the following discussion relates to sorghum line ‘Experi- mental 3’ and its derivatives. This paper brings together the results of research on colchicine-induced diploid mutants in Sorghum and, in the light of available information, proposes a new hypothesis to explain their origin. The significance of the hypothesis, if proved valid, is also considered. COLCHICINE TREATMENT OF SORGHUM LINE ‘EXPERIMENTAL 3’ Sorghum plants are treated with colchicine at the seedling stage. Shortly after germination, a mixture of 0.5 per cent colchicine in lanolin, heated to the melting point, is applied to cover the coleoptile. Treated seedlings have been planted in various media and kept under various con- ditions in greenhouse, laboratory, or rooms with controlled light and temperature. Both polyploids and diploid mutants have been obtained in the treated generation. There are indications that environmental con- ditions may influence the outcome of treatment. Polyploids Experimentally produced polyploids of ‘Experimental 3’ are, for the most part, tetraploid plants which exhibit the same qualitative characters as diploid plants, but are shorter and stockier with broader stems, heavier leaves, larger glumes and grain, and have a high degree of sterility. These characteristics are similar to those of polyploids in many other plant species. Self-progenies of these polyploids frequently include both diploid and polyploid plants, indicating that the treated plants were chimeras. The incidence of induction of polyploids, which has been as high as four out of 15 surviving treated seedlings, appears to be increased by conditions which favor the growth of treated seedlings (Franzke et al., 1960). Diploid Mutants Diploid mutants exhibit qualitative and quantitative characters which are radically different from those of ‘Experimental 3’ plants. Yet, the changes have been brought about by colchicine after the germination of an ‘Experimental 3’ seed. Uniform self-progenies of two diploid mutants are shown in Fics. 2 and 3; each plant repeats the characters of the original mutant. A segregating self-progeny of a nontrue-breeding mutant is shown in Frc. 4. In progenies of this kind, most characters may segregate, or some may segregate while others do not (Sanders et al., 1962). Thus, colchicine- induced diploid mutants may be true-breeding, nontrue-breeding, or intermediate between the two conditions. The incidence of induction of diploid mutants, especially true-breeding ones, which has been as high as [ VOL. XLV JOURNAL OF THE ARNOLD ARBORETUM ee REY SAN SE? ee % ¢ #2 EYES 1964 | SANDERS & FRANZKE, MUTANTS IN SORGHUM 39 four out of nine surviving treated seedlings, appears to be increased by conditions which retard the growth of treated seedlings (Franzke et al., 1960). The largest proportion of mutants among surviving treated plants was obtained under conditions which almost eliminated survival. Main- taining treated seedlings at a lower temperature (68° F.) increased survival while still permitting induction of mutants (Sanders et al., 1960). Some types of true-breeding mutants were found to recur thereby forming classes of almost identical mutant lines (Franzke & Sanders, in press). Some classes included as many as ten lines, most of them obtained within a five-year period of experimentation. Three lines from one class were intercrossed in all possible combinations and were each crossed to ‘Experimental 3’. The results indicated that these three mutant lines were genotypically as well as phenotypically alike (Chen et al., 1961). The original hypothesis proposed for the origin of the diploid mutants was “that such variant plants could originate through reductional grouping of the somatic chromosomes so that a concentration of chromosomes con- taining gene blocks originating from one of the ancestors of the polyploid species might occur in one cell’ (Franzke & Ross, 1952). When no irregularities were detected in chromosome-pairing relation- ships at pachytene in mutants or in F, plants from crosses between mutants and other lines including parental ones, the original proposal was discarded. It was then concluded “that no detectable rearrangement of chromatin occurred,” and “that the colchicine-induced variants have resulted from gene mutation or cryptic structural changes in the chro- matin” (Harpstead et al., 1954). Genetic studies of one of the mutants led to an estimate that “at least 12 (in all probability more) immediate mutations for the characters studied” had occurred (Foster et al., 1961). Current information, however, suggests that the origin of the mutants is more similar to the first proposal since the variations in them resemble major chromosome changes more than individual gene changes, despite the lack of confirming cytological evidence. PRESENT HYPOTHESIS FOR THE ORIGIN OF DIPLOID MUTANTS It is now proposed that colchicine-induced diploid mutants arise from the substitution of chromosomes of similar phylogenetic origin (analogous chromosomes), and that the substitutions have not been detected cyto- logically because there is a tendency for bivalent rather than multivalent s. 1-4. Progeny rows of sorghum line ‘Experimental 3’ and of three diploid mutants obtained from ‘Experimental 3’ after colchicine treatment. The yardstick indicates height. Fic. 1. ‘Experimental 3’. Fic. 2. Ir mutant. Class IV (Franzke & Sanders, in press). Fic. 3. True-breeding mutant. Class VB (Franzke & Sanders, in press). Fic. 4. Nontrue-breeding mutant; leat width, head type, height, and maturity can be seen to segregate. Height dif- ferences are indicated by the bags which cover the main heads. Maturity differences are indicated by late unbagged plants, i.e., the first in the row. and early plants with well developed tillers which account for many of the unbagged 40 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV pairing to occur in sorghums with 2m = 20, and because pairing may occur between analogous chromosomes. Information from the literature, as well as research on the mutants, lends support to this view. Major Chromosome Changes The phenotypic changes in the colchicine-induced diploid mutants appear to come about through major chromosome rearrangements since a large number of qualitative and quantitative characters are involved. Moreover, they are similar to natural variations found in sorghum varieties, and they include types which recur. Similarly, complex mutants in other species have been shown to result from the addition or loss of entire chromosomes, i.e., in Datura stramonium L. (Blakeslee & Belling, 1924) and in Evcopersicon esculentum Mill. (Rick & Barton, 1954). In these species, the same complex mutants recurred and were shown to result from the addition or loss of the same chromosomes, either original chromosomes as in primary trisomics or monosomics, or isochromosomes as in secondary trisomics. Vasek (1956) working with Clarkia unguiculata Lindl., a highly variable outcrossed species, concluded that aneuploid types probably have characteristic phenotypes only in highly homozygous diploid species. Otherwise effects of aberrant chromosomes may be either obscured by variability among the diploid plants, as in Zea Mays L. (McClintock, 1929), or overshadowed by the effects of duplicated genes in various genomes, as in hexaploid Triticum aestivum L. (Sears, 1944, as T. vulgare Vill.). Plants with extra chromosomes have been found in Sorghum (Endrizzi & Morgan, 1955; Hadley & Mahan, 1956; Price & Ross, 1957), but not in sufficient numbers in the same true-breeding line to establish whether or not the presence of a particular extra chromosome produces a particular phenotype. Since ‘Experimental 3’ is true-breeding and has a chromosome number of 2m = 20, chromosome mutants in this line might be expected to be phenotypically distinct. If the distinct mutant complexes depend on the distribution of whole chromosomes, they would be likely to recur since the haploid chromosome number in S. vulgare, ten, is relatively small. Substitution of Whole Chromosomes The mechanism for chromosome changes would have to be substitution since there is no change in chromosome number in the mutants. However, since substitution has not been detected cytologically, there would have to be additional factors which prevent the expected changes in chromosome pairing relationships. More detailed studies may reveal direct cytological evidence for chromosome substitution. Irregularities in the structure of some bivalents at metaphase in a nontrue-breeding mutant (Fics. 5 and 6) were originally disregarded as possible artifacts. However, their repe- tition from cell to cell suggests that the ten bivalents in the mutant may not be the same ten bivalents as in ‘Experimental 3’. It should be possible ‘o interpret such apparent differences by studying earlier meiotic stages. 1964 | SANDERS & FRANZKE, MUTANTS IN SORGHUM 41 . 5, 6. Sorghum chromosomes at metaphase I in pollen mother cells. Fic. Beocmmental 3’. Fic. 6. Nontrue-breeding mutant G (Sanders et al., 1962). Celis of the mutant have two relatively large aie bivalents in addition to a large bivalent are to the one in ‘Experimental 3 Three Requisite Assumptions ANALOGOUS CHROMOSOMES. The assumption that the haploid number of ten in Sorghum vulgare includes analogous chromosomes which can be substituted one for another without leading to inviability of the plant is basic. This assumption seems warranted as species of Sorghum with a haploid number of five are known, although not in the subgenus Sorghum * which includes S. vulgare. Several investigators have concluded that the 20-chromosome diploid sorghums are of polyploid origin. Huskins and Smith (1934) based their conclusion primarily on the sporadic multivalents found in all species with 2m = 20 studied, and on the frequent formation of multivalents in an asynaptic form of ‘Dakota Amber Sorgo’ in which normal pairing relationships were disrupted. Garber (1950) and Celarier (1958a), after examining the cytotaxonomic characteristics of the species of Sorghum, concluded that the ‘‘Eu-sorghums” are allopolyploids, i.e., S. vulgare with 2n = 20 is an allotetraploid. Damon (1962) also considered that culti- vated sorghums with 2m = 20 are allotetraploids and suggested that if they have two genomes with slight homologies between corresponding chromosome pairs, the “secondary associations” frequently reported for Sorghum chromosomes could account for the inheritance of blocks of This subgenus should be called Sorghum rather Ree Eu-sorghum since it includes the type species (Art. 22, Int. Code Bot. Nomencl. 1°61). 42 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV morphological characters frequently found following hybridization. He further pointed out that although Venkateswarlu and Reddi (1956) in their study of pachytene chromosomes of Sorghum vulgare var. subglabrescens (as S. subglabrescens (Steud.) Schwein. & Aschers.) identified ten distinct chromosome types, it seems feasible from their figures to arrange them as five pairs, the members of each pair differing primarily in arm length. A study of the paper itself reveals that, disregarding arm length, the chromo- somes can be matched fairly well on the basis of the morphology of the heavily staining regions on either side of each centromere. The chromo- somes have been numbered from 1 to 10 in order of decreasing length. The most probable pairs would appear to be: 1 and 3, 4 and 5, 2 and 9, 6 and 8, 7 and 10. Allotetraploids are generally described in terms of separate genomes, designated A and B, in which corresponding chromosomes (1A, 1B; 2A, 2B; etc.) are those of similar phylogenetic derivation, hence those with the greatest degree of homology. Huskins (1931) suggested that such chromosomes “‘might be referred to as homoeologous chromosomes, signi- fying similarity but not identity” as in the case of homologous. Because of the confusion occasioned by the similarity of the two words, we have discarded homoeologous in favor of analogous to designate phylogenetically similar chromosomes. The above indications of analogies among the ten haploid chromosomes of species of the subg. Sorghum are compatible with cytological studies which have been made on haploid plants of S. vulgare, including a haploid of line ‘Experimental 3’. All investigators agree that occasional bivalents are found at late diakinesis and metaphase I of meiosis in pollen mother cells (Brown, 1943; Kidd, 1952; Endrizzi & Morgan, 1955: Sanders & Franzke, 1963). Rare associations of three or four have also been reported (Kidd, 1952; Sanders & Franzke, 1963). Endrizzi and Morgan (1955) considered that a reciprocal translocation and other abnormalities in progeny plants obtained by pollinating a haploid with a diploid of the same line resulted from crossing over in duplicate segments of partially homologous chromosomes, thereby providing evidence for the reality of pairing observed between chromosomes of the haploid. Preliminary observations of pachytene chromosomes of sorghum hap- loids suggest that there are more paired than unpaired regions (Brown, 1943; Sanders & Franzke, 1963), and in haploid ‘Experimental 3’, cells at early diakinesis generally showed a multiple association including 5.5 chromosomes on the average. In some cells, all ten chromosomes were associated in a single complex in which most chromosomes were associated with at least two other chromosomes, some with more. The configuration seen in an individual cell appeared to depend on where separations had occurred in the complex. Cells at mid-diakinesis showed an average of 4.3 univalents and 2.1 bivalents with the remainder still in multiple associa- tions. The large number of bivalents at this stage, in some cells four and very rarely five, suggests that some of the associations in the complex are more persistent than others and may indicate analogous chromosomes. The 1964] SANDERS & FRANZKE, MUTANTS IN SORGHUM 43 occasional associations seen at late diakinesis and metaphase I may result from crossovers between like segments in analogous chromosomes, and the univalents may result from lack of crossovers due to the complexity of pairing relationships rather than from lack of pairing (Sanders & Franzke, 1963). These relationships between the chromosomes of an ‘Experimental 3’ haploid suggest that analogous chromosomes may differ by more than one translocation and agree with the interpretation suggested above for the work of Venkateswarlu and Reddi. PAIRING OF ANALOGUES AND BIVALENT PAIRING MECHANISM. Analogous chromosomes in Sorghum vulgare would have to be sufficiently alike to pair under certain circumstances, and control of chromosome pairing would have to include a mechanism which promotes bivalent and limits multi- valent pairing at the diploid level. These two assumptions are considered jointly because of interrelationships between them. Pairing of analogues is referred to by Waddington (1939, Chap. 2) and Stebbins (1947) as heterogenetic association. It is contrasted with homogenetic association (pairing of homologous chromosomes from the same genome), and is defined by Waddington as pairing between “homologous” chromosomes from different genomes when there is a considerable difference between them. The most direct evidence for pairing of analogous chromosomes in Sorghum is furnished by the data on haploid plants cited above. Evidence that normal pairing is under genetic control is furnished by cases in which gene mutations have resulted in abnormalities. A spontaneous asynaptic mutation in ‘Experimental 3’ appeared to be a Mendelian recessive in relation to the normal (Ross e¢ a/., 1960), and Huskins and Smith (1934) found an asynaptic strain of ‘Dakota Amber Sorgo’. Examination of pair- ing relationships in sorghums with different chromosome complements gives some indication that a bivalent pairing mechanism may operate at certain chromosome levels in Sorghum and suppress formation of multi- valents in some instances where they would otherwise be expected. The degree of pairing in Sorghum would appear to depend on some sort of genetic balance which shifts with the genomes present. No evidence has appeared for localized control as in hexaploid Triticum aestivum where the presence of the long arm of chromosome V maintains bivalent pairing apparently without regard for the rest of the genotype (Riley ef al., 1960). However, there are correspondences between certain wheat plants lacking chromosome V and certain sorghum plants. In both species, pairing be- tween analogous chromosomes occurs in haploids (Riley ef al., 1960; Sanders & Franzke, 1963) and in hybrids between species with different numbers of genomes, i.e., T. aestivum (2n = 42) x Secale cereale L. (2n = 14) (Riley et al., 1960), and Sorghum vulgare (2n = 20) X S. hale- pense (L.) Pers. (2n = 40) (Hadley, 1953; Endrizzi, 1957). Some multi- valent pairing occurs in S. halepense, but not enough to account for the degree of multivalent pairing in the hybrid (Celarier, 1958b). Addition of 44 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV chromosome V to the wheat plants eliminates analogous pairing, but con- trolling factors in Sorghum are not known. Cytological studies of extra-chromosomal types at the diploid level in Sorghum vulgare have shown that the extra chromosomes occur more frequently as univalents than in multivalent configurations. For a trisomic derivative from a haploid, Endrizzi and Morgan (1955) reported 53 pollen mother cells with ten bivalents and a univalent, 19 with nine bivalents and a trivalent, and one with nine bivalents and three univalents. Extra-chromosomal types studied by Price and Ross (1957) were derived from an unknown triploid which they believed to be S. vulgare rather than a species hybrid. In 17 singly trisomic plants, ten bivalents and a univalent were found more frequently than nine bivalents and a trivalent at diakinesis and metaphase I; in three plants with two extra chromosomes, ten bivalents and two univalents were usually found; and in a fourth plant with two extra chromosomes, eleven bivalents were usually found. Studies of multiple trisomics with from one to seven extra chromosomes in Clarkia unguiculata (2n = 18) showed an increase in trivalent forma- tion with an increase in extra chromosomes. Trivalents per cell per extra chromosome were 0.43 for 2n + 1, 0.46 for 2m + 3, 0.56 for 2n + 4, and 0.61 for 3m (Vasek, 1963). These results suggest that multivalent pair- ing may be influenced by a genetic balance which, in this case, shifts trivalent formation toward the frequency found in triploids when four or more extra chromosomes are present. Similar studies are needed in Sorghum. Sorghums with a complement of 30 chromosomes have shown a high degree of trivalent formation whether 3n S. vulgare (Kidd, 1952), an unknown triploid (Price & Ross, 1957), or a hybrid between S. vulgare and either S. halepense or S. almum Parodi (2n = 40) (Hadley, 1953; Endrizzi, 1957). The number of trivalents per cell was highest for S. vulgare and S. vulgare * S. almum (from 8.2 to 8.5), and lowest for S. vulgare X S. halepense (from 5.7 to 5.8). Trivalent formation was con- sistently higher than the sum of trivalents and quadrivalents in cor- responding 40-chromosome plants. Multivalent formation appears to be distinctly favored at the 30-chromosome level. Cytological studies on tetraploids of Sorghum vulgare have shown a majority of the chromosomes associated as bivalents. In tetraploid S, vulgare var. caffrorum (Retz.) Hubb. & Rehd. (as var. hegari without authority, Chin, 1946), only three quadrivalents occurred per cell on the average and trivalents and univalents were rare. (Even the octoploid of the same variety had almost half the chromosomes in bivalent configura- tions.) In a tetraploid ‘Experimental 3’ plant, means of 3.3 quadrivalents, 0.04 trivalents, 12.8 bivalents and 1.2 univalents per cell were found (Sanders & Franzke, 1962b). Although multivalents are present uni- formly in sorghum tetraploids, their formation is distinctly lower than in triploids, in spite of the increase in the number of homologous chromosomes. These cytological data suggest several generalities concerning pairing relationships in Sorghum which need to be investigated. (1) Pairing of 1964] SANDERS & FRANZKE, MUTANTS IN SORGHUM 45 analogous chromosomes may be limited chiefly to cases where a homologue is missing. Such a phenomenon could be designated as preferential pairing, the term used by Endrizzi (1957) to describe pairing relationships in S. halepense. (2) The tendency for bivalent pairing may be positively correlated with the presence of an even number of genomes, i.e., diploids and tetraploids. Configurations in S. vulgare haploids would have little bearing on bivalent or multivalent pairing as multiple associations appear to result from translocation differences between analogues. However, in triploids, trivalents are the rule. These relationships suggest that the bivalent pairing mechanism may be associated with genetic balance in some way. (3) Chromosome associations still present at diakinesis and metaphase I probably indicate only paired regions where crossovers oc- curred. Actual pairing relationships need to be studied at pachytene and diplotene. Patterns of Chromosome Substitution Based on the above assumptions, different patterns of chromosome substitution may be considered, and their results predicted. INDIVIDUAL CHROMOSOMES. If a single chromosome were lost and replaced by its analogue, the chromosome number would be unchanged, and, with analogous pairing and a tendency for bivalent formation, ten bivalents would be seen at diakinesis and metaphase I. Such a plant could be either phenotypically mutant or unchanged depending on the shift in genetic balance effected by the chromosomes involved. The self-progeny would segregate to a greater or lesser degree in direct proportion to the amount of crossing over which occurred between the paired analogous chromosomes. If more than one individual chromosome were lost and replaced by its analogue, the chances of the plant being phenotypically mutant would be increased, and the degree of segregation in the self- progeny would be increased. Such plants would be diploid nontrue- breeding mutants and would segregate for many characters. PAIRS OF CHROMOSOMES. If a pair of homologous chromosomes were lost and replaced by their analogues, the chromosome number would be unchanged, and ten bivalents would be seen at diakinesis and metaphase I. Such a plant would be phenotypically mutant and would produce a uniform self-progeny of the same mutant type. Replacement of more than one pair of homologous chromosomes by their analogues would increase the number of such types which could be obtained. Such plants would be diploid true-breeding mutants changed for fixed complexes of characters. INDIVIDUAL AND PAIRED CHROMOSOMES. If both single chromosomes and pairs of homologous chromosomes were lost and replaced by their respec- tive analogues, the chromosome number would be unchanged, and ten bivalents would be seen at diakinesis and metaphase I. Such a plant would 46 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV be phenotypically mutant and would produce a self-progeny which would be uniform for some characters present in the parent mutant and segregate for others. Substitution of different numbers of individual chromosomes and pairs of chromosomes would produce intermediate diploid mutants ranging from those which would produce self-progenies primarily segre- gating, to those which would produce self-progenies uniform for many mutant characters. EVALUATION OF THE HYPOTHESIS Relevant Effects of Colchicine That chromosomes may be lost and gained other than as entire genomes in cells of colchicine-treated tissues is indicated by the aneuploid types which have been obtained following colchicine treatment. In addition to polyploid types (some with aneuploid numbers) obtained from colchicine- treated Datura seeds, Bergner, Avery, and Blakeslee (1940) found 2n — 1 plants at about 70 times the rate, and 2m + 1 plants at about three times the rate that they were obtained from untreated seeds. Smith (1943) obtained aneuploids following colchicine treatment of two Nicotiana species and their F; hybrids. Both at diploid and polyploid levels, more of these plants resulted from chromosome deficiencies than from additions. One plant with the actual tetraploid number of 36 was off-type and appeared to have lost some chromosomes and gained others. The effects of colchicine on dividing cells can account for such phenomena. ARRESTED MITOSES. The primary observable cytological effects of colchicine result from the inactivation of the spindle mechanism in dividing cells (Levan, 1938; Eigsti & Dustin, 1955, Chap. 2). Mitoses are inter- rupted at metaphase, and, after a delay, chromosome divisions may occur without nuclear or cell division so that doubled or polyploid numbers of chromosomes may be built up within a single nucleus and cell. During recovery, processes which take place in dividing cells frequently resume more or less out-of-phase with one another and result in abnormalities other than polyploid nuclei. Arrested metaphases assume various patterns from a single clump with centromeres congregated at the center and chromosome arms extended, to an “exploded” type with chromosomes scattered throughout the cell, Multiple clumps and multipolar spindles are common also so that, when followed by subdivision of the cell, polyploid chromosome groups may be reduced to smaller ones (Eigsti & Dustin, 1955, Chap. 2). Delay in the division of the centromere accompanies inactivation of the spindle and results in characteristic “c-pairs” in which the daughter chromosomes remain attached only at the centromeres (Levan, 1938). After division of the centromeres, daughter chromosomes continue to lie parallel to one another and may remain so through several chromosome divisions so that an accumulation of the same chromosome is built up in 1964 | SANDERS & FRANZKE, MUTANTS IN SORGHUM 47 one position. Division of the centromeres is not simultaneous for all the “c-pairs” in one nucleus. There appears to be ample opportunity for two daughter chromosomes to act either as a unit or as two individuals during the irregularities of ‘‘c-mitoses” thereby making substitution of chromo- somes by homologous pairs as plausible as by single chromosomes. True- breeding, nontrue-breeding and intermediate mutants would all be expected. The suggestion of Bergner et al. (1940), that colchicine-induced delay of mitosis may exaggerate irregularities such as lagging and nondisjunction which in nature occur occasionally, should also be considered. REDUCTION OF CHROMOSOMES IN SOMATIC TISSUES. As indicated above, multiple clumps and multipolar spindles in cells undergoing ‘‘c-mitoses” provide mechanisms whereby chromosome numbers may be reduced as well as duplicated during colchicine treatment. Reduction of chromosomes in somatic tissues and its concomitant effects need further investigation. Isolated reports of chromosome reduction divisions and of haploid cells in somatic tissues led to a study of the phenomenon by Huskins (1948) and his associates. Colchicine was found to induce reductional groupings of chromosomes in root tips of Trades- cantia and Allium (Allen et al., 1950), and of Sorghum including line ‘Experimental 3’ (Atkinson et al., 1957). The concept of somatic reduction has been part of explanations offered for the origin of colchicine-induced mutants from the beginning (Franzke & Ross, 1952). Although haploid plants have been obtained from colchicine-treated diploid seedlings, they have been extremely rare, if anything less frequent than from untreated seedlings. In one progeny-test plot, haploids occurred at the rate of ap- proximately one per 1156 diploid plants. They were about seven times more frequent in progenies of untreated than of treated parents; the latter progenies were untreated plants a generation removed from colchicine treatment. On the other hand, following colchicine treatment of tetraploid ‘Experimental 3’ seedlings, four out of nine surviving plants were diploid and mutant (Sanders & Franzke, 1962b). A fifth plant was mixoploid and mutant. Phenotypic examination of 42 untreated plants from the same seed source and cytological examination of 37 of these indicated that all were still tetraploid and apparently unchanged. The diploid mutants obtained from tetraploids demonstrate that cells derived from reduction divisions in colchicine-treated tissues may survive and function in the production of new shoot apices. Reduction appears to take place more readily following treatment of tetraploids than of diploids, suggesting that reduction may occur more frequently in polyploid cells than at lower levels. It is suggested that duplication provides the extra chromosomes necessary for substitution and, perhaps, an unbalance which favors reduction in somatic cells, and that reduction effects the return to the diploid level and the simultaneous redistribution of the chromosomes. Both loss and gain of chromosomes are consistent with these known effects of colchicine. 48 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLV Substitution of analogous chromosomes in colchicine-induced mutants in Sorghum could be attributed to selection among the cells of the treated growing point which allows only viable and preferably vigorous chromo- some combinations to continue to reproduce themselves and organize a meristem capable of giving rise to a mutant plant. Observations of Levan (1938) are in accord with such an assumption. He noted, in root tips of Allium, that diploid cells are favored over polyploid cells during recovery after colchicine treatment and are usually more numerous near the apex. Relevant Information on Colchicine-induced Mutants The present explanation and the information that has been assembled on colchicine-induced mutants are generally in agreement. Several cases, however, remain controversial. SIMILARITIES BETWEEN TRUE-BREEDING AND NONTRUE-BREEDING MU- TANTS, There is no distinct separation between true-breeding and nontrue- breeding mutants since intermediate types appear which are true-breeding for some characters but not for others. All types may arise under the same conditions and in conjunction with one another (Sanders e¢ al., 1962). Such similarities suggest that all the diploid colchicine-induced mutants arise as variations of a single phenomenon. The present explanation ac- counts for all of the mutants as originating from different patterns of chromosome substitution. “REACTOR” AND “NONREACTOR” LINES. That sorghum lines differ in their response to colchicine has been demonstrated in a comparison be- tween the two lines ‘Experimental 3’ and ‘Norghum’ (Atkinson e¢ al., 1957). Lines such as ‘Experimental 3’ that produce a relatively high pro- portion of obvious diploid mutants following treatment have been termed “reactors,” whereas those which produce only polyploids and minor varia- tions have been termed “nonreactors.” According to the present explana- tion, a line could give rise to such mutants if its chromosome complement included analogous pairs which could be substituted without impairing plant viability, and if the analogues were sufficiently different that sub- stitution of one or both members of the pair resulted in obvious pheno- typic changes. Such lines would have the fixed heterozygosity postulated by Huskins and Smith (1934) for sorghums with 2m = 20 and would be “reactors.” On the other hand, if a line had analogous pairs which could not be substituted without impairing plant viability, or if the analogues were not sufficiently different for substitution to result in obvious pheno- typic changes, or if four homologues were present instead of two pairs of analogues in the five basic chromosome units, a line could not give rise to composite mutants by chromosome substitution and would be a “non- reactor.” The many different types of sorghums have been so intermingled by hybridization that possible combinations of chromosomes from different sources would appear to be exceedingly large. Further, if pairing of 1964 | SANDERS & FRANZKE, MUTANTS IN SORGHUM 49 analogues occurs in certain hybrids and is accompanied by crossing over, new chromosomes unlike any of the originals would have been formed. MUTANT INTERRELATIONSHIPS. Following colchicine treatment of one of the three Class I mutant lines shown to be genotypically as well as phenotypically alike (Chen e¢ a/., 1961), four mutants were obtained which belong to three other classes of true-breeding mutants derived directly from ‘Experimental 3’ (Franzke & Sanders, in press). Also two true- breeding plants of the same type in an F, self-progeny from a nontrue- breeding mutant belong to a class derived directly from ‘Experimental 3’. These results are consistent with the present explanation since a true- breeding or nontrue-breeding mutant derived directly from ‘Experimental 3’ could give rise to another true-breeding mutant type if chromosome substitutions which had taken place in the formation of the former did not preclude the substitutions necessary for the formation of the latter. If crossing over occurred between paired analogues in a nontrue-breeding mutant, chromosomes unlike any of the originals would be formed, and the probability of the self-progeny including types also derived directly from ‘Experimental 3’ would be decreased. SEGREGATING PROGENIES FROM PHENOTYPICALLY UNCHANGED PLANTS. All true-breeding mutants and the majority of nontrue-breeding mutants have been recognized by phenotypic changes in the treated plants. How- ever, some nontrue-breeding mutants have not been recognized until their self-progenies were grown. In some of these progenies there has appeared to be less segregation than in segregating progenies of recognized mutants (Sanders et al., 1959), but this has not always been the case, i.e., mutant E (Sanders et ae 1962). According to the present explanation, apparently unchanged treated plants could produce segregating self-progenies if only substitutions of individual chromosomes had occurred so that at least one of each of the ten haploid ‘Experimental 3’ chromosomes were still present, and if none of the substitutions shifted the gene balance sufficiently to be detected phenotypically. APPARENT MUTATIONS OF SINGLE LOCI FROM DOMINANT TO RECESSIVE ALLELES. Genetic analyses of both true-breeding and nontrue-breeding mutants have provided instances of apparent single-gene changes, homo- zygous in the former and heterozygous in the latter (Foster et al., 1961; Sanders et al., 1962). Mendelian ratios for several qualitative characters were obtained in Fy, populations from crosses between true-breeding mutants and the parent line ‘Experimental 3’, and in first generation self- progenies of nontrue-breeding mutants. Except where the mutants were intermediate, having both nontrue-breeding and true-breeding mutant characters, nontrue-breeding mutants behaved genetically much like F, hybrids between true-breeding mutants and ‘Experimental 3’. According to the present explanation, strictly nontrue-breeding mutants would be equivalent to such F, plants. 50 JOURNAL OF THE ARNOLD ARBORETUM | VOL, XLV If ‘Experimental 3’ is a fixed heterozygote for many genes as the result of an allopolyploid origin at the same time it is a true-breeding line, then the dominant alleles for these genes would be present on both homologues of one chromosome pair and the recessive alleles would be present on both members of the analogous pair. If the pair carrying the dominant alleles were replaced by an additional pair carrying the recessive alleles, the recessive phenotype would appear in the mutant and all its progeny. If such a mutant were crossed with ‘Experimental 3’, the F; would contain three homologous chromosomes carrying the recessive allele and one analogue with the dominant allele, and would exhibit the dominant pheno- type. With analogous and bivalent pairing, a 3:1 ratio for dominant: recessive phenotype would be obtained in the F, population. If an individual chromosome of ‘Experimental 3’ carrying a dominant allele were replaced by an additional chromosome carrying the recessive allele, the nontrue-breeding mutant would be genetically equivalent to the above F, plant for that particular unit of four chromosomes. It would seem that apparent mutations from dominant to recessive alleles may result from chromosome substitution rather than from point mutation. COMPLEX DIPLOID MUTANTS NOT INDUCED BY COLCHICINE. Diploid mutants have arisen without colchicine treatment in self-progenies both of unstable ‘Experimental 3’ polyploids (Franzke et al., 1962) and of ‘Experimental 3’ plants homozygous for an asynaptic gene (unpublished). These findings are in agreement with the present explanation since mutants should result from any interruptions of normal chromosome behavior which could result in new combinations. Multivalent pairing in a poly- ploid and lack of normal pairing in an asynaptic plant would result in gametes containing chromosome groups other than the normal complement. Diploid mutants could result from viable combinations either in 2n egg cells which developed parthenogenetically, or in haploid egg cells or pollen grains which effected fertilization. As would be expected, the three mutants from the asynaptic line were nontrue-breeding. The two diploid mutants found in polyploid self-progenies were true-breeding. Yet progeny plants in both groups showed an excess of univalents during meiosis I. CHROMOSOME IRREGULARITIES IN NONTRUE-BREEDING MUTANT PROG- entEs. Although irregularities of chromosome behavior have not thus far been found in true-breeding or nontrue-breeding mutants obtained from colchicine-treated diploid seedlings, some abnormalities were found in plants from nontrue-breeding mutant progenies (Sanders & Franzke, 1962a). The principal abnormality was an increase over untreated plants in the number of univalents at diakinesis and metaphase I in pollen mother cells. Rare trivalents and quadrivalents were also found, but since several investigators have reported occasional multivalents in various diploid sorghum lines (Huskins & Smith, 1934; Chin, 1946; Hadley, 1953), these cannot necessarily be attributed to the mutant condition. The increase in univalents in these plants is in agreement with the present explanation 1964 | SANDERS & FRANZKE, MUTANTS IN SORGHUM 51 since the pairing of analogous chromosomes in nontrue-breeding mutants would be expected to result in some chromosome irregularities in the progeny plants, especially if crossing over had occurred and interchanged segments of the analogues. Irregularities in pairing relationships should be sought in nontrue-breeding mutants, F; hybrids between true-breeding mutants and their parent lines, and plants from nontrue-breeding self- progenies and from F, populations. Crossovers between analogues would be equivalent to translocations between these chromosomes, but the usual translocation configurations would not be expected if there is a bivalent pairing mechanism. APPARENT MUTATIONS FROM RECESSIVE TO DOMINANT ALLELES. ‘Experi- mental 3’ exhibits the dominant phenotypes of the three observed charac- ters which might be attributed to single-gene differences. These are red as opposed to green seedling-base, dry as opposed to juicy stalk and midrib, and awnless as opposed to awned spikelets (Quinby & Martin, 1954). A Class I mutant line which exhibits the recessive phenotypes for all three characters, when treated with colchicine, gave rise to mutant lines belonging to three other classes which included at least one apparent mutation from recessive to dominant for each of the characters (Franzke & Sanders, in press). According to the present explanation, the Class I mutant would have resulted from the replacement of pairs of homologous chromosomes carrying the dominant alleles by their analogues carrying the recessive alleles. No further substitution would be possible in these units of four chromosomes since they would now consist of four homo- logues, and a source for the return to the dominant phenotype is not evident. However, sorghum genetics is not sufficiently understood to make a reliable interpretation of genotype on the basis of phenotypic observation alone. Similar phenotypes in sorghums have been found to result from different genotypes, alleles of certain genes have been found to have dif- ferent phenotypes in different genetic constitutions, and gene interaction has been found to be common (Celarier, 1958a). Class I mutants have been analyzed genetically in relation to ‘Experimental 3’ and to each other (Chen et al., 1961). It would be necessary to analyze mutant lines from the classes obtained both from Class I and from ‘Experimental 3’ in relation to both parental lines before any conclusions could be drawn. In other species, characters identified with particular genes have been found to change phenotypically in chromosome mutants without the chromosome carrying the known gene being involved, i.e., in Datura (Blakeslee, 1922) and in tomato (Lesley, 1928). Rick and Patton (1954) noted little or no relationship between the phenotype of a trisomic in tomato and the genes known to be on the extra chromosome. Such results indicate the complexity of phenotypic effects brought about by chromosome changes and suggest that apparent changes from recessive to dominant phenotypes might come about through chromosome substitution. 52 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV CHROMOSOME PAIRING RELATIONSHIPS IN COLCHICINE-INDUCED DIPLOID MUTANTS. Because of the presence of ten bivalent chromosomes at diakinesis and metaphase I of meiosis in pollen mother cells of colchicine-induced diploid mutants and their F, hybrids with other sorghum lines, the present explanation depends on analogous pairing and on a mechanism for bivalent pairing in diploid sorghums with 20 chromosomes. As cited above, in- stances of analogous pairing in Sorghum are known. The bivalent pairing mechanism provides the greatest difficulty since, as presented earlier, multivalents are found in Sorghum: in tetraploids, especially in triploids, sometimes in trisomics at the diploid level although a bivalent and a uni- valent are more usual, and very occasionally in straight diploids where they would seem to involve analogous pairing. It is on the basis of the rarity of multivalents in diploids and their relative infrequency in trisomics that it is proposed that bivalent pairing may be the rule and multivalent pairing the exception at the 20-chromosome level. Following a polyploid origin, the sorghums with 2n = 20 may have undergone “diploidization” which included the establishment of bivalent pairing even though the analogous chromosomes retain the ability to pair with one another, especially when uneven numbers of homologues are present. However, a clearcut understanding of the control of bivalent pairing, such as that worked out for hexaploid wheat, is needed. IMPLICATIONS OF THE HYPOTHESIS FOR THEORETICAL AND APPLIED RESEARCH If colchicine-induced diploid mutants in Sorghum result from substi- tution of analogous chromosomes, it should be possible to obtain the same results with other plants of allopolyploid derivation in which analogous chromosomes are sufficiently similar that they can be interchanged without an adverse effect on plant viability. If such mutants were to be easily recognizable, it would be necessary that the analogous chromosomes carry different alleles which would result in distinct phenotypes, and that the original line be true-breeding and preferably not too highly polyploid so that changes would not be obscured by the variation in the original popu- lation, or covered by the multiple effects of alleles of the same genes on other chromosomes, In Linum usitatissimum L., colchicine treatment of seedlings heterozygous for known genes resulted in one plant with branches homozygous for the gene markers (Dirks et al., 1956). Homologous chromosomes carrying different alleles may have been substituted for one another. Since this species (2n = 30) is thought to be of polyploid and perhaps of hybrid origin (Ray, 1944), it might also be expected to produce diploid mutants following treatment of seedlings of true-breeding lines. Meanwhile, the results suggest that, in heterozygous plants not necessarily of allopolyploid origin, it may be possible to induce substitution of homol- ogous chromosomes for one another by colchicine treatment, and hence to obtain homozygosity from heterozygosity without generations of inbreeding. 1964 | SANDERS & FRANZKE, MUTANTS IN SORGHUM 53 As stated earlier, there is some evidence that diploid mutants are obtained under conditions of colchicine treatment more stringent than those required to produce polyploids. Success in obtaining mutants de- pends on the skill of the researcher in insuring survival of treated seedlings under adverse circumstances. Environmental factors favorable for the production of mutants probably vary with the material, but if conditions necessary to obtain polyploids with colchicine are known, they could be taken as the starting point from which to develop more rigorous methods of treatment while still keeping the plants alive. Treated seedlings under favorable growing conditions appear to outgrow the effects of colchicine (Sanders et al., 1959). It would seem necessary to delay growth long enough to allow the drug not only to act (duplication of chromosomes to produce polyploid cells), but also to react (subdivision of polyploid cells to produce cells with viable new diploid complements) while still keeping the material alive. To what extent mutants can be obtained by colchicine treatment of plants other than Sorghum needs to be determined. In either a research or a breeding program, colchicine treatment would provide a method for obtaining recombinations of substitutable chromo- somes, sometimes accompanied by homozygosity, within the genetic complement either of an established line or of a hybrid. The method has been used in the sorghum-breeding program at South Dakota State College, and new true-breeding lines have been obtained and released as agronomic varieties following treatment both of seedlings from true-breeding lines, ie., ‘Winner’ (Class II, Franzke & Sanders, in press), and of hybrid seedlings, i.e., ‘Dual’ (Franzke, 1958). If substitution of individual chromosomes does occur, crossing over between paired segments of analo- gous chromosomes could result in recombination of parts of chromosomes as well as of whole chromosomes in the progeny. If the hypothesis is correct, the method in use in the sorghum program is basically similar to the new approach to plant breeding described by Riley (1963) for wheat. New possibilities are opening up following the discovery that elimination of chromosome V removes the limitations imposed by bivalent pairing and permits recombinations of chromosomes and chromosome segments between genomes. He writes, ‘In wheat we now have the first example of a system by which recombination between different parental patterns can be modified. The impact of this notion on plant breeding as a whole has still to be determined, but it could be con- siderable.” It may prove possible to obtain recombinations between different genomes in a plant by using colchicine to bring about chromo- some substitutions in somatic tissues, thus bypassing established bivalent pairing mechanisms. Use of asynaptic plants to interchange chromosomes between genomes should also be investigated. The degree of variability which might be obtained by using the method would depend on the genetic constitution of the original material. Treat- ment of a polyploid line which is a fixed heterozygote for many genes should provide a limited example of Stebbins’ (1940) description of possibilities for evolution within a polyploid complex: “It can produce 54 JOURNAL OF THE ARNOLD ARBORETUM [| VOL. XLV endless new species, but these are all or nearly all new combinations of the same supply of genic material; they are new variations on an old theme.” SUMMARY An hypothesis is presented that explains the diploid mutants which arise in certain lines of Sorghum following colchicine treatment of seedlings as resulting from the substitution of analogous chromosomes originally present in the genetic constitution of sorghums with 2n = 20. Nontrue-breeding mutants are formed if there are substitutions only of individual chromosomes; true-breeding mutants, if there are substitu- tions only of pairs of homologous chromosomes ; intermediate mutants, if there are substitutions of both individual chromosomes and pai These chromosome substitutions are not detected at hiakinesis or metaphase I since no abnormal pairing configurations result. Pachytene and diplotene chromosomes in nontrue-breeding mutants and in F, hybrids involving true-breeding mutant lines need to be carefully examined for evidence of analogous pairing. Sorghum and other species, particularly those of polyploid and hybrid origin, need to be investigated intensively in order to establish whether this hypothesis concerning colchicine-induced chromosome substitution is valid. LITERATURE CITED ALLEN, N. S., G. B. WILS & S. PoweLLt. 1950. Comparative effects of colchicine and sodium nacleeie a Hered. 41: 159-163. Atkinson, G. F., C. J. FranzkeE, & J. G. Ross. 1957. Differential reaction of two varieties of sorghum to colchione treatment. Jour. Hered. 48: 259-264. Beroner, A. D., A. G. Avery, & A. F. BLAKESLEE. 1940. Chromosomal defi- ciencies in Datura stramonium induced by colchicine treatment. Am. Jour. Bot. 27: 676-683 BLAKESLEE, A. F. 1922. Variations in Datura due to changes in chromosome number. Am. Nat. 56: 16-31. . 1937. Dédoublement du nombre de chromosomes chez les plantes par traitement chimique. Compt. Rend. Acad. Paris. 205: 476-479. & ELLING. 1924. Chromosomal mutations in the Jimson weed, Datura j a , TURKEY- | : ‘ iit oE1azig : \ an uz -_ va \ Golii | N A S | A ct Can 4 | (ANATOLIA) i /__1 38° Vv ae ? % eKonya steve li ry eMaras eee ~ Antalya ondan® Pais i. Pee el. See / 1RAQ e F / eo ais SYRIA a 36° ' MEDITERRANEAN SEA - ( kms. miles et . CYPRU'S 0 100 0 700 all l | l 26° 28° 30° 32° 34° 36° 38° 40° 42° 44 Text-FIG. 1. Map of Turkey, showing grid system used in citation of specimens. For explanation see text. 60 JOURNAL OF THE ARNOLD ARBORETUM | VOL. XLV Botanisches Institut und Botanischer Garten der Universitat, Wien, Austria WU). Botanischer Garten und Museum der Universitat, Zurich, Switzerland (z) (Anatolian collections of Markgraf). Special thanks are extended to Dr. J. E. Dandy, of the British Museum (Natural History), London; to Professor M. Geiger-Huber, of the Botanisches Institut der Universitat, Basel; and to Mr. B. L. Burtt, of the Royal Botanic Garden, Edinburgh, for their assistance regarding the typification of the binomial Alyssum minus and the dropping from usage of “Alyssum campestre.” Acknowledgment is due Mrs. E. C. Dudley, Miss J. Kroll, and Miss R. Smith for their excellent execution of the figure and plates incorporated in this paper. Thanks are due Miss Morag MacKenzie, for preparation of the original manuscript; Mrs. T. P. Walsh, for prepara- tion of the present one; and Dr. Bernice G. Schubert, for her time, patience, and many helpful editorial suggestions. The author gives sincere recognition to Dr. A. Huber-Morath, Basel, and to Dr. J. Cullen, University Department of Botany, Edinburgh, for their useful suggestions and criticisms. For his careful and constant attention and advice during the preparation of this manuscript, the author is indebted to his supervisor, Dr. P. H. Davis, University Department of Botany, Edinburgh. Alyssum Linnaeus, Sp. Pl. 2: 650. 1753; Gen. PI. ed. 5. 293. 1754. LecTOTYPE spEcIEs: Alyssum montanum L. (!) (See Britton & Brown, Ill. Fl. North. US. ed. 2. 2: 154. 1913; Hitchcock & Green, 1929). Meniocus Desv. Jour. Bot. Appl. 3: 173. 1814. Type spectes: Alyssum linifolium Steph. ex Willd. (!). Ptilotrichum Meyer in Ledebour, Fl. Alt. 3: 50. 1831, TYPE SPECIES: Alys- sum canescens DC. Syst. Nat. 2: 322. 1821 (!). Lectotype, Russia. In rupestribus montosis Siberiae, Mt. Altai, 1819, Fischer s.n. (c-Herb. DC., as A. sibiricum). Psilonema Meyer, ibid. LectoTYPE SPECIES: Alyssum dasycarpum Steph. ex Willd. Linn. Sp. Pl. ed. 4. 3(1): 469. 1800 (!). Holotype, Russia. In Sibiria ad Kamam et Volgam fluvium, Stephan (LE, non vidi); isotypes BM, G, K). Odontarrhena Meyer, ibid. 58. LecTOTYPE SPECIES: Alyssum tortuosum Willd. (!). Gamosepalum Hausskn. Mitt. Thiir. Bot. Ver. 11: 73. 1897. LECTOTYPE SPECIES: Gamosepalum lepidoto-stellatum Hausskn. & Bornm. ex Hausskn. (!). Triplopetalum Nyarady, Magyar Bot. Lap. 24: 97. 1925. TYPE SPECIES: Triplopetalum pinifolium Nyar. (!). Section Meniocus (Desv.) Hook. in Benth. & Hook. Gen. Pl. 1: 74. 1862. Type spectes: Alyssum linifolium Steph. ex Willd. Linn. Sp. Pl. ed. 4. 3(1): 467. 1800 (!). Syntypes, Crimea and Caucasus. 1964] DUDLEY, STUDIES IN ALYSSUM 61 Tauria et Armenia, Stephan s.n. (B, non vidi); isosyntype, ‘“Tauria”’ (Gc). Alyssum blepharocarpum Dudley & Huber-Morath, sp. nov. Holotype, Turkey, Huber-Morath 13722 (HM); isotype (E). Pv. I, Fics. 1-13. Affinis A, Auetit Boiss. sed habitu diverso, fructibus minoribus, stylo longiore, petalis maioribus, glandulis globosis (haud subulatis) differt. Herba annua, 7-15(-18) cm. alta et lata, e basi ramosa caulibus tenuibus sea vel patentibus parce foliatis. Planta ex toto virescens vel cinereo-cana, saepe ad maturitatem rubro-purpurea, indumento_pilis stellatis i ees minute punctatis brevibus ramosis 4—6 radiatis ut videtur multi-radiatis 0.2-0.4 mm. diametro vestita. Folia linearia, oblanceolata vel rare subspatulata, obtusa vel acuta, sessilia, vel ad ee sensim attenuata, (3.5—)8—20(—25) mm. longa, 0.5—2 mm. lata, summam versus decrescentia. Jnflorescentia 5—12(—20) cm. alta et lata, simplex vel corymbosa, ramulis simplicibus ascendentibus 2.5—5(—8) cm. altis. Pedicelli 3.5—7.5 mm. longi, subhorizontales vel divergentes inter se 4—7 mm. distantes, pilis stellatis minutis appressis paucibus radiis inaequalibus parce vel copiose provisi. Sepala elliptico-ovata vel lanceolata, (1.5—)2.5-3 mm. longa, 0.8—1 mm. lata, margine hyalina angusta, pilis stellatis appressis parce vel copiose vestita, interdum dorsaliter carinata 1—5 pilis stellatis basi manifeste tuberculatis provisa. Petala in sicco flava vel subflava, obovata, retusa vel subretusa rare integra, in unguem brevem angustum sensim attenuata, glabra vel solum in ungue pilis stellatis provisa, (2.5—)3-4(-5.5) mm. longa, 1—2.5(—3.5) mm. lata. Filamenta longa (1.5—)2—3.5 mm. longa, bilateraliter alata; ala una quam altera latior, 1—3-denticulata, apice libero 0.5 mm. longo. Filamenta brevia 1.5—2.5 mm longa, appendice ad basim connato acuto vel minute 2-—3-denticulato (0.5—)0.7-1(—1.5) mm. longo. Glandulae bene evolutae, globosae, lobatae. Antherae auriculatae, apice obtusae vel acutae, 0.5—0.6 mm. longae. Stylus glaber, rigidus, tenuis, (0.7—)1—1.5(—2) mm. longus, stigmate magno provisus. Silicula ovato-orbiculata vel elliptica, obtusa vel truncata, rare acuta vel emarginata, 3.5—4.5(—5) mm. longa, (2—)2.3—3.2(-3.8) provisa, valvis membranaceis manifeste nervosis aequaliter compressis minute papillosis vel glabris vel setis eis ad marginem similibus copiose provisa. Ovula in placenta lateraliter inserta, 2 or 3 (4) per loculum. Semina brunnea, 1.5—2 mm. longa, 1-1.5 mm. lata, ala (0.1—)0.2-0.3 mm. lata. Fl. Apr.—July, fr. May—July. DISTRIBUTION AND HABITAT: Anatolian endemic in disturbed situations, cultivated and neglected fields, Astragalus and salt steppe, gypsum out- crops and Tertiary rocks; alt. 500-16 20 m. urkey. A3: Prov. Ankara, Ankara-Sariyar, Sakaryatal, 139 km. westlich Peto auf Tertiarboden, 26 Apr. 1955, Walter 1359 (HM); westlich Beypazari, 62 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Yénikoy-Sakarya Strasse, Steppe, 500 m., 27 May 1957, Kiihne 267, 263a (stu); Beypazari, Nahe Abzweigung nach Tepekoy, Feld, 1 June 1957, Kiihne 496 (stu); Uytepe, Gipfelgrat, Steppe, 7 June 1957, Kiihne 733, 726 (stu). A8: Prov. Gumisane, dist. Bayburt, Gimusane-Bayburt, Linsenfeld, 21 km. vor Bayburt mit A. stylare, 1620 m., 15 July 1958, Huber-Morath 16485 (uM). B4: Prov. Konya, Cihanbeyli, steppe, fl. lutea, 7 June 1952, Davis 18636 (BM, E, K); Prov. Ankara, dist. Kadinhan, Sarayoni-Cihanbeyli, Weizenfeldrand, 13 km. nordostlich Saray6nii, 870 m., 1 June 1956, Huber-Morath 13722 (HM, E); Angora (Ankara), Frichés a Kodja-dagh, 8 July 1908, Fréres E. C. 458 (&, G); ee Dere, Steppe im Gelande der Deutschen Botschaft, Andesitlehm, 900 m., ay 1958, Markgraf sm. (z); Steppen-Probenfliche, 7 May 1958, Holes sm. (z); Bergsteppe, Hussein bei Ankara, 4 June 1932, Kotte 171 K); Wegrand am alten Flugplatz von Ankara, 850 m., 11 June 1949, Huber- Morath 9317 (4M); dist. Serefli Kochisar, versalzte Steppe am Tuz goli, 114 km. siidlich Ankara, 11 km. siidlich Sekerk6y, 920 m., 26 June 1959, Huber- ee 16079 (uM): Ankara-Golkarindan, soma ie Lenari, 20 June 1953, and & Zohary 2928 (ANK). B7: Prov. Erzincan, Baskoyukten, km. somas “be tarlan Lenari, 1020 m., 30 May 1956, Birand 141 (ANK). In central and eastern Anatolia, Alyssum blepharocarpum grows in small, scattered populations. It is closely related to A. Auetii Boiss. with which it is broadly sympatric. It differs from the latter by being a more delicate and spreading plant, having smaller and often orbicular or ovate fruits with sparser setae (or none) on the valves, longer styles, larger petals, and globose lobed glands. In the eastern part of its range Alyssum blepharocarpum is partially sympatric with A. stylare (Boiss. & Bal.) Boiss., which has its center of distribution farther south in the Cappadocian steppe and which may be distinguished from A. blepharocarpum by its longer styles, larger and bifid petals, and larger and elliptic fruits with longer, denser setae on the valves. In sect. Mentocus, Alyssum hueti, A. blepharocarpum, A. stylare, and A. heterotrichum Boiss. possess simple stiff setae on their fruits, but A. blepharocarpum is the only species among these which has such a wide range of variation in fruit indumentum. When the specimens which are cited above were examined individually it appeared that three different combinations of indumentum types were present on the fruit (PL. I, Fics. 2-5). These indumentum states are: 1) marginal setae 0.2—0.4 mm. long, the valve surfaces minutely papillose (Davis 18636); 2) marginal and surface setae 0.4-0.5 mm. long, the valve surfaces papillose or glabrous (Huber-Morath 13722); 3) marginal setae 0.2—-0.4 mm. long, the valve surfaces always glabrous (Huber-Morath 1359b). It was soon evident when the gatherings were compared with one another that the indumen- tum on the fruit could not be employed to distinguish three distinct taxa. No geographical pattern emerged to correlate with the three indumentum states which sometimes occurred, not only on different plants from the same population, but also on different fruits of the same plant. This new species was first pointed out to the author by Dr. A. Huber- Morath in his private herbarium in Basel, Switzerland; however, it was 1964] DUDLEY, STUDIES IN ALYSSUM 63 first collected near Ankara in 1908 by Fréres des E. C. (No. 458). The Latin and French descriptions which Aznavour attached to this collection were never published, and Dr. Huber-Morath suggested as appropriate the specific epithet blepharocarpum (Greek, blepharis, eye-lash), refer- ring to the short setae on the fruit. Section Alyssum. LEcTotyPE SPECIES: Alyssum montanum L. Sp. Pl. 2: 650. 1753 (!). Alyssum minus (L.) Rothmaler, Repert. Sp. Nov. 50: 77. 1941.— Breistroffer, Bull. Soc. Bot. Fr. 103 (82 Sess. Extra.): 153. 1956; Heywood, Repert. Sp. Nov. 64: 54. 1961. ms minor L. Flora Monspeliensis (Nathhorst, Dissert. [no. 70] 21. —non Amoenitates Academicae 4. 1759). Type: Magnol, Botanicum ee ed. 2. 251. 1688— “Thlaspi Alysson ae minus, capsulis majoribus rotundis, non foliatis.” Alyssum campestre L. Sp. Pl. ed. 2. 2: 909. 1763, pro parte — non 1759(!).— Turrill, Jour. Bot. (Lond.) 73: 261. 1935; Degen, Fl. Velebitica 2: 187. 1937; Maire, Bull. Soc. Hist. Nat. Afr. Nord 30: 258. 1939; Zohary, Palest. Jour. Bot. Jerusalem Series 2: 161. 1941; Rechinger f. Ark. Bot. II. 2: 345. Alyssum parsitiorwm F a as Fl. Taur. Cauc. 3: 434. 1819(!).— Turrill, loc. cit.; Mai Frequently, botanists confuse Alyssum minus (L.) Rothmaler (= 4. campestre auct.) with A. alyssoides (L.) L. in sect. PsttoneMa. In addi- tion to the subulate, wingless, edentate and unappendaged filaments of the latter, A. alyssoides can be separated readily from A. minus by having sepals always persistent. The specific epithet, a/yssoides, was furnished by Linnaeus in Species Plantarum (2: 652. 1753) under Clypeola No. 2, which he transferred to Alyssum in Systema Naturae, ed. 10. 1759. Most taxonomists today accept the binomial A. alyssoides, which is typified by a specimen extant in the Clifford herbarium (BM), though Linnaeus redescribed this specimen in the second edition of Species Plantarum (1763), as A. calycinum. The crux of the problem concerning the application of the binomial Alyssum minus or A. campestre is the identity and typification of Clypeola licen of Species Plantarum (1753). Though Clypeola No. 3, found n page 652 of Species Plantarum, is not furnished with a trivial epithet, the name campestris is found on page 1231 in the Errata. Linnaeus refers to Sauvage, Methodus Foliorum Monspeliensis ... 71. 1751, which reads ‘“‘No. 405 Clyp. annua siliculis bilocularibus dispermis calyce per- sistente. . . .” This Sauvage description must be regarded as the basis of Clypeola campestris. Though no specimens of Alyssum or Clypeola in the Linnaean herbarium are marked with an ‘‘M,” indicating their origin from Sauvage (cf. Stearn, Introduction to Species Plantarum, facsimile, 1957; Savage, 1945), it is logical to assume that Sauvage’s mention of 64 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV persistent sepals could apply only to Alyssum alyssoides, rather than to the other taxon commonly, but incorrectly, called Alyssum campestre, whose sepals are always early deciduous. That Linnaeus later altered Sauvage's circumscription in the second edition of Species Plantarum to read: “. . . calycibus caducis ...” is not relevant to the problem of typifying Clypeola campestris, or as it later became, Alyssum campestre (Systema Naturae, ed. 10. 1759) Further evidence is gathered from the C. Bauhin polynomial, ‘“Alysson dictum campestre minus” (Pinax, 107. 1623) which is referred to in- directly by Sauvage through Tournefort (/nstitutiones rei herbariae 1: 217. 1700—‘“‘Alyss. incanum serpilli folio minus”), and directly by Magnol in Botanicum Monspeliense, 251 (ed. of 1688). This Bauhin polynomial is cited by Linnaeus in Species Plantarum (1753) as a synonym of Clypeola campestris and as a possible synonym of Clypeola alyssoides in Hortus Cliffortianus (329. 1738). Linnaeus in Flora Monspeliensis (Nathhorst, Dissert. [no. 70] 21. 1756.) refers Clypeola campestris to No. 1137 in Magnol’s work. Linnaeus’s own copy of this work (at the Linnean Society, London) is numbered in his handwriting, and No. 1137 falls on the entry (page 251) which reads: “Thlaspi dictum campestre minus . . Dr. M. Gotoe Huber. of the Botanisches Institut der Universitat Basel, Switzerland, states (in a letter) that, although there are no authentic specimens of ‘‘Alysson dictum campestre minus” in the Bauhin Her- barium, C, Bauhin wrote in the Catalogus Plantarum circa Basileam . (1622) that it was common in the fields around Basel. As only two species of Alyssum are known to occur near Basel, the alpine A. montanum, and the common weed, A. alyssoides, it may be inferred that the Bauhin polynomial refers to the latter. Rather than increase the nomenclatural confusion which would be implicit if the binomial Alyssum campestre were substituted for A. alys- soides, it is preferable to retain A. alyssoides for the taxon with persistent sepals, and treat Alyssum campestre as a synonym of it. Now that it is clear that Alyssuwm campestre can be reduced to a synonym of A. alyssoides, the next unambiguous binomial must be found and applied to the plant which has long masqueraded as Alyssum campestre and which has early deciduous sepals. Turrill (1935), who recognized the confusion implicit in the Linnaean Alyssum campestre, claimed that Linnaeus himself used the name in dif- ferent senses and suggested that since this initial confusion has been a persistent source of error ‘A. campestre” should be rejected as a nomen confusum. He advocated the application of A. parviflorum Bieb. (1819), as did Maire in 1939. It is interesting to note that, in 1808, Marschall von Bieberstein used A. campestre in the sense of A. alyssoides. The view that A. parviflorum is the correct epithet for this species has been fol- lowed by a number of authors; however, none has mentioned Clypeola minor L. Flora Monspeliensis. 21. 1756. In his extensive enumeration of the Linnaean taxa found in the Dissertationes and the Amoenitates Aca- 1964 | DUDLEY, STUDIES IN ALYSSUM 65 demicae, Rothmaler (1941) pointed out that Clypeola campestris on page 21 of Flora Monspeliensis (1756) was to be referred to Clypeola alys- soides. What is more important is that Rothmaler recombined Clypeola minor, which occurs on the same page, as Alyssum minus, but he was wrong in equating it with A. campestre (1759), the basionym of which is Clypeola campestris (Sp. Pl. 2: 650 & 1231. 1753) (cf. Heywood, 1961). At this point it might prove useful to hypothesize that Linnaeus recog- nized the confusion surrounding his usage of the epithet campestre and attempted to rectify it. In the second edition of Species Plantarum he altered the original diagnosis of Clypeola campestris, which he had taken from Sauvage, Methodus Foliorum Monspeliensis. . . (1751), to read “calycibus caducis,” thereby hoping to exclude Clypeola (Alyssum) alys- soides, which has persistent sepals. In the same work, attempting to drop the epithet alyssoides entirely from use, Linnaeus redescribed Alyssum alyssoides as Alyssum calycinum from the same type in the Clifford herbarium. Unfortunately, the specimens which Linnaeus (LINN) used to describe Alyssum campestre for the second edition of Species Plantarum are mixed; among them is a plant (No. 828:10) of 4. alyssoides which is labelled in Linnaeus’s handwriting as A. campestre. Except for the type of A, alyssoides in the Clifford herbarium, there are no specimens of Alyssum now existing which Linnaeus could have used for the first edition of Species Plantarum. In the republished version of Flora Monspeliensis in the Amoenitates Academicae 4. 1759, Linnaeus deleted Clypeola minor, which appeared first in the Dissertationes, and for it substituted Clypeola alyssoides. Though Linnaeus tried to drop the binomial Clypeola minor at this point, priority rights must be adhered to, and Rothmaler’s combina- tion of Clypeola minor as Alyssum minus should be accepted as referring to the taxon with deciduous sepals. In the Dissertation edition of Flora Monspeliensis, Linnaeus refers Clypeola minor to No. 1138 in Magnol, Botanicum Monspeliense (ed. of 1688). Because of the discordant elements in the Linnaean herbarium, none of which can be directly applied as the lectotype of Clypeola minor, Magnol’s description is accepted as the basis of Alyssum minus. In Lin- naeus’s numbered copy of Magnol’s work (Linnean Society, London), No. 1138 is found on page 251 and reads: “Thlaspi Alysson dictum minus capsulis majoribus rotundis, non foliatis; Thlaspi montanum luteum I.B. in priori juxta capsulam quatuor adsunt foliola viridia, in isto vero nulla sunt, capsulae etiam multum differunt; in iisdem locis reperitur, & eodem tempore floret. Addit I.B. Thlaspi montanum minimum Monspeliensium folio laciniato, circa Boutonet collectum, quid sit ignoramus.” The two polynomials which Magnol refers to in his circumscription are to be found in J. Bauhin, Historia Plantarum 2. 1651. The first of these “Thlaspi montanum luteum .. .” is found on page 928 of Bauhin’s work and is accompanied by a figure which can be identified as the taxon commonly referred to as “Alyssum campestre,” with deciduous sepals and large fruits. This figure cannot be applied to A. alyssoides which always has persistent sepals and smaller fruits. The most important character in Magnol’s 66 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV description is, of course, “. . . capsulis majoribus rotundis, non foliatis ” (indicating deciduous sepals), which is clearly contrasted with Sauvage’ s description of that which became is diner campestre (Alyssum The ‘ campestre), reading: “. . . calyce persistente . . view is taken here that the Magnol description, “Thlaspi Alysson dictum minus capsulis majoribus rotundis, non foliatis .. .”’ (but not including his synonyms) constitutes the type of Clypeola minor (Alyssum minus), and that to take Bauhin’s figure as the type would be an unwise extension of the evidence, for this was not the material upon which Magnol’s description was based. It may only be taken as additional proof of identity. It is of incidental interest that Magnol’s herbarium went to Sauvage, whose herbarium in turn went to Linnaeus. The other polynomial in Bauhin, Historia Plan- tarum, “Thlaspi montanum minimum Monspeliensium folio laciniato,”’ also cited by Magnol, is found on page 929 of Bauhin’s work. It is im- possible to tell from this description the taxon to which it should be referred. There is no figure. The establishment of Clypeola minor as the basionym of the taxon with deciduous sepals necessitates reducing to synonymy the binomial Alyssum parviflorum, which has at one time or another been suggested as the correct name Alyssum campestre sensu lato as used by many authors (e.g., Busch, 1909) has long been an anomaly with regard to its specific limits. My own interpretation of A. minus is in the narrow sense, and is applied to the widespread European and Oriental annual weed with homomor phic in- dumentum (only one type of hair) on the fruits. The taxa with dimorphic fruit indumentum of small stellate hairs with appressed or divergent rays in addition to furcate or simple, long and tuberculate hairs (e.g., 4. strigosum Banks & Solander and A. hirsutum Bieb.) are excluded from A. minus. In addition to the homomorphic indumentum of fruit and style, the usually smaller, elliptic or ovate, and asymmetrically inflated fruits of A, minus var. minus distinguish it from A. strigosum. Some difficulty may be experienced in distinguishing A. minus var. micranthum (Meyer) Dudley from A. strigosum because of the similarity of their fruits, but the indumentum of the fruits and pedicels is always diagnostic for A. strigosum. Though the indumentum of the pedicel of A. minus var. micranthum frequently appears strigose, this condition is not due to the presence of two types of hairs, but to the spreading, divergent and unequal rays of the stellate hairs. The glabrous and shorter styles and the divergent-spread- ing pedicels of Alyssum strigosum are also of value in separating this taxon from A. minus var. micranthum. It must, however, be stressed that the most reliable diagnostic character of A. strigosum is the occurrence of dimorphic indumentum on its fruits and pedicels. This character is easily observed on immature fruits and ovaries, as well as on those which are fully developed. Although Alyssum minus is extremely polymorphic in its habit and fruit size, probably because of the diverse environments in which it grows, the specific limits are no longer as confused as indicated by Rechinger f. 1964 | DUDLEY, STUDIES IN ALYSSUM 67 (1953). The exclusion of A. strigosum with its dimorphic indumentum allows A. minus to be conveniently defined in a narrow sense. Alyssum minus (L.) Rothm. var. micranthum (Meyer) Dudley, comb. nov. A. micranthum Meyer, Ind. Sem. Petrop. 1: 22. 1835(!). Lectotype, Russia. In Campis ad Mare Caspium, Lenkoran, Meyer s.n. (LE, non vidi) ; isotypes (G, W). . campestre var. micranthum (Meyer) Boiss. Fl. Or. 1: 284. 1867 — pro a ) A. subsp. micranthum (Meyer) Briest. Bull. Soc. Bot. Fr. 103 (82 Ses ess. Ee: 154. 1956. DISTRIBUTION AND HABITAT: Sporadically in northern Greece, Crete, North Africa, Crimea, Caucasus, Transcaspian, Turkish and Russian Armenia, northern Turkey, Syria, Lebanon, Israel, Jordan and northern Iran: waste land, cultivated fields and pastures, steppic hillsides, river valleys, forestry plantations, dry south-facing igneous, limestone and gypsum slopes; alt. (30—)500—-1600(—2200) m. FI. (Feb) Apr—June. The major features distinguishing this taxon from the typical expression of Alyssum minus are its larger orbicular fruits, its longer styles, which are only sparsely pubescent, and the coarser and sparser stellate hairs with longer and unequal divergent rays on the fruits, as noted in the following key. Styles 0.7-1.3 mm. long, covered their entire length with dense, appressed and short-rayed stellate hairs; rays of stellate hairs on fruit overlapping, equal in length, 0.3-0.5 mm long: fruit elliptic-ovate, 3-4.5 (-5) 2.5-4 mm., obtuse or rarely emarginate: pedicels ascending; long filaments 0.1-0.8 7 BAW WIR 65a. hak bona I ee OO OPE EES Var. minus. Styles 1-2 mm. long with sparse, ‘long and divergent-rayed stellate hairs : their bases; rays of stellate hairs on fruit not overlapping, unequal in length, g; fruit orbicular, 4-7.5 mm. long an wide, emarginate; nedicels horizontal; long filaments 0.3-0.4 mm. wide. .. Var. micranthum. These differential characters of the varieties of Alyssum minus are ob- vious when typical material of each is examined, but in some geographical areas intermediate specimens are quite numerous, i.e., North Africa, Iraq and Iran. In the light of the occurrence of intermediates, and the fact that var. micranthum occurs sporadically as small populations, recognition at any rank other than variety would not be justified. Alyssum strigosum Banks & Sol. subsp. cedrorum (Schott & Kotschy) Dudley, stat. nov. A. cedrorum ea & Kotschy, Osterr. Bot. Wochenbl. 11: 169. 1854 (!). Holotype, Turkey, C5: Prov. Icel. In Tauro inter Cedros, Kotschy s.n. (cultivated ite seed in the Vienna Botanic Gardens, w); isotype (G). 68 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV A, strigosum var. cedrorum (Schott & Kotschy) Briest. Bull. Soc. Bot. Fr. 103(82 Sess. Extra.): 154. 1956 (!). DISTRIBUTION AND HABITAT: Found mainly in the Lycian and Cilician Taurus, and extending into Cyprus and northeastern Anatolia. Disturbed and ruderal situations, neglected and cultivated fields, limestone slopes and scree, volcanic plateaus, steppe; often in Quercus scrub and Pinus brutia woods; alt. 300-1700 m. FI. Apr.—Jul This taxon was originally thought to be a species distinct from Alyssum strigosum, but the presence of a large number of intermediate specimens from Anatolia and Cyprus (mostly occurring within the range of subsp. cedrorum) indicates that subspecific rank is more appropriate. The rela- tively limited distribution of subsp. cedrorum is completely contained within the much wider area covered by subsp. strigosum. In addition to being a common plant in practically every province of Turkey, subsp. strigosum is prevalent throughout southern and eastern Europe, North Africa, and the Levant, extending east into Afghanistan. The following key indicates the distinguishing characteristics of the subspecies of Alyssum strigosum. Sepals early pra Banani hairs on fruit stiff, 0.4-0.6(1) mm. long; ; petals entire or emarginate, 2.5-3 mm. long; fruit ong an re ide. eee bea Robe meee seed een Subsp. strigosum. Sepals alent bifurcate hairs on fruit + sericeous, 1-1.5 mm. long; styles —1.5 mm. long; petals bilobed, 4-4.5 mm. long: fruit 5-6.5 * 4-7 mm PRUE ES 4 eeu e DEES ERA EEG U5 HOdN ee bb seeeeaes es Subsp. cedrorum. Alyssum trichocarpum Dudley & Huber-Morath, sp. nov. Holotype, Turkey, Huber-Morath 9253 (HM); isotype (E). Pu. II, rics. 12-22. Affinis A. cephaloti Boiss. et A. xanthocarpo Boiss. sed a priore racemi elongati (haud globoso-congesti), habitu minus robusto, foliis summam versus decrescentibus haud involucratis, stylo pilis stellatis minoribus et parcioribus differt; a posteriore fructibus minoribus formae diversae et indumento densiore radiis pilorum stellatorum longioribus recedit. Planta annua, patens, e basi multiramosa, (8—) 15—20(—30) cm. alta et lata. Caules decumbentes vel ascendentes basi haud ramosi, rubro-pur- purei. Folia post anthesin decidua, sessilia vel breviter petiolata, obovata vel oblanceolata, ad basim sensim attenuata, acuta, 10-15(—20) mm. longa 2.5—4(—7) mm. lata, summum versus decrescentia, pilis stellatis parcis vel densis semper basi tuberculatis radiis antrorse suberectis paucis longis in- aequalibus rare ramosis vestita. Racemus 5-10 cm. longus, simplex, cylindricus, erectus, in statu florescenti 1-1.5 cm. lato, deinde dense fructiferus apice globoso. Pedicelli rigidi, divaricati vel ascendentes, re- curvi, 4-7 mm. longi, indumento dimorpho parco vel denso pilis stellatis radiis longis antrorse suberectis inaequalibus et radiis longissime sinuatis sericeis basi tuberculatis inaequalibus tecti. Sepala subpersistentia, basi argulato-saccata, subcarinata, margine angusta hyalina, ovata, 3.5-4 mm. 1964 | DUDLEY, STUDIES IN ALYSSUM 69 longa, 1 mm. lata, indumento denso dimorpho pilis stellatis radiis longis parcis suberectis et praesertim apicem versus pilis parcis vel multis longis- simis antrorse erectis simplicibus vel bifurcatis basi manifeste tuberculatis tecta. Petala obovata, basi cuneata, retusa vel bifida, (4—)5—6 mm. longa, 1-2 mm. lata, in unguem sensim attenuata, indumento denso solum pilis stellatis radiis crassis parcis rare ramosis tecta. Filamenta longa inaequal- iter bilateraliter alatis, 3.5-4.5(—5) mm. longa, ala una quam altera 2-plo latior dentibus minutis 2-3 supra medium provisa, ala altera apicem versus abrupte contracta vel dentata. Filamenta brevia 3—4(—4.5) mm. longa, appendice 0.8—1(—1.3) mm. lato, in dimidio inferiore connata. Glandulae valde conspicuae globosae. Stylus subrostratus, basi manifeste dilatatus, 2-3.5 mm. longus, dense pilosus in dimidio inferiore pilis stellatis minutis radiis paucis brevibus divergentibus praeditus, stigmate magno globoso provisus. Silicula duoseminata, pallide flavo-viridia, orbiculata, acuta vel ad basim attenuata (3.5—)4—-5(-6.5) mm. longa, 3-4.5 mm. lata, valvis cartilagineis difficiliter separabilibus subcompressis + aequaliter inflatis indumento denso dimorpho pilis stellatis appressis vel subappressis ramosis radiatis et pilis simplicibus vel bifurcatis (1—)2—3 mm. longis antrorse ascendentibus sinuatis sericeis basi tuberculatis (tuberculis globosis inflatis (1.5—)2.5-3 mm. altis!) provisa. Ovula dua per loculum. Semen fuscum vel subnigrum, 2.5-3.5 mm. longum, 2—2.5 mm. latum, ala 0.05—0.1 mm. lata, valde crenulata. Fl. May—June, fr. June—July. ey. B6: Prov. Malatya, Darende-Akcadag, ca. 50 km. Ostlich Se ee 1450-1480 m., 20 June 1949, Huber-Morath 9253 (HM; E). Prov. Yozgat, Akdagmadeni, Aktas, June 1960, Curtis 124 (£). This rare and unusual annual belongs in the species complex of sect. AtyssuM which includes Alyssum strigosum Banks & Sol., A. xanthocarpum Boiss., A. hirsutum Bieb., A. bulbotrichum Hausskn. & Bornm. and A. cephalotes Boiss. The new species shares with all of these the annual habit and dimorphic indumentum on the fruit. It may be distinguished from A. cephalotes, one of its closest allies, by the elongate fruiting raceme, more reduced form of growth, uppermost leaves which are not involucrate but are smaller than the basal leaves, and indumentum on the style com- posed of minute stellate hairs only. The new species may be separated from A. xanthocarpum by smaller, orbicular fruits with denser indumentum composed primarily of much longer and more strongly tuberculate hairs. A, trichocarpum may further be distinguished from all species in the complex by the outwardly arcuate pedicels and the very thick fruit valves separated only with difficulty. Alyssum praecox Boiss. & Bal. var. albiflorum Dudley, var. nov. Holotype, Turkey, B5: Prov. Kirsehir, Sifegol, near Mucur, salt marsh in zone nearest lake, in stiff clay, leaves succulent, flowers white, 17 June 1954, Davis 21795 (£); isotypes (BM, K). A typo floribus albis et habitu halophilo differt. 70 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV This new variety is a local and extremely halophytic low altitude race which formed a large population in the clay zone nearest the edge of a small salt lake in central Anatolia. It differs from the high altitude typical variety in possessing white flowers and somewhat more succulent leaves (features originally noted on the label). Its distinctness is supported by the additional information that the nearest station of the yellow-flowered typical variety of A. praecox lies ca. 100-150 miles to the south in the high mountains of the Anti-Taurus. The habitat of var. praecox is pre- dominantly alpine, at altitudes of 1500-2600 meters. The plants of var. praecox, more often than not, are found very near the snow line. The distribution of var. praecox is that of a typical alpine or montane eastern Anatolian species, extending from the Cilician Taurus to the Armenian Highlands, and occasionally into Turkish Kurdistan. The altitude of the type locality, in the steppe, of var. albiflorum is approximately 600 meters. The vegetation present there is predominantly of Irano-Turanian ele- ments, and is similar to that found at most Anatolian salt lakes. The substrate on which var. praecox is found is calcareous in nature, or rarely igneous. Section Gamosepalum (Hausskn.) Dudley, comb. & stat. nov. Gamosepalum Hausskn. Mitt. Thiir. Bot. Ver. 11: 73. 1897 (!). LecrotyPE SPECIES: Alyssum lepidoto-stellatum (Hausskn. & Bornm.) Dudley. iii Pram: Dudley, ser. nov. Type spectes: Alyssum lepidoto- llatum (Hausskn. & Bornm.) Dudley Filamenta longa connata. Indumentum e > pilis stellatis pauciradiatis (radiis ramosis) compositum. Petala albida vel pallide flava, limbo basi saepe purpureo-venoso. Alyssum lepidoto-stellatum (Hausskn. & Bornm.) Dudley, comb. nov. Gamosepalum lepidoto-stellatum Hausskn. & Bornm. Mitt. Thir. Bot. Ver. 11: ' 7 (!). Holotype, Turkey, A/B 6: Prov. Sivas, in Ponto Australi in ip abi apricis inter Sivas et Mt. Yyldisdagh (Yildis dag), 1300-1400 m., 6 June 1890, Bornmuller 1671 (JE, non vidi); isotypes (BM, G). Gamosepalum confine Hausskn. op. cit. 74 (!). Lectotype, Turkey, B7: Prov. Erzincan, Chama on Euphrates, 1890, Sintenis 2161, (JE, non vidi) ; isolecto- DISTRIBUTION AND HABITAT: Turkish endemic from the central and southern steppe areas and calcareous hills; alt. 1300-1524 m. Fl. May-— June The reasons for recognizing Gamosepalum as a section within Alyssum, rather than as a distinct genus, will be considered in a later paper. The type species, Alyssum lepidoto-stellatum, is easily separated from A. paphlagonicum (Hausskn.) Dudley, with which it has sometimes been confused, by its merely retuse and sparsely pubescent petals, shorter styles, 1964 | DUDLEY, STUDIES IN ALYSSUM 71 more spreading usually horizontal pedicels and few-fruited condensed racemes. The habit of A. lepidoto-stellatum is more dwarf and caespitose than the mostly erect, luxuriant, and coarse habit of A. paphlagonicum. Schulz (1927) reduced Gamosepalum confine to synonymy under G. lepidoto-stellatum Hausskn. & Bornm. The lectotype collection of G. confine (Sintenis 2161) matches Alyssum lepidoto-stellatum (Hausskn. & Bornm.) Dudley in all characters except its petals (ca. 6 mm. long), which are nearly as large as the largest found in A. paphlagonicum. The other specimen of G. confine cited and collected by Haussknecht (Iran, Monte Singara Mesopotamiae, May 1867) has not been seen. Although Hauss- knecht claims to have been unable to distinguish his specimen from the type of G. confine (= A. lepidoto-stellatum), from the diagnosis in which he contrasts the Iranian plant with the type of A. lepidoto-stellatum, I think he may possibly have had A. sulphureum Dudley & Huber-Morath. Alyssum paphlagonicum (Hausskn.) Dudley, comb. nov. Gamosepalum paphlagonicum Hausskn. Mitt. Thir. Bot. Ver. 11: 74. 1897 (!). Holo otype, Turkey, A4/5: Prov. Kastamonu, Paphlagoniae, dist. Kastambuli, in montosis supra Kissiltscha (Kisilia-Kisilca) pr. Tossia, 24 May 1892, Sintenis 3865 (JE, non vidi); isotypes (BM, G, K DISTRIBUTION AND HABITAT: Endemic to central and northeastern Anatolia, in vineyards, serpentine and noncalcareous areas, and Artemisia steppe; alt. 800-1372 m. Fl. May—June. This species is partially sympatric with Alyssum lepidoto-stellatum. In addition to the differential characters cited in the discussion of A. lepidoto-stellatum, A. paphlagonicum possesses several other distinctive characteristics. The sepals of A. paphlagonicum show a tendency to be slightly inflated and are separable with difficulty at maturity. Those of A. lepidoto-stellatum do not illustrate any obvious inflation and are easily separated at maturity. The overall indumentum of A. paphlagonicum is composed of larger and coarser stellate hairs, and the floral parts are, on the average, larger than those of A. lepidoto-stellatum. However, on individual specimens the range of variation is too great to allow these species to be safely distinguished on such quantitative characters. Alyssum paphlagonicum appears to have an edaphic preference for serpentine or igneous rubble, whereas A. l/epidoto-stellatum is confined almost entirely to calcareous soils. Alyssum thymops (Huber-Morath & Reese) Dudley, comb. nov. ee thymops Huber-Morath & Reese, Repert. Sp. Nov. 52: 3 (!). Holotype, Turkey, C/5: Prov. Nigde, Cappadocien, Steppe bel ae Vilajet Nigde, 10 June 1937, Reese s.n. (HM); isotype (£). DISTRIBUTION AND HABITAT: Endemic to central Anatolia in steppe. wheat fields, eruptive stones; in association with Pinus nigra; alt. 980- 1560 m. Fl. June. 72 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Huber-Morath (1943) correctly points out that his Pt#ilotrichum thymops is morphologically closer to Gamosepalum lepidoto-stellatum than any oriental species of Ptilotrichum sensu Meyer. He noted, however, that the connate sepals and connate long filaments characteristic of Gamosepalum were not found in his Ptilotrichum thymops. Examination of the type material of Ptilotrichum (Alyssum) thymops revealed that the fusion of its long filaments, the “pseudo-connation” of its sepals, and its indumentum of few-rayed stellate hairs (not lepidote as indicated by Huber-Morath) were consistent. These features ally it directly to the complex in sect. GAMOSEPALUM (series Connata) ee A, tetrastemon Boiss., A. lepidoto-stellatum, and A. paphlagonic Alyssum thymops, which is partially gdbebien Sette in ie central Anatolian steppe with two of its closest allies, A. tetrastemon and A. paphlagonicum, may be distinguished from these by the weakly dimorphic and sparse short-rayed hairs on the sepals; the obovate, glabrous, pale cream petals; the very narrow linear-oblanceolate leaves; and the noninvolucrate cauline leaves. The narrow wings of the long filaments of A. thymops led Huber- Morath to describe “stamina non alata.” These wings, which are narrower than those of any other species in sect. GAMOSEPALUM, are of further distinguishing value. — mee Dudley, ser. nov. Type sPECIES: Alyssum baumgariner- m Bornm. ex. Baumg. Jahresb. Kaiser Franz Josef-Land. Gymn. Oberenlacts Baden bei Wien 48: 16. 1911. Lectotype, Lebanon. Libanon, Dschebel Barik ad nives, 2100 m., 15 June 1910, Born- miiller 11405 (w); isolectotypes (BM, E, G, K). A serie Connatis filamentis longis contiguis liberis (non connatis), in- dumento pilis lepidotis multo-radiatis, petalis sulphureis haud purpureo- venosis recedit. Alyssum corningii Dudley, sp. nov. Holotype, Turkey, Dudley (D. 35911) (£); isotypes (A, HM). Pr. III, rics. 1-11. A. tetrastemon Boiss. var. cappadocicum Boiss. Fl. Or. 1: 278. 1867 (!). Holotype, Turkey, Balansa 493 (G); isotype (GH) — non var. cappadocica Boiss. 1842 A. tetrastemon sensu Baumg. Jahresb. Kaiser Franz Josef-Land. Gymn. Oberrealsch. Baden bei Wien 48: 5. 1911 (!) —non Boiss. 1842. Affinis A. tetrastemon Boiss. sed habitu diverso, indumento ex toto fere monomorpho pilis stellatis paucis radiatis brevibus, foliis haud involucra- tis, racemi elongati multifructi differt. Herba perennis suberecta vel patens, basi multiramosa et suffruticosa, 7-12 cm. alta, 5—-10(—15) cm. lata. Planta ex toto indumento cinereo-cano e pilis stellatis densis punctatis semper stipitatis interdum manifeste basi tuberculatis 0.2—0.4 mm. latis radiis brevibus vel longis 5—10 radiatis crassiusculis suberectis (rare appressis) inaequalibus multiramosis vel 1964 | DUDLEY, STUDIES IN ALYSSUM 73 lepidotis composito. Caules floriferi ascendentes vel erecti, 2—7(-12) cm. longi. Surculi steriles dense foliati, basi caulium floriferorum conferti, 1-3 cm. longi. Folia caulium floriferorum linearia, oblonga vel oblanceo- lata, acuta, sensim in petiolum attenuata, 6-10 mm. longa, 1.5—2 (—2.5) mm. lata; superiora longiora et latiora, ad marginem saepe subhirsuta, non involucrata. Folia caulium sterilium \ineari-oblanceolata, acuta vel rare obtusa, 6-10 mm. longa, 0.5—-1(-1.5) mm. lata, pilis stellatis quam eis foliorum caulium floriferorum densioribus et minoribus. Racemus simplex, brevibus, saepe confertus, rare subumbelliformis, post anthesin elongatus, saepe corymbosus ramulis aequalibus 1.8—2(—2.5) cm. longus. Pedicelli erecto-patentes, stricti, rigidi, subrecurvi, 3-6 mm. longi, pilis stellatis furcatis et suberectis sicut sepala obsiti. Sepala basi subsaccata, lanceo- lata vel ovata, acuta, 3—-3.5 mm. longa, 1-1.5 mm. lata, obscure pseudo- connata ob pilis intertextis cohaerentia, facile separabiles, extus indumento pilis stellatis appressis et rare apicem versus pilis furcatis basi tuberculatis radiis delicatissimis sinuatis inaequalibus sericeis sparse vel dense provisa, intus pilis stellatis radiis delicatulis inaequalibus sericeis obsita. Petala in sicco pallida, late spatulata, integra vel subretusa, 4.5-5.5 mm. longa 1-1.5 mm. lata, infra constrictionem medianem ungue alata (ala denticulata 0.4-0.6 mm. lata) munita, nervo mediano unguis pilis: stellatis parcis radiis haud ramosis proviso. Filamenta longa 3-4 mm. longa, libera, contigua, bilateraliter anguste marginata, superne sensim contracta, basi vel etiam ad medium rare conniventia sed facile pga pilis stellatis parcis minutis provisa. Filamenta brevia (1.8- )2-3.5 mm. longa, basi dente brevissimo 0.2-0.3 mm. longo provisa. Stylus rostratus vel subro- stratus, 1.5—2.5(-3) mm. longus, in dimidio inferiore pilis stellatis minutis multiradiatis parce vel dense provisus. Silicula uniseminata vel rare duo- seminata, aaa ovata, truncata vel subemarginata, 3.5—4 mm. longa, .5—3.5 mm. lata, valvis cartilagineis aequaliter vel inaequaliter inflatis ae cano dense vestitis. Ovula dua per loculum. Semen immaturum, alatum. FI. et fr. June. DISTRIBUTION AND HABITAT: Entirely confined to the central Anatolian steppe; alt. 1050-1500 m. Turkey. B4: Prov. Ankara, Giarrkale bei Haymana, Berg Steppe, Kotte 1165 (x). BS: Prov. Nigde, Tuff bei Nigde, 1100 m., 4 June 1898, Siehe 60 (B, BM G, K, W). C4: Konya to Sultanhani on new road, 18 miles from Konya, 1050 m., in new enclosed forestry plantation, steppic habitat, pale lemon flowers, rare, 16 June 1962, Dudley (D. 35911) (£; A, HM); Steppen bei Divle (Karaman- Eregli), 1300 m., June 1906, Siehe 241 (£); Sarayonii, civari Step yel Lenari, May 1943, Birand s.n. (ANK). C5: Prov. Seyhan, Partie superieure de la Vallee du Kamechly-Tchai (Keamsli), 1490-1500 m., 16 June 1856, Balansa 493 (G, GH). Alyssum corningii was first described as A. tetrastemon var. cappadoct- cum by Boissier (1867), who differentiated it from the typical A. tetraste- mon because it possessed “. . . pili omnes lepidoti et caules saepe ramu- losi.” This description may be contrasted with “basi suffrutescens lepi- 74 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV a dotum et insuper breviter tomentoso-hirtum caulibus humilibus . . . given by Boissier (1867) for A. tetrastemon sensu stricto. Boissier’s var. cappodocicum, of 1867, is antedated by his var. cappadocica, of 1842, which was based on a different type and is obviously a synonym of A. tetrastemon sensu stricto. The indumentum of Alyssum corningii is for the most part monomorphic and appressed and is lepidote in the lower portions of the plant. That of A. tetrastemon is dimorphic and distinctly hirsute and is tomentose especially on the sepals, pedicels, and upper stems and leaves. Baum- gartner (1911, pp. 5, 6) stated that the indumentum difference did not justify describing a new species. In fact, Baumgartner had only the Siehe 60 collection of A. corningii at his disposal. However, examination of all available material indicates that the indumentum character is of differential value when combined with the other consistent diagnostic characters. Accordingly, Boissier’s var. cappadocicum of 1867 should be maintained as a distinct species. In addition to the characters detailed by Boissier, Alyssum corningii may be separated from A. tetrastemon, the taxon with which it is most easily confused, by the closely adjacent and overlapping, but free long filaments (varely connate at their bases), the narrower cauline leaves which are not involucrate, and the many-fruited, elongate racemes. The distribution and general ecology of Alyssum corningii and A. tetrastemon do not appear to be mutually exclusive, though they have never been found growing together. Both species probably represent branches from a common ancestor within the perennials of sect. ALYSSUM and are closely allied to A. iranicum Baumg., A. aizoides Boiss., A. born- muelleri Hausskn. ex Degen, A. doerfleri Degen, A. taygeteum Heldr., and A. caespitosum Baumg. The specific epithet of this new taxon honors a friend and benefactor, the Honorable Erastus Corning II, of Albany, New York, whose support and interest made my investigations of Alyssum at Edinburgh a reality. Alyssum sulphureum Dudley & Huber-Morath, sp. nov. Holotype, Turkey, Huber-Morath 11969 (uM); isotype (rE). PL. IV, Fics 13-26. Gamosepalum paphlagonicum sensu Bornm. Symb. Fl. Anat. 59. 1936, pro parte quoad plantam Mesopotamiam a Handel-Mazzetti lectam — non Hausskn. (!). Affinis A. bornmuelleri (sect. ALyssum), A. aizoidi (sect. ALYSSUM) et A. harputico Dudley (sect. GAMosEPALUM), sed ob omnibus speciebus habitu diverso, sepalis angustioribus margine membranaceo, appendice filamentorum brevium multo breviore differt; insuper a primo forma squamorum lepidotorum, petalis retusis vel bilobatis recedit; a secundo filamentis longis edentatis, unguibus petalorum pilis stellatis provisis divergit; a tertio indumento diverso, petalis majoribus, pilis stellatis forma foliorum distinguitur. 1964 | DUDLEY, STUDIES IN ALYSSUM 75 Herba perennis, suffruticosa, dense caespitosa, pusilla, multiramosa, sed ad basim efoliata, (5—)10—20 cm. alta. Planta ex toto indumento denso appresso cinereo-argenteo squamis minute punctatis oe (0.2—)0.3-0.5 mm. diametro vestitis, sed superne et in marginibus (rare in paginis ambobus) foliorum superiorum indumento ut in sepalis dimorpho obsita, pedicellis et sepalis pilis stellatis radiis longis sericeo-villosis vestitis. Caules floriferi erecto-ascendentes, parce foliati, (2.5—)8-10 cm. longi. Surculi steriles numerosi, densissime conferti, patentes, dense foliati, 1.2—5 cm longi. Folia caulium floriferorum \ineari-oblanceolata, sessilia, acuta, (5—) 10-15 mm. longa, 1.5—2(-—2.5) mm. lata, ascendentes, superiora sensim increscentia, post anthesin decidua; summa involucrata. Folia caulium sterilium oblanceolata vel subspatulata, acuta, (2—)6—-10 mm. longa, 1.5—2 mm. lata. Corymbi umbelliformes capitato-rotundati, congesti, 1.5-2 cm longi et lati, floribus 5-15. Pedicelli in statu fructifero 3-4.5(-6) mm. longi, crassi, rigidi, ascendentes vel subhorizontales, indumento dimorpho eo sepalorum simili dense obtecti. Sepala subpersistentia, apice cucullata, inaequalia, acuta, pseudoconnata, (3.5—)4.5—5.5 mm. longa, (1.5—)2-3 mm. lata, margine hyalina angusta, 0.1-0.2 mm. lata squamis lepidotis tecta in dimidio inferiore intertexto connata, extus squamis lepidotis minute punctatis argentis 0.3-0.4 mm. diametro apicem versus indumento dimorpho e pilis stellatis basi tuberculatis sericeo-villosis radiis inaequali- bus longis paucis vel numerosis 0.5-1 mm. longis squamis lepidotis com- posito, intus indumento dimorpho parce vel dense obtecta e pilis stellatis radiis sericeis valde inaequalibus duobus longis antrorse appressis et pilis stellatis radiis aequalibus brevibus composito. Petala sulphurea, late spatulata, retusa vel bilobata, (5.5-)6-8 mm. longa, 2.5-3 mm. lata; unguis ad medium constrictus alis latis membranaceis saepe integris, nervo mediano pilis stellatis sublepidotis radiis paucis vel multis suberectis dense vel sparse obsito. Filamenta longa 4.5—5.5(—6) mm. longa, libera, non connata sed verum contigua, edentata, anguste bilateraliter alata, ad basim sensim dilatata. Filamenta brevia 34.5 mm. longa, angustissime alata ad basim appendice brevissima minuta praedita. Antherae auriculatae, apice acutae, 1 mm. longae. Stylus in statu fructifero, tenuis ad basin dilatatus, (2—)3-3.5(—4) mm. longus, glaber vel in dimidio inferiore pilis stellatis radiis suberectis obsitus. Silicula uniseminata, rare duoseminata, orbiculata 5(- m argenteo e squamis appressis minute punctatis lepidotis vel sublepidotis multiradiatis 0.3-0.5 mm. diametro composito obsitis. Ovula dua per loculum. Semen pallide brunneum, 3.5—4 mm. longum, 3 mm. latum, ala angusta (0.05—)0.1—0.2 mm. lata. Fl. June, fr. July. DISTRIBUTION AND HABITAT: Southeastern Anatolia and northern Syria and Iraq, on limestone and marl ledges, steppe hillsides, and scree; alt. 600-1650 m. ey. B6: Prov. Malatya, Kalkmergelhiigel ostlich ob Darende, 1000-1050 m., 28 June 1953, Huber-Morath 11969 (11M; E); Mergelhiigel, 2 km. westlich 76 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV ob Darende, an der Strasse nach Giiriin, 1090 m., 28 June 1953, Huber-Morath 11970 (aM); Mergelhiigel nérdlich ob dem Dorf Darende, 950-980 m., 20 June 1949, Huber-Morath 9252 (HM); entre Derindeh et Kavak Aghatch (Akcadag), 14 June 1906, Post 52 (Gc). C9: Prov. Hakari, mons Bestabije (Beytiissebap) prope Hoz, Amadia (Imfiya)-Sattak (Catak), 1650 m., July 1910, Ndbélek 1264 (BRA). Northern Syria. Desert nr. Marbat Antar, 3 May 1900, Post s.n. (BM). Northern Iraq. Kalk-Hange bei Dorfruine Chara (Bir Garan) im Dschebel Abd’elasis (Debel-Abdul Aziz), 600 m., Handel-Mazzetti s.n. (w). The closest ally to this new species in sect. GAMOSEPALUM is probably Alyssum harputicum Dudley. Alyssum sulphureum may be separated from it easily by the caespitose, dense cushion-forming habit, broader and shorter leaves, smaller lepidote scales with longer peripheral rays on the sterile shoot leaves and fruits, larger and usually bilobed petals, and the narrower membranous sepal margins. In addition, the stellate hairs on the upper cauline leaves, pedicels, and sepals often appear strigose with long, unequal and spreading rays. This form of indumentum is shared with A. bornmuelleri and A. aizoides in sect. ALyssum, from which A sulphureum may be separated by the same characters which distinguish it from A. harputicum. The edentate long filaments and the presence of indumentum on the petals are also of differential value in separating the new species from A. aizoides. Considerable difficulty was experienced in placing Alyssum sulphureum in the correct section due to its resemblance to some species of sect. AtyssuM. The problem was clarified by observation on A. sulphureum of dimorphic sepals which, because of their interlocking indumentum, appear fused; copious indumentum on the inner surface of the sepals; and eden- tate, free but adjacent and overlapping, long filaments. Haussknecht (1897) was bewildered by this resemblance of Gamose- palum to Alyssum and discussed it at length when he proposed the bino- mial Gamosepalum alyssoides. However, in any case, the specific epithet alyssoides may not be used for any taxon in Alyssum other than A. alyssoides (L.) L. (1759), a combination based on Clypeola alyssoides L. (1753). Haussknecht proposed that the name Gamosepalum alyssoides should be applied if, with the examination of additional material, it became evident that the three species he described (G. lepidoto-stellatum, G. con- fine, and G. paphlagonicum) did not warrant specific recognition. These three species would, according to him, then be reduced to the rank of varieties under G. alyssoides. It is interesting that the many additional collections of Gamosepalum, since the time of Haussknecht’s descriptions, indicate that G. lepidoto-stellatum and G. paphlagonicum are distinct species and well differentiated from one another. Gamosepalum confine, however, as Schulz (1927) points out, must be treated as synonymous with G. lepidoto-stellatum. Alyssum harputicum Dudley, sp. nov. Holotype, Turkey, B7: Prov. Elazig, Armenia Turcica, Kharput (Harput) in summo monte Kisil 1964] DUDLEY, STUDIES IN ALYSSUM 77 Depe, 10 May 1889, Sintenis 323 (c); aise (kK, w) (all as A. aizoides). L. IV, Fics. 1-12. A, aizoides sensu Baumg. Jahresb. Nied.-Ost. Land.-Lehrers. Wien.-Neust. 36: 24. 1909 et sensu Bornm. Symb. ad FI. Anat. 54. 1936—non Boiss.(!). Affinis A. lycaonico (Schulz) Dudley sed squamis lepidotis duplo minori- bus, caulibus longioribus ascendentibus, foliis surculorum sterilium angus- tioribus, sepalis haud inflatis minoribus, petalis minoribus indumento parciore differt. erba perennis e basi ramosa suffruticosa. Planta ex toto squamis ap- pressis disciformibus albo-lepidotis punctatis 0.2-0.3 mm. latis dense vestita. Caules floriferi erecti vel arcuato-ascendentes, 5-11 cm. longi. Surculi steriles numerosi eoque caespitosi, fastigiati, ascendentes, 1-4 cm. longi, dense foliati. Folia caulina sessilia, acuta; inferiora obovato- oblan- ceolata, ca. 4 mm. longa, 1--1.5 mm. lata: mediana linearia vel lineari- cuneata, ca. 10 mm. longa, 0.5-1 mm. lata. Folia surculorum sterilium lineari-oblanceolata, 5-10 mm. longa, 0.5-1 mm. lata. Corymdz simplices vel pauci-ramosi, congesti, rotundati, 1.3-1.5(—2) cm. lati, foliis caulinis summis subinvolucrati. Pedicelli patentes vel erecti, ca. 5 mm. longi. Sepala persistentia, pseudoconnata, subcarinata, subinaequalia, apice cucullata, acuta, 2.5—4 mm. longa, 1—1.5 mm. lata, margine hyalina squamis tecta, in dimidio inferiore ob indumento intertexto cohaerentia, extus dense lepidota, intus pilis radiis valde inaequalibus duobus longis antrorse appressis et pilis stellatis radiis aequalibus brevibus vestita. Petala lutea, obovata vel spatulata, 5—5.5 mm. longa, 2 mm. lata, apice integra vel leviter retusa, unguis margine hyalina in parte superiore dilatata, 0.2—0.3 mm. lata, pilis stellatis paucis obsitus. Filamenta longa ca. 4 mm. longa, libera, sed verum contigua, edentata, bilateraliter membranaceo-marginata, superno sensim contracta. Filamenta brevia 3-3.5 mm. longa, basi appendice brevissima praedita. Glandulae prominentes, globosae. Antherae 0.8-1 mm. longae. Ovarium orbiculare, 2 mm. longum, 1.5 mm. latum, dense lepidoto- squamatum, apice emarginatum. Ovula dua per loculum, non marginata. Stylus in statu florescenti 2.5-3 mm. longus ad basin dense lepidotus stigmate globoso capitato provisus. Fructus ignotus. Fl. May The diagnostic characters separating Alyssum harputicum from A. lycaonicum are easily observed, and, although these species are widely separated geographically, they appear to be closely related. Alyssum lycaonicum has large and extremely inflated sepals enclosing the ovary or fruit (presumably of dispersal value), whereas A. harputicum has per- sistent but considerably smaller sepals which are never inflated. The petals and lepidote scales of both species are similar in form, but those of A. harputicum are always smaller. The narrow, almost needle-like, involute leaves of A. harputicum are unique in sect. GAMOSEPALUM. Similar leaves, however, are possessed by Alyssum (Triplopetalum) pinifolium (Nyar.) Dudley (sect. ODONTARRHENA), an example of parallel evolution in mor- phologically and taxonomically widely separated species. 78 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Baumgartner (1909, p. 24) erroneously thought the type collection of Alyssum harputicum represented A. aizoides Boiss. in sect. ALYSSUM. Unfortunately he did not examine any material in addition to Sintenis 323 and did not know A. aizoides sensu stricto, but relied on a Stapf determi- nation. Bornmiiller (1936) repeated this error but stated “folia linearia,” accordingly including A. Aarputicum within A. aizoides. Alyssum harputi- cum is so distinct morphologically that it should never be confused with any other species, even its closest ally. Alyssum niveum Dudley, sp. nov. Holotype, Turkey, B3: Prov. Eskisehir, distr. Sivrihisar, nordéstlich von Yaka Pinar, Steppe auf Kalk, 18 May 1941, Romieux s.n. (HM). Pi. III, rics. 12-20. Affinis A. harputico Dudley sed habitu valde diverso, foliis latioribus, petalis longioribus et squamis lepidotis maioribus differt. Herba perennis, caespitosa, conferta, 3 cm. alta, 5 cm. lata. Planta ex toto indumento manifeste albo-argenteo squamis lepidotis multiradiatis appressis disciformibus manifeste punctatis (0.1—)0.2—0.5(0.7) mm diametro dense vestita. Folia imbricata, summam versus accrescentia, sessilia, acuta, squamis lepidotis 0.4-0.5 mm. diametro dense obtecta; superiora involucrata, linearia, oblanceolata vel spatulata, 10-15 mm. longa, 2-3 mm. lata; inferiora lanceolata vel obovata, 7-10 mm. longa, 2—2.5 mm. lata. Corymbi simplices, conferti, umbelliformes, 1—2 cm. lati et longi, floribus 8-15. Pedicelli in statu florescenti subhorizontales vel divergentes, interdum deflexi, 3-4 mm. longi, indumento denso squamis lepidotis 0.4-0.5 mm. diametro tecti. Sepala lanceolata vel elliptica, obtusa vel subacuta, basi subsaccato-angulata, leviter dimorpha, pseudoconnata sed facile separata, in dimidio inferiore ob indumento intertexto cohaeren- tia, persistentia, (3—)3.5—4.5 mm. longa, (1.5—)2.5-3 mm. lata, margine hyalina 0.2—0.3 mm. lata, extus indumento denso squamis lepidotis 0.2—0.3 mm. diametro provisa, intus pilis stellatis appressis multiradiatis sublepi- dotis parce vel copiose munita. Petala spatulata, retusa vel subretusa, rare integra, flava, 6.5—-7 mm. longa, 3-3.5 mm. lata, in ungue sensim attenuata vel ad medium constricta margine unguis saepe denticulata, squamis lepidotis 0.1-0.2 mm. latis in ungue (rare in lamina) obsita. Filamenta longa libera sed verum contigua, interdum omnino conniventia, angustissime bilateraliter alta, ad basim leviter dilatata, edentata, 3.5—4 mm. longa. Filamenta brevia 2.5-3 mm. longa, ad basim leviter dilatata, basi appendice brevissimo provisa. Glandulae minutae, globosae, lobatae. Antherae 1 mm. longae, apice obtusae vel subacutae, connectivo prolongato 0.2-0.3 mm. longo. Ovarium orbiculato-ovatum, apice emarginatum, 1.5— 2.5 mm. longum, 1-2 mm. latum, indumento denso albo squamis lepidotis 0.4—0.5 mm. diametro provisum. Ovula dua per loculum, ut videtur solum una vel dua maturantes. Stylus in statu florescenti 2—2.5(—3) mm. longus, rigidus, ad basim dilatatus, in dimidio inferiore (rare ex toto) squamis lepidotis densis provisus, stigmate magno globoso provisus. Fructus ignotus. Fl. May 1964 | DUDLEY, STUDIES IN ALYSSUM 79 A detailed comparison of the species in sect. GAMOSEPALUM shows that Alyssum niveum, a very rare steppe species, is morphologically closest to A. harputicum. Alyssum niveum deviates from A. harputicum in its extremely reduced habit, wider leaves, and larger lepidote scales. The petals of A. niveum are also distinctive in being very large and usually retuse, and in frequently having denticulate wings on the claws. The petals of A. harpu- ticum are smaller and have entire claw wings. Unfortunately, the species is known only from the type collection of one plant, but the numerous morphological discontinuities which distinguish it from the other species in sect. GAMOSEPALUM indicate that specific rank is appropriate. It is hoped that future collectors will rediscover this beau- tiful species. I was not successful in finding it in an area near the type locality. Alyssum lycaonicum (Schulz) Dudley, comb. nov. Gamosepalum lycaonicum Schulz, Notizbl. Bot. Gart. Mus. Berlin 10: 110. 1927 (!). Holotype, Turkey, B3: Prov. Konya, Wilajet Konia, Steppe bei Korasch (Korasi) in Lykaonien, 1400 m., June 1906, Siehe 274 (B, non vidi); isotype (BM). DISTRIBUTION AND HABITAT: Rare Anatolian endemic, confined to the dry steppe near Konya; alt. 1400 m. FI. June. The morphological discontinuities which define the limits of this species are of such significance that there should be no confusion with any other. The homomorphic indumentum of very large disciform lepidote scales (up to 2 mm. in diameter), the extremely inflated and enveloping calyx, the erect and tall growth, and the noninvolucrate cauline leaves which decrease in size upwards are sufficient to distinguish the species clearly from Alyssum harputicum and A. niveum, its closest allies. The distribution of these three species is completely allopatric. This ‘‘“Gamosepalum” was described as having deep yellow petals, a character which invalidated the original generic criterion of white petals (Schulz, 1927). However, Alyssum baumgartnerianum, designated by me as the type species of series Libera in sect. GAMOSEPALUM, also has deep yellow petals. Section Odontarrhena (Meyer) Koch, Synop. Fl. Germ. Helv. 59. 1836. LECTOTYPE SPECIES: Alyssum tortuosum Waldst. & Kit. ex Willd. Linn. Sp. Pl. ed. 4. 3(1): 466. 1800. [Waldst. & Kit. Pl. Rar. Hung. 1: 94. ¢. 91. 1802.| Lectotype, Hungary. In arenosis sterilissimis Hungariae (pr, Herb. Waldstein no. 502369). Subsection Inflata, subsect. nov. TYPE species: Alyssum tortuosum Waldst. & Kit. ex Willd. Silicula dehiscentia, pedicello rigido suffulta, valvis fere inflatis (aequali- bus vel inaequalibus) vel raro subcompressis non undulatis vel conspicue nervosis. 80 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Alyssum discolor Dudley & Huber-Morath, sp. nov. Holotype, Turkey, Huber-Morath 11755 (uM); isotype (E). Pr. V, Fics. 1-11. Affinis A. syriaco Nyar. et A. chondrogyno Burtt, sed ab omnibus specie- bus habitu altiore et magis diffuso, inflorescentia maiore ramis longioribus divergentibus provisa, foliis maioribus petiolatis longioribus, et fructibus late ellipticis vel obovatis coriaceis rugosis glabris basi longe attenuatis stipitatis differt; a priore appendice diversa filamentae longae recedit; a posteriore (cui probabiliter magis affinis) forma et amplitudine fructuum distinguitur. Planta perennis, pulchra, erecta, suffruticosa, supra basin multiramosa, ramis tenuibus subflexuosis. Caules floriferi erecti, 25-40 cm. longi, parce foliati, basi flavescenti-rubri, pilis stellatis parcis multiradiatis brevibus inaequalibus tecti. Surculi steriles erecti, e basi lignosi caulium floriferorum ascendentes, basi efoliati, 5-10 cm. longi, supra medium foliati, indumento denso cano obtecti. Folia conspicue bicolorata; pagina superior eiroyitens pilis stellatis punctatis minutis appressis multiradiatis Sparsis vestita pagina inferior indumento denso albo-cano e pilis stellatis 15-20 radi atis tenuibus delicatulis appressis punctatis 0.2—0.3(—0.4) mm. diametro com- posito vestita. Folia caulium floriferorum divergentes rare subhorizontales, superne sensim decrescentia; inferiora obovata, spatulata, obtusa (15—) 20-35 mm. longa, (4—)6—-8 mm. lata, petiolis 5-10 mm. longis; superiora oblanceolata vel anguste obovata, breviter petiolata (7—)8—-10(—15) mm. longa, 1-2 mm. lata; summa subbracteata, oblanceolata, acuta, 8—-10(-—15) mm. longa, 1-1.5 mm. lata. Folia caulium sterilium stricte divergentia interdum deflexa, obovata, spatulata, obtusa, 1-20 mm. longa, 3-5 (-6) mm. lata, petiolis 4-8 mm. longis. Corymbi 7-15(-18) cm. alti et lati, e ramis multis divergentibus vel stricte ascendentibus 3—5(—8) cm. longis compositi; inflorescentiae ultimae subumbelliformes vel in statu fructifero elongatae. Pedicelli tenues saepe filiformes, divergentes, 4—6.5 mm. longi, pilis stellatis appressis punctatis minutis multiradiatis valde parce muniti. Sepala membranacea, manifeste nervosa, lanceolata vel anguste ovata, basi subsaccata, obtusa, 3 mm. longa, | mm. lata, indumento pilis stellatis vel multis ca. 0.2 mm. diametro radiis saepe ramosis. Petala obovata, clavata, integra vel emarginata vel subemarginata, 3.5—4 mm. longa, 1 mm. lata, glabra. Filamenta longa 3-4 mm. longa, haud alata, appendice uni- dentata 1—1.5 mm. longa, basi connata. Filamenta brevia 2—2.5(-—3) mm. longa, appendice libera lineari-lanceolata acuta vel minute bi- vel tridentata 1.5—2 mm. longa. Glandulae inconspicuae, minutae. Antherae 0.8-1 mm longae, auriculatae, apice acutae vel subobtusae. Ovarium 1-1.5 mm. unum per loculum. Stylus erectus tenuis, 2—2.5 (—3) mm. longus, glaber, stigmate magno capitato 0.2-0.3 mm. diametro provisus. Silicula late elliptica vel obovata (rare orbiculata) (5.5—)6—6.5 mm. longa, (2.7—)3—4 mm. lata, stipitata, basi longissime attenuata, stipite (0.5—)1—1.5 mm. longo, valvis coriaceis aequaliter inflatis, glabris manifeste rugosis promi- 1964] DUDLEY, STUDIES IN ALYSSUM 81 nenter nervosis. Semen immaturum ut videtur anguste alatum. Fl. Apr.— May, fr. May—June DISTRIBUTION AND HABITAT: Rare endemic of southwest Anatolia found on limestone and sandstone cliffs and scree (rarely serpentine); often associated with Pinus brutia and Smilax; alt. 0-300 m. Turkey. C2: Prov. Mugla, Marmaris, 30 m., Pinetum brutiae on serpentine, 24 Mar. 1956, Davis 25246 (£); Hisaroénii, Marmaris-Datca, 80-100 m., saxatile on dry sandstone rocks, uncommon, 6 June 1962, Dudley (D. 35410) (k£); Mugla-Marmaris, 5 km. from Marmaris on descent from high pass, 150-200 m., common saxatile, forming dense thick clumps on limestone cliffs and scree, 3 June 1962, Dudley (D. 35391) (£); hills on south side of Marmaris village, 100 m., common forming dense clumps under Pinus brutia and Smilax, 7 June 1962, Dudley (D. 35458) (£). C3: Prov. Antalya, Bucht von Adrasan am Siidfuss des Tahtali dag (as “Calbali dag”) zwischen Cirali und Finike, Macchie auf Kalk, 0-300 m., 27 May 1950, Huber-Morath 11755 (HM; E). Alyssum discolor is in the complex of perennial species including A. corsicum Duby, A. masmenaeum Boiss., A. syriacum Nyar., A. troodi Boiss., and A. chondrogynum Burtt, all of which have glabrous or glaucous fruits (rarely papillose and never with an indumentum of stellate hairs) and broadly spathulate or obcuneate, often bicolorous, sterile shoot leaves. The new species, Alyssum discolor, differs from its allies, A. syriacum and A. chondrogynum, in being a more spreading plant, larger in all parts, and with broadly elliptic or obovate, rugose, basally long-attenuate and stipitate fruits. Furthermore, A. discolor may be separated from A. syria- cum by the long basally connate appendage of the long filaments; and from A. chondrogynum, to which it is probably most closely related, by the shape and size of its fruits which are lacking in papillae A recent gathering of Alyssum discolor (Dudley, D. 35410) has broadly elliptic fruits somewhat larger than the obovate ones of the type material (PL. V, Fics. 1 & 2). In all other characters, however, this collection corresponds exactly with the type gathering of Huber-Morath. An addi- tional gathering (Dudley; D. 35458), collected in close proximity, has fruits exactly matching those of the type specimen. These collections in- dicate that the shape and size of the fruits of this species are subject to marked variation. This beautiful chamaephyte is endemic to only a few localities in south- western Anatolia and exhibits a preference for calcareous situations and low altitudes. One gathering (Davis 25246) indicates ‘‘serpentine,” but it is notable that these specimens are poorly developed and somewhat de- pauperate compared to populations from calcareous substrates. Most of the plants examined in the field formed large, woody, and saxatile clumps on limestone cliffs near the sea or on calcareous rubble under Pinus brutia. Alyssum davisianum Dudley, sp. nov. Holotype, Turkey, Davis 36799 (E); isotypes (A, HM). Pr. I, Fics. 14-22. 82 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLV Affinis A. sibirico Willd. sed in omnibus partibus minor, fructibus glabris, foliis angustioribus, indumento foliorum pauciori dissimili differt. Planta perennis, caespitosa, 4-10 cm. alta et lata, in sicco lutescens, e basi multiramosa aphylla, lignosa, rubra, cicatrice prominenti foliorum, ex toto indumento virescenti vel cinereo, e pilis stellatis appressis puncta- tis 4-6 radiatis radiis ramosis aequalibus 0.3—-0.6 mm. diametro composito. Caules floriferi tenues, ascendentes vel erecti, 4-8 cm. longi, parce foliati, ad basin efoliati, pilis stellatis radiis -- longis ramosis sparse vel dense tecti. Surculi steriles apici dense foliati, e basi lignoso caulium floriferorum ascendentes, basi efoliati, 1-2.5 cm. longi. Folia caulium floriferorum oblanceolata, obtusa vel subacuta, descrescentia summam versus, post anthesim decidua, involucrata, flavo-virescentia, 7-10 mm. longa, 1-2 mm. lata. Folia surculorum sterilium \ineari-oblanceolata vel spatulata; in- feriora superioribus longiore, 5-8 mm. longa, 1—1.5 mm. lata; superiora dense imbricata, 3—5 mm. longa, 0.5—-1 mm. lata. Corymbi ramosi, parvi- flori, ca. 2~5 cm. longi et lati, ramulis stricto ascendentibus vel patentibus, corymbis secundariis subumbelliformibus. Pedicelli tenues, rigidi, diver- gentes vel horizontales, rare deflexi, inter se 1-2 mm. distantes, 3-4 mm. longi, pilis stellatis appressis vel subappressis parce vel copiose provisi. Sepala decidua, membranacea, ad apicem carinata, cucullata, ovata, obtusa, 1—1.5 mm. longa, 0.5—0.8 mm. lata, anguste hyalino-marginata, pilis stella- tis sparsissimis provisa. Petala in sicco aurea, obovata, integra vel retusa, in unguem sensim attenuata, cae vel solum in ungue pilis stellatis pro- visa, 2 mm. longa, 0.5—0.7 mm. lata. Filamenta longa ca. 1 mm. longa, recurva ; appendix bidentata aA acuta, in dimidio inferiore ad filamentum connata, apice libera, 0.3-0.5 mm. longa. Filamenta brevia 0.5—-1 mm. longa, recurva; peer re libera, lanceolata, filamentum subaequans, apice bidentata vel acuta, 0.2-0.3 mm. lata. Glandulae conspicuae, globosae. Antherae minutae, 0.3-0.6 mm. longae. Ovarium orbiculatum, emargina- tum, 0.5—0.7 mm. longum et latum, glabrum sed minute papillosum, trans- lucens, flavo-viride. Stylus glaber, tenuis, rigidus, 0.5-1 mm. longus, stigmate leviter dilatato provisus. Silicula ut videtur uniseminata, orbicu- lata, emarginata, 2-3 mm. longa et lata, valvis membranaceis prominenter nervosis glabris laevigatis valde inaequaliter inflatis. Ovuwlum unum per loculum. Semen immaturum non alatum, FI. June—July. Turkey. B2: Prov. Kutahya, Murat dag, above Gediz, above Kesit Sogit, 1900 m., rocky igneous slopes, flowers bright yellow, 5 July 1962, Davis 36799 (A, E, HM); between Kesit Sogiit and Hamam, 1400 m., on metamorphic rocks, 5 July 1962, Davis 36862 (A, E). Alyssum davisianum does not appear to be closely allied to any known species in sect. ODONTARRHENA, although it resembles A. sibiricum Willd. in the orbicular and extremely asymmetrically inflated fruit which has an S-shaped configuration in cross section. It differs strongly from A. sibiricum and the closest allies of that species (A. borzaeanum Nyar. and A. caliacrae Nyar.), in the reduced cushion-forming habit, the prominent 1964] DUDLEY, STUDIES IN ALYSSUM 83 lignose base with copious and raised leaf scars; the much shorter and fewer-fruited corymb; the smaller and always glabrous fruit; the smaller and linear leaves; the indumentum of few-rayed, minute and appressed stellate hairs; and the yellowish-green aspect of the plant. Furthermore, A, davisianum, as a restricted endemic, has a preference for igneous or metamorphic substrates, whereas A. sibiricum is an extremely widespread and polymorphic species usually found in calcareous situations. This new species is named in honor of the collector, Dr. P. H. Davis, of the University Department of Botany, Edinburgh, Scotland, whose energy and careful attention as the author’s supervisor has always been inspiring. Alyssum huber-morathii Dudley, sp. nov. Holotype, Turkey, Khan, Prance & Ratcliffe 256 (£); isotype (kK). PL. V, Fics. 12-23. Affinis A. fallacino Hausskn. (syn. A. baldaccii Vierh. ex Nyar.) et verosimiliter A. murali Waldst. & Kit. (sensu 4. chlorocarpum Hausskn.) sed ab ambobus habitu valde suffruticoso, foliis caulium fertilium et steril- ium manifeste conduplicatis et valde argenteis, fructibus minoribus ellip- ticis vel anguste obovatis indumento parciore differt; a priore corymbis ultimis umbelliformibus (haud elongatis), pedicellis brevioribus tenuioribus recedit; a posteriore stylo tenuiore, seminibus angustissime alatis (haud late alatis) divergit. Planta perennis, valde suffruticosa, a basi multiramosa. Caules floriferi stricti-ascendentes, parce foliosi, 18-25(—35) cm. alti, a basi rubro- purpurei indumento parce summam versus magis conspicuo. Surculi steriles dense foliati, erecto-ascendentes vel laxe patentes, 8-16 cm. longi. Folia caulium floriferorum lineari-oblanceolata, acuta vel subobtusa, con- duplicata, recurva, virescentia, post anthesin decidua, 10-15(—20) mm. longa, 1.5—2 mm. lata. Folia caulium sterilium lineari-oblanceolata, acuta, manifeste conduplicato-sulcata, (4—)10—20(—25) mm. longa, 1—1.5(—2.5) mm. lata, indumento dense lepidoto e pilis albis argenteis minute punctatis 15—20(—30) radiatis multo ramosis 0.4-0.5 mm. diametro composito. Corymbi constricti umbelliformes, (2.5—)8-10 cm. longi et lati, ramulis stricte ascendentibus, floribus in sicco pallide flavis. Pedicelli in statu fructifero rigidi, divergentes, ascendentes, 2.5—4.5 mm. longi, indumento appresso eo fructuum similiter dense obtecti. Sepala 1-1.5 mm. longa, 0.5— 0.6 mm. lata, membranacea, cucullata, anguste ovata, obtusa, margine angusta hyalina, pilis stellatis solum 1—7 munita. Petala 2—2.5 mm. longa, (0.5—)0.6-0.7 mm. lata, clavata vel obovata, integra vel emarginata, glabra vel raro lamina pilis stellatis munita. Filamenta longa 2—2.5 mm. longa, appendicibus 1—1.5-plo longiora; appendix bidentata vel ad apicem minute multidentata, in dimidio inferiore ad filamentum connata, parte libera 0.2—0.3 mm. longa. Filamenta brevia 1-1.5 mm. longa; appendix libera filamentum subaequans, apice acuta vel bidentata. Glandulae minutae, haud conspicuae. Antherae minutae, 0.4-0.5 mm. longae. Ovarium obovatum, 1—1.5 mm. longum, 0.5—0.8 mm. latum, sursum apicem 84 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV dense piliferorum. Stylus in statu florescenti tenuis, glaber, 1.2-1.6 mm. longus, stigmate globoso provisus. Silicula uniseminata, elliptica vel an- guste obovata, (2—)3-3.5(—4) mm. longa, 1.5—2 mm. lata, pallido-vires- cens, obtusa vel subacuta, valvis inaequaliter inflatis, indumento parce apicem versus magis conspicuo e pilis stellatis appressis minute punctatis 0.2-0.3(—0.4) mm. diametro 4-8 radiatis multo ramosis composito obsitis. Ovulum unum per loculum. Semina (0.5—)1—1.5(—2) mm. longa et lata, pallide brunnea, anguste alata, 0.05—0.1 mm. diametro. Fl. June, fr. July. Turkey. C2: Prov. Denizli, Suleymanlar deresi between Acipayam and Abbas Koy, 18 July 1947, Davis 13461 (BM, E, K); Prov. Burdur, Yesilova-Denizli, 20 miles from Yesilova, 900 m., Quercus scrub and neglected fields, common in one area, 1 June 1962, Dudley (D. 35324 & 35325) (A, £). C3: Prov. Antalya, Antalya-Finike, 1350 m., stony ground, flowers yellow, 26 July 1960, Khan, Prance & Ratcliffe 256 (£, K); dist. Kemer, (Lycia) between Ovacik yayla on Teke Dag and Bogut yayla near Calbali Dag, dominant on metamorphic slopes, 13 July 1949, Davis 15225 (A, BM, E, K). The affinity of Alyssum huber-morathi with any species occurring in Turkey is difficult to assess. Considering all representatives of sect. ODONTARRHENA, this new species appears most closely allied to A. fal- lacinum Hausskn., a rare species found in Crete and Greek Thessaly. In its habit, A. huber-morathii resembles A. chlorocarpum, a minor variant of A. murale Waldst. & Kit. which occurs in the Pindus region of northern Greece. The much woodier habit, the very silvery, linear and always conduplicate leaves, and the smaller, usually elliptic fruits with sparser indumentum distinguish Alyssum huber-morathii from A. fallacinum and A. murale. In addition, the new species may be differentiated from A. fallacinum by its umbellate fruiting branches and shorter, more fragile pedicels, and which also has conduplicate leaves and, frequently, similar fruits and indumentum. However, A. cypricum differs in its smaller spathulate- obovate leaves, usually larger emarginate or truncate fruits with a dense indumentum; larger petals, wingless seeds and a distinctly cushion-forming type of growth. The populations of Alyssum huber-morathu are frequently quite large and show a preference for serpentine or metamorphic rubble, but are dis- tributed over a relatively limited range in southwestern Anatolia. It is a pleasure to name this distinct species in honor of Dr. A. Huber- Morath, of Basel, Switzerland, a friend and untiring worker on Anatolian Flora, who put his voluminous Alyssum collection at my disposal. Alyssum pateri Nyar. subsp. prostratum (Nyar.) Dudley, comb. & stat. nov A. prostratum Boiss. & Huet ex Nyar. Bul. Grad. Bot. Cluj 18: 98. fig. 3, 1. 5, fig. 5, n. 4, t. 4, f. 4. 1938 (!). Holotype, Turkey, B8: Prov. Erzurum, 1964 | DUDLEY, STUDIES IN ALYSSUM 85 Armenia, Circa Erzerum, Apr. a June 1853 (Bayburt- rege, Kocha- pinar, 1829-2134 m.) gag mn, (JE); isotypes (BM, G, GH, A, reli pet op. cit. 93. fig. in . 6, t. 6, f.9. 1938 (!). ee Turkey, : . Sivas, Siete (Divriti), 1000 m., June 1893, Bornmiiller 3247 . E, non ” vidi) isotype A. erzerumi Nyar. op. cit. = syn. (!). DISTRIBUTION AND HABITAT: Replacing the typical subspecies of central Turkey in the Armenian Highlands and found in disturbed areas, corn fields, gravelly plains, and south-facing igneous and calcareous slopes and screes; alt. 1350-2591 m. Fl. June—July. This taxon was originally described as a species by Nyarady, but examination of the type material, its duplicates, and additional gatherings shows that it has a maximum of three differential characters which can separate it from Alyssum pateri. These are: the wings on the long filament almost as long as the filaments; prominently dentate, retuse petals; and styles with a sparse indumentum. Ny4arddy himself stated that A. pros- tratum was Closely allied to his A. pateri subsp. squarrosoramosum, now regarded as a synonym of A. pateri subsp. pateri. Though subspecies prostratum replaces the typical subspecies in the Armenian Highlands, their distributions overlap in Turkish Kurdistan. A few intermediates occur in the eastern and southeastern range of the typical subspecies Nyarady allies Alyssum divrikii to the Iranian A. lanigerum; however’ the type of indumentum and the flower and fruit morphology of A. divrikii indicate that it should be treated as synonymous with A. pateri subsp. prostratum., Alyssum eee Boiss. & Hausskn. subsp. flexibile (Nyar.) Dudley, stat. no A, nee Pha as Grad. Bot. Cluj 7: 157. pl. 18, t. 5, f. 122, t. 6, f. 21, , f. 33. 1927(!). Holotype, Syria. Circa Zebdaine pr. ae scu a inter saxa ine inter Bludam et ome 1896 m., 17 June 1855, Kotschy 147 (cL, non vidi); isotypes (BM, 6, be A. alpestre L. [var.] 8B. ie esate Boiss. Fl. Or. 1: 268. (oe parte quoad plantam a Kotschy lectam (No. 147)(!). A. surculosum Shott & Kotschy ex Nyar. op. cit. 150. t. 7, f. 14, t. 9, f. 26('). Lectotype, Turkey, C5: Prov. Icel, Bulgar dagh, Gusguta valley (nr. Kiilek bog.), 2508 m., 26 June 1853, Kotschy 10 (w); isolectotypes (G, kK). A. venustum Nyar. var. rosulatum Nyar. f. ari thine Anal. Acad. R Pop. Rom. Sect. Stiinte Geol. Geog. Biol. ser. A. mem. 3. 1: 61. 1949(1). Lectotype, Turkey, C5: Prov. Adana, ane Beret (nr. Pozanti), 3000 m., July 1909, Siehe 309 (JE); isolectotypes a A, condensatum var. typicum Nyar. op. cit. 72. t. 2, f. _ (!). Holotype, Turkey, B6: igs ae Berit dag, 2338-2743 m., , Aug. 1865, Haussknecht s.n. (JE); isot A. condensatum var. Se eae (Nyar.) Nyar. op. cit. 73 (‘). DIsTRIBUTION AND HABITAT: In Anatolia from the Cilician Taurus, Anti-Taurus and rarely in Turkish Kurdistan. Also known from Syria 86 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLV and Lebanon, A montane and alpine taxon on limestone scree and slopes; often in Pinus nigra woods, and mixed Abies and Cedrus forests; alt. (1400—) 2000-3000(—3500) m. Fl. June—July. The holotype of subsp. flexibile (given full specific rank in 1927 and varietal status under Alyssum condensatum in 1949 by Nyarady) was re- ferred by Boissier to his polymorphic and indefinable taxon 4. alpestre L. [var.] 8. suffrutescens. The typical expression of subsp. flexibile is well distinguished from subsp. condensatum, with which it is partially sympatric; but intermediates occur in the area of overlap. The major features differentiating subsp. flexibile from the typical subspecies are the wider petals with an indumen- tum, and the fruits‘ with usually attenuate apex and sparser and easily displaced indumentum. Subsection Compressa Dudley, subsect. nov. Typr species: Alyssum murale Waldst. & Kit. Pl. Rar. Hung. 1: 5. ¢. 6. 1799 (1) A subsect. /nflatis siliculis compressis vel subinflatis -- undulatis semper conspicue nervosis differt. Series Integra Dudley, ser. nov. TypE species: Alyssum murale Waldst. & Kit. Silicula marginibus apteris integris pedicellis rigidis patulis suffulta. Alyssum murale Waldst. & Kit. Pl. Rar. Hung. 1: 5. ¢. 6. 1799. Lecto- type, Romania. In muris vetustae arcis Devae in Transylvania atque in rupibus apricis vicini Banatus (pr, Herb. Waldstein N. 495746a). (Authentic specimens originally from the Waldstein herbarium are present in the herbaria of BM and GH.) Due to the extreme polymorphism of Alyssum murale, approximately 40 specific and infraspecific epithets have been proposed for minor variations within the taxon which is now designated as subsp. murale. The full synonymy of A. murale will be published later. When considered through- out the whole morphological distribution pattern, A. murale seems to represent one highly variable and actively evolving genetic unit. Within this species, however, there are six reference points of variation which are morphologically and geographically correlated, and merit taxonomic dis- tinction. The following synopsis cites the diagnostic characters and dis- tribution of the recognized infraspecific taxa: 1. Subsp. murale: fertile stems few, ascending or lax; stellate hairs on fruits + dense, 0.2-0.5(-1) mm. diameter; appendages of short filaments always free: seed wings 0.2-1 mm. wide. Widespread in central and southern Europe, the Levant, and extending as far east as Iraq, Caucasia, and Trans- caspla. a) Var. murale: vegetative indumentum bicolorous with equal and + short- rayed, appressed stellate hairs, 0.2-0.5 mm. diameter; seed wings 0.4—1 1964 | DUDLEY, STUDIES IN ALYSSUM 87 mm. wide; leaves flat, midvein on upper surface not conspicuously de- pressed, uppermost cauline leaves not bracteate or involucrate. Distribu- b) Var. alpinum Boiss. ex Nyar.: vegetative indumentum and leaves as in var. murale; seed wings 0.2-0.3 mm. wide; uppermost cauline leaves bracteate and involucrate. Predominantly ee. pitas the Cilician Taurus to Lazistan and extending into central Cauca Var. pichleri (Velen.) Dudley: vegetative Hee aes bicolorous with stellate hairs 1-1.5 mm. diameter of coarse, long, strigose and unequal rays; seed wings and leaves as in var. murale. Forming local populations in Bulgaria, Romania, pate Greece and Yugoslavia. Var. haradjianii (Rech. f.) Dudley: vegetative ee concolorous with appressed equal- ae stellate hairs, 0.2—0.5 diameter; seed wings as in var. murale; leaves frequently eric midvein on upper surface conspicuously depressed, uppermost cauline leaves not bracteate or involucrate. The Amanus in southern Anatolia, Syria and Lebanon. bsp. stojanoffii (Nyar.) Dudley: fertile stems many, decumbent or de- flexed; stellate hairs on fruits sparse 0.2 mm. penne ee of short filaments always adnate to filament; seed wings ca. 0.2 mm. in diameter or less. Rare on the Greek island of Samothrace eee in the eta Rhodope Mountains of Bulgaria. Qa Q bo Nn oc wm 1-) Var. pichleri (Velen.) Dudley, stat. nov. A. pichleri Velen. Fl. Bulg. 38. 1891 (!). Holotype, Bulgaria. In collibus ad Philippoplin (Plovdiv), 1890, Pichler s.n. (prc). DISTRIBUTION AND HABITAT: Rare on calcareous and siliceous sub- strates in Yugoslavia, northern Greece, Bulgaria, Romania and (?) Tur- key-in-Europe; alt. (120—)400—900 m. Fl. May—July. A number of speci- mens collected from the Dalmatian coast in Yugoslavia are intermediate between var. pichleri and var. murale in regard to the size and form of the stellate hairs. Recognition of this taxon at specific, or even subspecific rank, cannot be recommended, because its larger stellate hairs (with unequal and divergent rays) is the only character (cf. Nyarady, 1927, p. 84) which will dis- tinguish it from the typical expression of Alyssum murale. The stability of the configuration of its stellate hairs throughout the rather limited range of its distribution (excluding Yugoslavia) make it seem advisable to assign varietal status to A. pichleri, rather than to recognize it at any other rank Var. haradjianii (Rech. f.) Dudley, comb. & stat. nov. A. pune so f. Ark. Bot. II. 5: 172. 1959 (!). Holotype, Turkey. D6 Hatay: Cassius (Akra dag) 300-700 m., June 1909, Haradjian 3021 ie ee s (E, K, W A. constellatum sensu Nyar. Anal. tae eas Pop. sae ae Stiinte Geol. Geog. Biol. ser. A. mem. 3. 1: 9 —non Boissi 88 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV The characters of Alyssum haradjianii by which it may be separated from the typical variety of A. murale are not of an order indicating taxo- nomic distinctness beyond the rank of variety. The height of the fertile stems of the Haradjian gatherings, which caused Rechinger (1959) to interpret this taxon as intermediate between Nyarady’s artificial categories “Hymiliories” and “Elatiories,” is not constant in the other specimens of the variety. The plants may be as tall or taller than many collections of var. murale. The small local populations of var. haradjianii are sympatric with var. murale in the Amanus of southern Turkey and in montane Syria and Lebanon. However, the concolorous and frequently conduplicate leaves, which always have conspicuously depressed mid-veins, of var. haradjianii permit it to be separated from var. murate. Subsp. stojanofhii (Nyar.) Dudley, comb. nov. A. pichleri subsp. stojanofii Nyar. Repert. Sp. Nov. 27: 395. ¢. 103, f. 3. 1930. Holotype, Bulgaria. Macedonia bulgarica, in graminosis alpinis sub cacumine Kalabak, m. Belasica, alt. c. 2000 m., 20 July 1920, Stojanoff s.n. (som, non vidi). A. degenianum Nyar. Bul. Grad. Bot. Cluj 7: 87. ¢. 3, f. 45, 46. 1927 (!). Holotype, Greece. Samothrake, Monte Phenogari, 28 June 1890, Degen Sn. ; A. degenianum {. subcaespitosum Nyar. loc. cit.(!). Holotype, Greece. Samothrake, Monte Phenogari, 28 June 1890, Degen s.n. (w). DISTRIBUTION AND HABITAT: Montane in grassy meadows on the Greek island of Samothrace and the southern Rhodope Mountains in Bulgaria; alt. (600—) 1000-2000 m. FI. June. Although the type gathering of this subspecies has not been seen, its original description and accompanying photograph clearly establish it as the same taxon which Nyarady had previously regarded as Alyssum degenianum. Additional support for this conclusion is that specimens of A. degenianum, determined by Nyarady, have been collected from very near the type locality of subsp. stojanoffii. This close proximity and the correlation of the specimens with the literature establishes without much doubt that the two taxa are conspecific. It is evident that because 4. pichleri differs from A. murale only by its very large (1-1.5 mm. in diameter) and coarse stellate hairs, it was necessary to reduce A. pichleri to varietal rank. Though subsp. stojanoffii was assigned originally to A. pichleri, the characters which distinguish it from all the other infra- specific taxa of A. murale are of a significance to merit subspecific recogni- tion. Therefore, it would be unrealistic to reduce subsp. stojano ffir to the rank of form under var. pichleri. The specimens of A. degenianum cited by Nyarady (1949) from Lebanon, all refer to A. murale var. haradjianit. Subsection Samarifera Dudley, subsect. nov. TypE species: Alyssum samariferum Boiss. & Hausskn. in Boiss. Fl. Or. 1: 272. 1867. Holotype, Turkey, C6: Prov. Maras, in monte Berytdagh Cataoniae 1964] DUDLEY, STUDIES IN ALYSSUM 89 (Berit dag), 1829-2338 m., 8 Aug. 1865, Haussknecht s.n. (Gc); isotypes (BM, K, W). A subsect. /nflatis siliculis indehiscentibus samaroideis undulatis pendulis papyraceis semper compressis pedicello capillari flexuoso deflexo suffultis differt. Alyssum peltarioides Boiss. subsp. virgatiforme (Nyar.) Dudley, Stat. nov. A. peltarioides var. virgatiforme Nyar. Anal. Acad. Rep. Pop. Rom. Sect. Stiinte Geol. Geog. Biol. ser. A. mem. 3. 1: 84. 1949 (1). Lectotype, Turkey, B7: Prov. Erzincan, Sipikordagh, 28 July 1890, Sintenis 3126 (JE); isolectotypes (BM, E, G, K, W). DISTRIBUTION AND HABITAT: North and central Anatolia, Armenian Highlands, Anti-Taurus and Amanus; on north-facing igneous slopes; alt. 914-2600 m. FI. June-July. KEY TO SUBSPECIES pina Rp mie ately d da A earnapees ek a ale teeters ep eurane aac: Subsp. peltarioides. Inflorescence widely spreading, strongly branched, 10-20 cm. long, and wit 0 (or more) fruits; stems arcuate, ascending or erect, 25-50(-75) cm. NOON ea thas Ae e sa aan oc Wet el einige. a Subsp. virgatiforme. The characters distinguishing the infraspecific taxa of this species are constant over its whole range and are coupled with altitudinal and geo- graphical preferences. The most westerly stations of the species are com- posed entirely of subsp. virgatiforme; however, in some localities in the eastern part of the range (Armenia), where the two subspecies are sym- patric, they are confined to different altitudes and have different flowering times. The typical subspecies, which flowers later than subsp. virgati- forme and has a very reduced habit, is found near melting snow at altitudes from 2000 to 3580 meters. The consistency of their morphological expres- sion in conjunction with the facts of geographical and altitudinal replace- ment indicates that for these taxa subspecific rank is more realistic than varietal rank. Alyssum caricum Dudley & Huber-Morath, sp. nov. Holotype, Turkey, Huber-Morath 12824 (uM); isotypes (A, E). Po, 11, ries. 1-11. A. floribundum Boiss. & Bal. var. latifolium Nyar. Anal. Acad. Rep. Pop. Rom. Sect. Stiinte Geol. Geog. Biol. ser. A. mem. 3. 1: 82. 1949 (!). Lectotype, Turkey, Huber-Morath 5566 (uM). Affinis A. peltarioidi Boiss. et A. trapeziformi Bornm. ex Nyar., sed ab ambobus fructibus dissimilibus leviter obcordatis vel late obovatis, stylo longiore tenuiore, habitu magis ramoso, surculis sterilibus crassioribus 90 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV fere maioribus differt; insuper a priore fructu minus undulato obscure nervoso et ramulis corymbis longis rigidis ascendentibus minus laxis recedit; a posteriore forma pilorum stellatorum dissimili divergit. Frutex perennis, supra basin multiramosus, basi lignosa efoliata 3-13 cm. alta. Planta ex toto (imprimis folia caulium sterilium) indumento albo-argenteo e pilis stellatis radiis (10—) 15-25 (—35) undulatis delicatulis longis vel brevibus saepe ramosis composito dense vestita. Caules florifert erecti-ascendentes, e basi parce ramosi, parce foliati, 17-30(—37) cm. alti, basi rubro-purpurei, pilis stellatis appressis multiradiatis superne densiori- bus tecti. Surculi steriles erecti, e parte inferiore caulium floriferorum ascendentes, rare longiores caulibus floriferis (2.5—)5—13 cm. alti (subinde multiramosi et efoliati, ramis 25-30 cm. altis), dense foliati. F olia caulium floriferorum anguste oblanceolata vel oblongo-obovata, acuta vel obtusa, 10-15(—20) mm. longa, 1.5—2(—3) mm. lata, petiolis 3-5 (—7) mm. longis, in apice bracteis inflorescentiae simulantibus, 7—10(—15) mm. longa, 1-2(—3) mm. lata. Folia caulium sterilium bene evoluta bicolora, obovato-oblanceolata, spatulata, basi cuneata, obtusa vel rare subacuta, (7—) 10-20 mm. longa, (3—)5—7 mm. lata, petiolis 5-10 mm. longis; pagina superior indumento 3—4-plo minore eo paginae inferioris et radiis pau- cioribus brevioribus obtecta. Corymbi magni, multiramosi, 5—10(-15) cm, alti et lati, ramulis ultimis rigidis stricte ascendentibus 5—10(-15) mm, longis corymbis ultimis umbelliformibus fructibus 4-10(-15). Pedi- celli fragillimi, tenues, flexuosi, deflexi, (3—)4-5(-S.5) mm. longi pilis stellatis appressis vel subappressis 10—15-radiatis pauce vel dense obsiti. Sepala lanceolata vel ovata, acuta, apice cucullata, 2—2.5 mm. longa, (0.5—)0.8-1 mm. lata, margine angusta hyalina et parcis vel copiosis pilis stellatis multiradiatis radiis + brevibus provisa. Petala obovata, spatulata, integra vel subretusa, 3-3.5 mm. longa 1-1.5 mm. lata, glabra vel rare in ungue pilis stellatis paucis provisa. Filamenta longa 2.5—-3 mm. longa, ala unilaterali in dimidio inferiore connata, apice libero 1-2 dentato 0.3-0.5 (-0.7) mm. longo. Filamenta brevia 2 mm. longa, appendice lata libera lanceolata vel oblanceolata acuta vel minute denticulata, (1—)1.5—2(—2.5) mm. longa, 0.4-0.5 mm. lata praedita. Glandulae minutae. Ovarium 1-1.5 mm. longum, 0.5-1 mm. latum, glabrum, transparens, infime obcordatum, retusum, subflavo-viride. Stylus glaber, tenuis, inaequaliter insertus, 1-1.5(—2) mm. longus. Silicula uni- vel duoseminata, brunneo- viridia, late obovata vel leviter obcordata, retuse vel rare truncata, basi breve attenuata (5—)6-7.5(-8) mm. longa, 4.5-6(—6.5) mm. lata, saepe pendula, margine parce papillosa imprimis apicem versus, valvis membran- aceis glabris laevigatis haud undulatis ad medium subinflatis inconspicue nervosis. Ovulum unum per loculum. Semina oblonga, compressa, 2.5 mm. longa, 1.5 mm. lata, ala 0.1-0.2(-0.3) mm. lata. Fl. Apr—June, fr. June—July. Hasirat: Saxatile on serpentine outcrops, cliffs, ledges and scree; often in Quercus scrub and Pinus brutia or P. nigra woods; alt. (40—) 100-900 (1219) m. 1964 | DUDLEY, STUDIES IN ALYSSUM 91 Turkey. C2: Prov. Mugla, Mugla-Fethiye, Kalkgeroll linkes Ufer des Namlam Cay, 42 km. siidéstlich Mugla, 100 m., 20 June 1954, Huber-Morath 12824 (uM; A, E); 8 miles s. of Mugla, 590 m., forming dense compact clumps, common in roadside scree and serpentine outcrops, 29 May 1962, Dudley (D. 35132) (A, E); 25 miles from Mugla, roadside serpentine ledges and scree, 70 m., 29 May 1962, Dudley (D. 35137) (A, £); Mugla-Marmaris, Poterium Macchie, 45 km. siidlich Mugla, 20 km. nordlich Marmaris, 40 m., 19 June 1954, Huber- Morath 12823 (k, HM); 20 km. from Marmaris, 60 m., very dry roadside, and serpentine outcrops, forming dense suffruticose, saxatile, spreading clumps, often leafless up to 2 ft., 5 June 1962, Dudley (D. 35390) (a, E); Kara boyurtleri Koyu, 30 April 1958, Kayacik & Yaltirik sn. (z); Sandras dag above Agla on serpentine, 1219 m., 22 July 1947, Davis 13563 & 13621 (kg); Mugla-Fethiye, rechtes Ufer des Namlam Cay, Macchie, 42 km. siidéstlich Mugla, 7 June 1938, Huber-Morath 5149 (um); Pinetum 116 km. siidéstlich Mugla, 6 km. ob. Golcuk, 300 m., 7 June 1938, Huber-Morath 5566 (HM); Kale Tavas-Mufla, Pinus brutia-Wald, 47 km. nach Kale, 5 June 1938, Huber-Morath 5567 (uM); Mugla-Kale, 35 km. from Mugla, 900 m., serpentine outcrops in deep gorge, open Pinus nigra woods, 28 May 1962, Dudley (D. 35128) (A, E); 30 km. from Mugla, 850-870 m., steepest descent on road, serpentine outcrops and scree, dominant plant in shade of Pinus brutia-Pinus nigra overlap, 28 May 1962, Dudley (D. 35123) (A, £); 30 miles from Mufla, ca. 25 miles from Kale, 460 m., serpentine outcrops and scree, saxatile on cliffs and ledges, 9 June 1962, Dudley (D. 35531) (A, E). Alyssum caricum shows the closest affinity with A. peltarioides and A. trapeziforme Bornm. ex Nyar., from which it is readily distinguished by its distinctive obcordate fruits, shorter styles, extremely woody and branched base (often leafless from the base for a foot or more) and the dense cushion of sterile shoots (often 3-4 feet across). I have observed that Alyssum caricum grows with A. corsicum Duby in the lowland areas of Caria, in southwestern Anatolia but is restricted to serpentine outcrops and completely replaces A. corsicum in montane and alpine regions. Whereas A. caricum has a very restricted distribution of about 100 square miles, A. corsicum has a much wider range in western Anatolia and is primarily a plant of neglected fields and disturbed areas. Nyarady (1949) erroneously described this plant as a variety of Alyssum floribundum (var. latifolium) to which A. caricum is not closely allied, and from which it differs in the shape and indumentum of the leaves and in the strikingly dissimilar fruits. Although a lectotype has been chosen from among the several syntypes of Alyssum floribundum var. latifolium, it certainly seems more satisfactory to describe this plant as a new species, rather than to apply a nomen novum. The description of A. floribundum var. latifolium is built into the structure of Nyarady’s very long key and for this reason lacks the essential details which distinguish A. caricum from A. floribundum or any other species in subsect. SAMARIFERA. In any event, the epithet /atifolium could not be applied at the specific level because of the earlier A. latifolium Vis., a synonym of A. minus ie.) Rothm. 92 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Alyssum pinifolium (Nyar.) Dudley, comb. nov. Triplopetalum pinifolium Nyar. Magyar Bot. Lap. 24: 97. ¢. 1, f. 1-14. 1925 (!). Holotype, Turkey, Al (A): Prov. Canakkale, in monte Ulu Dagh prope Renkoi (Erenkéy) Anatoliae, 24 Apr. 1883, Sintenis 292 (LD, non vidi); isotypes (BM, E, G, K, UW, W). DistripuTION: An Anatolian endemic collected only twice from the Dardanelles. The decision to treat this taxon, originally described as the type of a monotypic genus, Triplopetalum Nyar., as a species of Alyssum was difficult. Apart from the problem of assessing the value of the characters used to delimit Triplopetalum in a generic sense, the material extant in the numerous herbaria visited consisted only of duplicates of the type collection, all without fruits. Unfortunately, the only other collection (Kirk, ©), made many years earlier, was also in the flowering state. It is established by overall resemblances and floral morphology, however, that Triplopetalum falls within a natural species complex containing all the taxa of sect. ODONTARRHENA with indehiscent fruits and brittle, deflexed or sigmoid pedicels. The characters distinguishing Alyssum pinifolium from the other species of the complex are essentially those which characterized the genus Triplo- petalum, i.e., the consistent presence of “‘sphaeroid crystals’? appearing as pellucid dots on the sepals, petals, filaments, and ovary; the unusual needle-like, almost cylindrical leaves; and the petals with basal appendages (whence the name of the genus). That Alyssum lesbiacum (Candargy) Rech. f. (1943) possesses con- sistently appendaged petals, as noted originally by Candargy (‘‘petala . unguiculata basi bialata’’), cannot be ignored when determining the fate of Triplopetalum pinifolium. Without doubt, the petal appendages of A, lesbiacum are homologous with those of Triplopetalum. The samaroid, indehiscent fruits on fragile sigmoid pedicels of A. pinifolium and A. lesbiacum clearly ally both to such species as A. samariferum Boiss. & Hausskn., A. peltarioides, A. dubertretii Gomb., etc. To consider Triplo- petalum as a distinct genus would defeat the purposes of the natural grouping of species within sect. ODONTARRHENA and would cause it to be separated from those species with which it is very closely allied. Further- more, it would be equally unwarranted to regard A. lesbiacum as a species of Triplopetalum, thereby also cleaving it from its closest allies. LITERATURE CITED AND IMPORTANT REFERENCES BAUHIN, J. 1651. Historia plantarum 2. Ebroduni. BAUMGARTNER, J. 1907. Die ausdauernden Arten der Sectio Eualyssum aus der Gattung Alyssum, pt. 1. Beil. Jahresb. Nied.-Ost. Land.-Lehrers. Wiener- 35. . 1909. Ibid. Pt. 3. 36: 1-38. 1964] DUDLEY, STUDIES IN ALYSSUM 93 1911. Pt. 4. Jahresb. Kaiser Franz Josef-Land.-Gymn. Oberrealsch. Baden bei Wien. 48: 1-18. BrrAND, H. 1952. Plantae Turcicae. Ankar Botssier, E. 1842. Plantae Aucherianae orientale enumeratae cum novarum specierum descriptione. Ann. Sci. Nat. II. 17: 150-205. 1867. Flora orientalis. 1. Genéve & Bale. BORNMULLER, J. 1936. Symbolae ad floram Anatolicam. Repert. Sp. Nov. Beih. 89(1): 1-64. BROTERO, F. 1826. Nocgodes botanicas das especies de Nicociana. Lisbon Buscu, N. 1909. Jn: Kuznetsov, Busch & Fomin, Flora Caucasica critica 3(4). Erivan. CANDOLLE, A. P. DE. 1821. Systema naturale 2. Paris. 1824. Prodromus systematis naturalis 1. Paris. Czeczott, H. 1937. The distribution of some species in northern Asia Minor. Mitteil. Konigl. Naturw. Instituten Sofia 10: 43-68. . 1939. A contribution to the knowledge of the flora and vegetation of Turkey, II. Repert. Sp. Nov. Beih. 107: 137-281 Davis, P. H. 1951. Cliff vegetation in the eastern Mediterranean. Jour. Ecol. > 63-93. 1953. Old and new place names used in studies on the Turkish flora. Notes Bot. Gard. Edinb. 22: 587-591. DeceENn, A. 1937. Flora Velebitica 2: 187-190. Budapest. Dub Ley, T. R. 1962. Some new Alyssa from the Near East. Notes Bot. Gard. Edinb. 24: 157-165. GrossHEIM, A. A. 1950. Flora Kavkaza, ed. 2. 4. Baku HAUSSKNECHT, C. 1897. Drei neue Cruciferen-Gattungen der orientalischen Flora. Mitt. Thir. Bot. Ver. 11: 68-76. Hayek, A. 1925. Prodromus florae peninsulae Balcanicae 1. Repert. Sp. Nov. Beih. 30: 383-512. HeEywoop, V. 1961. bi flora of the Sierra de Cazorla, S. E. Spain I. Repert. Nov. 64: 28-7 Hitcucock, A. & ae M. 1929. Standard-species of Linnaean genera of Phanerogamae (1753-54) in International Botanical Congress, Cambridge (England), 1930, “Nomenclature — Proposals by British Botanists,” 110- 112. London. Huser-Moratu, A. 1940. Novitiae florae Anatolicae II. Repert. Sp. Nov. 48: 273-292. 1943. Novitiae florae Anatolicae III. /bid. 52: ie 1943. Ein Beitrag zur Kenntnis der Anatolischen Flora I. /bid. 52: 179-229. Komarov, V. L., et al. 1939. Flora U.R.S.S. 8. Moscow. Lanjouw, J. (ed.). 1961. ue oe code of botanical nomenclature. egnum Vegetabile 23. Utre LINNAEUS, C. 1753. Species aide 2. Stockholm 1756. Flora Monspeliensis . . piemaienes): defert. T. E. Nath- horst. Uppsala. ———. 1759. Systema naturae, ed. 10. 2. Stockholm. 1759. Amoenitates academicae 4. Stockholm. . 1763. Species plantarum, ed. 2. 2. Stockholm. MacnoL, P. 1688. Botanicum Monspeliense, ed. 2. Monspelii MARSCHALL VON BIEBERSTEIN, F. A. 1808. Flora Tauro-Caucasica 2. Charkoviae. 94 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV . 1819. Supplementum flora Tauro-Caucasica 3. Charkoviae. NyArApy, E. J. 1925, iat al novum genus e familia Cruciferarum. ae Bot. Lap. 24: ——_ 6. In schedae sd ae Romaniae exsiccatum.’ ed. Borza, Bul. =a Bot. Cluj 6: 81- 1927. ee ie einige Arten der Section Odontarrhena der Gattung Alyssum, Bul. Grad. Bot. Cluj 7: 1-51, 65-160. Tab. 1-10. 1928. Jbid. 8: on 156. ——, 1929. Ibid. 9: 1-68. . 1930. Neue beri zur Kenntnis der balkanischer Alyssum Arten. Repert. Sp. Nov. 27: 392-395, . 1931. Les | vraies et fausses de l’espéce Alyssum alpestre. Bul. Grad. Bot. Cluj 11: 69-78. . 1932. Die Klarstellung zweier zweifelhafter Alyssum-Arten. Notizbl. Bot. Gart. Berlin 11: 631-635 . 1932. eee einige Westmediterrane Alyssum-Arten. Bul. Soc. Stiinte ane 6: 446-4 . 1938. ai Alyssum- sir und Formen aus der Odontarrhena-Sektion. Bul. ee Bot. Cluj 2-99. ——. 1949, nt aa ata variatonum et formarum sectionis Odon- tarrhenae. Generis Alyssum. Analele Academiei Republicii Populare Romane, Sectia de Stiinte Geologice, Geografice Si Biologice Ser. A. Mem. 3. 1 ae 1-133. Tab. 1-6. ———. 1955. Flora Republicii Populare Romine, ed. Savulescu. 5. Budapest. Nyman, C. F. 1878. Conspectus florae Europaeae 1: Orekro Post, G. E. & Dinsmore, J. E. 1932. Flora of Syria, ie and Sinai, ed. 2. 1. Beirut. RECHINGER, K. H. 1943. Flora Aegaea. Denkschr. Akad. Wien Math. Naturw. 105: xx + 924 1953. Zur Flora von Palastina und Transjordanien. Ark. Bot. II. 2: 271-273. . 1959. Zur Flora von ain Libanon und den angrenzenden tiirkischen Gebieten. Ark. Bot. II. 1-488 1961. Die Flora von gsaen: ‘Bot. Jahrb. 80: 294-382. RoTHMALER, E. 1941. Nomenklatorisches, meist aus dem westlichen Mittle- meergebiet. III. Repert. Sp. Nov. 78. SAVAGE, S. 1945. A catalogue of the Linnaean herbarium. London ScHuLz, O. 1927. Uber me Gattung Gamosepalum Hausskn. Notizbl. Bot. Gart. Berlin 10: 109- 1936. In: Bauer : Prantl, Die Naturlichen Pflanzenfamilien II. 17b: TCHIHATCHEFF, P. 1860. Asie Mineure, Bot. 1(3). Paris TurRILL, W. B. 1935. Alyssum campestre L. Jour. Bot. 73: 261-272 WILLDENow, C. L. 1800. Linnaeus, Species plantarum, ed. 4. 3(1). Berlin. EXPLANATION OF PLATES PLATE I Fics. 1-13. Alyssum blepharocarpum: 1, silicle (side aie < 6; 2-5, silicles alas variability of indumentum (face view), X 6; seed, X 6: 7, sepals, 8, long filaments, X 9; 9, short filaments, “ 9; th petal, x 9; 11 & 12, 1964 | DUDLEY, STUDIES IN ALYSSUM 95 leaves of two different gatherings, X 2; 13, stellate hairs from upper surface of af, 0. Fics. 14-22. A. davisianum: 14, silicle (face view), X 6; 15, silicle (side view), X 6; 16, long filaments, X 9; 17, short filaments, X 9; 18, petal, x 9; 19, sepal, x 9; 20, cauline leaf, poy 21, stellate hairs from upper surface of a cauline leaf, X 40; 22, woody base of plant with a sterile shoot, 1 PLATE II Fics. 1-11. A. caricum: 1, ultimate inflorescence, X 1; 2, silicle (face view), x 6; 3, silicle (side view), ; 4, seed, X 6; 5, long filaments, x 9; 6, short filaments, x 9; 7, petal, X 9; 8, sepals, X 9; leaf of sterile shoot, ge nO} Fics, 12-22. A. trichocarpum: 12, silicle (face view), X 6; 13, silicle (side - 14, pares hair from silicle, * 40; 15, tuberculate hairs from silicles, 12; 16, seed, X 6; 17, long filaments, X 9; 18, short filaments, x 9; 19, petal, x 9: a, Sepa, x 9; 21, leaf, x 2; 22, stellate hairs from upper surface of a alee PEATE Ti Fics. 1-11. Alyssum abe ee 1, silicle (face view), X 6.5; 2, silicle Pei view), X 6.5; : seed ort filament, : 6.5; 5, long filament, a ? 6, petals, ee : ntral view of outer sepal, x 8; 8, dorsal view of inner ae 8; 9 ie hair from upper surface of a an leaf, X 100; 10, stellate hair Fa ‘ilile, 100; 11, stellate hair from ra stem 100. Ics. 12- iveum: 12, se als, X 6.5; 12a, detail of base of sepals show- h Pe ing interdigitating hairs, X 26; 13, short filament, an eo 14, long filament, 5; 15, petal, X 8; 16, young ovary, X 6.5; 1 i xX 33; 18, cauline leaf, & 4.5; 19, lepidote hair from upper surface a ane leaf, X 65; 2 lepidote hair from dorsal surface of sepal, X 65. PLATE IV Fics. 1-12. Alyssum series 1, silicle Mia nade 9; 2, lepidote hair from silicle, X 40; 3,s ment, X 9; filaments, X 9; 5, petal, x 9; 6, sepals, X 9; 7, aN ae of a sepal with stellate hairs, x 9; 8, stigma, Fics. 13-26. A. sulphureum: 13, ultimate inflorescence, X 1; 14, silicle (face view), X 6; 15, silicle (side view), X 6; 16, stellate hair from silicle, x 40; o short filaments, X 9; 18, long filaments, x 9; 19, pet a is 9; 20, sepals, 21, ventral surface of a sepal with stellate hairs, x 9; tuberculate and pe rayed stellate be from apex of a sepal, x 40; 23, nee stellate hair from oo zs a sepal, X 40; 24, seed, x 6; a5 ‘sterile shoot, X 2; 26, cauline leaf, PLATE V Fics. 1-11. Alyssum discolor: 1 & 2, silicles (face views), < 7; 3, silicle on view), X 7; 4, sepal, X 9; 5, petals, x 9; 6, long filament, x 8: 7, shor ; x leaf of a sterile shoot, x 2: 22, ge hair from upper surface of a leaf of a sterile shoot, * 20; 23, seed, x 26. PiaTE I Jour. ARNOLD Ars. VoL. XLV DupLEy, STUDIES IN ALYSSUM Jour. ARNOLD Ars. VoL. XLV Prat_E II DupLEY, STUDIES IN ALYSSUM Piate III Jour. ARNOLD Ars. VoL. XLV STUDIES IN ALYSSUM DUDLEY, Jour. ARNOLD Ars. VoL. XLV PLaTE IV DupDLey, STUDIES IN ALYSSUM PLATE V Jour. ARNOLD Ars. VoL. XLV * tr «*ne © Fe Seeger? + “+ + Ke KS ns pan Aa tre nt RO eg HRN gli oe a “+ DupLey, STupIEs IN ALYSSUM 1964] JAYAWEERA, MUSSAENDA IN THE PHILIPPINES 101 THE RUBIACEOUS GENUS MUSSAENDA: THE SPECIES OF THE PHILIPPINE ISLANDS Don M. A. JAYAWEERA IN PREVIOUS PAPERS! the morphology of the rubiaceous genus Mus- saenda has been discussed, and the species of India and Ceylon have been treated. In the Philippine Islands Mussaenda is represented by a larger number of species than in the areas considered earlier. Many of the Philip- pine species are endemic to a single island or are of very limited distribu- tion; only a few may be found throughout the island chain. The Philippine species seem to have some affinity with those of the mainland of China and of the Malayan Peninsula, and to differ rather remarkably from the species of India in the pubescence characters of the inner surface of the corolla tube. Students of any group of Philippine flowering plants are hampered in their work by the unfortunate loss of the types during World War II. In most cases isotypes are available in one or more of the herbaria of the United States. In such instances a lectotype has been chosen from among them. In some cases it has, however, been necessary to choose a neotype to represent the species. The institutions from which material has been borrowed for my use were cited in the first of these papers. I should like to express again my appre- ciation to the officers in charge of these collections for their courtesy, and to the members of the staff of the Arnold Arboretum who have made publication of these studies possible. The interrelationships of the Philippine species of Mussaenda are so complex that an arrangement by affinities seems impracticable at present. For convenience the key which follows is an artificial one. The species have been arranged in alphabetical order. ARTIFICIAL KEY TO THE SPECIES OF MUSSAENDA IN THE PHILIPPINE ISLANDS A. Calyx lobes deciduous almost immediately after flowering. B. Stipules entire or faintly bifid at apex. C. Corolla tube less than 2 cm. in length. D. Corolla tube campanulate, 1.1-1.3 cm. long, corolla lobes broader than long; calyx lobes lanceolate, 5.5—6.5 mm. long, hairy on both surfaces; berry ovoid, 1.2-1.7 cm. long, scantily pubescent; seeds 1 The rubiaceous genus Mussaenda: Morphology of the Asiatic species. Jour. Arn. Arb. 44: 111-126. 1963. The rubiaceous genus Mussaenda: The species of India and Ceylon. Ibid. 232-267. 102 iS) JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV 0.43-0.56 mm. long with 3-6 foveae in the areoles Me the testa. . Corolla tube tubular, broader at the top, 1.7-2 cm. Tong, corolla lobes broadly ovate, as broad as long; calyx lobes 6-12 mm. long, lanceolate, glabrous or pubescent Nias berry (immature) obo- void, 1 cm. long, glabrescent; seeds 1-0. 83 mm. long with 4-12 foveae in the areoles of the testa. ........... 6. M. chlorantha. C. Corolla tube 2-3 cm. long, glabrous or straggly pubescent; stipules 7.5-16 mm. long; calyx lobes narrowly oblong-lanceolate or linear, m. long, sparsely pubescent; seeds 0.5-0.67 mm. long with 3-6 or 4-12 foveae in the areoles of the testa. |. 10. M. magallanensis. B. Stipules bifurcate from the apex for about % their length. E. Midrib of the leaf sunken on the upper surface. F. Calyx lobes 3-5 mm. long, setaceously pointed; corolla tube slenderly tubular, lobes elliptic, ese on the outer surface; stipules broadly lanceolate, 9-12 mm. long; berry or red- dish strigose- pubescent; seeds 0.6- i 7 mm OR -6) foveae in the areoles of the testa......... : - ie F. Calyx lobes 11-16.5 mm. long, linear with 4 a pana midvein; corolla tube subfalcate, lobes ovate-lanceolate, granular on the inner surface; stipules ovate-acuminate, 4-5.5 mm. long; berry ellipsoid or globu lar, 1 cm. long, glabrous, eo 0.46- 0.6 mm. long, with 4-12 foveae in the areoles of the E. Midrib ae the leaf not sinker: on the upper surface. G. Fruits with persistent, conical, nectariferous discs, nearly glabrous, without lenticels; leaves oblanceolate, 7-20 cm. long, recurved at the apex and decurrently attenuate at base; stipules deciduous, 5-11 mm. long, hairy on both surfaces; calyx lobes lanceolate, 3 mm. long; corolla tube 2.7—4 cm. long; berry ellipsoidally elon- gated, 1.2-2 cm. long; seeds 0.64-0.77 mm. long, with 3-6(-9) foveae in the areoles of the testa... 4. M. attenutfolia. G. Fruits without persistent, conical, nectariferous discs. H. Corolla tube hairy within 4% way down, slender, 2—2.5 cm. long: leaves ovate-elliptic, 6-13.3 cm. long; stipules 7-8 mm. long, hairy on the inner surface; calyx lobes linear-oblong, 3.5—5.5 mm. long, sharply acuminate; berry ellipsoid, 1-1.3 cm. long; seeds spiny, 0.83-0.9 mm. long, with 3-9 foveae in the areoles 18 of the testa. ....... . M. scandens. H. Corolla tube hairy withan Yu's way ap wn. I. Flower buds curved and club shaped; leaves 9-25.5 cm. long, ferruginous-hairy on both surfaces; stipules 6.5-10.5 mm. long, glabrous on the inner surface; corolla tube 2.2-3 cm. long, hirsute with spreading hairs; berry ellipsoid, 1.5— 2.2 cm. long, ferruginous-pubescent; seeds 0.53-0.73 mm. long, with 3-8 foveae in the areoles of the testa. a 20. M. vidalii. I. Flower bute seis curv ved: nor aii ha aped. J. Corolla lobes 10-10.5 mm. long, linear-lanceolate, corolla tube very slender, 2.6— a cm. long, pubescent with short, 1964] JAYAWEERA, MUSSAENDA IN THE PHILIPPINES 103 scattered hairs; leaves 11.5-29 cm. long, pubescent on both surfaces with long, scattered hairs; stipules entire or faintly bifid 4% their length, hairy on both surfaces; berry (immature) obovoid, 1-1.3 cm. long; seeds 0.67- 0.83 mm. long, with 2-8 foveae i in the gan - the testa. Bee C2 aed we hy tied See M. albiflora. : Corolle lobes less than 10 mm. long. K. Leaves glabrous on the upper surface, minutely pubescent or hirsute beneath, recurved and strongly lar, 1-1.6 cm. long, lenticellate; seeds 0.6-1 mm. long with 2-5(—7) foveae in the areoles of the testa re Ne. 4's ans een 15. M. philippica. . Leaves pubescent on both surfaces L. Calyx lobes linear, 6-15 mm. long, hairy; leaves 8.5-30 cm. long, long decurrent on petiole from the base; stipules ovate, 9-17 mm. long; corolla tube 2.5-3 cm. long, hirsute, hairs spreading; berry ellipsoid, 1.4—2 cm. long, scantily pubescent; seeds 0.63-0.73 mm. long, with 2-6 foveae in the areoles Ol Te eRe ic Be hth we ec 3. M. anisophylla. . Calyx lobes lanceolate or oblong-lanceolate. M. Corolla tube infundibuliform, 2—2.2 cm. long, hairs appressed on the outer surface, lobes acuminate; stipules lanceolate, bifurcate from apex more than % their length; calyx lobes 6.7-10.5 mm. Say. slenderly acuminate; berry ovoid or obovoid, 1.2-1.4 cm. long, pubescent; seeds 0.6-0.77 mm. long, with 3-6(—8) foveae in the areoles of the eae eee fe 4 ates g M. palawanensis. Corolla tube tubular, 2- a cm. long, hairs hor- izontally spreading on the outer surface, lobes apiculate; stipules triangular-ovate, bifurcate “a = = lobes 3.2-8 mm. long, acute; berry globose with very large nectariferous scars; seeds 0.67— 0.87 mm. long, with 2-6 foveae in the areoles of the testa. ......... eka th oh 9. M. macrophylla var. brevipilosa. A. Calyx lobes persistent on the ovary after flowering. N. Inflorescence at the same level or below the terminal leaves O. Corolla tube glabrous within, 2.9 cm. long, densely hirsute on the outer surface; calyx lobes short, 1.2-2.5 mm. long, lanceolate; leaves 16-32 cm. long, pubescent on both surfaces; stipules small, slenderly acuminate, bifurcate at the apex, hairy on both surfaces; berry ellipsoid, 1 cm. long, pubescent; seeds 0.64-0.77 mm. long, with 1-4 foveae in the areoles of the testa Se ee 7. M. grandifolia. O. Corolla tube hairy within. 104 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV P. Stipules entire or faintly bifid at apex Q. Calyx lobes 7.7-8 mm. long, linear- subulate; leaves small, 4— 8.3. cm. long, glabrous except on veins beneath; stipules lanceolate, 5—8.5 mm. long, faintly bifid at apex, hairy on the outer surface, ‘ane within at base and apex only; corolla tube 2-3 cm. long, ae on the outer surface except at the base; DOITY MOL SOON. 2k had eee heed 7. M. pin natubensis, . Calyx lobes 10.5 15 mm. poe. narrowly lanceolate; leave 9.5-20 cm. long, glabrate on upper surface, hirsute ee stipules 9-13 mm. long, hairy on both surfaces, entire or faintly bifid at apex; corolla tube 2—2.5 cm. long, hairy; berry elliptic, 1.3-1.5 cm. long, rugose; pate 0.6—0.67 mm. long, with 3-8 foveae in the areoles of the testa. ...... 16. M. philippinensis. P. Stipules bifurcate from apex for ‘“ their length, 7-10 mm. long, oblong-ovate; leaves pubescent on both surfaces, 5-15 cm. long; calyx lobes 5—9.5 mm. long, ae arp gradually tapering to an acute apex; corolla tube 2.7—3.6 . long, hairy; berry ellipsoid-fusiform, 1.5-2.5 cm. long, strigose, with calyx segments subpersistent; seeds 0.7-0.77 mm. long, with 4-6(-8) foveae in the areoles of the testa. .................., 5. M. benguetensis. N. Inflorescence extending beyond the terminal leaves. R. Leaves with 15-18 pairs of lateral veins, 12-25 cm. long, strigose on both surfaces; stipules ovate, 6-8 mm. long, bifurcate pia apex for about % their length; calyx lobes linear-lanceolate, 4-7 mm. long, hairy on both surfaces; corolla ay 3 cm. long, hirsute; ber lipid 1.5 cm. long, sparingly RUTSUUG, ss gages oe yea eee . M. setosa. R. Leaves ‘with less than 15 pairs ie lateral veins. S. Calyx lobes broadly lanceolate, 6.7-28 mm. long, 1-6 mm. broad, hairy; leaves 9.5-30 cm. long, densely hirsute on both surfaces; stipules ovate, 12.5-14 mm. long, bifurcate at apex; corolla tube 2.5-3.3 cm. long, densely hirsute, hairs spreading; berry oblong or elliptic, 1.2-2 cm. long, lenticellate; seeds 0.53-0.67 mm. long, faint- ly spiny, with 4-11 foveae in the areoles of the testa. hes nk yaaa be eee +e ads oe ene 12. M. multibracteata. S. Calyx lobes linear-lanceolate, 11(-15) mm. long, 1.4 mm. brvad, hairy; leaves 9-16 cm. long, villous on both ete stipules oblong, 15 mm. long; corolla tube 2.5 cm. long, ferruginous- villous on the outer surface; berry (immature) obovoid, 1.4-1.5 cm. long sparingly pubescent ; seeds 0.46-0.6 mm. sia with 4-8(-—11) foveae in the areoles of the testa. . . 1. M. acuminatissima. © 1, Mussaenda acuminatissima Merr. Philip. Jour. Sci. Bot. 17: 436. 1920 (Type: Ramos 33133); Merrill, Enum. Philip. Fl. Pl. 3: 517. 1923. Fic. 4, 0, p, q. Erect shrub about 3 m. high with reddish-brown, lenticellate, glabrous branches and densely ferruginous-villous branchlets. Leaves ovate, 9-16 cm. long, 5—8.5 cm. broad, subcaudate-acuminate at the apex, abruptly long attenuate at base, decurrent on petiole, villous on both surfaces, more densely so on the midrib and veins of the lower surface, lateral veins 10-12 pairs; petiole 1.3-2 cm. long, villous. Stipules oblong, 15 mm. long, acute 1964 | JAYAWEERA, MUSSAENDA IN THE PHILIPPINES 105 or acuminate, hirsute. Inflorescence a terminal, dichotomously branched, pubescent, many-flowered cyme, primary cyme short, axillary cymes from the terminal pair of leaves longer and extending beyond the leaves; bracts and bracteoles linear-lanceolate, about 7 mm, long, acuminate, hairy on both surfaces. Flowers yellow, on stout, pubescent pedicels shorter than the ovaries. Calyx lobes linear-lanceolate, 11(—15) mm. long, 1.4 mm. broad, hairy on both surfaces with one pair of glands at the base of each; petaloid sepal white, ovate or elliptic-ovate, 4—6 cm. long, acuminate, cuneate at base, pubescent on both surfaces: “petiole’’ about 1 cm. long, hirsute. Corolla tube 2.5 cm. long, ferruginous-villous on the outer surface, hairs within short (long-styled form), extending as far as the bases of the anthers, below glabrous, not tufted at the mouth; corolla yellow, the lobes ovate, 3.5 mm. long and as broad, apiculate, hairy on the outer sur- face, papillate within. Stamens with short filaments, epipetalous above the middle of the tube, 3/5 way up; anthers linear, dorsifixed, introrse, 4.5 mm. long, bifid at the base. Ovary broadly fusiform, 5 mm. long, villous, 2-locular with numerous ovules on axile placentae; style 2 cm. long, stigma lobes 7 mm. long, stout, not protruding beyond the corolla tube. Berry (immature) obovoid, 1.4—1.5 cm. long, sparingly pubescent with persistent calyx segments; seeds (immature) minute, reticulate, irregularly oblong or ovoid, 0.46-0.6 mm. long, 0.43-0.56 mm. broad, with 4-8(—11) foveae in the areoles of the testa. DiIsTRIBUTION. This species is endemic on the dry slopes of Mt. Nagapatan in Luzon at an elevation of 700 meters above sea level. Philippine Islands. Luzon: Ilocos Norte, Mt. Nagapatan, Ramos 33133 (NY- lectotype). Only the isotype from the New York Botanical Garden was available for examination, and I designate it the lectotype. This species is closely allied to Mussaenda philippinensis in its elongated, persistent sepals but differs from it in the indumentum, leaves, oblong stipules, and larger ovaries. Mussaenda acuminatissima is distinguished from the other species by its ovate, villous leaves which are acuminate at base and apex, caudate at the apex, abruptly contracted at the base, and decurrent on the petiole, its oblong stipules, short, terminal, primary cymes, pubescent corolla tube, sparingly pubescent fruits with persistent calyx segments, and its minute seeds with 4—8(—11) foveae in the areoles of the testa. 2. Mussaenda albiflora Merr. Philip. Jour. Sci. Bot. 5: 241. 1910 (Type: Curran 17358); Merrill, Enum. Philip. Fl. Pl. 3: 518. 1923. Pao. 1, hy i: Fre. 5,2; b,c, d, ¢. Erect shrub 2-5 m. tall with terete branches and densely hirsute branchlets. Leaves ovate, oblong-ovate or lanceolate-elliptic, 11.5—-29 cm. long, 5.5-9.5 cm. broad, slenderly acuminate at apex, decurrent-acuminate at base, pubescent on both surfaces with scattered, long, spreading hairs, 106 JOURNAL OF THE ARNOLD ARBORETUM [| VOL. XLV (i ut \! My nh ui i iy Nh iu a 4d {I MA ip Ma es 1), X 3. e, f, M. attenuifolia (Elmer 13304), * 3%. g, M. scandens (Elmer 11291), x 4. h, i, M. albiflora: Ramos & Edano 31107), X 3; i (Dias 29885), X 2%. i, 'M. philippica ),xX 3 Sulit & Conklin 17652), K 2%; 0 (Escritor 21240), * 3; p (McGregor 32449), * 2%. q, M. grandifolia (Edano 77424), X 5, an old stipule, the apical lobes reduced to stumps. 1964] JAYAWEERA, MUSSAENDA IN THE PHILIPPINES 107 lateral veins 10-14 pairs; petiole 1-3 cm. long, densely hirsute. Stipules broadly ovate or lance-ovate, 1.1 cm. long, 6—7.5 mm. broad at the base, entire or faintly bifid from apex for 1/3 their length, hairy on both surfaces with numerous glands in groups at the base within. Inflorescence a termir- al, densely pubescent (with long or short spreading hairs), trichotomously branched, compact cyme; bracts and bracteoles lanceolate, about 7 mm. long, acuminate, hairy on the outer surface, glabrous within. Flowers heterostylous, on stout, pubescent pedicels shorter than the ovaries. Calyx lobes linear-lanceolate, 7-8 mm. long, 0.7-0.9 mm. broad, gradually nar- rowing to the acuminate apex, hirsute on the outer surface, glabrous within with 2 or 3 pairs of glands at the base of each; petaloid sepal white, ovate to oblong- or elliptic-ovate, 5—7.2(—9) cm. long, 2.7-4.3(—5.5) cm. broad, sharply acute or acuminate at apex, rounded or acute at base, glabrous on both surfaces except on veins beneath, “petiole” 1-2 cm. long, hirsute. Corolla tube white, very slender, 2.6-3.6 cm. long, covered with short, scattered hairs on the outer surface, hairy within as far as the bases of the anthers in long-styled forms, 1/2 way down from the top in short-styled forms, hairs not tufted at the mouth and shorter in long-styled forms; corolla lobes linear-lanceolate, 10-10.5 mm. long, 1.5—2.5 mm. broad, gradually narrowed to an acute apex, hairy on the outer surface, papillate within. Stamens with short filaments, epipetalous just below the middle of the tube in long-styled forms and about 2/3 way up from the base in short-styled forms; anthers linear, dorsifixed, introrse, 3.5-4 mm. long, acute at apex, bilobed at the base. Ovary turbinate or cylindric-turbinate, 3-4 mm. long, hairy, 2-locular with numerous ovules on cushion-shaped, axile placentae; style and stigma lobes 2 cm. and 6.5 mm. long respectively in long-styled forms, each 3 mm. long in short-styled forms. Berry (imma- ture) obovoid, 1-1.3 cm. long with few, long, scattered hairs, calyx lobes deciduous; seeds minute, reticulate, 0.67—0.83 mm. long, 0.43- 0.53 mm. broad, with 2—8 foveae in the areoles of the testa. ILLUSTRATION, Sulit, Philip. Jour. Forestry 2(1): 43. pl. 2, fig. 1. 1939. DIsTRIBUTION. This species grows under humid conditions in gullies to an altitude of about 300 meters above sea level in the islands of Panay and Negros where it is endemic. It has been collected in flower from April to June and in September; in fruit in September. Philippine Islands. NEGRos. Curran 17358 (Ny-lectotype), 17359 (us); Occ. Negros, Iglamgam, Dias 29885 (a, NY). Panay: Capiz, Jamindan, Ramos & Edano 31107 (A), 31382 (us); Libacao, Martelino & Edano 35391 (A, NY); Antique, McGregor 32450 (A) There do not seem to be any close relatives of Mussaenda albiflora in the Philippine Islands. It may be distinguished from other species by its large, pubescent leaves; large, ovate or lance-ovate stipules (entire or faintly bifid at the apex); long, slender corolla tubes; linear-lanceolate, white corolla lobes; obovoid, scantily pubescent fruits with deciduous calyx segments; and large seeds with 2 to 8 foveae in the areoles of the testa. 108 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV 3. Mussaenda anisophylla Vidal, Phan. Cuming. Philip. 178. 1885 (Type: Cuming 918); Merrill, Enum. Philip. Fl. Pl. 3: 518. 1923. Fic. 3, j-s. Shrub 2-8 m. high with rather lax, hairy branchlets. Leaves membra- nous, oblong-ovate or elliptic, 8.5-30 cm. long, 3.2—-16 cm. broad, acute or acuminate, cuneate, long decurrent on petioles from the base, hairy on both surfaces with 7-15 pairs of lateral veins; petiole 0.5—5.5 cm. long, stout, hirsute. Stipules ovate, 9-17 mm. long, 4.5-10 mm. broad at the base, bifurcate from apex for about 1/2 their length, lobes straight or diverging, densely hirsute on the outer surface, glabrous within or pubescent with numerous glands in a continuous band. Inflorescence a terminal, di- or trichotomously branched, densely hirsute, diffuse, many-flowered cyme; bracts and bracteoles lanceolate, about 6-9 mm. long, hairy on both sur- faces; bracteoles usually in opposite pairs, broader at the base with a few glands and trilaciniate to about the middle, lateral lobes shorter. Flowers heterostylous on stout, densely hairy pedicels shorter than the ovaries. Calyx lobes linear, 6-15 mm. long, 1-2 mm. broad, covered with sharp- pointed, straight, stiff hairs on the outer surface, hairs short within with 2 pairs of glands at the base of each sepal; petaloid sepal white, elliptic- ovate, 7.5-10 cm. long, 3.7—-6 cm. broad, acute, cuneate at the base, hirsute on both surfaces, 5-veined, “petiole” 1-2 cm. long, hirsute. Corolla orange to yellow, the tube 2.5-3 cm. long, densely hairy on the outer surface, hairs long and spreading, hairy within as far as the bases of anthers or lower; hairs of the inner surface long in short-styled forms and short in long- styled forms, not tufted at the mouth; corolla lobes broadly ovate, 5 mm. long, 6 mm. broad, apiculate, hairy on the outer surface, papillate within. Stamens with short filaments, epipetalous about 1/2 way on the tube in long-styled forms and 3/5 way up in short-styled forms; anthers linear, dorsifixed, introrse, 5.5—6.5 mm. long, bilobed at the base. Ovary turbi- nate, 4.5 mm. long, hirsute, 2-locular with numerous ovules on axile pla- centae; style and stigma lobes 2.2 cm. and 9 mm. long respectively in long- styled forms, 3.5—4 mm. and 2—2.5 mm. long in short-styled forms. Berry ellipsoid, 1.4-2 cm. long, scantily hairy, lenticellate, calyx segments deciduous; seeds minute, reticulate, oblong or orbicular-ovate, 0.63—-0.73 mm. long, 0.33—0.63 mm. broad, with 2—6 foveae in the areoles of the testa. ILLUSTRATION. Sulit, Philip. Jour. Forestry 2(1): 43. pl. 1. 1939. DisTRIBUTION. This species grows near water courses in deeply shaded ravines in the mossy forests at medium elevations up to 800 meters in the islands of Luzon, Alabat, Mindoro, Negros, and Mindanao. It has been collected in flower in January, February, May, June to August, and October; in fruit in February, June, and November. Philippine Islands. Luzon: Albay Prov., Cuming 918 (BM-lectotype); Pan- gasinan Prov., Labrador, Mt. San Isidro, Fenix 29851 (Ny, Us); Bataan Prov., Mt. Mariveles, Lamao River, Meyer 3020 (ny, us), Merrill 2508 (Ny, us), Bor- romeo 25602 (A, US), Batangas Prov., Vidal 801 (a); Laguna Prov., Los Banos 1964 | JAYAWEERA, MUSSAENDA IN THE PHILIPPINES 109 (Mt. Maquiling), Elmer 17481 (a, GH, Ny, US); Mt. Makiling, Canicosa 9730 (PNH), Sulit 8232 (A, PNH), 7087 (A, PNH); San Antonio, Ramos 21988 (us), . Q ‘ iS) oR 8 ma LS) Ww S S Qn —_~ > q n ww ae be) oO p& Se pf wn 4°) =“ co) rs a c ie) ion pt) 5 bs ~ 5S i) | © ~— ww wR — > 4 a Gq mn aa? e 8. Prov., Antipolo, Merrill 1341 (ny); Kalinga Subprov., Balbalan, Celestino 7819 (A, PNH). ALABAT: Ramos & Edano 48110 (A, Ny, US). Mrnporo: Mt. Halcon, Ramos & Edano 40703 (a), Edano 3500 (a, PNH); Mt. Ilong, Edano 3290 (a, PNH). NecRos: Gimogaan River, Danao & Aspillera 5218 (US). MINDANAO: Dinagat Island, Ramos & Convocar 84044 (a). Probably the only isotypes of this species are in the British Museum (Nat. Hist.) and in the Conservatoire et Jardin botaniques, Genéve, since the holotype (Cuming 918 |PNH]) and other isotypes apparently have been destroyed. The sheet in the British Museum has two specimens of the same species mounted on it. The specimen at the top agrees with the original description and it is therefore selected as the lectotype. This species approaches Mussaenda magallanensis in the size and pubescence of its stems, leaves, and inflorescence, but differs from it in the deeply bifur- cate stipules, longer calyx lobes, hairy corolla tube, larger siamens, and larger seeds containing fewer foveae in the areoles of the testa. It agrees completely with Wenzel 442 and Ramos 15328 from Leyte belonging to M. magallanensis in the structure of the seed and the number of foveae in the areoles of the testa. The stipule of Elmer 17481 resembles typical M. macrophylla Wall. except that it is glabrous within. The species is distinguished by the almost horizontal, long hairs on the stems, leaves, and flowers, the large leaves long decurrent on the petiole, the linear or linear-lanceolate sepals, the large, scarcely pubescent, lenti- cellate berries with deciduous calyx segments, and the small seeds with 2 to 6 foveae in the areoles of the testa. Uses. The fresh leaves of this species are used medicinally, in the form of a decoction, as a cure for asthma in the Philippine Islands. 4. Mussaenda attenuifolia Elmer, Leafl, Philip. Bot. 5: 1874. 1913 (Type: Elmer 13304); Merrill, Enum. Philip. Fl. Pl. 3: 518. 1923. Pic, 1, f. Shrub about 5 m. high with grayish-white mottled bark and slender, lax, dark brown, white-lenticellate branchlets, the young suberect portions densely soft ferruginous. Leaves opposite, ascending, submembranous, oblanceolate, 7-20 cm. long, 2—7 cm. broad, abruptly short-pointed, acu- inate at the recurved apex, decurrently attenuate at base, sparsely pubes- cent on the upper surface, more densely soft pubescent along the midrib and veins on the lower surface, veins 10-14 pairs, curved-ascending, obscurely reticulate; petiole 0.5-1.5 cm. long, stout, hirsute. Stipules oblong-ovate, membranous, deciduous, 5-11 mm. long, 4.5—5.5 mm. broad, occasionally split or regularly bifurcate at the apex more than 1/2 way, strigose-hairy on the outer surface, pubescent within with numerous glands in a continuous band at the base. Inflorescence a terminal, pubescent, erect or suberect, dichotomously branched, few-flowered cyme not produced be- 110 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV yond the leaves, lower branches curved-ascending, opposite, subtended by slender linear or lanceolate hairy bracts and bracteoles, terminal flowers subsessile. Flowers probably heterostylous on stout, pubescent pedicels shorter than the ovaries. Calyx lobes lanceolate, acuminate or sharply acute, about 3 mm. long, pubescent on both surfaces; petaloid sepal creamy white or white, ovate or elliptic, 8-13 cm. long, abruptly pointed at apex and abruptly attenuate at base, 5-nerved, “‘petiole” 2—3 cm. long, pubescent. Corolla tube greenish to yellowish-white, 2.7—4 cm. long, pubes- cent on the outer surface, hairy within as far as the middle of the tube in long-styled forms, hairs short and not tufted at the mouth (short-styled forms not seen); corolla lobes orange to yellow, ovate or more or less oblong, 4 mm. long, 2.5 mm. broad, long-apiculate or acuminate, pubescent on the outer surface, papillate within. Stamens with short filaments, epi- petalous a trifle above the middle of the corolla tube; anthers linear, dorsifixed, introrse, 5—5.1 mm. long, slender, blunt at apex, basal sterile portion constricted and bilobed. Ovary turbinate, 3—-3.5 mm. long, hairy, 2-locular with numerous ovules on axile placentae; style slender, 1.8 cm. long, stigma lobes flattened, fleshy, 7.5-8 mm. long. Berry ellipsoidally elongate, 1.2—2 cm. long, pubescent when young, becoming nearly glabrous, without lenticels, calyx segments deciduous but the nectariferous disc persistent as a blunt, conical knob; seeds minute, reticulate, blackish brown, 0.64-0.77 mm. long, 0.43—0.63 mm. broad, with 3-6(—9) foveae in the areoles of the testa. DIstriBUTION. This species is endemic to Mindanao where it grows at an elevation of 150 to 450 meters above sea level. Elmer collected it in flower and fruit in July, 1912, and no more recent collections seem to have been made, although he stated that it is “fairly well scattered in wet earth and upon rock ledges of the wooded banks of the Catangan Creek.” Philippine Islands. M1inpanao: Agusan Prov., Cabadbaran (Mt. Urdaneta), Elmer 13304 (a-lectotype; GH, Ny, US—iso types). Only the isotypes were available for examination. From these the speci- men deposited in the herbarium of the Arnold Arboretum is designated as the lectotype. Elmer describes the stipules as ‘‘quite variable in size, brown, membranous, deciduous, strigose hairy on the outside, 5 to 8 mm. long, gradually tapering from the expanded base to the setiform apex, occasion- ally split.” The stipule in fact is much longer, 10-11 mm. in length, bifurcate more than 1/2 way from the apex in all examples examined. This species seems to be allied to Mussaenda palawanensis, though it differs much in the form of the leaf, inflorescence, etc. It agrees in the character of the stipule except for its hairiness and the banded glands inside. Other common characters are the deciduous calyx, ovate-acuminate corolla lobes, size of the seed, and foveae in the areoles of the testa. It is distin- guished from other species by the obovate-oblong leaves with recurved, abruptly acuminate apices and the decurrent-attenuate bases; deeply bifurcate, oblong-ovate stipules with glands at the base within a continuous band; acuminate-apiculate, ovate corolla lobes; glabrous, ellipsoid berries 1964 | JAYAWEERA, MUSSAENDA IN THE PHILIPPINES 111 (not lenticellate) with a pointed, conical, persistent, nectariferous disc at the apex, and minute seeds with 3—6(-9) foveae i in the areoles of the testa. 5. Mussaenda benguetensis Elmer, Leafl. Philip. Bot. 1: 13. 1906 (Type: Elmer 5935); Merrill, Enum. Philip. Fl. Pl. 3: 518. 1923. Fic. 1, 1; Fic. 3, x, y, z, aa. Shrub, sometimes tree-like, 1-4 m. tall with ferruginous-pubescent branches. Leaves clustered toward the ends of branches, ovate, oblong or oblong-lanceolate, 5-15 cm. long, 2.3-5.2 cm. broad, acute or acuminate, abruptly cuneate at base, rounded or decurrent-acuminate, appressed pubescent on both surfaces, scantily on the upper surface, more densely especially on the prominent veins of the lower surface, lateral veins 8-11 pairs, ascending; petiole 0.5—2.5 cm. long, densely hirsute. Stipules sub- persistent, broadly oblong-ovate, 7-10 mm. long, 6.8-8 mm. broad, bi- furcate from apex for about 1/2 their length, rami gradually tapering and diverging into 2 acuminate lobes, hairy on the outer surface, glabrous within except for a few hairs at the apices of the divergent lobes and at the base between numerous glands occurring in 2 groups. Inflorescence a terminal, subcorymbose cyme barely exceeding the leaves in length, pubescent, ultimate branches terminating in clusters of 3 to 5 flowers; bracts and bracteoles lanceolate, 4-13 mm. long, 0.7—2 mm. broad, pubes- cent on the outer surface, scantily hairy or glabrous within, bracteoles bifid or trifid, lateral lobes shorter. Flowers heterostylous, fragrant, on stout, pubescent pedicels shorter than the ovaries. Calyx lobes linear- acuminate gradually tapering to an acute apex, 5—9.5 mm. long, pubescent on the outer surface, scantily pubescent or glabrous within with 2 or 3 pairs of glands at the base of each; petaloid sepal whitish, small, ovate or elliptic, 3.8-5.2 cm. long, 2—2.8 cm. broad, scantily pubescent on the upper surface, densely pubescent especially on the veins of the lower surface; “petiole” short, stout, 0.5 cm. long, hirsute. Corolla yellow, the tube 2.7— 3.6 cm. long, hairy on the outer surface, hairy within 1/2—5/9 way down from the top in long-styled forms, 1/4—1/3 way in short-styled forms, hairs not tufted at the mouth and shorter in long-styled forms; corolla lobes ovate or suborbicular, 4-5 mm. long, 3.4-6 mm. broad at the base, acute, hairy on the outer surface, papillate within. Stamens with short filaments, epipetalous about 1/2 way on the tube in long-styled forms, 2/3-3/4 way up from the base in short-styled forms; anthers linear, dorsifixed, introrse, 5-7 mm. long, bilobed at the base, lobes diverging. Ovary broadly fusiform, 5-6 mm. long, pubescent, 2-locular with numerous ovules on axile placentae; style and stigma lobes 1.75—2.2 cm. and 8.7—12 mm. long respectively in long-styled forms, 4-6 mm. and 3-—3.5 mm. long in short-styled forms. Berry ellipsoid-fusiform, 1.5—2.5 cm. long, strigose, lenticellate with subpersistent calyx segments; seeds minute, reticulate, 0.7-0.77 mm. long, 0.53—0.63 mm. broad, with 4-6(—8) foveae in the areoles of the testa. DISTRIBUTION. This species is endemic to Luzon and grows on open 112 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV slopes and in thickets or ravines from 900 to 2300 meters elevation. Elmer comments in his field notes that the plant grows “. . . in ravines but usually in open places; or skirting dense hardwood jungles of the limestone region at 4500 feet; . . .” It has been collected in flower from January to June, October, and November; in fruit from January to June and in December. Philippine Islands. Luzon: Mountain Prov., Baguio, Steimer 22585 (PNH), 35835 (PNH, US), 41617 (pNH), Elmer 5935 (pNH-holotype, Ny-isotype), 8431 (A, NY, US), Williams 1006 (GH, Ny, US), Sulit 7722 (pnH), Walker 7501 (us), Topping 64 (us), Santos 34 (a); Mancayan to Baguio, Ramos & Edano 40504 (a); Mt. Santo Tomas, Sinclair & Edano 55365 (PNH), Quisum- bing 2193 (A, PNH), Garcia 34977 (pNH), Elmer 6536 (Ny, US), Clemens 51845 (a, us); Barrio Agawa, Besao, Santos 5497 (us); Mt. Libbing, Mendoza 40936 (pNH); Mt. Polis, Steiner 41580 (pNH); Benguet Subprov., Papuai, Santos 31880 (NY), 32026 (A), Merrill 713 (us); Bontoc, Mt. Santo Tomas, Walker 7543 (us); Mt. Polis, Ramos & Edano 37633 (a); Lepanto Dist., Balili (Mt. Data), Merrill 4639 (us); Ifugao, Mt. Polis, McGregor 19779 (us). CuttivaTeD: Trinidad, Loher 1523 (us). As in the case of Mussaenda pinatubensis, the holotype of M. bengue- tensis was not destroyed in World War II. The collection Vanoverbergh 190 (a) is interesting, for it possesses the stipules and leaf characters of Mussaenda lanata and the floral characters of M. benguetensis. Its stipules are triangular-acuminate, 13 mm. long, 7.5 mm. broad, not bifid at the apex, hairy on both sides, with numerous glands at the base within. The corolla tube is 2.5 cm. long, larger than that of M. lanata but smaller than in M. benguetensis. It was collected in Bontoc Subprovince, Luzon, where the two species overlap. One sheet of Williams 1006 bore two abnormal flowers one of which was dissected. It had a calyx and corolla of 10 segments each but 15 stamens and 3 ovaries fused together with 7 stigma lobes. The filaments of the innermost stamens were longer, about 5 mm. in length. It has been observed that several fruits with mature seeds bore calyx segments while in others the calyx segments have been shed. Elmer refers to the petaloid sepal as a “persistent bract’’ but does not comment on the other sepals. It is probable that the calyx is subpersistent on the fruit. Mussaenda benguetensis is distinguished from other species by its smaller leaves confined to the ends of branches with prominent veins on the lower surface, oblong-ovate stipules with diverging apical lobes, larger elliptic-fusiform, lenticellate berries with or without calyx segments, and somewhat larger seeds with 4—6(—8) foveae in the areoles of the testa. 6. Mussaenda chlorantha Merr. Philip. Jour. Sci. Bot. 8: 47. 1913 (Type: Merrill 770); Merrill, Enum. Philip. Fl. Pl. 3: 518. 1923. Fic. 1, b, c; Fic. 2, d, e. Tree 6-8 m. tall with glabrous, reddish-brown, lenticellate branches, the younger ones appressed pubescent. Leaves broadly ovate, 9.5-16 cm. 1964 | JAYAWEERA, MUSSAENDA IN THE PHILIPPINES 113 long, 6-9.5 cm. broad, sparsely hirsute on the upper surface, puberulous beneath and appressed hirsute on the midrib and lateral veins, short acu- minate at apex, abruptly cuneate and somewhat decurrent at base; lateral veins 7-12 pairs, arcuate; petiole 1-4 cm. long, hirsute. Stipules 8.5—10 mm. long, 4.5-6.5 mm. broad at the base, triangular-ovate or oblong-ovate, entire or bifid at apex, densely hirsute on both surfaces with numerous glands in 2 groups or in a continuous band at the base within. Inflorescence a terminal, hirsute, many-flowered, corymbose cyme; bracts and bracte- oles linear-oblong to oblong-lanceolate, 4-7 mm. long, hairy on both sur- faces, bracteoles larger, trifid, the lateral lobes reduced to stubs. Flowers on stout, pubescent pedicels shorter than the ovaries. Calyx lobes 6-12 mm. long, 1.2—2.2 mm. broad, lanceolate, acuminate, pubescent on the outer surface, glabrous or pubescent within with 1 or 2 pairs of glands at the base of each; petaloid sepal white, broadly ovate, 5-8 cm. long, glabrous on the upper surface, hairy on veins beneath, “petiole” about 2 cm. long, hirsute. Corolla tube green, 1.7—2 cm. long, hirsute on the outer surface, clothed with short hairs (0.5 mm. long) within for 3/4 the length of the tube, basal 1/4 glabrous, hairs not tufted at the mouth; corolla lobes yellow, broadly ovate, 3-4.5 mm. long, 3.5-4.5 mm. broad, hairy on the outer surface, papillate within. Stamens with short filaments, epipe- talous 1/2 way or higher up on the tube; sometimes filaments free (for about 2.5 mm.) as short loops midway between the base of the corolla tube and the stamens; anthers linear, dorsifixed, introrse, 3.5—-5 mm. long, acute, bifid at base. Ovary oblong-turbinate, 5-6 mm. long, appressed pubescent, 2-locular, with a thick wall and axile placentation; style and stigma lobes 1-1.4 cm. and 5-6.5 mm. long respectively; stigmas stout, linear, diverging, above the anthers or placed at the same level as the anthers and surrounded by them. Berry (immature) obovoid, 1 cm. long, glabrescent, calyx lobes deciduous; seeds minute, reticulate, 0.7-0.83 mm. long, 0.5—0.56 mm. broad with 4-12 foveae in the areoles of the testa. DIsTRIBUTION. This species was collected in flower in May by Merrill and by Elmer in Luzon at an altitude of about 1800 meters; it has not been recollected. Philippine Islands. Luzon: Benguet Subprov., Mt. Tanglon, Merrill 770 (us-— lectotype), May 1911; Pampanga Prov., Camp Stotsenburg (Mt. Pinatubo), Elmer 21940 (GH, PNH), May 1927; Twin Peaks, Elmer 6318 (us), May 1904. Elmer’s collections differ slightly from the type in the bifid stipules with glands in two groups at the base within and sepals hairy on both surfaces. Otherwise the collections resemble each other closely. Mussaenda chlo- rantha is allied to M. macrophylla Wall. var. brevipilosa Jayaweera, par- ticularly Fenix 3770 (Ny, us), but differs in its shorter corolla tube and the position of the stigma lobes. Mussaenda chlorantha is distinguished by its broadly ovate leaves; en- tire or bifid, densely hirsute, triangular-ovate stipules; oblong-lanceolate sepals; short corolla tube with broadly ovate lobes; large, oblong-turbinate ovary, stout stigma lobes, and 4 to 15 foveae in the areoles of the testa. 114 JOURNAL OF THE ARNOLD ARBORETUM ‘VOL. XLV flowers of some s seen from within, : , M. x 2; b (idem), calyx lobes, X 214; c (idem), corolla lobe, X 6. d, e, M. chlorantha: d (Elmer 21940), X 2; e (idem), calyx lobes, X 3. f-h, M. magallanensis: { (Elmer 1964 | JAYAWEERA, MUSSAENDA IN THE PHILIPPINES 115 7. Mussaenda grandifolia Elmer, Leafl. Philip. Bot. 1: 12. 1906 (Type: Merrill 817); Merrill, Enum. Philip. Fl. Pl. 3: 518. 1923. Fic. 1, q; Fic. 4, a, b, c. Shrub 3—4 m. high with densely pubescent younger branchlets. Leaves ovate, oblong or oblanceolate, 16-32 cm. long, 7.5—15 cm. broad, acute at apex, cuneate or long attenuate at base, pubescent on both surfaces, densely so beneath and on the 11—14 pairs of lateral veins; petiole 1-2.5 cm. long, villous. Stipules small, slenderly acuminate from a broad base, bifurcate at the apex, hairy on both surfaces with numerous glands in 2 groups or in a continuous band at the base within. Inflorescence terminal, shorter than leaves, pubescent, trichotomously branched, branches straight, divari- cate, diffuse, terminal flowers sessile and others subsessile; bracts and bracteoles small, ovate, 1.5-2 mm. long, 0.5—-1.5 mm. broad, acuminate, hirsute on both surfaces. Calyx lobes small, lanceolate, 1.2—-2.5 mm. long, 0.7-1 mm. broad, acuminate, hairy on both surfaces with 1-3 pairs of glands at the base of each; petaloid sepal yellowish white, broadly ovate, 4.5—-7 cm. long, 44.5 cm. broad, puberulous on the upper surface and hirsute beneath especially on veins, “petiole” 1 cm. long, hirsute. Corolla yellow, the tube 2.9 cm. long, densely hairy on the outer surface, glabrous within except for a few short hairs in 5 groups at the bases of the corolla lobes; corolla lobes ovate, 4 mm. long, acuminate, hairy on the outer sur- face. Stamens with short filaments, epipetalous at slightly different levels about 1/2 way on the tube or a little below in the long-styled forms; anthers linear, dorsifixed, introrse, 3.2—4.2 mm, long, of different sizes in the same flower, acute at apex and bifid at the base. Ovary broadly fusi- form, 4.5 mm. long, densely hirsute, 2-locular with numerous ovules on axile placentae; style 2.2 cm. long, stigma lobes 1 cm. long and stout in long-styled forms. Berry ellipsoid, 1 cm. long, pubescent with small, persistent calyx segments; seeds (immature) minute, reticulate, 0.64—0. 77 mm. long, 0.43-0.53 mm. broad, with 1-4 foveae in the areoles of the testa. DIsTRIBUTION. This species is endemic to Palawan where it grows in forests at lower elevations. It has been collected in flower and fruit in February and March. Philippine Islands. PALAWAN (Paragua): Point Separacion, Merrill 817 (us- lectotype); Panalingajan River, Edano 77424 (a). Only two collections were available for examination, of which one was 12451), * 1%; g (idem), calyx lobes, & 4%; h (idem), corolla lobe, x 4¥4, i-ka, M. nervosa: i (Elmer 10510), X 2%4; j (Ramos & Edano 26422), ; ; ; : Ww scandens: 1 (Wenzel 3354), X 1%; m (Elmer 11291), X 1%; n (idem), calyx obes, & 4; 0 (idem), corolla labe. 6 (idem), seed, X 23. q-x, M. vidalii: ‘ ae & Edano 39035), * 1% . (Sulit 6098) K 13s Ci calyx lobes, t (idem), corolla lobe, < re u-x, stipules, u (Sulit 6280), K 3%; : (dimer 11309), & 3%; w (Sulit 6098), « 3; x (Ramos & Edano 39035), X 116 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV an isotype (which now becomes the lectotype). Both collections were long-styled forms, and heterostyly of the species is not established, though it is quite likely that a short-styled form exists which has not been col- lected up to now. Elmer’s description of the pubescence of the calyx segments and the corolla tube is not accurate. He says that the calyx segments are “glabrous on the inner surface” and the corolla tube ‘“‘lanose on the interior.’ An examination of the lectotype specimen shows that the calyx segments are pubescent on the inner surface and the corolla tube glabrous on the interior. The species does not seem to be allied to any of the Philippine mussaendas though it resembles M. vidalii in its inflorescence and sessile terminal flowers.. It is distinguished from the other species by the large, pubescent leaves; small stipules; sessile terminal flowers in the forks of the peduncles; corolla tube glabrous on the interior surface; pubescent fruits with small, persistent calyx segments; and minute seeds with 1 to 4 foveae in the areoles of the testa. 8. Mussaenda lanata C. B. Robinson, Philip. Jour. Sci. Bot. 6: 357. 1911 (Type: Merrill 6681); Merrill, Enum. Philip. Fl. Pl. 3: 518. 1923. Fic. 1, a; Fic. 2, a, b,c. Shrub or small tree 2-5 m. high with greenish-gray, pubescent branch- lets. Leaves unequal at nodes, ovate, oblong or orbicular, 10.5-30 cm. long, 6.5—16.5 cm. broad, abruptly acute or acuminate at the apex, cuneate at the base, hirsute on both surfaces, more densely so beneath, with 8-13 pairs of lateral veins; petiole 0.8—-7.5 cm. long, stout, hairy. Stipules ovate, long acuminate, 10-18.5 mm. long, 6-8.5 mm. broad at the base, faintly bifid at apex, hairy on both surfaces, with numerous glands in 2 groups or in a continuous band at the base within. Inflorescence a terminal, dichotomously branched, densely pubescent, many-flowered, subcorymbose cyme; bracts and bracteoles lanceolate, 3-7 mm. long, hairy on both surfaces. Flowers heterostylous, on stout, densely pubescent pedicels shorter than the ovaries. Calyx lobes lanceolate, 5.5—6.5 mm. long, 1-1.2 mm. broad, acute, hairy on both surfaces, with a pair of thin glands at the base of each lobe; petaloid sepal white, oval, 4.5—5.5 cm. long, 3—3.7 cm. broad, hirsute on both surfaces, 5-veined, “petiole” 1.4 cm. long, hirsute. Corolla tube 1.1-1.3 cm. long, campanulate, yellow-tomen- tose on the outer surface, hairy within to 1/2 the length of the tube in long- styled forms and 1/3 the length of the tube from the top in short-styled forms, glabrous at base; hairs of the throat long (1.5—1.7 mm.) in short- styled forms and short (1 mm. long) in long-styled forms, not tufted at the mouth; corolla lobes broader than long, somewhat truncate, 1.2—1.5 mm. long, 2.2-3 mm. broad, mucronate or abruptly apiculate at the apex, brown-tomentose on the outer surface, papillate within. Stamens with short filaments epipetalous 3/4 way up on the tube in short-styled forms, (lower down) about t/2 way up in long-styled forms; anthers linear- lanceolate, dorsifixed, introrse, 3-4 mm. long, acute at the apex, bilobed at the base. Ovary turbinate, 3.5-4.5 mm. long, densely hairy, 2-locular, 1964] JAYAWEERA, MUSSAENDA IN THE PHILIPPINES 117 with numerous ovules on cushion-shaped, axile placentae, nectariferous disc well developed; style and stigma lobes 5-7 mm. an mm. long respectively on long-styled forms, 1.6 mm. and 1.8 mm. long in short- styled forms. Berry ovoid, with a fruity fragrance, 1.2-1.7 cm. long, scantily pubescent, lenticellate, with deciduous calyx segments; seeds minute, reticulate, angularly rounded, 0.43-0.56 mm. long, 0.36—0.43 mm. broad, with 3-6 foveae in the areoles of the testa. DISTRIBUTION. This species is endemic to Luzon and grows in thickets along streams from 900 to 1600 meters elevation. It has been collected in flower in March, April, May, August, October, and December; in fruit in October, November, and December Philippine Islands. Luzon: Abra, Ramos 7254 (Ny, us); Bontoc, Ramos & Edano 38134 (a); Benguet, Sablan, Elmer 6195 (nv, us), Williams 1558 (cu, NY, US), Merrill 6681 (us-lectotype); Baguio, Mabesa 35292 (pnw), Elmer 8976 (A, NY, US), Curran 5078 (us); Sabang, Fenix 12588 (us); Itogon, Williams 1011 (GH, Ny, US), Merrill 1768 (GH, Ny), Loher 1524 (us), Clemens 17249 (ny); Ilocos Spur, Mt. Tirad Concepcion, Santos 5678 (us); Lepanto, Cervantes Trail, Ramos & Edano 38103 (GH); Zambales, Ramos 5044 (Ny, US). All collections of Mussaenda lanata are similar in their distinguishing characters, having large, membranous, densely pubescent leaves, acuminate stipules which are hairy and faintly bifid at the apex, crowded inflorescence with short campanulate flowers and small corolla lobes, and larger fruits bearing small seeds with 3 to 6 or 2 to 5 foveae in the areoles of the testa. The usual variations are in the size and shape of the leaf and in the fruit. The seeds are small and characterized by 3 to 6 foveae in the areoles of the testa. The collection, Ramos 5044, however, differs slightly in that its seeds are larger (0.64—-0.67 mm. long, 0.5-0.56 mm. broad) and contain to 5 foveae in the areoles of the testa. Robinson (1911) stated that this species is allied to Mussaenda aniso- phylla and M. villosa Wall. It seems to me that its affinities are rather with M. magallanensis and to a lesser degree with M. anisophylla and not at all with M. villosa. The species agrees with M. magallanensis in the important characters of the stipule, calyx lobes, stamens, ovary, and seed. Mussaenda villosa, on the other hand, is a climber with broadly lanceolate bracts and bracteoles, recurved calyx lobes, and slender corolla tube, bearing smaller, glabrous berries with oblong, spiny seeds containing 2 to 8 foveae in the areoles of the testa. None of these characters are common to the two species. 9. Mussaenda macrophylla Wall. var. brevipilosa, var. nov. (Type: Fenix 3770). M. macrophylla sensu Merr. Philip. Jour. Sci. Bot. 3: 437. 1908 and Enum. Philip. Fl. Pl. 3: 518. 1923, non Wall. A typo differt stipulis et calycibus minoribus; pilis ad faucem corollae 118 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV longistylae brevibus, ad orifaceum non_penicillatis; ovario obconico; fructibus lenticellatis; seminibus minoribus. A large, subscandent shrub with stout branches densely pubescent with grayish hairs becoming rufous-brown when dry. Leaves ovate, elliptic, oblong, or elliptic-lanceolate, 10-26 cm. long, 5-15 cm. broad, glabrous on both surfaces but hairy on the veins beneath or pubescent on the upper surface and hirsute below, short-acuminate, cuneate at base, somewhat decurrent on the petiole, lateral veins 8-13 pairs, conspicuous beneath; petiole 0.6—4.2 cm. long, hairy. Stipules large, ovate or acuminate from a broad base, 7-12 mm. long, 3.5—7 mm. broad at the base, bifurcate to 1/4-2/3 their length from the apex, lobes straight, hairy on the outer surface, glabrous or pubescent within with numerous glands in 2 groups or in a continuous band. Inflorescence a terminal, spreading or somewhat compact, trichotomously branched, pubescent cyme; bracts and bracteoles lanceolate or ovate, acuminate, hairy, bracteoles deeply trilaciniate into lanceolate, acuminate lobes and broader at the base. Flowers large, heterostylous, nearly sessile or on very short, stout, pubescent pedicels shorter than the ovaries. Calyx lobes triangular-lanceolate or oblong- lanceolate, 3.2—-8 mm. long, 1—-2.5 mm. broad, acute or acuminate, hairy on both surfaces or on the outer surface only, glabrous within except at apex, with 1—5 pairs of glands at the base of each; petaloid sepal white, ovate, oblong or orbicular-ovate, 6.5-9 cm. long, 3.3-7 cm. broad, short- acuminate, acute or obtuse at apex, cuneate at base, glabrous on both surfaces except on veins beneath or puberulous on the upper surface and hirsute beneath, ‘petiole’? 1-4 cm. long, hirsute. Corolla tube 2-3 cm. long, hairy on the outer surface, hairy within as far as the bases of anthers, hairs long in short-styled forms, short in long-styled forms and not tufted at the mouth of the tube; corolla lobes broadly ovate or orbicular, acute or apiculate, pubescent on the outer surface, papillate on the inner. Stamens with short filaments, epipetalous about 1/2 way or a little below the middle of the tube in long-styled forms, 3/5—2/3 of the way up in short-styled forms; anthers linear, dorsifixed, introrse, 5.5—6.7 mm. long, bilobed at the base. Ovary 4-5.5 mm. long, fusiform or obconical, hairy, 2-locular, with numerous ovules on cushion-shaped, axile placentae; style and stigma lobes 1.9 cm. and 0.7 cm. long respectively in long-styled forms, 3.8-8 mm. and 3-3.5 mm. long in short-styled forms. Berry ovoid or oblong-ovoid, 1.1—1.2 cm. long, lenticellate, sparsely hirsute with very broad nectariferous scars, calyx lobes deciduous; seeds minute, reticulate, 0.67-0.87 mm. long, with 2—6 foveae in the areoles of the testa. DIstRIBUTION. This variety of Mussaenda macrophylla is confined to the islands of the Philippine group. It has been collected in flower from April to July and in fruit from August to October. Philippine Islands. Batan: Fenix 3770 (Nny-holotype, us—isotype), Ramos 80071 (A). CALAYAN: Velasco 26650 (us). CamicuIn: Edano 79343 (A) Luzon: Edano 79448 (a), Ramos & Edano 29719 (A, us); Bohol, Ramos 1964 | JAYAWEERA, MUSSAENDA IN THE PHILIPPINES 119 42577 (A). Minpanao: Ramos & Edano 85008 (A); Camiguin de Mindanao, Ramos 14604 (us). This variety differs in a number of characters from Mussaenda macro- phylla var. macrophylla which is distributed from Nepal and Assam to Yunnan. It has smaller stipules, flowers, and calyx segments; the hairs inside the corolla tube in long-styled forms are short and not tufted at the mouth as compared to the long hairs which are tufted at the mouth in the typical form of var. macrophylla; the seeds are smaller, although the ‘number of foveae in the areoles of the testa is the same in both varieties. 10. Mussaenda magallanensis Elmer, Leafl. Philip. Bot. 3: 996. 1911 (Type: Elmer 12451); Merrill, Enum. Philip. Fl. Pl. 3: 519. 1923. Fic. 1, d; Fie. 2, f, g, h. M. macrophylla sensu Vidal, Phan. Cuming. Philip. 118. 1885, non Wall. Shrub about 2-3 m. high with terete, densely pilose branches, hairs almost at right angles to the stem. Leaves oblong or elliptic, 6.5—28.5 cm. long, 3.5-13(—19.5) cm. broad, abruptly acuminate, base cuneate to obtuse, decurrent on petiole, hirsute on both surfaces with 10-13 pairs of lateral veins conspicuous beneath; petiole 0.6-3.5(—7.5) cm. long, stout. pilose. Stipules triangular-ovate, acuminate from a broad base, 7.5—16 mm. long, 4.5-9 mm. broad at the base, apex entire or bifid, densely hairy on the outer surface, pubescent within with numerous glands in a continuous band at the base. Inflorescence a terminal, hairy, di- or trichotomously branched corymbose cyme; bracts and bracteoles lanceolate, 5.5-11 mm long, 1—2.2 mm. broad, hairy on both surfaces or on the outer surface only, bracteoles trilaciniate, lobes subulate. Flowers heterostylous on stout. pubescent pedicels shorter than the ovaries. Calyx lobes narrowly oblong- lanceolate or linear, 3.5-6 mm. long, 0.6-1.5 mm. broad, gradually tapering to a point, 3-striate towards the base, hairy on both surfaces with 1 or 2 pairs of glands at the base of each sepal. Petaloid sepal white, orbicular- oblong, elliptic or broadly ovate, acute, cuneate at base, glabrous, except on veins, on both surfaces, “petiole” 1.2-3.5 cm. long, hirsute. Corolla tube 2-3 cm. long, glabrous or with only straggling hairs on the outer surface, hairy within as far as the bases of the anthers or to about 1 mm. lower: hairs not tufted at the mouth, long in short-styled forms covering about 1/3 the tube from the top, short in long-styled forms extending to about 1/2 the length of the tube; corolla lobes orange, suborbicular-ovate, 3—3.5 mm. long, 2.7-4 mm. broad, rounded or apiculate, hairy on the outer surface, papillate within. Stamens with short filaments, epipetalous 2/3 way up on the tube in short-styled forms and 1/2 way up in long-styled forms; anthers linear, dorsifixed, introrse, 3—4.5 mm. long, bifid at the ase. Ovary turbinate, 3.2-4.5 mm. long, densely hairy, 2-locular with numerous ovules on axile placentae; style and stigma lobes 1.6 cm. and 6.5 mm. long respectively in long-styled forms, 1.5—6 mm. and 1.5—2.7 mm. long in short-styled forms. Berry ellipsoid, 1.2—2.2 cm. long, hairy when 120 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. XLV ——— 7s SSP» >5: 25 SSSI " | a Way vit i Fic. 3. Longitudinal sections of the long-styled and short-styled flowers of some Philippine species of Mussaenda, hein and paso lobes seen from within, and variations in the stipules of M. anisophylla. ora: a (Ramos & Edano 31107), X 1; b (Dias 29885), x l; c (idem), apical portion of a flower 1964] JAYAWEERA, MUSSAENDA IN THE PHILIPPINES 121 young, becoming sparsely pubescent, lenticellate, calyx lobes deciduous; seeds minute, reticulate, 0.5—0.67 mm. long, 0.33-0.53 mm. broad, oblong or oblong-ovate with 3-6 or 4—-12 foveae in the areoles of the testa. DIstRIBUTION. Mussaenda magallanensis, up to now considered as en- demic to Mindoro and Sibuyan Islands, seems to extend to the neighboring islands of Luzon, Panay, and Leyte, growing along rivers, On mountain slopes, and in brushland from sea level up to an elevation of 300 meters. It has been collected in flower from January to May, in August, and from October to December; in fruit in January and from August to November. Philippine Islands. Luzon: Laguna Prov., Mt. Banajao, Ocampo 27899 (us), Sulit 18877 (PNH, US); Isabela Prov., San Mariano, Ramos & Edano 46977 (Ny). Mrnporo: Puerta Galera, Britton 19422 (PpNH), Santos 5212 (vs), Ramos 46385 (Ny); Baco River, McGregor 131 (us, Ny), Merrill 4072 (us, (Mt. Giting-giting), Elmer 12451 (a-lectotype; cH, NY, US—isotypes). PANAY: Capiz Prov., Edano 46114 (ny), Escritor 21240 (us). Leyte: Wenzel 442 (a, ) Mussaenda magallanensis has affinities with M. anisophylla Vid. in the pubescence of the stem and leaves but differs from it in the entire or slightly bifid stipules, the 3-striate, linear or oblong-lanceolate sepals tapering to a point at the apex, and the glabrous or straggly-hairy corolla tube. Also it seems to be allied to M. chlorantha Merr. by the character of its stipules. Wenzel 442 and Ramos 15328 from Leyte differ from the type in the large leaves with longer petioles and seeds with testa bearing 3 to 6 foveae in the areoles. The specimen Ramos & Edano 46977 differs slightly in its large leaves, the petaloid sepals, and the smaller calyx lobes. This species is distinguished by the more or less horizontal pubescence on its stems and branches, the characteristic sepals and stipules, the large fruits with deciduous calyx segments, and the small seeds. Uses. The leaves of this species are used as a substitute for tobacco. 11. Mussaenda milleri Elmer, Leafl. Philip. Bot. 10: 3780. 1939. Laxly branched undershrub about 3 m. high with few, widely spreading, subterete, gray to brown brittle branches, rebranching toward their ends. bud, X 2%4; d (idem), calyx lobes, X 3; e (idem), corolla lobe, X 134. f-i, M. philippica: f (Elmer 7045), X 1; g (Ramos 17447), X 1%: h (idem), calyx lobes, X 3%; i (idem), corolla lobe, X 4. j-s, M. anisophylla: j (McGregor 23006), X 14; k (idem), corolla lobe from back, X 2%; 1 (Elmer 17481), X 1; m (idem), calyx lobes, X 2%; n-s, stipules, n (Merrill 2508), K 234; 0 (Mc- 2%; ; . x-aa, M. benguetensis: x (Santos 5497), x 1; y (idem), calyx lobe, X 3%; z (Loher 1523), X 1; aa (idem), calyx lobes, xX 2%. 122 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Leaves more or less clustered at the ends of branchlets, elliptic to ovate- or oblong-elliptic, 7-12 cm. long, 3-6 cm. broad, sharply acute or acu- minately recurved at apex, cuneate at base, pubescent on both surfaces, densely so on veins beneath, midrib straight, sunken along the upper sur- face and filled with reddish-brown hairs; veins 5—11 pairs, ascending, curved, cross bars numerous, very evident on the lower surface; petiole 1-2 cm. long, reddish brown, hirsute. Stipules broadly lanceolate, 9-12 mm. long, 5-5.5 mm. broad, bifurcate 1/2 way or more, lobes lanceolate, covered densely on both surfaces with reddish-brown hairs, the base within somewhat glabrous with large, linear glands in 2 groups covered with finer hairs, base of the outer surface red streaked along with the hairs. In- florescence terminal, corymbosely paniculate, hirsute, basal pair of branches subtended by foliaceous bracts, the upper branchlets with linear bracts similarly pubescent. Flowers on short, pubescent pedicels; calyx segments 3-5 mm. long, setaceously pointed; petaloid sepal occasional, yellowish white, subelliptic on a_ slender “petiole”; corolla slenderly tubular, 2-3 cm. long, hairy on the outer surface, lobes elliptic, orange-red, glabrous on the upper surface. Berry ellipsoid, narrowing towards the pedicel, 1.5 cm. long, lenticellate, reddish strigose-pubescent, calyx segments deciduous ; seeds minute, reticulate, triangular-ovate, broader than long, 0.6—0.7 mm. long, 0.67—0.8 mm. broad, with 2-4(—6) foveae in the areoles of the testa. DistTRIBUTION. This species is endemic to Luzon, and grows upon the summit mountain ridge at about 1067 meters above sea level. It has been collected in flower in March and in fruit in August and September. Philippine Islands. Luzon: Nueva Ecija Prov., Mt. Umingan, Ramos & Edano 26468 (NY, US), Aug—Sept. 1916, in fruit only. The isotypes of Mussaenda milleri apparently were not distributed to herbaria in the U. S. A. and were unavailable for examination. Ramos & Edano 26468 is nearest to the original description, agreeing with it in the form and pubescence of the leaves and stipules, but differing from it in the number of pairs of lateral veins in the leaf. As in the original description, the midrib is sunken on the upper surface and filled with reddish-brown pubescence. This collection was made much earlier than the type but not described, probably because of insufficient material. The holotype was destroyed in World War II; if isotypes are not located Ramos & Edano 26468 (Ny) should become the neotype. Mussaenda milleri is similar to M. nervosa in its small leaves, midrib sunken on the upper surface, short corolla lobes, and fruits with deciduous calyx segments, but differs from that species in the characters of the stipules, calyx segments, and pubescence of the fruit. The species is dis- tinguished from others by the smaller leaves with sunken midrib on the upper surface filled with reddish-brown pubescence, recurved leaf apices; characteristic broadly lanceolate stipules, hirsute with reddish-brown pubescence and streaked red on the outside; ellipsoid, lenticellate, strigose berries with deciduous calyx segments. 1964 | JAYAWEERA, MUSSAENDA IN THE PHILIPPINES beg 12. Mussaenda multibracteata Merr. Philip. Jour. Sci. Bot. 11: 34. 1916 (Type: Ramos 23585); Merrill, Enum. Philip. Fl. Pl. 3: 519. 1923. Fic. 4, f-n. Shrub or small tree, 2-4 m. high with terete stems and densely hirsute branches, leaves, and inflorescences. Leaves membranous, ovate, oblong- ovate, or lance-elliptic, 9.5-30 cm. long, 4-16 cm. broad, acuminate at apex, cuneate at base or long attenuate and decurrent on petiole, promi- nently hirsute with spreading hairs on both surfaces especially along the veins, lateral veins 8-11 pairs, conspicuous beneath; petiole 1—9.5 cm. long, hirsute. Stipules ovate, 12.5-14 mm. long, 8-10 mm. broad at the base, acuminate, bifurcate from the apex for about 1/2—2/3 their length, hairy on the outer surface, usually glabrous within or hairy at the base only, with numerous glands in 2 groups. Inflorescence a terminal, di- or trichotomously branched, diffuse, densely hairy, cymose panicle; bracts and bracteoles subpersistent, numerous, crowded, lanceolate, 5-13 mm. long, acuminate, entire, bifid or trifid at apex, lateral lobes much shorter than the central one, hirsute on the outer surface, glabrous within with a tuft of hairs at the base. Flowers heterostylous, on stout, densely hirsute pedicels shorter than the ovaries. Calyx lobes broadly lanceolate, 6.7—28 mm. long, 1-6 mm. broad, acuminate, hirsute on the outer surface, spar- ingly pubescent or glabrous within with a tuft of hairs and 1 or 2 pairs of glands at the base of each; petaloid sepal ovate or elliptic, 5.5-10.5(—12.5) cm. long, 3-6(—9) cm. broad, acute or subacute at apex, cuneate at base, hirsute on both surfaces and 5-veined, “petiole” 2 cm. long, densely hirsute. Corolla yellow, the tube stout, 2.5-3.3 cm. long, densely hirsute on the outer surface with hairs almost horizontally spreading, hairy within 3/4 way down from the mouth of the tube in long-styled forms and 1/2 way in short-styled forms; hairs within not tufted at the mouth and shorter in long-styled forms; corolla lobes ovate or orbicular, 3.5—-6 mm. long, 4.5—6 mm. broad, hairy on the outer surface, glabrous within. Stamens with short filaments, epipetalous below the middle of the tube 2/5—1/2 way up in long-styled forms and above the middle (2/3 way up) in short-styled forms; anthers linear, dorsifixed, introrse, 4.5-5 mm. long, bifid at the base. Ovary broadly or cylindrically turbinate, 4.5-6 mm. long, densely hirsute, 2-locular with numerous ovules on cushion-shaped, axile placentae; style and stigma lobes 1.7—2 cm, and 8.5—10 mm. long respectively in long- styled forms, 3 mm. and 2.5 mm. long in short-styled forms. Berry oblong or elliptic, 1.2-2 cm. long, hairy, lenticellate with persistent calyx seg- ments; seeds minutely obnapiform, reticulate, 0.53—0.67 mm. long, 0.46— 0.56 mm. broad, faintly spiny but not obvious, with 4-11 foveae in the areoles of the testa. DistTRIBUTION. This species grows by streams or on damp slopes in the primary forests of Luzon and Catanduanes from 150 to 2100 meters elevation. It has been collected in flower from May to August, in fruit from July to September, November, and December. Philippine Islands. Luzon: Albay, Mayon Volcano, Mendoza 18431 (PNH, 124 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV us), 18376 (pNH); Tabayas, Mt. Binuang, Ramos & Edano 28687 (A, US), 28783 (A, us); Mt. Tulaog, Ramos & Edano 29093 (a, us); Casiguran, Ramos & Edano 45480 (Ny); Camarines Sur, Her-it River, Edano 76425 (ny); Pili, Mambugna (Mt. Isarog), Convocar 2820 (A, PNH); Cagayan, Mt. Dos Cuernos, Ramos 77006 (pNH); Camarines Norte, Minasag River, Edano 40230 (PNH); Isabela, Clemens 17011 (Ny); Sorsogon, Ramos 23715 (us), Ramos 23585 (ny-lectotype). CATANDUANES: Ramos 30509 (A, Us), 30262 (us); Mt. Tagmasuso, Ramos & Edano 75259 (Ny), 75159 (ny); Bato Trail to Viga, Ramos & Edano 75143 (NY, US). This species resembles Mussaenda philippinensis in its persistent, lan- ceolate sepals which are hairy on both surfaces. It differs considerably, however, in its characteristic pubescence, deeply bilobed stipules, longer corolla tube with lobes which are glabrous within, and the larger stamens, ovaries, fruits, and smaller seeds. Mussaenda multibracteata is charac- terized and distinguished from other species by its pubescence; large leaves; numerous, crowded, subpersistent bracts and _bracteoles; large stipules which are glabrous within; broadly lanceolate, persistent sepals; hirsute petaloid sepals; corolla lobes glabrous within; large, hairy, len- ticellate fruits and small, faintly spiny seeds with 4-11 foveae in the areoles of the testa. 13. Mussaenda nervosa Elmer, Leafl. Philip. Bot. 3: 994. 1911 (TyPE: Elmer 10510); Merrill, Enum. Philip. Fl. Pl. 3: 519. 1923. Fic. 1, e; Fic. 2, i, j, k, ka. Shrub 3 m. high with much-branched stems and erect branchlets. Leaves ovate-oblong or lance-elliptic, 6-15 cm. long, 2.5—5 cm. broad, acute or acuminate at apex, cuneate, acute or obtuse, and sometimes decurrent at the base, rugose and glabrous or pubescent on the upper surface, hairy on veins on the lower surface with 10-15 pairs of parallel curved-ascending lateral veins, very prominent beneath but sunken into the blade on the upper surface; petiole 0.5—4 cm. long, pubescent. Stipules deciduous, ovate- acuminate, 4-5.5 mm. long, 3.5 mm. broad at the base, bifurcate from the apex for about 1/2 their length, lobes slightly diverging, hirsute on the outer surface, glabrous within with a few glands in 2 groups at the base. Inflorescence an erect, terminal, dichotomously branched, pubescent cyme; bracts and bracteoles setose-acuminate, 0.7—1,mm. long, glabrous within. Flowers heterostylous, not clustered, on very short, stout, pubescent pedi- cels. Calyx lobes linear, laciniate, 11-16.5 mm. long, 0.7—1 mm. broad, with a prominent midvein terminating in a fine point, pubescent on the outer surface, glabrous or pubescent within, with or without a tuft of hairs, and with 1 or 2 pairs of glands at the base; petaloid sepal ovate, 5—8 cm. long, 3-5.2 cm. broad, glabrous on the upper surface, hirsute on veins beneath, cuneate at the base, “petiole” 1.2-3 cm. long, hirsute. Corolla tube 2.5-3 cm. long, subfalcate, hairy on the outer surface, hairy within as far as the bases of anthers or a little lower; hairs long in short-styled forms and short in long-styled forms, not tufted at the mouth. Corolla lobes 1964] JAYAWEERA, MUSSAENDA IN THE PHILIPPINES 125 deep yellow, ovate-lanceolate or elliptic-acuminate, 8.5—9 mm. long, 4.5—5 mm. broad, hairy on the outer surface, granular within and 3- or 5-parallel- veined. Stamens with short filaments, epipetalous on the tube about 1/2 way up in short-styled forms, below the middle of the tube (about 2/5 way up from the base) in long-styled forms; anthers linear, dorsifixed, introrse, 6.5—7.5 mm. long, acute at apex and bifid at the base. Ovary broadly turbinate, 4—4.5 mm. long, hairy, 2-locular with numerous ovules on cushion-shaped, axile placentae; style and stigma lobes 1.7 cm. and 8.5 mm. long respectively in long-styled forms, 1.45 cm. and 2 mm. long in short-styled forms. Berry ellipsoid or globular, 1 cm. long, hairy when young, glabrous at maturity, lenticellate, calyx segments deciduous; seeds minute, reticulate, angularly globose, 0.46—0.6 mm. long, 0.36—0.43 mm. broad, with 4-12 foveae in the areoles of the testa. DIsTRIBUTION. This species grows on the islands of Luzon and Min- danao in forests up to an elevation of about 1200 meters above sea level. Elmer’s type specimen was gathered from the densely forested basin of Mt. Apo in Todaya District. The species has been collected in flower in May, June, and July; in fruit in August and September. Philippine Islands. Luzon: Nueva Ecija Prov., Ramos & Edano 26422 (A, NY, US). Mrinpanao: Bukidnon Subprov., Ramos & Edano 39077 (a, US); Davao Dist., Todaya (Mt. Apo), Elmer 10510 (a-lectotype; GH, NY, US- isotypes). Mussaenda nervosa seems to be distantly allied to M. philippinensis with which it agrees only in the long calyx segments, and from which it differs considerably in the characters of the stipules, petals, and fruits. On the other hand it shows affinity to M. milleri in that both species are shrubs of comparable size, with the veins of the leaves sunken on the upper surface, which occur in the same locality in Luzon. Mussaenda nervosa differs from M. milleri, however, in the ovate-acuminate, short stipules, longer calyx segments, subfalcate corolla tube, smaller and glabrous fruits, and smaller seeds with a greater number of foveae in the areoles of the esta. Mussaenda nervosa is distinguished from other species by the small leaves, with the midrib and 10 to 15 pairs of lateral veins sunken on the upper surface; small stipules glabrous within and with fewer glands; laciniate calyx segments; subfalcate corolla tube; glabrous and globular or ellipsoid berries with deciduous calyx segments, and small seeds with 4 to 12 foveae in the areoles of the testa. 14. Mussaenda palawanensis Merr. Philip. Jour. Sci. Bot. 10: 103. 1915 (Type: Fenix 15531); Merrill, Enum. Philip. Fl. Pl. 3: 519. 1923. Fic: 1, k. Erect shrub 2—3 m. tall, more or less ciliate-hirsute in all parts with reddish-brown, terete, glabrous branches, the younger ones hirsute. Leaves oblong or oblong-ovate, 7-18 cm. long, 3-10 cm. wide, abruptly acuminate 126 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV at apex, rounded or cuneate at base, hairy on both surfaces, upper surface with scattered stiff hairs, lower surface with paler hairs also scattered, margins ciliate, lateral veins 9-14 pairs; petiole stout, 0.5—-0.9 cm. long, hirsute. Stipules lanceolate, 8-10 mm. long, 4.5—-8.5 mm. broad at the base, acuminate and bifurcate more than 1/2 their length from the apex, * 5, note kate oa ad c (idem), corolla oe x 4%. d, e, M. setosa d (Merrill 9496), ovary with persistent calyx lobes, X 214; e (idem), eee ithi 414, f-m, M. multibracteata: f-h, stipules; f (Convocar 2820), * 2%; g (Ramos & ares 28783), X 2%;h (Ran mos &¥ Edano 29093), x 2; 1 (Convocar 2820), * 1; j (Ramos & Edano 28783), * 1; k (idem), calyx lobes, & 2°4, glabrous ae except for band of hairs at base; l (Ramos & Edano 29093), aap lobes, X 2, hairy within; m (Convocar 2820), calyx lobes, ~ Zz, me _ i hairiness (hairs anitted nen 1 lobe); n (idem), corolla lobe, acuminatissima: 0 (Ramos 33133), * 2; p (idem), caly lobes. ay ae q a. corolla lobe, & 414. 1964] JAYAWEERA, MUSSAENDA IN THE PHILIPPINES 127 lobes subulate and diverging, hirsute on the outer surface, glabrous within except at the base among the few glands occurring in 2 groups. Inflores- cence of terminal, few-flowered, densely hirsute cymes, flowers usually in crowded terminal clusters; bracts and bracteoles lanceolate, 3.5—6 mm. long, hairy on both surfaces, bracteoles trifid, lateral lobes very short. Flowers heterostylous on stout, pubescent er shorter than the ovaries. Calyx lobes lanceolate, 6.7—10.5 mm. long, 2—2.5 mm. broad, slenderly acuminate, hairy on both surfaces, hairs pen within and scantier, with 3 or 4 pairs of glands at the base of each; petaloid sepal white, broadly ovate, 6-12 cm. long, 4-9 cm. broad, bluntly acuminate at the apex, cuneate at the base, puberulous on the upper surface, hirsute beneath; “petiole” 1—1.5 cm. long, hirsute. Corolla white to bright orange, the tube 2—2.2 cm. long, infundibuliform, appressed hirsute on the outer surface, hairy within 1/2 way in long-styled forms and 1/3 way from the top in short-styled forms, hairs not tufted at the mouth and shorter in long-styled forms; corolla lobes ovate, 6 mm. long, 4 mm. broad, prominently acumi- nate or caudate, hairy on the outer surface, papillate within. Stamens with short filaments, epipetalous on the tube just above the middle in long- styled forms and 2/3 way up in the broader portion in short-styled forms; anthers linear, dorsifixed, introrse, 5—6.5 mm. long, bifid at the base. Ovary cylindric-turbinate, 4.5—6 mm. long, densely hirsute, 2-locular with numerous ovules on axile placentae; style and stigma lobes 1.4 cm. and 6.5 mm. long respectively in long-styled forms, 3.5 mm. and 3.2 mm. long in short-styled forms. Berry ovoid or obovoid, 1.2—1.4 cm. long, pubescent, lenticellate, calyx segments deciduous; seeds minute, reticulate, broadly triangular-ovate, broader than long, 0.6—0.77 mm. long, 0.73—0.83 mm. broad; with 3-6(—8) foveae in the areoles of the testa. DiIsTRIBUTION. This species is endemic to Palawan where it occurs in thickets and old clearings up to an elevation of about 150 meters above sea level. It has been collected in flower in February, July, August, and December; in fruit in December. Philippine Islands. PALAWAN: San Antonio Bay, Merrill 867 (us); Baraki, Fox 13354 (A, PNH); Mt. Ibusi, Ebalo & Conklin 1227 (a); Point Separacion, Fenix 15531 (us-lectotype); Puerto Princesa, Kondo & Edano 36685 (PNB); Mt. Kabangaan, Edano 77718 (A); Brooke’s Point, Edano 244 (a, PNH). Of the collections examined Ebalo & Conklin 1227 differs from the lectotype in that its leaves are larger (18 cm. long), petaloid sepals 12 cm. long, stipules broader, and it bears orange flowers while it agrees in all other characters. The fruits of Fox 13354 are larger (1.2—1.4 cm. long) than described for the type. Kondo & Edano 36685 also possesses larger fruits but its leaves are obovate-lanceolate, resembling those of Mussaenda attenuitfolia. The species is closely allied to Mussaenda philippica from which it differs in the pubescence of the leaves, size and form of calyx and corolla segments, and in the seeds which, however, resemble those of M. bengue- tensis. It is distinguished from the other species by the characteristic 128 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV ciliation of its short-petioled leaves; lance-acuminate, bifurcate stipules with diverging, subulate lobes; broadly lanceolate sepals; short, thick, appressed-hirsute corolla tube; and scantily pubescent, ovoid berries with deciduous calyx segments. 15. Mussaenda philippica A. Rich. Mém. Soc. Hist. Nat. Paris 5: 245. 1834 (Type: Perrottet); Merrill, Enum. Philip. Fl. Pl. 3: 519. 1923. Fic. 1, j; Fie. 3, f, g, bh, i. M. acutiflora Bartling ex DC. Prodr. 4: 370. 1830. Calycophyllum grandiflorum Meyen, Reise 2: 234. 1835. M. abe sensu Blanco, FI. Filip. 167. 1837; ed. 2.118. 1845; ed. 3. 1: 211. pl. 58. 1877, non M. glabra sensu Fernandez- Villar, Novis. App. 108. 1880, non Vahl. M. grandiflora Rolfe, Jour. Linn. Soc. Bot. 21: 311. 1884. Shrub or small tree 2—8 m. tall with terete or subterete, more or less pubescent or nearly glabrous, brown or gray, ascending, lenticellate branches, Leaves very variable, ovate, oblong-ovate or -lanceolate or obovate-oblanceolate, strongly conduplicate on the upper side, 3.5—23.5 cm. long, 1.7-12 cm. broad, acuminate at apex, cuneate or long attenuate at base, glabrous or scantily pubescent on the upper surface, glabrate on the lower surface, minutely pubescent or hirsute especially along the veins, lateral nerves 6-15 pairs, arcuate; petioles 0.3—2.5(—7.5) cm. long, minutely appressed pubescent or hirsute. Stipules oblong-ovate or acumi- nate from a broad base, 3—10.5(—12.5) mm. long, 2.2-7 mm. broad at the base, bifurcate from apex 1/4—-3/4 their length, lobes not diverging and somewhat straight, pubescent on the outer surface, glabrous within, hairy throughout or at the base only and among the glands occurring in 2 groups. Inflorescence a terminal, di- or trichotomously branched, many- or few- flowered cyme; bracts and bracteoles lanceolate or ovate, 3-5 mm. long, hairy on the outer surface, glabrous within. Flowers heterostylous on stout, pubescent pedicels shorter or longer than the ovaries. Calyx lobes triangular or lance-acuminate, 1.2—7.5 mm. long, 0.5—1.5 mm. broad, hairy on the outer surface, glabrous or pubescent within with 2—S pairs of glands at the base of each; petaloid sepal ovate or orbicular-ovate, 4-13 cm. long, 3—8.5 cm. broad, acute or subacute at apex, abruptly cuneate at base, glabrous on both surfaces or puberulous on the upper surface and hairy beneath, “petiole” 1—2 cm. long, pubescent. Corolla tube pale green, 2—3.5 cm. long, minutely appressed pubescent or hirsute on the outer surface, hairy within as far as the bases of the anthers, hairs within longer in short-styled forms than in long-styled forms, not tufted at the mouth; corolla lobes orange, lanceolate, ovate, or orbicular, 2.5-8(—12) mm. long, 2—5.5 mm. broad, apiculate, acuminate or somewhat caudate, hairy outside, papillate within. Stamens with short filaments epipetalous on the corolla tube 1/2—3/4 way up in short-styled forms and at about the middle in long-styled forms; anthers narrowly linear or lance-linear, dorsifixed, introrse, 6—8.5 mm. long in short-styled forms, 4-6 mm. long in long- 1964 | JAYAWEERA, MUSSAENDA IN THE PHILIPPINES 129 styled forms, acute or blunt at apex, bilobed at the base. Ovary broadly fusiform, oblong-turbinate or oblong, 3.5—7.5 mm. long, minutely appressed pubescent or hirsute; style and stigma lobes 1.5—2.9 cm. and 3—7.5 mm. long respectively in long-styled forms, 2.5—-8.5 mm. and 2—4.5 mm. long in short-styled forms. Berry globular, 1-1.6 cm. long, glabrous or scantily pubescent, lenticellate, calyx segments deciduous; seeds minute, reticulate, not spiny, 0.6-1.0 mm. long, 0.39-0.8 mm. broad, testa with 2—5(-7) foveae in the areoles. ILLUSTRATIONS. Blanco, FI. Filip. ed. 3. 1: pl. 58. 1877; Vidal, Synop- sis de familias. Atlas p/. 56, f. 1883 (not seen); Sulit, Philip. Jour. Forestry 2(1): 43. pl. 3, fig. 2. 1939. DistRIBUTION. This species has a very wide distribution. It occurs in almost all the islands of the Philippines, extending from Luzon to Palawan and Mindanao, eastward to Palau and Yap islands of the Caroline Group, and southward in the Solomon and Fiji islands, from sea level to an altitude of about 700 meters. It is a common shrub found along forested ridges, slopes and streams in old second-growth jungles and also in rugged, open, rolling country. It has been collected in flower and fruit throughout the year. Philippine Islands. BATan: Mt. Iraya, Ramos 80283 (A). Luzon: Cagayan Prov., Littoc, Adduru 142 (a, us); Lagum, Adduru 218 (A, us); Velasco 24852 (ny); Rizal Prov., Antipolo, Vidal 389° (a), Ramos & Edano 29536 (A, US), 45292 (ny), Reillo 19245 (us); west of Famy, Brenner, Jr. 2584 (PNH), Ramos 34 (us), 13573 (us); San Mateo, Ahern’s Collector 119 (us); Bosoboso, Merrill 2655 (ny, us), Vidal 14574 (a); Camarines Prov., Nueva Caceres, Vidal 1457s (a); Dalupaon-Paracao, Ahern 818 (us), 802 (us); Dact, Hallier (NY); Kamugong River, Edano 75879 (Ny); Paracale, Ramos & Edano 33520 (us), 33469 (Ny); Carambola, Pili, Convocar 2945 (a, pNH); Mt. Isarog, Merrill, Species Blancoanae 413 (A, GH, NY, US); Sorsogon Prov., Curran wes (ny, us); Irosin, Edano & vg elge 37744 (PNH), 38555 (A, PNH); Bulusan, Elmer 14388 (A, C, GH, NY, US); Tabayas Prov., Manuel 23486 ae Guinayangan, Vidal 801 (a); samt ‘Escritor 21192 (us); Lucena, Merrill 2890 (ny, Us); Sariaya, Whitford 559 (us); Bulacan Prov., Quingua, Vidal 1457¢ Sue Angat, Vidal 389¢ (A); Malinto, Robinson & Merritt 6128 (GH); Laguna , Los Banos, Elmer 8134 (ny), Steiner 40063 (PNH), Holman 43 (GH); Mt. Makiling, Elmer 17670 (A, GH, NY, US), Sulit 9747 (A, PNH), Mendoza 12224 (us); Makiling National Park, Salvoza 3214 (PNH); Pangil, Ramos 39882 (PNH); Bataan Prov., Pascual 28682 (A), 23071 (A); Mt. Mariveles, Elmer 6673 (NY); Lamao River, Williams 127 (Ny, us), Whitford 388 (Ny, US), 524 (ny, us), Borden 1220 (Ny, us), Bartlett 14653 (A); Bagac, ae 27292 (aA); Albay Prov., Mayon Volcano, Mendoza 18350 (PNH); bela Prov., Palanan, McGregor 10726 (Ny); Apayao Subprov., Fenix ee (A, US); Cavite Prov., Ramos & Deroy 22518 (us); Manila, San Francisco, Loher 1520 (us); Novaliches, Loher 1519 (us). ALABAT: Ramos & Edano 48089 (ny). CATANDUANES: Ramos 30452 (us). MARINDUQUE: Vidal 1457 (a). Mtnporo: Mansalay, Merrill 912 (GH, NY, US); Mt. Yagaw, Sulit & Conklin 16840 (A), 16875 (PNH); Sablayan, Ligaya, Reed 40895 (PpNH); Bulalacao, Padam Mt Ebalo 254 (A). Samar: Vidal 3895s (a), Ramos 17447 (us); Catubig River, 130 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Edano 24806 (us), Sablaya 45 (a), Ramos 24174 (a, us); Tubabau Island, ree tea 2045 (a); Matuguinao, Baruz, Gachalian 15483 (pNH); Oras, d (pro), Alcasid & Oano 39863 (PNH); Laquilacon, McGregor 43797 (NY), Sulit 6125 (A, PNH); Borongan, pees Castro & Anonuevo 5810 (A), Pinamgasan, Castro & Anonuevo 5853 (PNH). BILIRAN : McGregor 18813 (a, us). BUSUANGA: Ramos 41262 (A), 41236 (A, US). CULION: Bermejos 182 (GH, NY, PNH, US). Panay: Capiz Prov., Edano 46220 (ny); Dumarao, Taleon 22279 (pnu), 33828 Sola Libacao, Martelino & Edano 35331 (Ny); Antique Prov., McGregor 32540 (us). GUIMARAS: Gammill 261 (NY, US). LEYTE: Wenzel 57 (A, GH, US), 179 (A, GH, US); Palo, Elmer 7045 (a, NY). NEGROS: Mt. Katugasan, Edano 21809 (PNH), 21782 (pit): Dumaguete (Cuernos Mts.), Elmer 10121 (A, NY, US). PALAWAN: Foxworthy 605 (GH, NY, PNH, US), 691 (NY, US), Prrighia tes 350 (GH, NY, US), Curran 4517 (NY, US); Taytay, Merrill 9332 (A, NY, US); Puerto Princesa, Cenabre spe (A); ene eer (NY, US); Mt. ee Elmer 12792 (A, GH, NY, US). Minpanao: Devore & Hoover 172 (us), 163 (us); Agusan Prov., Cabadbaran (Mt. Urdaneta), Elmer 13301 (A, GH, NY, US); Davao Prov., Madam, Edano 1312 (A, pNH): Catalnan, Kanehira 2497 (Ny); Magdug River, pete 11074 (A); Santa Cruz, Williams NY, US); Cotabato- Dansalan, Zwickey 34 (A, NY); Davao Dist., Copeland 355 ine us); Surigao Prov., Ahern 339 . Mt. Cantugas, Ramos é& Convocar 83500 (a); Butuan Subprov., Miranda 20580 (us); Misamis Prov., Mt. Malindang, Mearns & Hutchinson 4716 (Ny, us); Zamboanga Prov., Margosa Tubig, Guerrero 29549 (a); Dikus, Frake 38115 (a, PNH), Frake 36058 (PNH). Barapac: Vidal 3894 (A); Mangubat 481 (us). CAGAYAN: Sulu, Mearns 36 (us), 37 (us), Warburg 14896 (a). Patmas: Merrill 5338 (Ny, US). Caroling Islands. Yar: Mt. Matade, Fosberg 25555 (Ny); Balabat, Taka- matsu 1872 (A). PALAU: Fosberg 25769 (ny); Takamatsu 1151 (A). Solomon Islands. BOUGAINVILLE: Kupei Gold Field, AKajewski 1666 (A). Santa YSABEL: Meringe, Brass 3538 (A). OWA RIKI: Brass 3075 (A). Fiji Islands. WayA: Fosberg 18008 (A). From an examination of over 200 collections of this species from the different islands of the Philippines, the Caroline Group, Fiji, and the Solomon Islands it is evident that some are allied to Mussaenda glabra, others to M. palawanensis, and still others rather distantly to M. macro- phylla, M. philippinensis, M. pinatubensis, and M. anisophylla. The collections Elmer 12792, from Palawan; Brass 3538 and 3078, from the Solomon Islands; Mepill 912, from Mindoro: Elmer 14388, Curran 10535, Fenix 28234, Merrill 413, Convocar 2945, etc., from Luzon; and Fosberg 25555, 25769, Takamatsu 1151, 1872, from the Caroline Is- lands, bear resemblance to Mussaenda glabra in that their leaves are glabrate, minutely appressed pubescent on veins beneath, their stipules small and lanceolate with few glands at the bases within, and the corolla tubes are also minutely pubescent with short, lanceolate lobes. They differ, however, in the size and ios venation of the leaves, the length of the corolla tubes, petals, anthers, The collections Adduru 142, ae 17670, Whitford 524, Ramos 29882, Robinson & Merritt 6128, etc., are allied to Mussaenda palawanensis. Their leaves are pubescent on both surfaces, densely so on the lower sur- 1964 | JAYAWEERA, MUSSAENDA IN THE PHILIPPINES 131 face, the sepals are broadly lanceolate, acuminate, hairy on both surfaces, and the petals are ovate and prominently acuminate. They differ in their longer petioles, shorter calyx lobes, and less hirsute corolla tubes. One collection, Merrill 5338, from Palmas Islands, bears sepals similar to those of Mussaenda philippinensis but differs in numerous other char- acters which establish that it is M. philippica. Edano 46220 and Warburg 14896 have some resemblance to M. macrophylla in the pubescence of their stems and inflorescences, and in their orbicular-apiculate petals. Merrill 2655, from Luzon, bears affinities to M. pinatubensis in that its branchlets are erect, opposite, and terminated by leaves and the flower-bearing por- tion. It has small leaves, and the older pairs of stipules are fused around their nodes. The specimen Edano & Gutierrez 37744 has spreading hairs on the stem and deeply bifurcate stipules hairy on both sides, characters of M. anisophylla. The collections Brass 3075 and 3538, from the Solomon Islands, which resemble Mussaenda glabra in external characters, are related by other characters to M. frondosa L. in that the throat hairs of the corolla tube are long in the long-styled forms and extend below the bases of the anthers. Throughout the species the seed character seems to be constant, though the seeds themselves vary in size and shape. The testa is foveolate with 3-5(-7) foveae in the areoles. Cortez & Fernandez 34388 bears smaller fruits (0.8 cm. long) and somewhat spiny seeds with 3 to 10 foveae in the areoles of the testa, while Frake 36058 also bears smaller fruits but has larger seeds (0.73-0.93 mm. long) with rugose testa. From the foregoing it is evident that at least two genetic lines are in- volved in the evolution of this species: M. glabra coming down from the mainland of China and M. palawanensis from Palawan. Mussaenda philippica is distinguished from other species by its ascend- ing, freely rebranching branchlets; its glabrous or pubescent leaves re- curved and strongly conduplicate; its erect green inflorescences and corolla tubes; its large anthers; its glabrous, lenticellate fruits with deciduous calyx segments; and its small, smooth seeds with reticulate testa bearing 2—5(—7) foveae in the areoles. Uses. The plant is used medicinally for stomach-ache in the Fiji Islands. The juice of the bark is used as a cure for headache in Mindoro. The leaf is employed as a substitute for tobacco, while in Mindanao the whole plant is used in agricultural rituals. Mussaenda philippica forma aurorae (Sulit), stat. nov. M. philippica A. Rich. var. aurorae Sulit, Philip. Jour. Forestry 2(1): 39. pl. 3, fig. 1. 1939 (Type: Mabesa 24876). Shrub 1-3 m. high with terete, pubescent branches. Leaves ovate or elliptic, 8-16 cm. long, 3.2-8.5 cm. broad, abruptly acuminate, cuneate at base, glabrate on the upper surface, pubescent beneath with 8-10 pairs of lateral veins; petiole 0.7—1.5 cm. long, hirsute. Stipules ovate, 7-9 mm. long, 5-6 mm. broad at the base, hairy on the outer surface, glabrous 132 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV within except below the glands, acuminate, bifurcate from apex 1/3-1/2 their length. Inflorescence a terminal, pubescent, few-flowered cyme; bracts and bracteoles lanceolate, 3.5-4 mm. long, hairy. Calyx lobes when un- expanded linear or lanceolate, 3-6 mm. long, 1 mm. broad; when expanded white, oblong, “petioled,” 3.5—-9 cm. long, + 1.5—5.5 cm. broad, pubescent on both surfaces, 5-veined, stipe + 1 cm. long, hairy. Corolla tube 2.5—3 cm. long, broader at the top, hairy on the outer surface, hairy within as far as the bases of the anthers; corolla lobes yellow, ovate, 4.5—6 mm. long, 4—5 mm. broad, apiculate, hairy on the outer surface, papillate within. Stamens with short filaments, epipetalous 2/3-3/5 way up on the corolla tube; anthers linear, dorsifixed, introrse, 7-8 mm. long, blunt at apex, bilobed at the base. Ovary turbinate, 3.5 mm. long, hairy, 2-locular and with axile placentation; style and stigma lobes each 3.5 mm. long, the latter appressed on their receptive surfaces. DIsTRIBUTION. Mussaenda philippica forma aurorae was first collected by Mabesa at the foot of Mt. Balong-bulo, Laguna Province, in 1915. It was collected in 1930 by Curran in Sitio Buot, Makiling National Park, growing at an elevation of 100 meters above sea level. Now it is much cultivated as an ornamental shrub. It has never set seed and all plants in cultivation have been raised vegetatively from Curran’s plant which was transplanted into the Forestry Nursery and propagated. Philippine Islands. Luzon: Laguna Prov., Mabesa 24876 (a-lectotype; vs— isotype) ; Los Banos, Forestry College, Walker 7456 (us), Steiner 1743 (pNH); anila Garden, Quisumbing 2101 (pNH), Steiner 2950 (a); Malacanan Grounds, Quisumbing 4334 (A) Forma aurorae differs from Mussaenda philippica forma philippica in the one to five expanded calyx segments in the flowers of an inflorescence but resembles it in all other characters. It is a short-styled form and has never produced fruit. Though the ovary contains ovules, the stigma lobes seem to remain appressed on their receptive surfaces, preventing pollina- tion. Cuttings of this plant sent by Dr. L. H. McDaniels to the Fairchild Tropical Garden, Florida, from the Philippine Islands in 1959, were suc- cessfully rooted and multiplied. As a male parent it has been used to good effect in the Philippine Islands in crosses with M. erythrophylla Schum. & Thonn., an African species with the petaloid sepal crimson, producing a number of cultivars. The two most outstanding of these are called Mussaenda ‘Luz’ or ‘Dona Luz’ and Mussaenda ‘Alicia’ or ‘Dona Alicia,’ both of which are much cultivated for their ornamental qualities. In the former all five calyx segments are deep rose-purple and petaloid while in the latter only one segment is petaloid. — On Mussaenda philippinensis Merr. Philip. Jour. Sci. Bot. 3: 264. 1908 (Type: Merrill 4139); Merrill, Enum. Philip. Fl. Pl. 3: 520. 1923. Fic. 1 ,n, 0, p; Fic. 3, t, u, v, w. Shrub 1-4 m. high with more or less hirsute, terete branches. Leaves 1964] JAYAWEERA, MUSSAENDA IN THE PHILIPPINES 133 membranous, oblong-elliptic, 9.5-20 cm. long, 3.8-10 cm. broad, nearly glabrous or scantily pubescent on the upper surface, hirsute on veins of the lower surface with scattered hairs on the lamina, acuminate at apex, cuneate at base or long decurrent on the petiole, veins 8-12 pairs, petiole 0.3-4 cm. long, hairy. Stipules lance-ovate, acuminate, 9-13 mm. long, 4.5 mm. broad at the base, hairy on both surfaces, entire or faintly bifid at the tip with numerous glands in a continuous band at the base, ascending in the middle. Inflorescence a terminal, hirsute, dichotomously branched, somewhat compact, cymose panicle; bracts and bracteoles deciduous, linear-lanceolate, hirsute; bracteoles trilaciniate, lobes narrow, acuminate. Flowers heterostylous on stout, pubescent pedicels shorter than the ova- ries. Calyx lobes narrowly lanceolate, 10.5-15 mm. long, 1 mm. wide, hairy on the outer surface, glabrous within or pubescent with 2 or 3 pairs of glands at the base of each sepal; petaloid sepal white, ovate or elliptic, 5-9 cm. long, 3—6 cm. broad, acute or subacute at apex, cuneate at the base, glabrous on the upper surface, hirsute on veins beneath, ‘‘petiole” 1.7—2.5 cm. long, hirsute. Corolla tube yellow, 2—2.5 cm. long, pubescent on the outer surface with short appressed hairs mixed with few long spreading ones, hairy within as far as the bases of the anthers, not tufted at the mouth; hairs shorter in long-styled forms than in short-styled forms, glabrous below the anthers; corolla lobes ovate, 1.5—3 mm. long, 2.2-3 mm. broad, acute or acuminate, hairy on the outer surface, papillate within. Stamens with short filaments, epipetalous on the tube below the middle, 2/5 way up in long-styled forms and 3/5 way up in short-styled forms; anthers linear, dorsifixed, introrse, 3-5 mm. long, abruptly acute at apex, bifid at base. Ovary obconical or turbinate, 3-3.5 mm. long, scantily appressed hairy, 2-locular with numerous ovules on cushion-shaped, axile placentae; style and stigma lobes 1.9 cm. and 6.5 mm. long respectively in long-styled forms, 1-2 mm. and 1.5—2 mm. in short-styled forms. Berry elliptic, 1.3-1.5 cm. long, rugose, scantily pubescent with persistent calyx segments; seeds minute, reticulate, broadly oblong or angularly globose, 0.6—-0.67 mm. long, 0.46—0.6 mm. broad, with 3-8 foveae in the areoles of the testa. DISTRIBUTION. This species grows in partial shade along creeks and ridges of second-growth forests from 10 to 300 meters elevation above sea level. It is confined to the islands of Luzon, Alabat, Mindoro, Semirara, and Panay. It has been collected in flower from April to July; in fruit in June, November, and December. The original description is based on a single collection from Semirara. Philippine Islands. Luzon: Apayao Subprov., Mt. Duraragan, Edano 19875 (PpNH), alt. 300 m. ALaABaT: Merrill 10437 (A, NY, PNH, US). MINDORO: Mt. Yagaw (eastern slope), Sulit & Conklin 17652 (a, PNH), alt. 267 m., 16875 (A), alt. 300 m.; Oyon Creek, Paniza 9436 (A, PNH), alt. 10 m., Ebalo 193 (a, PNH). SeMIRARA: Merrill 4139 (ny-lectotype). Panay: McGregor 32449 (a, US). Several collections have been erroneously placed in this species which should be referred to Mussaenda magallanensis. The collection Merrill 134 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV 10437 (A, NY, PNH, US) from Alabat Island possesses large, lance-elliptic leaves with the characteristic pubescence and hairy, lenticellate berries with persistent calyx segments, apparently resembling this species, but an examination of the seed indicates that it is nearer M. magallanensis. Its seeds are black, smaller (0.46—0.6 mm. long), obtusely 3- or 4-sided, testa faintly spiny with 4-12 foveae in the areoles. McGregor 32449 (A, Us) from Panay seems to be a variation from the type collection, differing from it in the shorter calyx segments (8 mm. long) and larger, acuminate petals (6 mm. long). Mussaenda philip pinensis seems to be allied to M. magallanensis, on the one hand, and to M. wrayii King, from the Malay Peninsula, on the other. It is distinguished by its entire or almost entire stipules, compact inflo- rescence, long, persistent calyx lobes, large fruits with small seeds, and 3-8 foveae in the areoles of the testa. 17. Mussaenda pinatubensis Elmer, Leafl. Philip. Bot. 9: 3210. 1934 (Type: Elmer 21978). Shrub with spreading main branches dividing to terminate in glabrous, subangular, erect, short branchlets. Leaves opposite, often unequal, 4—8.3 cm. long, 2—3.7 cm. broad, slenderly or sharply acuminate at the apex but obtuse in many of the smaller leaves, cuneate at base, glabrous or puberu- lous on the upper surface, short-hairy on veins and venules beneath, lateral veins 6-11 pairs, ascending; petiole 3-10 mm. long, grooved above, sparse- ly strigose beneath. Stipules lanceolate, 5—8.5 mm. long, 2.5—-5 mm. broad at the base, bifid from the apex for about 1/5 their length, teeth terminating in a gland, hairy on the outer surface, hairy within at the base and apex, glands many in 2 groups, opposite pairs of older stipules fusing at their edges at the base to form a ring around each node. Inflorescence a terminal, dichotomously branched, many-flowered, minutely appressed-pubescent cyme almost equalling the leaves in length; bracts and bracteoles lance- olate or linear, 5—6 mm. long, hairy. Flowers probably heterostylous on stout, appressed-pubescent pedicels about 3 mm. long. Calyx lobes persis- tent, linear subulate, 7.7-8 mm. long, 1 mm. broad at the base, hairy on the outer surface, glabrous within; petaloid sepal yellowish-white, ovate or broadly lanceolate-elliptic, 5-8 cm. long, 2.5—3.6 cm. broad, acuminate, cuneate at base, glabrous on both surfaces, pubescent on the veins beneath; “petiole” 0.5—-1.3 cm. long, hairy. Corolla deep yellow or orange, the tube 2-3 cm. long, appressed pubescent outside except at the subglabrous base, inside hairy as far as the bases of stamens, hairs long in short-styled forms (long-styled forms not seen), not tufted at the mouth: corolla lobes ovate, 4.5 mm. long, 4.2 mm. broad, apex acuminate-subcuspidate, strigose-hairy outside, papillate within. Stamens with short filaments, epipetalous 2/3 way up on the tube in short-styled forms; anthers linear, sub-basifixed, introrse, 5.5-10 mm. long, blunt at the apex and bifid at the sterile base. Ovary oblong, 4 mm. long, appressed pubescent, 2-locular with numerous 1964 | JAYAWEERA, MUSSAENDA IN THE PHILIPPINES 135 ovules on axile placentae; style 4 mm. long and stigma lobes 2.5 mm. long in short-styled forms. Berry not seen. DISTRIBUTION. This species is endemic to Luzon where it grows on hot and dry river embankments at an elevation of about 1065 meters above sea level. It has been collected in flower in May. Philippine Islands. Luzon: Pampanga Prov., Zambales Mountains (Mt. Pina- tubo), Elmer 21978 (pNH-holotype; A, GH, Ny-—isotypes). At first no trace of Elmer’s original collection was found, but a closer examination of the material available revealed that it had been placed under an unpublished name. Happily, the holotype of Mussaenda pina- tubensis is still in existence, apparently having been on loan during World War II. It is allied to M. nervosa by its leaf and stipule characters and also to M. philippica but differs from the latter in the smaller leaves, longer calyx segments, and faintly bifid stipules. The species is distinguished from others in the small, glabrous or subglabrous leaves; lanceolate stipules faintly bifid at apex, each pair fusing at the base to form a ring around the node: linear-subulate, persistent calyx segments; ovate, sub- cuspidate corolla lobes; and oblong, scantily appressed-pubescent ovaries. 18. Mussaenda scandens Elmer, Leafl. Philip. Bot. 3: 992. 1911 (Type: Elmer 11291); Merrill, Enum. Philip. Fl. Pl. 3: 518. 1923. Fic. 1, g; Fic. 2,1, m, n, 9, p. M. villosa sensu Merrill, Philip. Jour. Sci. Bot. 5: 243. 1910 and Elmer, Leafl. Philip. Bot. 3: 992. 1911, non Wall. Climber with numerous, curved, interlaced, lenticellate branches. Leaves ovate-elliptic or broadly oblong, 6—13.3 cm. long, 3-8 cm. broad, abruptly acute at apex, cuneate at base, obtuse or subrotund, pubescent on the upper surface, hairs short with few longer scattered ones, finely pubescent on the veins beneath with 6-10 pairs of lateral veins; petiole 0.3-1.5 cm. long, subglabrous or pubescent. Stipules deciduous, ovate-acuminate, 7-8 mm. long, 5 mm. broad at the base, bifurcate from apex more than 1/2 their length, lobes subulate, diverging, hairy on both surfaces, hairs within shorter with few glands in 2 groups at the base. Inflorescence a terminal, di- or trichotomously branched, appressed-pubescent, many-flowered, corymbose cyme; bracts and bracteoles lanceolate, acuminate, hairy on both surfaces, bracteoles broader, longer, and trilobed at apex. Flowers heterostylous, on stout, finely pubescent pedicels shorter than the ovaries. Calyx lobes linear-oblong, recurved, 3.5—5.5 mm. long, 0.8-1 mm. broad, sharply acuminate, hairy on both surfaces with 1 or 2 pairs of glands at the base of each sepal within; petaloid sepal ovate-elliptic, 6—7 cm. long, 3.84.5 cm. broad, puberulous on both surfaces, hairy on veins beneath, “petiole” 1-1.5 cm. long, pubescent. Corolla tube slender, 2—2.5 cm. long. hairy on the outer surface, hairy within 4/5 way down, glabrous at base; 136 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV hairs long in short-styled forms and short in long-styled forms, not tufted at the mouth; corolla lobes yellow, ovate, 2.5 mm. long, 2—2.5 mm. broad, acute, hairy on the outer surface, papillate within. Stamens with short filaments, epipetalous on the tube about 1/2 way in long-styled forms and a little higher in short-styled forms, adherent filaments free as loops on the tube for about 2 mm. midway between the base and the stamens; anthers linear, dorsifixed, introrse, 3.5—4.5 mm. long, bilobed at the base. Ovary turbinate, 3-3.5 mm. long, finely appressed pubescent, 2-locular with numerous ovules on cushion-shaped, axile placentae; style and stigma lobes 1.3 cm. and 6 mm. long respectively in long-styled forms, 4 mm. and 2.5 mm. long in short-styled forms. Berry ellipsoid, 1-1.3 cm. long, glabrous, calyx lobes deciduous; seeds minute, reticulate, ovoid, 0.83-0.9 mm. long, 0.67—-0.7 mm. broad, spiny, with 3—9 foveae in the areoles of the testa. DIsTRIBUTION. This species is endemic to the island of Mindanao and grows in forests along streams at an altitude of 600 to 800 meters. It has been collected in flower in May, July, and August; in fruit in August, October, and November. Philippine Islands. MinpAnao: Davao Dist., Todaya (Mt. Apo), Elmer 11291 (GH-lectotype; A, NY, Us—isotypes) ; Surigao, Wenzel 3354 (a, GH, NY); Zam- boanga Dist., Malangas, Ramos & Edano 37307 (A). In both habit and appearance this species is closely allied to Mussaenda villosa Wall., of Siam, Malaya, Sumatra, and Borneo. The resemblance is so great that Clemens 562 from Mindanao was identified by Merrill as M. villosa, an error later rectified. Close examination of the collections Ramos 1836, 1783 (us), etc., of M. villosa from British North Borneo, shows that the characters of the Philippine plant are sufficiently distinct for its treat- ment as a separate species. Both M. villosa and M. scandens are climbers with elliptic, pubescent leaves, slender flowers in terminal, pubescent cymes, glabrous fruits with deciduous calyx segments and spiny seeds. The flower buds of M. villosa are rounded at the top, somewhat con- stricted at the neck between the corolla tube and the lobes. Its sepals are lanceolate and corolla tube glabrous on the outer surface below the level of the stamens. Mussaenda scandens on the other hand bears flower buds which are elliptic at the top and not constricted at the neck. Its sepals are linear-oblong and the corolla tube entirely hairy on the outer surface. The hairs on the leaves of the former are long. Wenzel 3354 from Surigao differs slightly from the type in its longer sepals and corolla tube. The species may be distinguished from others by its climbing habit, ovate- elliptic or -oblong, pubescent leaves; slender corolla tube pubescent on the outer surface, hairy within 4/5 way down; sharply acuminate, linear- oblong sepals; ovate-acuminate stipules with diverging lobes; glabrous berries with deciduous calyx segments, and spiny seeds with 3 to 9 foveae in the areoles of the testa. 1964 | JAYAWEERA, MUSSAENDA IN THE PHILIPPINES 137 19. Mussaenda setosa Merr. Philip. Jour. Sci. Bot. 10: 104. 1915 (Type: Merrill 9496). Fic. 1, m; Fic. 4, d, e. Shrub or small tree, 3-4 m. high with terete, pale grayish-brown branches, the younger ones hirsute. Leaves in equal or subequal pairs, oblong-elliptic, elliptic, or ovate, 12-25 cm. long, 6-13 cm. broad, acumi- nate, rounded or decurrent-acuminate at base, setose on both surfaces, the margins ciliate, lateral veins 15—18 pairs, prominent on the lower surface; petiole 0.5-3 cm. long, hirsute. Stipules ovate, acuminate, 6-8 mm. long, 6.5 mm. broad at the base, bifurcate from apex for about 1/2 their length, lobes subulate, diverging, hirsute on the outer surface, glabrous within except at the base and among numerous glands occurring in 2 groups. Inflorescence a terminal, divaricate, few-flowered, hirsute-ciliate cyme; bracts and bracteoles small, lanceolate, about 3.5 mm. long, hairy on both surfaces, tufted at the base within, bracteoles broader, trifid. Flowers probably heterostylous, subsessile or on stout, pubescent pedicels shorter than the ovaries. Calyx lobes persistent, linear-lanceolate, 4-7 mm. long, 1 mm. broad, hairy on both surfaces with 2 pairs of glands within at the base of each; petaloid sepal white, broadly ovate, 7 cm. long, slightly acuminate, puberulent on the upper surface, hirsute on veins beneath, “petiole” 1 cm. long, hirsute. Corolla tube white, cylindric, 3 cm. long, hairy outside, inside hairy as far as the bases of the anthers, hairs not tufted at the mouth and short in the long-styled forms (short-styled form not seen); corolla lobes yellow, broadly ovate, about 2.5 mm. long, abruptly acuminate, hairy on the outer surface, glabrous within. Stamens with short filaments, epipetalous on the tube a little below the middle in long-styled forms; anthers linear-lanceolate, dorsifixed, introrse, 4.5 mm. long, blunt at the apex and bilobed at the sterile base. Ovary broadly fusiform, setose-ciliate, 2-locular with numerous ovules on axile placentae; style 1.6 cm. long, stigma lobes stout, 7.5 mm. long in the long-styled form. Berry ellipsoid, 1.5 cm. long, sparingly hirsute, black when dry, calyx lobes persisting until nearly ripe. DIsTRIBUTION. This species is endemic to Palawan where it grows on forested ridges at an altitude of about 700 meters. It has been collected in flower in April. Philippine Islands. PALAWAN: Mount Capoas, Malampaya Bay, Merrill 9496 (us-—lectotype), April 1913. The only specimen available for examination was the isotype from the United States National Museum belonging to the long-styled form and this is chosen as the lectotype, the holotype having been destroyed. The species seems to be distantly allied to Mussaenda palawanensis in the stipule and petal characters while differing from it in numerous other char- acters. It is, however, distinguished from other species by its characteristic setose indumentum, large leaves with 15-18 pairs of lateral veins which are prominent on the lower surface; ovate, bifurcate stipules; long, slender 138 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV corolla tubes, and sparingly hirsute, ellipsoid berries with persistent calyx segments 20. Mussaenda vidalii Elmer, Leafl. Philip. Bot. 3: 993. 1911 (Typr: Elmer 11309); Merrill, Enum. Philip. Fl. Pl. 3: 520. 1923. Fic. 2, q, r,s, t, u, Vv, W, x. Scandent shrub or small tree, 3-4 m. tall with lax, somewhat drooping, lenticellate, ferruginous pubescent branches. Leaves ovate, oblong-ovate or elliptic, 9—-25.5 cm. long, 4.3-15.2 cm. broad, abruptly acute at apex, broadly obtuse, cuneate or rounded at base, ferruginous hairy on the upper surface, densely so on the lower surface with 6-11 pairs of lateral veins; petiole 1-5 cm. long, hirsute. Stipules deciduous, triangular-lanceolate, 6.5—-10.5 mm. long, 6—9 mm. broad at the base, bifurcate from the apex 1/3-1/2 their length, lobes erect or slightly diverging, hirsute on the outer surface, glabrous or scantily hairy within at the base with few or numerous glands. Inflorescence a terminal, dichotomously branched, hairy, many- flowered cyme; bracts and bracteoles linear, or broadly lanceolate, 3—5 mm. long, hairy on the outer surface, glabrous or pubescent within, bracte- oles broader, trifid at the apex. Flowers heterostylous on stout, densely hairy pedicels shorter than the ovaries. Calyx lobes oblong or ovate, 4-10 mm. long, 1—-3.2 mm. broad, tapering to an acute apex, hairy on both surfaces, hairs on the outer surface longer and more dense; petaloid sepal whitish, 5-10 cm. long, 2.5—8.5 cm. broad, subacute, short cuneate at base, hirsute on both surfaces, 5-veined, “petiole” 2—2.5 cm. long, hirsute. Corolla tube yellowish green, 2.2-3 cm. long, curved, broad above, taper- ing to a narrow base, densely hairy on the outer surface, hairs long and spreading more or less at right angles to the tube, hairy within as far as the bases of anthers or lower, not tufted at the mouth; hairs extending as far as 1/2 the length of the tube in long-styled forms and 2/5 the length of the tube from the top in short-styled forms; corolla lobes pink to orange- red or yellow, broadly orbicular, 2-3 mm. long, 4-6 mm. broad, abruptly acute, hairy on the outer surface, papillate within. Stamens with short filaments, epipetalous 1/2—2/3 way up on the tube in long-styled forms and 3/5 way up in short-styled forms; anthers linear, dorsifixed, introrse, 5—6.2 mm. long, slightly bent, bilobed at the base. Ovary turbinate, 3.5- 5.5 mm. long, densely hairy, 2-locular, with numerous ovules on cushion- shaped, axile placentae; style and stigma lobes 1.7—2.8 cm. and 6-8 mm. long respectively in long-styled forms, 2.5 mm. and 2 mm. long in short- styled forms. Berry ellipsoid, 1.5—2.2 cm. long, ferruginous pubescent, lenticellate, calyx lobes deciduous; seeds minute, reticulate, 0.53—0.73 mm. long, with 3-8 foveae in the areoles of the testa. DIsTRIBUTION. This species grows along the margins of humid forests or in the secondary growth of open rolling country at elevations between 150 and 600 meters above sea level in the islands of Mindanao, Leyte, and Samar. The type collection was made at 1140 meters elevation. Material 1964 | JAYAWEERA, MUSSAENDA IN THE PHILIPPINES 139 has been collected in flower between March and August; in fruit in March, June, and July. Philippine Islands. Samar: Catubig River, Ramos 24485 (a); Loquilocon, Sulit 6098 (a); Mt. Cansayas, Sulit 14366 (A); Bagacay, Sulit 6280 (A). LEYTE: Wenzel 667 (A); Ormoc, Antilao River, Edano 11873 (a, PNH); Lake Danao, Edano 11926 (PNH). MuInpANAO: Bukidnon Prov., Pigtaoranan, Anonuevo 13516 (A, PNH); Tangculan and vicinity, Ramos & Edano 39050 (a), 39186 (A, US), 39035 (GH, US); Davao Dist., Todaya (Mt. Apo), Elmer 11309 (cH-lectotype; A, NY, US-isotypes); Mt. Apo, Clemens 15279 (ny, Us). This species shows the greatest variation in the collections from Leyte. They differ from the type collection in that they are small trees, with broader bracts and bracteoles in the inflorescence, ovate sepals somewhat resembling those of Mussaenda macrophylla, broader petals, and larger anthers and fruits. The seeds are identical with those of the typical form both in size and in the number oi foveae in the areoles of the testa. Mussaenda vidalii may be distinguished from other species by the spread- ing, ferruginous pubescence on stems, leaves, and flowers, the oblong or ovate sepals, short petals, large pubescent fruits with deciduous calyx segments, and the small seeds with 2 to 8 foveae in the areoles of the testa. Uses. The water in which the leaves of Mussaenda vidalu have been soaked for a few muintes is used as an eye wash; and the macerated leaves are applied to the head to allay drunkenness. RoyaL BOTANIC GARDENS, PERADENIYA, CEYLON 140 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV COMPARATIVE ANATOMY OF THE LEAF-BEARING CACTACEAE, XI THE XYLEM OF PERESKIOPSIS AND QUIABENTIA I. W. BatLey ! IT IS OF INTEREST from a taxonomic point of view that Pereskiopsis and Quiabentia may be differentiated anatomically from Pereskia (1) by the characteristic development in young stems of a highly differentiated crystalliferous hypodermis in these genera (Bailey, 1961b); (2) by the absence of diagnostic forms of sclereids in their secondary phloem (Bailey, 1961la); (3) by the sporadic occurrence in their stems and leaves of the curious short cells with broad annular or helical thickenings which have attracted so much discussion in the mammillarias and other excessively succulent forms of the Cactaceae (Bailey, 1960); and (4) except in com- parisons with Pereskia conzattii Britt. & Rose, Pereskia autumnalis (Eichlam) Rose, and Pereskia nicoyana Web., by the characteristic pseudo- palmate or palmate vasculature of their leaves (Bailey, 1960). At present uncertainties regarding the status of the numerous described species of Pereskiopsis and Quiabentia are fully recognized by many taxonomists. Owing to these uncertainties and to the wide ranges of structural variability, not only in the same clone when grown under different environmental influences, but also in different parts of a single adult plant, it seems advisable in a preliminary investigation of the xylem to concern oneself largely with ranges of anatomical variability at a generic level rather than with possible structural differences between putative species and geographical races of the two genera. The anatomical data presented in the following pages are based upon the investigation of 33 collections of Pereskiopsis from various parts of Mexico, kindly preserved for me largely by Norman Boke, R. L. Dressler, Efraim Hernandez X., Myron Kimnach, and R. V. Moran. Twenty-nine of these collections exhibited more or less reliable affinities to the following putative species: P. aquosa (Web.) Britt. & Rose three collections, P. blakeana Ortega two, P. chapistle Britt. & Rose five, P. diqueti Britt. & Rose one, P. gatesii Baxter one, P. pititache Britt. & Rose three, P. porteri (Brand.) Britt. & Rose seven, P. rotundifolia (DC.) Britt. & Rose one, P. scandens Britt. & Rose one, P. spathulata (Otto) Britt. & Rose two, and P. velutina Rose three. There is no reliable clue at present to the affinities of four other collections. In the case of Quiabentia, I have material of Q. aff. chacoensis Backbg. * This investigation was financed by a grant from the National Science Foundation. I am indebted to the American Philosophical Society for the loan of a Wild microscope. 1964] BAILEY, LEAF-BEARING CACTACEAE, XI 141 collected by Carenzo and Legname in the Jujuy Province of Argentina, of QO. pereziensis Backbg. collected by Cardenas in Bolivia, and of Q. zehntnert (Britt. & Rose) Britt. & Rose obtained from a cultivated clone by Boke. From the point of view of the phylogeny of the dicotyledons as a whole, the cambium and xylem of Pereskiopsis and Quiabentia have attained a high general level of evolutionary specialization comparable to that which has occurred in Pereskia and in a number of dicotyledonous families (Bailey & Srivastava, 1962). The fusiform initials of the cambium are comparatively short and tend to become arranged in stratified or “storied” patterns. The vessel members are short and have simple porous perforation plates; the wood parenchyma strands are short and of scanty paratracheal distribution in denser parts of the secondary xylem; the libriform fibers which may be septate or non-septate function in the storage of starch; and the rays are multiseriate, uniseriate ones having been phylogenetically eliminated. It should be noted in this connection, however, that in Pereskiopsis, as at times in Pereskia, the vessels which eommonly are diffusely distributed either singly or in small clusters frequently exhibit aggregation into concentric zonal patterns in association with varying proportions of wood parenchyma (Fics. 1— It is in divergent trends of specialization (i.e., excessive broadening of multiseriate rays and the elimination of secondary walls and lignification within them) that stems of Pereskiopsis and Quiabentia differ most sig- nificantly from those of most species of Pereskia.-In the case of arborescent forms of Pereskia, incipient evidence of the suppression of lignification, when present, is confined in P. sacharosa Griseb., P. grandifolia Haw., and P. bleo DC. to the innermost part of the first-formed multiseriate rays of roots (Bailey, 1963c). In arborescent forms of P. colombiana Britt. & Rose, P. guamacho Web., P. cubensis Britt. & Rose, and P. portulacifolia Haw., the phenomenon is accentuated in roots but not in stems (Bailey, 1963d). In the arborescent P. conzattii Britt. & Rose, P. autumnalis (Eichlam) Rose, and P. nicoyana Web., it occurs in a curiously modified form in the innermost parts of multiseriate rays in basal parts of large trunks (Bailey, 1963b). In contrast to these larger arborescent forms of pereskias, the smaller commonly more shrubby forms of P. humboldti Britt. & Rose, P. weberiana K. Schum., and P. diaz-romeroana Card., with their not infrequent decumbent, scrambling or scandent branches, exhibit highly accentuated trends of divergent specialization in their roots, but only slight incipient evidence at times of the elimination of secondary walls and lignification in their stems (Bailey, 1963a). Within Pereskzia, it is the highly scandent P. aculeata Mill. which differs most markedly from other species in having advanced trends of divergent specialization in both its roots and stems (Bailey, 1962). In contrast to the arborescent species of Pereskia, the numerous putative species of Pereskiopsis are relatively low plants of shrubby form with stems or branches which frequently tend to be more or less decumbent, scrambling, or scandent. It should be noted that collectors find it difficult at present to locate plants of unmodified form in the wild. In many cases 142 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV specimens are obtainable only from plants cultivated in hedges or gardens, or, in remoter regions, from plants that have been cut back or variously mutilated by the activities of man. However, in all investigated collections, the ranges of variability in width, distribution, and internal structure of the multiseriate rays are extensive in stems from different parts of a single plant or clone. Furthermore, in the larger stems of all putative species of Pereskiopsis, there are conspicuous evidences of divergent trends of struc- tural specialization, i.e., those leading toward more or less extensive elimination of secondary walls and lignification in rays and wood paren- chyma. In parts of some stems of Pereskiopsis, the first-formed secondary xylem resembles the wood of Pereskia sacharosa and of other arborescent peres- kias. In such dense tissue (Fic. 1) comparatively narrow multiseriate rays, related to parenchymatous gaps of the eustele, are fully lignified, and relatively uniformly distributed as seen in transverse sections (Fics. 1 and 3). In subsequently formed secondary xylem (Fics. 3 and 4) the broadening rays exhibit more or less extensive patches of unlignified tissue which contain abundant druses of calcium oxalate. In addition, more or less extensive arcs or zones of unlignified wood parenchyma are of common occurrence (Fics. 4 and 5) not infrequently in association with zonal aggregations of vessels (Fic. 5 In other stems of Pereskiopsis, having broader parenchymatous gaps in the eustele, the rays of the first-formed secondary xylem are broader (compare Fics. 3 and 9). The inner parts of such rays, related to inter- fascicular parts of the eustele, may be lignified, or unlignified (as in Fic. 9). Furthermore, in such rays the cells frequently are broader tangentially than radially. In subsequently formed secondary xylem, more or less extensive patches of unlignified ray tissue and unlignified zonal wood parenchyma tend to occur (Fic. 7). In transverse sections of some stems, aggregations of excessively broadened, lignified, unlignified, or partly unlignified rays may occur in wedges of secondary xylem which alternate around the circumference of a stem with intervening wedges of wood of denser, more normal structure, i.e., having narrower rays and a higher proportion of lignified vessels and libriform fibers. In these stems, the denser precociously flaring wedges of secondary xylem are related internally to the fascicular parts of the eustele. Such structural tendencies are diagrammatically illustrated in Fic. 2, a transverse section of a stem of Pereskiopsis porteri. In this stem, there are five conspicuous wedges of softer secondary xylem that are depressed below the outer boundary of denser, structurally more normal, intervening ones. The number of depressed wedges in the transverse section of the stem suggests a possible correlation with a 2/5 or 3/8 phyllotaxy. That aggregations of broadened rays actually are related to phyllotaxy may be demonstrated by examining decorticated stems after treatment with phloroglucin-HCl (TExt-Fric. B), or by studying tangential longi- tudinal sections of the secondary xylem. As shown in the case of Pereskiop- sis aff. aquosa (Fics. 11 and 12), the connecting vascular tissues of the 1964 | BAILEY, LEAF-BEARING CACTACEAE, XI 143 leaf and areole are subtended in the secondary xylem by broad unlignified rays and relatively few tenuous strands of lignified libriform fibers and vessels. In passing downward from such a node through successively lower internodes to the next node of an orthostichy, the rays sooner or later become narrower and the strands of intervening lignified libriform fibers and vessels become correspondingly more numerous and broader. Thus, the structural differences between wedges of secondary xylem, visible in transverse sections of these stems, are due largely to the level at which the orthostichies are sectioned below their nodes. In some larger stems of Pereskiopsis aff. aquosa the aggregations of multiseriate rays subtending the node of an orthostichy, not only are very broad and unlignified as in Fic. 12, but also at times tend, just below the node, to fuse forming a broad wedge of unlignified tissue as seen in trans- verse sections (Fic. 13). In such wedges of unlignified parenchyma, narrow strands of lignified libriform fibers and vessels may be eliminated, their place being occupied by more or less vertically elongated, unlignified derivatives of the cambium between the broad fusing rays (Fic. 13). It should be emphasized that in Pereskia the connecting vascular tissues of the leaves and areoles are jacketed in the secondary xylem by a very limited amount of unlignified parenchyma and are subtended by fully lignified tissue. When decorticated the stems resemble TExT-FIc. A, except that the rays are lignified throughout their radial extension outward in the secondary xylem. In the case of Pereskiopsis, some stems exhibit comparable lignification in the earlier formed secondary xylem. In other stems or parts of stems, lignification tends to be suppressed in the secondary xylem for varying distances below the connecting vascular tissues of the leaves and areoles (TEXxT-FIG. B) In basal parts of some of the largest stems of Pereskiopsis aff. aquosa, cells with broad annular thickenings occur, not only in elements of the primary body adjacent to vessels of the fascicular parts, but also in unlignified multiseriate rays of the secondary xylem. In the primary body, the cells are more or less elongated vertically and of fusiform outline (Fic. 16), and have more than one annular thickening per cell. In the rays of the secondary xylem, on the contrary, the cells are of rectangular form and contain a single broad annular thickening per cell (Fics. 15 and 17). The xylem in roots of Pereskiopsis commonly exhibits divergent trends of anatomical specialization comparable to those that occur in roots of Pereskia aculeata (Bailey, 1962), P. cubensis, and P. portulacifolia (Bailey, 1963d). The inner parts of multiseriate rays related to protoxylem poles of the primary body tend to be unlignified. In the outward extension of these rays, and in those subsequently formed, there tend to be alternating patches of lignified and unlignified tissue. In addition, more or less exten- sive arcs or zones of unlignified wood parenchyma are of not infrequent occurrence. However, in general, the divergent tendency toward suppres- sion of lignification tends to be more highly accentuated in stems than in roots of Pereskiopsis. This is in contrast to what happens in most species 144 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Text-Fics. A-C. Decorticated stems treated with phloroglucin-HCl, natural size. A, Quiabentia aff. chacoensis [Tucuman], B, Pereskiopsis porteri | Moran 7349|, C, Quiabentia zehntneri [Boke]. of Pereskia where incipient trends of divergent specialization commonly are more advanced in roots than in stems. In an earlier paper of this series (Bailey, 1961b), I noted briefly that where crystals of calcium oxalate are deposited in the fully lignified rays of arborescent pereskias, they occur as single large crystals or a few smaller independent ones, aggregations into typical druses being absent. Con- versely, where multiseriate rays or parts of them are unlignified, druses of calcium oxalate are conspicuously developed. Since writing this paper additional collections of roots and stems of Pereskia and Pereskiopsis have been obtained. Although conspicuously numerous druses occur charac- teristically in unlignified ray tissue of these collections, minor aberrations may occur at times in rays that are composed of alternating patches of lignified and unlignified tissue. For example, in the case of Pereskia abundant in the lignified rays of stems, a few crystals of this form may occur in association with druses in unlignified patches of ray tissue in roots. Conversely, a few druses may occur at times in lignified parts of the rays in stems of Pereskiopsis. Most exceptional in this connection, however, is the occurrence of a few druses in the fully lignified rays of stems of the Peruvian and Bolivian pereskias. According to descriptions and illustrations in the literature, Quiabentia 1964 | BAILEY, LEAF-BEARING CACTACEAE, XI 145 chacoensis, Q. pereziensis and Q. zehntneri are characterized by their erect stems bearing a number of upstanding coarse branches from which rela- tively short smaller ones diverge. All of the aerial parts of these plants (stems, branches, and leaves) are highly succulent. In my limited collec- tions of the three species, the stems and larger branches have a greatly expanded pith and a relatively much reduced volume of secondary xylem. For example, in a stem of Q. pereziensis six centimeters in diameter, the pith has a breadth of three centimeters; in a stem of Q. aff. chacoensis, having a diameter of three and one half centimeters, the pith is two centi- meters in diameter; and in a vigorous branch of Q. zehntneri, having a diameter of three centimeters, the pith is 18 millimeters in breadth. It should be emphasized in this connection, however, that the accen- tuated succulence in stems and branches of Quiabentia is not due solely to excessive expansion of the soft tissues of the primary body with con- comitant reduction in cambial activity. There is an obvious tendency in the secondary xylem, as in many stems of Pereskiopsis, toward increasing the proportion of unlignified tissue and concomitantly reducing the propor- tion of lignified libriform fibers and vessels. As indicated in Fic. 10, part of a transverse section of a vigorously growing branch of Q. zehntneri, the conspicuously broadened multiseriate rays are unlignified and contain numerous druses of calcium oxalate. This is true even in wedges of sec- ondary xylem that are subtended by fascicular parts of the eustele and results in the formation of relatively widely spaced, exceedingly narrow wedges of lignified tissue composed of libriform fibers and vessels (com- pare Fic. 10 and Text-Fric. C A similar tendency toward reduction of lignification occurs in stems of Quiabentia pereziensis. In Fic. 8, part of a transverse section cut near the base of a large stem, i.e., nearer the level of the ground, the narrower strands of secondary xylem are less diagrammatically distributed, being form fibers and vessels. The contortions of the grain and the excessive breadth of the unlignified rays in this stem are shown in tangential longi- tudinal section in Fic. 14. The stems of Quiabentia pereziensis and Q. zehntneri in my limited collections, when decorticated and treated with phloroglucin-HCl, have conspicuous unlignified parenchymatous lacunae in the secondary xylem subtending the leaves and areoles at nodal levels (TExt-Fic. C), as in some stems of Pereskiopsis aff. aquosa and P. porteri (TExtT-FIc. B). In the case of Quiabentia aff. chacoensis, the only stem of comparable diameter available to me at present has relatively long internodes and broad wedges of lignified xylem between its unlignified multiseriate rays, i.e., aS seen in transverse sections of the stem. Furthermore, although there is some tendency for aggregation of broadened rays in the secondary xylem subtending the bases of the leaves and areoles at subnodal levels (TExtT-FIc. A), these rays do not fuse to form large conspicuous unlignified parenchymatous lacunae comparable to those that occur in secondary xylem 146 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV of QO. pereziensis and Q. zehntneri. This suggests that the tendency toward increasing succulence of the wood in Quiabentia is less advanced in Q. aff. chacoensis than in the other two species of the genus. The suggestion is strengthened by the structure of one available root of Q. aff. chacoensis. As illustrated in Fic. 6, the xylem is of normal fully lignified structure and closely resembles that which occurs in the roots of Pereskia sacharosa (Bailey, 1962, Fig. 2) and other dicotyledons at equivalent levels of phylogenetic anatomical specialization. However, until extensive collec- tions become available for studying the ranges of structural variability in different plants of the same clone or species, conclusions regarding anatomi- cal differences between putative species of Quiabentia, as between those of Pereskiopsis, must remain largely tentative. SUMMARY AND GENERAL CONCLUSIONS Evidence presented in this and preceding papers reveals incipient stages of divergent anatomical specialization in Pereskia, Pereskiopsis and Quia- bentia. One of the most significant of these, for a clearer understanding of greatly accentuated succulence in stems of the Opuntieae and Cereeae, is the tendency toward increasing the circumference of the eustele with concomitant expansion of the pith during later ontogenetic stages of the development of the primary body. Incipient stages of this phenomenon are detectable in larger immature stems of various pereskias (Bailey, 1962, 1963c, 1963d). In such stems expansion of the parenchymatous inter- fascicular parts of the eustele and enlargement of the pith may occur after initiation of cambial activity in fascicular parts of the primary body. Enlargement of the pith is much accentuated in the basal parts of the large trunks of Pereskia conzattii, P. autumnalis, and P. nicoyana (Bailey, 1963b). In these pereskias, multiseriate rays related to unlignified paren- chymatous gaps of the eustele remain unlignified and their cells have a capacity for tangential expansion, thus facilitating additional increase in the circumference of the eustele and enhanced enlargement of the pith. As indicated in this paper, similar phenomena are of commoner occurrence in stems of Pereskiopsis and Quiabentia (Fics. 7, 9, and 10). However, as indicated earlier in this paper, enhanced succulence in stems of Quiabentia and Pereskiopsis is not due solely to increasing the propor- tion of soft tissue in the primary body, with concomitant reduction in the volume of secondary xylem. There is obvious evidence of divergent trends of specialization in the secondary xylem itself, leading to increasing succu- lence by suppression of secondary walls and lignification in multiseriate rays and wood parenchyma. Although this phenomenon is largely confined in Pereskia to roots and to stems of the scandent P. aculeata, it becomes more or less highly accentuated in stems of Pereskiopsis and Quiabentia. Furthermore, there is a tendency in some stems of the latter genera for aggregations of broad unlignified rays to subtend the connecting vascular tissues of leaves and areoles at nodal levels. This phenomenon is par- ticularly significant in attaining a clearer understanding of the origin and 1964 | BAILEY, LEAF-BEARING CACTACEAE, XI 147 development of huge parenchymatous lacunae in the secondary xylem of various representatives of the Cereeae and Opuntieae. The breadth, longi- tudinal extension, and structural details of such subtending succulent strands of secondary xylem vary more or less markedly at times even in different stems of the same clone or plant. Many of the more conspicuous structural differences appear to be correlated at least in part with varia- tions in the diameter of the primary body, in the number of fascicular strands in the eustele, in the width of parenchymatous gaps between them, with changes in phyllotactic patterns from 2/3 to 3/8 and possibly to 5/13, and with variations in vigor of growth involving conspicuous dif- ferences in internodal elongation. There are additional trends of divergent anatomical specialization that merit attention. One of these is the tendency to form zones of wood parenchyma as seen in transverse sections of stems. Such zonal patterns of wood parenchyma commonly are associated with concentric aggregations of vessels and appear to be correlated with seasonal variations in cambial activity and in the maturation of cambial derivatives. Although zonal aggregations of vessels and lignified wood parenchyma occur in the later- formed secondary xylem of large stems of several arborescent pereskias, incipient stages of the suppression of lignification in wood parenchyma of stems is largely confined in Pereskia to the scandent P. aculeata. In con- trast to this, the suppression of lignification in wood parenchyma is of common occurrence in stems of the shrubby representatives of Pereskiopsis and Quiabentia. A trend of divergent specialization, which does not occur in any of my numerous collections of Pereskia, is the sporadic occurrence in leaves and stems of Pereskiopsis and Quiabentia of the curious short cells with broad annular or helical thickenings which are so abundantly developed in more succulent representatives of the Cactaceae. The occurrence, distribution, and structure of these cells in Pereskiopsis and Ouiabentia merit additional detailed investigation, since the study of more comprehensive collections of adequate material may yield clues regarding the origin, function, and phylogenetic significance of such curious cells in the Cactaceae as a whole. It should be emphasized in conclusion that in the leaf-bearing Cactaceae accentuations of divergent trends of anatomical specialization are more or less closely associated with changes in the stature and habits of growth of the plants in which they occur. In arborescent forms of Pereskia (i.e., those having a more typically woody dicotyledonous habit of growth) incipient stages of divergent specialization in the xylem are rarely detectable in P. sacharosa, P. grandifolia, and P. bleo; more or less conspicuous in roots of P. colombiana, P. guamacho, P. cubensis, and P. portulacifolia; and confined in P. conzattii, P. autumnalis, and P. nicoyana to the first-formed secondary xylem in the basal parts of large trunks of the trees. On the contrary, the phenomena are greatly accentuated in roots of the compara- tively low, shrubby forms of Peruvian and Bolivian pereskias, and in both roots and stems of the scandent P. aculeata. In contrast to this, all stems in available collections of the shrubby representatives of Pereskiopsis and 148 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Quiabentia, regardless of variations in their habit of growth and in different degrees of succulence, exhibit more or less accentuated divergent trends of specialization in their xylem. Thus, the various representatives of Pereskia, Pereskiopsis, and Quiabentia may be arranged in an ascending evolutionary series of increasing divergent specialization. In this ascending phylogenetic sequence, the arborescent forms of Pereskia occur at the lowest level and the highly succulent forms of Quiabentia at the apex. Furthermore, the evolutionary series at least suggests that among surviving representatives of the family Pereskia sacharosa, P. grandifolia, and P. bleo have most closely retained the vegetative habit of growth and internal structure of ancestral Cactaceae. From a purely taxonomic point of view, the structure of the xylem provides additional diagnostic criteria for differentiating Pereskia from the genera Pereskiopsis and Quiabentia. However, it does not at present pro- vide a convincing argument for separating Rhodocactus from Pereskia or Pereskiopsis from Quiabentia. LITERATURE CITED Bartey, I. W. 1960. Comparative anatomy of the leaf-bearing Cactaceae, I. Foliar vasculature of Pereskia, Pereskiopsis and Quiabentia. Jour. Arnol Arb. 41: 341-356. 1961a. II. Structure and distribution of sclerenchyma in the phloem of Pereskia, Pereskiopsis and Quiabentia. Ibid. 42: 144-156. . 1961b. III. Form and distribution of crystals in Pereskia, Pereskiopsis and Quiabentia. Ibid. 42: 334-347. 1962. VI. The xylem of Pereskia sacharosa and Pereskia aculeata. Ibid. 43: 376-383. . 1963a. VII. The xylem of pereskias from Peru and Bolivia. /bid. 44: 127-137. . 1963b. VIII. The xylem of pereskias from Southern Mexico and Central America. /bid. 211-221. —. 1963c. IX. The xylem of Pereskia grandifolia and Pereskia bleo. Ibid. 222-231. 1963d. X. The xylem of Pereskia colombiana, Pereskia guamacho, Pereskia cubensis, and Pereskia portulacifolia. Ibid. 44: 390-401. L. M. Srtvastava. 1962. Comparative Anatomy of the leaf-bearing Cactaceae. IV. The fusiform initials of the cambium and the form an structure of their derivatives. bid. 43: 187-202. EXPLANATION OF PLATES (The black spots in unlignified tissue of Fics. 2-4, and 9-13 are druses. Those in Fic. 14 are starch.) PLATE I Fics. 1 & 2. Transverse sections of stems, 1, Pereskiopsis aff. chapistle [ Boke B-3], inner xylem of a vigorously growing shoot 25 millimeters in diameter with a pith six millimeters in diameter, X 14. 2, Pereskiopsis porteri [Moran 7349), 1964 | BAILEY, LEAF-BEARING CACTACEAE, XI 149 t of a stem 18 millimeters in diameter with a pith three millimeters in eee xX 11 PLATE II Fics. 3 & cana sections of stems, & 11. 3, Pereskiopsis aff. chapistle | Boke B. 3], inner part of a stem seven centimeters in diameter with a pp four millimeters in cancer 4, The same, part of outer secondary xylem PLATE III Fics. 5 & 6. Transverse sections of secondary xylem, X 11. 5, Part of outer xylem from the same stem as Fics. 3 & 4. 6, Quiabentia aff. ae section of a root PLATE IV FIG & 8. Transverse sections of xylem, 11. 7, Pereskiopsis aff. aquosa [Dressler], section of basal part of stem six centimeters in diameter with a pith eight millimeters in sencice 8, Quiabentia pereziensis | Cdrdenas]|, large root. PLATE V Fics 10. Transverse sections of stems, X 11. 9, Pereskiopsis aff. aquosa Dreier, xylem from a stem four centimeters in diameter with a pith 12 millimeters in diameter. 10, Quiabentia zehntneri [ Boke], part of xylem from a stem four centimeters in diameter with a pith 14 millimeters in diameter PLATE VI Fics. 11 & 12. eee ie bid es of stems of ee aff. aquosa Boke B-31]|, X 11, From a stem 23 millimeters in diam pith seven ee in diameter, shoe a leat traces and cylin se vascular tissue of a 2, From a stem five centimeters in diameter with a pith n ea repr in ee eter. PLATE VII Fics. 13 & 14. Transverse and tangential longitudinal sections. . oe aff. one [ Boke B-31], transverse section of larger : . 14, Quiabent pereziensis | Cdrden as], tangential section of outer secondary Ls rom a stem six centimeters in diameter with a pith 33 millimeters in acer xX 43. PLATE VIII Fics. 15-17. Sections from the largest basal part of a stem of Pereskiopsis aff. aguosa [Dressler], * 88. 15, Transverse section of ray cells showing broad annular thickenings. 16, Longitudinal section of the outer part of the eustele showing fusiform cells with broad annular thickenings. 17, Radial section of a multiseriate ray showing single broad annular thickening in ray cells. PLATE I XLV ARNOLD ARB. VOL. i JOUR. che a # > fe e* t, 8, * EAE, CACTAC LEAF-BEARING BAILEY, Jour. ARNOLD ArB. VoL. XLV PiaTE II BAILEY, LEAF-BEARING CACTACEAE, XI Pirate III] Jour. ARNOLD Ars. VoL. XLV Be Za nes ¢ z4 wee He i ye o> pet e ~ aa #. eee Se ae ad sae fs ts ey" : ie «is 05 + %S - ee te. ss 4, Mines ipa Dy 2 Ms re BAILEY, LEAF-BEARING CACTACEAE, XI Jour. ARNOLD Ars. VoL. XLV PLATE IV Seats. CS Rei - aloe SEPT Aes oe BAILEY, LEAF-BEARING CACTACEAE, XI PLATE V Jour. ArNotp Ars. VoL. XLV . seer : we oa ESAS a Tons a «a VW cI fa CEAE LEAF-BEARING CACTAC BAILEY, PLaTE VI Jour. ARNoLD Ars. VoL. XLV Pe. ey BN. Say «f es < +" rt . ee ie’ om Eee es * Sep ae ee ox XI G CACTACEAE LEAF-BEARIN BAILEY Jour. ARNOLD Ars. VoL. XLV PuaTeE VII ren aR jigs ? ea a) 7, RS coe iP ae: WY) ws # PE 4 Pd i jhe pa o® $3: 27% : nH «7 ~ tat Oe D eas “\ A et. INC A ka ~ sell ‘ ake antl RU RO oP LA POM ers Sa Pee {seh Los? eos, Wisk | ake NE ier’ Fre oe & AN 4 oes Yi ie a 4 mS Pa BAILEY, LEAF-BEARING CACTACEAE, XI Piate VIII Jour. ARNOLD Ars. VoL. XLV XI BAILEY, LEAF-BEARING CACTACEAE, 158 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV TYPIFICATION IN DIRCA Lorin I. NEVLING, JR. IN THE THYMELAEACEOUS GENUS Dirca L., which is restricted to Canada and the United States, two species, D. palustris L. and D. occidentalis Gray, are known. Dirca palustris has an extensive range in eastern North America; D. occidentalis is restricted to six counties in California. It is of interest that there has been no taxonomic confusion in the genus since it was described as a Thymelaea by Gronovius in 1743 (Flora Virginica 155), and there has been only one nomenclatural synonym for the genus during this period, which presents no problem, being based directly on Dirca L. (Dofia Adans. Fam. 2: 285. 1763). Although with this virtually unblemished history the question of typification has never been of critical importance, there are some interesting ries concerning typification of the two species which seem worth discussin The type species of the genus, Dirca sien was described, validly, by Linnaeus in Species Plantarum (358. 1753) with the citation ‘“palus- tris. 1. Dirca Gen. Nov. 1078.*,”’ the asterisk indicating that in the volume cited there was a good description. The citation refers to Nova Plantarum Genera . . . Leonhard Johan Chenon, 1751, one of a long series of dissertations written by Linnaeus but defended, as theses, by his students. In both the dissertation and the Species Plantarum the descrip- tion of Dirca palustris is taken verbatim from Gronovius’ Flora Virginica (155. 1743), and from this description alone it is impossible to tell whether or not Linnaeus had actually seen the plant described. In the dissertation, under Dirca, Linnaeus cited “fig. 7.,” but in the two copies of it which I examined the figures were lacking. In his Genera Plantarum (ed. 4. 167. 1754) Linnaeus again used Gronovius’ description for Dirca. In 1756, the portion of the Linnaean dissertation of 1751 concerning Dirca was republished (Linn. Amoen. Acad. ed. 1. 3: 12, 13). Here the description of Gronovius is repeated once more, but Linnaeus also pro- vided an original diagnosis and an illustration of a flowering branchlet and inflorescence (presumably the same as “fig. 7.”’ of the dissertation). The illustration of the flowering branchlet agrees reasonably well with specimen no. 501.1 in the herbarium of the Linnean Society of London. The sheet apparently was seen by Linnaeus before 1753 and bears the handwritten notation ‘1. palustris.” It is listed by B. D. Jackson (Index to the Linnean Herbarium in Suppl. Proc. Linn. Soc. London, 124th Sess., 26, 69. 1912) as being present in the herbarium of Linnaeus in 1753. Specimen number 501.1 must, therefore, be considered the holotype rather than a specimen Gronovius might have had, in spite of the number of times Linnaeus chose to repeat the Gronovian description. According to 1964] NEVLING, TYPIFICATION IN DIRCA 159 the citation of Gronovius of 1743 the plants he examined for Flora Vir- ginica were collected by John Clayton. In the edition of Flora Virginica of 1762 (p. 60) the citation “Clayt. n. 858.” is appended to the descrip- tion of Dirca palustris. In addition, Linnaeus (see Jackson ibid. 11) is quoted as saying, “When I assisted Dr. Gronovius in examining plants from Virginia I got Duplicates of most of them.” The holotype specimen in the Linnean Herbarium (LINN) is very likely of Clayton’s collecting and given by Gronovius to Linnaeus. The first report of Dirca in California was published by John Torrey in 1857 (Botany, in Reports of Exploration and Surveys 4: 133). This reference is to D. palustris L. in which Torrey says, ‘Mountains near Oakland, California; April 4, (with flowers and young fruit.) We have never before received this plant from any part of the United States west of the Mississippi.” Dirca palustris is now well known as far west as eastern Oklahoma but Torrey’s reference can only be to the plant later described by Asa Gray (Proc. Am. Acad. 8: 631. 1873) as D. occidentalis. The collection of April 4, 1854, from Oakland, was made by Dr. J. M. Bigelow and has been cited by some as the holotype (McMinn, H. E. Manual of California Shrubs p. 365. 1951) which it clearly is not. Gray, following his description of D. occidentalis, and remarking on the Bigelow collection says, ‘there are only vestiges of the former [the flowers and young fruit] in my specimens. If they had been in good condition, Dr. Torrey would have noticed the characters of the species, which are now manifest.”’ The additional specimen making the characters “manifest” is “Dr. A. Kellogg & W. G. W. Harford n. 895 of distribution.” This collec- tion was made March 13, 1869, in the Oakland hills. Both the Bigelow and Kellogg & Harford collections are mounted on a single sheet now on deposit in the Gray Herbarium (GH). On this sheet are three branchlets with the left-hand (immature fruit) and center specimens (sterile ? and insect damaged) comprising the Bigelow collection while the right-hand specimen (flowering) is the Kellogg & Harford gathering. The Bigelow label contains only the name ‘“Dirca” while the Kellogg & Harford label bears, in Gray’s handwriting, ‘‘Dirca occidentalis n. sp. A. G.” Therefore, there is little question that the latter, and later, collection should be con- sidered the holotype of D. occidentalis. JOURNAL OF THE ARNOLD ARBORETUM VoL. XLV APRIL 1964 NUMBER 2 THE GENUS CHAENOMELES (ROSACEAE) CLAUDE WEBER THE JAPANESE OR FLOWERING QUINCE, botanically, the genus Chae- nomeles of the family Rosaceae, has been under cultivation in temperate areas of the world for over 400 years. Selections have been made and named in nearly every country where the plants are grown. Through the years, horticulturists have selected and propagated over 500 named varie- ties (cultivars). The classification of these many cultivars has been diffi- cult, for unresolved problems exist in the morphology and taxonomy of the plants at the generic, specific, and varietal levels. Individual species exhibit a plasticity of form which is troublesome to botanists but of value in commercial horticulture. Hybrids are known involving two or all three species of the genus in both simple and in multiple crosses. One of the goals of this study was to determine whether such a large group of cultivars of woody plants, largely of unknown parentage, could be placed in a systematic arrangement. The late L. H. Bailey, in his Manual of Cultivated Plants, did not attempt to include the cultivated varieties of Chaenomeles. He stated “To describe the main varieties of cultivated plants, even those customarily meriting Latin names, requires long and painstaking study, collections for the most part not in existence, historical perspective, a special application of nomenclature, and an appre- ciation of systematic values yet little developed.” I have tried in the present study to meet these qualifications. The Arnold Arboretum has an excellent representation of the common cultivars of Chaenomeles in its living collections. Through the cooperation of many other botanical gardens and commercial nurseries, representatives of most available cultivars were assembled for comparative studies. Di- rectors and curators of herbaria and arboreta in many parts of the world have graciously loaned or given herbarium specimens of Chaenomeles species, varieties, and cultivars, and these have been an important aid in the historical and descriptive studies of the genus. The taxonomic, floristic, and horticultural literature in many libraries has supplied an “historical perspective.” Of particular value in assembling the hundreds of “fancy names” applied to the selections of Japanese Quinces has been 162 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV the collection of nursery catalogues of the Massachusetts Horticultural Society. The “special application of nomenclature” and the “appreciation of systematic values” which Bailey pointed out were “little developed” in 1948, have both matured with the acceptance of the category of “cultivar” in the 1953 and subsequent editions of the /nternational Code of Nomen- clature for Cultivated Plants. The genus Chaenomeles comprises only three species of moderate-sized shrubs adaptable to cultivation in temperate areas. By observing the plants through several years the seasonal changes in morphology could be observed; growing them in close proximity or in isolation, the natural and man-induced breeding habits could be studied. Collections of fruits, seeds, cuttings, and scions allowed experimentation on methods of germination, propagation, and subsequent development. Living plants were the subject of this study when current observations could be compared with the descriptions published in a scattered support- ing literature, and with an assemblage of herbarium specimens. Regret- tably, a study of the basic species in China and Japan has not been possible and the known ecology of Chaenomeles is that of the plants in cultivation. The taxonomic study of the genus Chaenomeles would have been im- possible without examining numerous specimens dispersed in the herbaria of North America, Europe, and Asia. I wish to thank the directors and curators of the following institutions for the loan of herbarium specimens. They are cited according to the standard abbreviations of /ndex Her- bariorum. A Arnold Arboretum of Harvard University, Cambridge, Massachusetts AAH Herbarium of cultivated plants, Arnold Arboretum, Jamaica Plain, Massa- chusetts BM. British Museum (Nat. Hist.), London, England G Conservatoire et Jardin Botaniques, Genéve, Switzerland cH Gray Herbarium of Harvard University, Cambridge, Massachusetts K Herbarium, Royal Botanic Gardens, Kew, Surrey, England L Rijksherbarium, Leiden, Netherlands LE Botanical Institute of the Academy of Sciences, U.S.S.R., Leningrad, U.S.S.R Ny New York Botanical Garden, New York P Muséum National d’Histoire Naturelle, Paris, France us U.S. National Museum, Smithsonian Institution, Washington, D. C. TI Botanical Institute, University of Tokyo, Hongo, Tokyo, Japan upsv Vaxtbiologiska Institutionen, Uppsala Universitet, Uppsala, Sweden Among the historical herbarium specimens, the “cultivars” were little represented. To fill this gap, numerous persons sent fresh material or took the time to press flowering branches, and to compile data. To the staffs of the following arboreta and nurseries, goes my appreciation for the gifts of living plants and herbarium specimens which added to both the living collection and the herbarium of cultivated plants of the Arnold Arboretum in Jamaica Plain, Massachusetts: 1964 | WEBER, CHAENOMELES 163 Beal-Garfield Botanic Garden, East Lansing, Michigan Belmonte Arboretum, Wageningen, Netherlands Dawes Arboretum, Newark, Ohio Department of Parks, Rochester, New York Du Pont Arboretum, Winterthur, Delaware George Landis Arboretum, Esperance, New York Holden Arboretum, Mentor, Ohio Ida Cason Callaway Gardens, Pine Mountain, ei Longwood Gardens, Kennett Square, Pennsylva Morris Arboretum, ange an ea Morton Arboretum, Lisle, Illin Mount Airy Arboretum, Cincinnati, Ohio New York State Agricultural Experiment Station, Geneva, New York New York Botanical Garden, New York Orland E. White Arboretum, Boyce, Virginia Planting Fields Arboretum, Oyster Bay, New York Proefstation, Boskoop, Netherlands Royal Botanic Gardens, Kew, Surrey, England Stanley M. Rowe Arboretum, Cincinnati, Ohio Station of Medicinal Plants, Kasukabe-Shi, Japan Tudor House, Ripley, England United States National Arboretum, Washington, D. C. Oe haa Dept. of Agriculture, Plant Introduction Station, Glenn Dale, and ian of Connecticut, Storrs, Connectic University of Washington Arboretum, Seale Tiesto: Villa Taranto, Verbania-Pallanza, Italy Bunyard Nursery, Maidstone, England Darthuizer Nursery, Leersum, Netherlands Fruitland Nursery, Augusta, Georgia Harrison Nursery, New Zealand Monrovia Nursery, Azusa, California ee Winchester, Tennessee Wayside Gardens Nursery, Mentor, Ohio Willig Nursery, Ottawa, Kansas The staff of the Chinese-Japanese Library of Harvard University was most helpful in checking Chinese and Japanese bibliographical notes or geographic names. The work necessary for the completion of this study was rendered pleasant by the interest and the kindness of the staff of the Arnold Arboretum in Cambridge, Jamaica Plain, and Weston, and of many graduate students in botany, who shared their knowledge or provided any help which was needed. My thanks go especially to Dr. Donald Wyman for sharing his observations on ornamental shrubs, and to Mr. Alfred Fordham for the care he took in propagating material, sometimes received in poor condition. Their horticultural experience was invaluable. My deepest appreciation goes to Dr. Richard A. Howard, who first sug- gested the genus Chaenomeles as a thesis subject, and whose enthusiasm never failed to be inspiring during the pursuit of this study. 164 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV SYSTEMATIC POSITION OF CHAENOMELES IN THE MALOIDEAE The genus Chaenomeles Lindley is assigned to the subfamily Maloideae ! of the family Rosaceae. This natural division, under the name Pomaceae was suggested as a “‘fragmentum”’ by Linnaeus in 1763, who listed under it several genera such as Pyrus and Crataegus, but included also Punica and Ribes which are now placed in the Punicaceae and in the Saxifragaceae respectively. No description was given at that time. Since then, the justi- cation for this taxon has not been contested by taxonomists. Some have treated the Maloideae as a separate family,” others as a tribe.? Currently it is widely accepted as a subfamily of the Rosaceae.‘ The morphological characteristics of the Maloideae are: Trees or shrubs, unarmed or spiny. Leaves petiolate, alternate, simple or compound, with free stipules (at least on young shoots). Flowers regular, hermaphrodite (rarely unisexual by abortion), solitary, fasciculate or in terminal racemes, white, pink or red. Calyx campanulate-urceolate with 5 persistent or de- ciduous lobes. Petals 5, unguiculate, inserted in the throat of the calyx. Stamens 10-60, inserted at the same level as the petals, filaments free, anthers introrse with 2 locules. Disc laminate or fleshy, dry or nectari- ferous. Ovary formed of (1) 2—5 carpels, more or less adherent to the calyx on their abaxial side, completely fused with it at maturity. Ovules anatropous, usually 2 (1 by abortion) in each locule, to indefinite (Chae- nomeles, Cydonia). Styles 1-5, terminal, free or connate at the base. Pollen tricolpate. Fruit fleshy, usually a pome (exceptionally a berry), formed of the hypanthium, the carpels becoming chartaceous or carti- laginous, and sometimes dehiscing on the adaxial suture, to bony and indehiscent. Seeds ascending (or horizontal when numerous), testa coriaceous, or less often mucilaginous, raphe and chalaza apical (or subapical), endosperm lacking, embryo orthotropous with short and conical radicle and imperceptible plumule. According to Metcalfe and Chalk (1950), the Maloideae can not be separated by anatomical characters from the other subfamilies of the " According to the International Code of Botanical Nomenclature, the name Pomoi- deae, in current use, is incorrect as it is not based on a genus. The ila name Pomoideae should be replaced by Maloideae, based on Malus Miller Maloideae Weber, stat. nov. Type: ae Miller. Malaceae Small, F]. Southeast. U.S. *Pomaceae: Lindley, 1822, 1847; Baring Pome), 1830; Spach, 1834; Meisner, 1836-43; Endlicher, 1841; Walpers, 1843; Hoffmann, 1846; Roemer, 1847; Decaisne, 1874; Dippel, 1893; Renae. 1896a & b, eee & b. Malaceae: Small, 1903, 1905; penn seen Jussieu, 1789. Pomaceae: De Candolle, 1825. Pomeae: Bentham & Hooker, 186 ‘Pomeae: ‘Gray, 1842; Koehne, 1893. Pomoideae: Focke, 1894; Dalla Torre & Harms, 1900-07; Ascherson & Graebner, 1906-10; Rendle, 1925; Wettstein, 1935; Diels, 1956; Skottsberg, 1940; Rehder, 1949; Lawrence, 1951; Benson, 1957; Ember- ger, 1960. 1964 | WEBER, CHAENOMELES 165 Rosaceae. The genera of the Maloideae have few characteristics in com- mon, as is true in the other subfamilies of the Rosaceae. Variation in the stem structure occurs even in individual species of Chaenomeles com- parable to that recorded by Aubertot (1910) in the genera Pyrus, Cratae- gus, and Prunus. The Maloideae (except Osteomeles) have exclusively solitary vessels, while most of the other Rosaceae have a tendency to have a few together. They have fibers with fewer bordered pits on the tangen- tial wall than on the radial walls, while the other Rosaceae, except some species of Spiraea have pits equally numerous on both walls. The only leaves with glandular teeth are found in Chaenomeles, Pyrus, Sorbus, and Crataegus, all members of the Maloideae. If the Maloideae do not stand apart anatomically from the other sub- families of the Rosaceae, the genera do so cytologically. According to Sax (1931, 1932, 1934) the basic chromosome number is 17 in all genera of the Maloideae, while it is 8 and 9 in the Spiraeoideae, 7 and 9 in the Rosoideae, 8 in the Prunoideae. This has been confirmed by all subse- quent authors, Moffett (1931), Darlington (1945), Delay (1950-51), Cave (1958, 1959, 1960) and Léve and Live (1961). Only Ouillaja brasiliensis Mart., which undoubtedly belongs (by its morphology) to the subfamily Spiraeoideae, and has always been placed there by botanists, presents an exception by having 17 pairs (Bowden, 1945). Polyploidy occurs in the whole family Rosaceae, being limited to tri- and tetraploidy in the Maloideae. Darlington and Moffett (1930) suggested that the 17 pairs of chromosomes in Pyrus are made up from a basic number of 7 by duplication of 4 pairs and a triplication of 3 pairs of chromosomes. Sax (1932) rejected this idea and proposed that the Maloideae originated by hybridization of members of two different subfamilies, most likely belong- ing to the Spiraeoideae and to the Rosoideae. The Maloideae would thus be allotetraploid. In geographic origin, the Maloideae are nearly restricted to the northern temperate hemisphere. Sax (1931) notes that 15 of the 18 genera enu- merated by Rehder (1927) are represented by Asiatic species. He con- cluded that the Maloideae must have originated in Asia before starting their migration toward America (1 genus Osteomeles in Hawaii), Europe (1 genus Chamaemeles localized in Madeira), and North Africa. Only Hesperomeles reached South America, following the Andes as far as Chile. There is no argument among taxonomists about the genera to be in- cluded or excluded from the Maloideae. Difficulty arises, however, when one has to determine the limits of the genera of the Maloideae, for many intergeneric hybrids are reported. The genera recognized by Rehder (1949) are the following: Amelanchier Medicus, Aronia Medicus, Chaenomeles Lindl., Cotoneaster Ehrhart, Cra- taegus L., Cydonia Miller, Docynia Decaisne, Eriobotrya Lindl., Malus Miller, Mespilus L., Osteomeles Lindl., Peraphyllum Nuttall, Photinia Lindl., Pyracantha Roemer, Pyrus L., Rhaphiolepis Lindl., Sorbus L., Stranvaesia Lindl. To these should be added the genus Chamaemeles Lindl. and Hesperomeles Lindl. which are not cultivated, and the hybrid 166 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Aronia Hesperomeles Fic. 1. Sexual intergeneric hybrids in the Maloideae. The genera are not arranged according to their affinities as such an arrangement would make the plate difficult to read. genera Amelasorbus Rehder (Amelanchier >< Sorbus), Crataegomes pilus Simon-Louis (Crataegus « Mespilus), Pyracomeles (Osteomeles < P yra- cantha), Pyrocrataegus Daniel (Crataegus Pyrus), Pyronia Veitch (Cydonia & Pyrus), Sorbaronia Schneider (Aronia < Sorbus), Sorbopyrus Schneider (Pyrus < Sorbus). According to Rehder (1940), the cross Cydonia oblonga Malus is reported to have been raised by I. V. Michurin, while Sax (in litt.) mentioned that the cross Malus Pyrus was produced in Australia (Fic. 1). Within the subfamily Maloideae, the genus Chaenomeles is most closely related to the genus Cydonia, with which it shares the character, unusual in the Maloideae, of an indefinite number of seeds in each locule. C ydonia and Chaenomeles have been placed in the same genus, namely Cydonia Miller, by numerous authors.» By others, Chaenomeles is considered a >Persoon (1807), Loiseleur-Deslongchamps (1815, 1817), Sweet (1818, 1827), Hayne (1822), Guimpel, Otto & Hayne (1825), Siebold (1830), Loudon (1838), Endlicher (1840), Miritzi (1845-46), Van Houtte (1849), Planchon (1849), Lemaire (1856), Verschaffelt (1856) Courtin (1857), Hoffman & Schultes (1864), Wenzig Chaenomeles, 1914), Burgerstein (1901), Hemsley (190la & b), Schumann (1901), Muth (1902), Bean (1903, 1914, 1930; Chaenomeles, 1951), Grignan (1903), Veitch (1903-04), Beckett (1907, 1909-10), Makino (1908), Nakai (1908, 1909; Chaenome- 1964 | WEBER, CHAENOMELES 167 section of Cydonia.6 Chaenomeles is also related to Pyrus, and to Malus if this taxon is separated from Pyrus. Chaenomeles japonica (Thunb.) Lindl. ex Spach, the type species. of the genus, was described first by Thunberg (1784) as a species of Pyrus. As late as 1902, some authors still joined Chaenomeles (and Cydonia) to Pyrus." Andrews (1807) transferred Pyrus japonica Thunb. to Malus. CHAENOMELES 2-4-(5)-6-7-10-11-14- 15-18-19 MALUS 1-2-4-6-7-(8)-(9)- \, 10-11-13-16-18-20 DOCY NIA 2-4-(5)-6-7-9- 11-13-15-17-19 ¢ 2-4-6-7-(9)-11-13 RUS CYDONIA PY 2-3-5-8-10-12 5-8-12-13- 1-4-5-8-9-12-13- 13-15-17-19 GENERIC CHARACTERS 16-17-20 . Vernation conduplicate. e . Vernation involute. . Aestivation contorted. 2 4, Aestivation imbricate. 6. Flowers on old wood. 8 3 5, Flowers on leafy shoots. 7. Sepals erect at anthesis. . Sepals reflexed at anthesis. 9. Petals white. 10. Petals usually colored (pink to red). 11. Stamens in two rows (15-60). 12.’Stameng in one row (20-30). 13. Styles free (sometimes coalescent). 14. Styles fused in a column. 15, Stigma decurrent. 16. Stigma discoid. 18. Grit-cells absent’ (or in a row ir the exocarp 17. Grit-cells numerous and in the whole h thium. 19. Numerous ete ovules in each 20. Two ovules only in each locule,. locule. . Affinities between the genera meee gen to riety The of the species. having even les, 1916a), Mottet (1917), Cardot (1918), Richards (1923), Spath Paes Diels (1948), Jex- or (1950), Emberger (1960). the nus. The fae (1803), Jacquin (1809), Aiton (1811), L (1826), Morren (1851), Miguel (1867), Koch (1869), Franchet & Savatier (1873-75 168 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLV Another closely related genus, Docynia, shares with Chaenomeles the character of numerous stamens in two rows (40-60 in Chaenomeles, 30-50 in Docynia). The type species of the taxon, Docynia indica (Cole- brooke) Decne., was described first as Pyrus indica by Colebrooke (in Wallich, 1831), then transferred to Cydonia by Spach (1834), before Decaisne (1874) established the genus Docynia.® The genus Pseudochaenomeles Carriére (1882), based on an unreliable fruiting characteristic of Chaenomeles japonica (Thunb.) Lindl. ex Spach (as C. maulei) as opposed to C. speciosa (Sweet) Nakai (as C. japonica), was not maintained by its author. The genus Pseudocydonia Schneider, which was first established as a section of Chaenomeles by Schneider in May 1906, was raised by him to generic rank in November of the same year, to include Cydonia sinensis (Dum.-Cours.) Thouin, will not be considered here. These two genera are not at present recognized by any taxonomist. In Ficure 2 an attempt is made to show the affinities between the genera related to Chaenomeles. The characters chosen are arranged to offer an alternative. The one given the odd number is supposed to be more primi- tive than the character having an even number. If it were possible to show the entire subfamily Maloideae on a multidimensional system, it would help to understand the generic limits and the phylogeny of the group. This is not practical for the genera are too numerous and many are not known well enough at the present time. Because of the wide divergence of opinion among taxonomists as to how many genera should be recognized in the Maloideae, it is presumptuous to attempt to discuss the evolution of the different genera without having studied the whole group carefully. From its geographic distribution, as shown above, from the morphological characters of the inferior ovary and the syncarpic fruit, and from a consideration of the high basic chromosome number, the subfamily Maloideae appears to be of more recent origin than the other subfamilies of the Rosaceae. The Maloideae is closest morphologically to the Spiraeoideae and the Rosoideae, and closest anatomically to the Prunoideae. The genera of the Maloideae are closely related, as shown by their morphology, and sup- ported by the graft compatibilities (F1c. 3) and numerous sexual hybrids. This was shown by Decaisne in 1872, and Sax in 1931, each of whom wrote that botanists having a tendency for lumping could go as far as to consider all the Maloideae one genus. Burgerstein (1896a) suggested on the basis of the anatomical structure: “Alle untersuchten Pomaceen — 13 Arten (inclusive Hybriden und Varietéten) — die sich auf 16 Gattungen Franchet in Chaenomeles, 1883), Hemsley (1873; Cydonia, 1901), Kurz (1873), Masters (1874a), Neumann (1875), Smith (1875), Hooker (1884), Tanaka & Ono (1 891), Arnott (1902 * Abnormal flowers of Chaenomeles speciosa observed in the spring of 1961 had the styles fused in a hairy cone protru gre above the insertion of the stamens and petals, very much like Docynia, and i extreme cases, no ovules were present. Lobed leaves, — in Chaenomeles reins occur regularly in juvenile leaves of Decynia indic 1964 | WEBER, CHAENOMELES 169 Aronia (l Pyracantha i\| | i | Mespilus Hesperomeles Fic. 3. Grafting affinities of the Maloideae. The arrows indicate the direction from understock to scion. vertheilen, zeigten im Wesentlichen einem Uebereinstimmenden Holzbau.” The genera of the Maloideae are thus delimited mostly for the sake of convenience, and there will always be disagreement among taxonomists as to how many genera and which ones should be recognized. The position of the genus Chaenomeles within the Maloideae is not clear. The numerous seeds in each locule indicate that Cydonia and Chaenomeles are primitive in this respect, and Pyrus and Malus, with only two seeds in each locule, are more advanced. The characteristic of the number of seeds induced Decaisne (1847) and Wenzig (1874 to 1883), in accordance with most of the authors recognizing only the two genera Cydonia (including Chaenomeles) and Pyrus (including Malus), to consider Chaenomeles most closely related to the genus Cydonia. On the other hand the mor- phology of the flower (stamens and styles) and of the fruit, and the presence or absence of grit cells placed Chaenomeles closer to Malus, and Cydonia closer to Pyrus. Koehne (1890) suggested an evolutionary scheme of the Maloideae, in which the genera evolved from Sorbus in six distinct lines, and he is also followed by Dippel (1893). Unfortunately, the results provided by anatomical investigations do not agree with the conclusions reached by observations on the gross morphol- ogy. The anatomical investigations are also contradictory, for so far only one characteristic has been considered at a time. If the size of the vessels and the number of cells in the width of the vascular rays are considered 170 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV alone (Burgerstein 1896a, 1896b, 1898a, 1898b), Chaenomeles is closer to Pyrus, and Cydonia to Malus. If the presence or absence of a tertiary thickening in the vessels (Burgerstein 1898a) or the anatomy of the leaf (Gérard, 1884) are the only characters chosen, Chaenomeles and Cydonia are more closely related, as are Pyrus and Malus. The divergence of con- clusions reached by independent study of the gross morphology or of the anatomy has been pointed out by Burgerstein (1896a), and Folgner (1897). The taxon Cydonia sinensis presents difficulties in the distinction of the four closely related genera Chaenomeles, Cydonia, Pyrus and Malus (Docynia not being well enough known). This species has been placed successively in the four genera mentioned, in Malus by Dumont de Courset (1811), in Cydonia by Thouin (1812), in Pyrus by Poiret (1816), and in Chaenomeles by Koehne (1890). Schneider (1906a) established a special section for C. sinensis which fitted so poorly in Chaenomeles that within the same year he (Schneider, 1906b) raised it in rank to form the monotypic genus Pseudocydonia. According to Burgerstein (1898b), C. sinensis is “eine Briicke” between Cydonia oblonga Miller and Chaenome- les speciosa. The basis for the transfer of Cydonia sinensis to Chaenomeles by Koehne is not clear. He wrongly stated that the styles are fused. The only characteristics Cydonia sinensis shares with Chaenomeles are the glabrous fruits with deciduous sepals, serrate leaves, and an Asiatic origin (extinct in the wild), while it shares more than 35 other generic characters with Cydonia. Moreover, all the three species of Chaenomeles have fertile hybrids between them, while crosses of Chaenomeles species with Cydonia sinensis have been attempted several times but have always failed. The fact that Cydonia oblonga has entire leaves while C. sinensis and Chae- nomeles have serrate leaves must not be given much weight, even if con- venient in a key. The first few leaves of Cydonia oblonga seedlings are often serrate and later appear with entire margins. The fact that the calyx of Cydonia sinensis is deciduous, in contrast to the persistent calyces of Cydonia oblonga, is also unimportant for this character varies on indi- vidual shrubs of Chaenomeles. Rubsow (1944) noted that all the primitive species of Pyrus have a deciduous calyx, while the more advanced species have a persistent calyx. Burgerstein (1896a, 1896b, 1898b) studied the wood structure of Cydonia sinensis and found it to correspond more closely with the struc- ture of Chaenomeles wood than with that of Cydonia wood. From his wood study, this author considered Cydonia and Chaenomeles as distinct, and the latter as including C. sinensis (as suggested by Koehne). In 1898 he observed for the first time the flowers of Cydonia sinensis on a tree growing in the Vienna Botanic Garden. Burgerstein was forced to admit that the styles are free and not fused, as he had previously believed. This fact must have disturbed him, for in 1901 he published a paper, accom- panied by a good colored plate, entirely devoted to this species, which he reassigned to Cydonia. His anatomical key to the genera of the Maloideae 1964 | WEBER, CHAENOMELES 171 did not correspond to the morphological affinities he observed in the styles, and he united Cydonia and Chaenomeles again. Burgerstein is the only botanist so far who arrived at this conclusion after having studied both genera Excluding Burgerstein’s opinion of the validity of the genus Chaenome- les, all other taxonomists who have studied Cydonia and Chaenomeles, namely Bean (after 1951), Carriére, Decaisne, Koehne, Koidzumi, Natal (after 1916), Rehder (after 1914), and Wenzig (after 1883) have con- cluded that they are distinct genera. This conclusion might have been reached more easily if the limits of the two genera were not somewhat blurred by the incorrect position in which Koehne placed Cydonia sinensis. There is a tendency often to delay a change of name in a nursery catalogue, sometimes even against horticultural knowledge. It is shown here in the reluctance of Goldring, Grignan, Hemsley, Mottet, Planchon, Spath, and Sweet to accept the conclusion that Chaenomeles and Cydonia are distinct; and some even persisted in placing Chaenomeles species in the genus Pyrus. In spite of the reservations of some botanists and horticulturists, the recognition of the genus Chaenomeles is justified. The three species it includes, C. japonica, C. speciosa, and C. cathayensis form a good taxono- mic unit, based on their geographic distribution, their morphology, and their breeding behavior. The three species do cross with each other but do not cross with species of other genera. Within the Maloideae, Chae- nomeles possesses characteristics more primitive than the other genera, such as numerous seeds and numerous stamens in 2 rows, but it also shows a tendency toward unisexual flowers, which is to be considered an advanced character. The genus Chaenomeles has not been found as a fossil in Japan (Tanai, 1961), in spite of the fact that several Rosaceae were recorded in the Tertiary flora of the island and that the leaves of Japanese Quinces should be easily recognized by their serrations. Chaenomeles is perhaps a rela- tively recent immigrant in Japan, having evolved in China in the Yunnan region, where two of its species are still found. MORPHOLOGY AND ANATOMY The morphology of the genus Chaenomeles was studied on living speci- mens observed over a period of three years. These observations showed the unsuspected extent of variability existing in each individual and per- mitted consideration of the characters commonly used to distinguish the genus Chaenomeles from closely related genera, as well as those used to separate its species. A survey was also made of the anatomical characters of Chaenomeles to determine if any feature would distinguish this genus among the Maloideae. Roots. The roots are numerous, fasciculate, and relatively superficial. In an old shrub, there may be several main roots, reaching up to 3 cm. in diameter. 172 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV André (1872) reports having observed in the Simon-Louis Nursery, Metz, France, roots which when dug up to make root cuttings, were bearing flower buds. One of these root fragments, left on a table, produced flowers which opened within a few days. André’s figure shows flowers with long peduncles, but on an otherwise normal spring inflorescence. Twice, at the Arnold Arboretum, we observed flowers coming up through the ground, without being able to ascertain whether they were attached to a root or to a rhizome. The inflorescences were of a special type, and will be dis- cussed further on. Habit. There are three growth-forms in Chaenomeles corresponding to the three species. C. japonica has numerous spreading branches which completely cover the ground, and when in contact with it have a tendency to form adventitious roots. C. speciosa possesses many branches which are erect at first, and then spreading. Plants of this species increase by new suckers which, when left without control, may cover extensive areas. C. cathayensis, in contrast, possesses only a few straight, erect branches, with numerous short lateral branchlets terminated by a spur. These plants have a stiff appearance and can be trained as small trees. The interspecific hybrids usually show an intermediate shape, but there are a few exceptions. C. « superba ‘Crimson and Gold’ tends to spread, forming a mat up to 50 cm. high on the ground, as would a pure C. japonica. Two cultivars of C. speciosa, ‘Spitfire’ and ‘Starlight’ have a characteristic vase shape due to numerous straight branches. Several cultivars show abnormal branching as indicated by their names ‘Fastigiata’, ‘Pendula’, ‘Contorta’, and ‘Tortuosa’. In ‘Contorta’ the mutation affecting the shape of the branches (which change direction at each node) is carried into the progeny, or at least part of it, even though the other parent has normal branching. The method of training Japanese Quinces along a wall, in common use in England, indicates how the shape of a shrubby plant may easily be changed, even with a minimum amount of pruning. The natural shape seems also to be modified through the influence of heavy snow cover. The hybrids of C. cathayensis often lose their branches completely, due to cold. The parts protected by the snow usually survive and start new shoots the following spring. On old plants, the snow may produce an artificial spreading form by repeated elimination of the straight, erect branches. In very cold winters, flower buds of Chaenomeles freeze above the snow line. The hardiness of C. japonica is probably due to its low habit rather than to intrinsic physiological qualities. Buds. There are three kinds of buds, leaf buds and flower buds, both well formed, and adventitious buds. The pointed and round leaf buds, usually 1 to 2 mm. long, are covered by a few pubescent, loosely imbricate scales. There are no true terminal buds since the long shoots are either terminated by a spine, or keep growing intil the fall, when they are killed back by frost. The vernation is con- 1964 | WEBER, CHAENOMELES 173 duplicate, in contrast to the involute vernation of Pyrus and of some species of Malus. When the leaf buds are dissected, numerous brown imbricate scales, ciliate at base, or with tufts of trichomes at base and apex, and varying in size, may be seen. Within are about 4 to 8 embryonic leaves, with petioles already visible by the end of the summer and, in the center, meristem which will produce the next year’s long shoot. The glabrous, embryonic leaves appear entire, but are bordered by narrow, ovate, transparent glands, lined side by side, and directed toward the tip. When the leaves unfold, the petioles grow faster than the blades, repre- senting at one time 14 to 1% of the leaf, but in mature leaves only 14 to 1% of the length. Before the inner leaves have finished unfolding, the glands may be seen to secrete liquid. The flower buds are formed early in the fall. The 1 to 6 flowers con- tained in each bud start to expand at the end of the summer, breaking up the bud. Ciliate scales surround the young flowers which spend the winter in a very advanced stage. By October, the flower color can often be recognized. In fact, the flowers are ready to open in a few days, at about any time from November, and often do, if the cold weather does not interrupt their development. This is why branches of Japanese Quinces are so easily forced for sale as cut flowers. In spite of this precocity of pigmentation of chlorophyll and anthocyanins, meiosis for most buds occurs in the spring. The buds do not show specific differences in Chaenomeles. In order to identify the shrubs in winter condition, the shape of the branches, the spines, and the warting of the twigs normally present good characters for distinguishing the species. In cultivation, the abundance of interspecific hybrids showing intermediate characters renders this method very un- certain, however. Shoots and spines. There are two main types of shoots in Chae- nomeles, as in the closely related genera, which represent long and short shoots. These appear at different times of the year and spend the winter in distinctive stages, the long shoots as shoot “primordia,” and the short shoots as embryonic leaves. The short shoots develop first, nearly simultaneously with the flowers, usually the first week of May in the climate of Boston. They appear on old branches as well as on the long shoots of the preceding year, without any indication of future development into long shoots. They bear four to eight fasciculate leaves varying widely in size, the outer ones being 49 to 1% the size of the inner ones. The outer leaves are frequently malformed, subentire (but with glands), and often emarginate or with a dry tip. They are usually promptly caducous. As the inner leaves unfold, the glands terminating the veins form the tip of each serration, alternating with a gland in each indentation. On the long shoot of the preceding year, a spine axillary to the leaves with a bud on each side may often be observed. In C. cathayensis, one of these buds develops into a short shoot which may be located on the right, 174 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV or on the left of the spine, but it will be on the same side all along the branch. The buds on the other side of the spines will remain dormant under a large scale, for one or a few years, and may later produce flowers. In C. japonica, buds produced basally on both sides of the spines usually produce short shoots, giving a characteristic bushy appearance to this species. C. speciosa normally does not have spines axillary to each leaf of a long shoot, but rather has one bud produced at each node. The leaves of the short shoots are produced within a few days. The shoots terminate in a bud, which may or may not produce a spine. After the leaves have expanded on short shoots, and by the middle of May in New England, a second flush of growth takes place. From this time on, long shoots only will be produced until the following spring. Leaves of long shoots do not differ in shape or size, but they are accom- panied by large, paired reniform stipules, contrasting with the exstipulate leaves of the short shoots. New long shoots appear at the apex of last year’s long shoots, if these have been interrupted and partially killed back by frost. The last living axillary bud, or many of the upper ones, may at once start to produce long shoots. New long shoots rarely appear if the long shoot of the preceding year was terminated by a spine. Long shoots located at the tip of branches do not usually ramify during the first season. The straight lateral branchlets’ occurring at right angles with the main branches, typical of C. cathayensis and its hybrid C. x cali- fornica, are formed on long shoots during the second season. Each axillary bud will produce a long shoot which is terminated by a strong spine or spur. Very often flowers will be borne on these short branchlets and in C. X californica long shoots will appear as a continuous mass of flowers along most of their length. Long shoots, when expanding, are usually covered with a short and scabrous tomentum. The trichomes are unicellular, unbranched, and slightly bulbous at the base. They fall off the first summer without leaving a scar in C. cathayensis and in C. speciosa. In C. japonica where the trichomes are more abundant, the bulbous hair bases persist and appear as dark warts on the shoots during the second year. As the twigs enlarge in girth, the warted epidermis is shed during the second summer, and is replaced by a smooth bark. The presence of warts on the second year twigs is a characteristic of C. japonica and its hybrids. The old warted epidermis persists longer around the nodes, and can sometimes be found after the second year. Temporary warts occur exceptionally in C. speciosa, and have been observed in at least one herbarium specimen from China, where hybridization with C. japonica is very unlikely. Spines may be formed very early, axillary to a leaf in a long shoot, especially in C. cathayensis. Most of the time, however, they appear only in the summer, terminating either a short shoot, or a long shoot. They are slender at first, straight, and probably increase in diameter during the second year. Young spines show by their occasional pubescence, or the presence of buds, that they are modified long shoots. The spines are stouter and more abundant in C. cathayensis, more slender in C. japonica, 1964] WEBER, CHAENOMELES 175 and show an intermediate condition in C. speciosa and in the interspecific hybrids. A few cultivars have been named for a lack of spines, however, there are usually a few on old shrubs of these spineless cultivars. C. speciosa ‘Contorta’, and C. X superba ‘Tortuosa’, in addition to unusual branch form, possess hooked spines, the points of which are directed toward the base of the branch. The curvature is evident as early as the spines ap- pear from the buds. Abnormal production of long shoots has been observed in the following cases: after the shrubs have been killed back to the ground or severely damaged by frost; after application of a weed-deterrent; and after an ex- tremely severe attack of insects. After a killing frost, long shoots are produced by adventitious buds at the base of the plant. These shoots are abnormally vigorous. They grow and ramify very rapidly, reaching as much as 1 meter in length within a few weeks. The pubescence is reduced or completely absent, the epidermis shedding off soon after being produced. Spines often appear immediately, axillary to each leaf, at least in C. cathayensis. The leaves and stipules of these shoots attain greater size than those in the crown of the shrubs. Following the application of a weed-deterrent (Simozine), in September 1961, an abnormal production of long shoots was observed on many young plants of Chaenomeles in the Arnold Arboretum. The reaction seemed to be similar in the different cultivars which were affected. As an immediate response, long shoots were produced from adventitious buds located at the base of the shrubs, some reaching 80 centimeters or more three weeks after the application of the weed-deterrent. These shoots were often ramified, but could be distinguished at once from shoots produced after a frost effect by their abnormal leaves and stipules which were, in most cases, extremely narrow and showed discoloration. The leaves appeared varie- gated and colored pink by anthocyanin pigments. Green pigmentation, chlorophyll, was present in many only along the midribs. T hese abnormal shoots were especially sensitive to subsequent frosts. In Cincinnati, Ohio, a very severe attack of seventeen-year locusts occurred in the summer of 1959. In the spring of 1962, Chaenomeles shrubs badly damaged three years before were observed at the Stanley M. Rowe Arboretum in different stages of recovery, In many branches the tips were affected so much that they had died back. In such cases, new long shoots were formed from the buds which otherwise would have pro- duced short shoots. Since the branches attacked by the seventeen-year locusts have a tendency to crack open, the living part was often limited to one side. When this happened, all long shoots were produced on this same side, changing radically the ramification which normally alternates along the axis. Anatomy of the stem. In a young long shoot, a large part of the stem is occupied by pith consisting of round cells containing starch grains. Each of the numerous vascular bundles, distinct at this stage, has a cap of 176 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV fibers. The rays are uniseriate. In the cortex, the cells of the ground tissue are loosely arranged and contain chloroplasts and starch grains. The epidermis is brown, as are the simple unicellular trichomes. In a one-year old stem of C. japonica, the trichomes have been shed and a multicellular wart has developed replacing the bulbous base of the deciduous hair. The epidermal layers consist of about four rows of cells arranged tangentially to the axis. Older stems show a star-shaped pith formed by loosely arranged round cells. The proportion of pith is much smaller in older stems than in young shoots. Diffused through the pith are patches of vertical parenchyma. The uniseriate or sometimes biseriate rays originate in the metaxylem. The annual rings of irregular size are well marked. The wood is diffuse porous but more vessels are formed in the spring. The vessels are usually solitary, or rarely two to three together. The diffuse vertical parenchyma is abundant among the fiber tracheids. The region of functioning sec- ondary phloem is very narrow. The metaphloem consists of a few rows of crushed phloem elements. The protophloem appears as isolated patches of primary phloem fibers. The cortex is formed of 8 to 10 rows of cells tangentially arranged, with many intercellular spaces. Forming the peri- derm, are the phelloderm, the phellogen, and the phellem consisting of smaller rectangular thick-walled cells, tightly packed together. In Pyrus, according to Esau (1953), the periderm originates partly in the epidermis, and partly beneath it. The same is true of Chaenomeles in which the wood is very similar to that of Pyrus. The cells of the epidermis are light brown in color. In radial and tangential sections the rays can be observed to be homo- geneous, formed of round to oval cells, and non-storied. They can be from 25 to 45 cells high. The vessel elements show simple perforations. The walls are reinforced by spiral thickenings. Between the vessels and the fiber tracheids are alternate to opposite bordered pits. Simple pits are found between the ray cells. The only specific difference observed in the stems is the difference in growth rate between Chaenomeles japonica on one hand, and C. speciosa and C. cathayensis on the other. For a given stem diameter, C. japonica may show twice as many growth rings as the others. C. japonica is the only species to show warts on the epidermis of the one-year shoot. The first epidermis is shed during the second year, and after that the stems of C’. japonica appear as smooth as those of the other species. Node and petiole. In a long shoot of Chaenomeles, the node shows three traces coming from three gaps. In the pulvinus, the two lateral traces divide, one branch of each going into the stipule, the other into the petiole. The three traces inside the petiole fuse together and form one trace along the length of the petiole. According to Howard (1962), this pattern represents three traces from three gaps, with bundles fusing to form a simple arc by simple marginal fusion of the traces. In the pulvinus the petiole has the shape of a half circle. Under the epidermis are one to 1964] WEBER, CHAENOMELES 177 two rows of large brown cells, tangentially arranged. The rest is occupied by parenchyma and by the vascular bundle appearing as a smaller half circle in the center with xylem on the adaxial side and phloem on the abaxial side. At the base of the petiole there are no fibers. By the middle of its length, fibers occur in patches forming a broken arc outside of the phloem. Higher up the petiole changes shape as small wings are formed at the adaxial corners. A patch of fibers is found in each wing. In the upper part of the petiole, the traces appear which form the pinnately arranged veins. The first two veins do not leave the midrib at exactly the same point. One appears first, then the other, followed rapidly by the third and the fourth veins on alternating sides. Unicellular trichomes are occasionally found on petioles, and are a regular feature on the midribs of C. speciosa and on the abaxial surface of C. cathayensis. In contrast to Cydonia, there are no glands on the petioles of Chaenomeles. Leaves and stipules. In Chaenomeles, the leaves offer excellent char- acteristics for distinguishing species (Fic. 4). C. japonica possesses obovate to spathulate leaves often terminated by an indentation, and coarsely crenate on the margins. In contrast, the leaves of the other two species are serrate. In C. cathayensis, the leaves are elliptic to lanceolate, finely and sharply serrate, each serration terminating in an awn-like tip. C. speciosa has ovate to oblong, sharply serrate leaves. The type and amount of pubescence is a characteristic of the species, and is as useful as the serration for recognizing the parents of a hybrid. In C. japonica, the leaves are usually completely glabrous, even when unfolding, and only rarely have a few short hairs on the midrib of the under surface. C. speciosa is usually pubescent on the midrib of the under surface (excep- tionally so on the upper one). C. cathayensis has leaves often completely covered, when young, by a fulvous tomentum underneath but when the leaves expand, the tomentum may become thinner or fall off. C. cathayensis var. wilsonii (Rehd.) Bean, was based on the presence of this tomentum, which in reality occurs in variable abundance in individual adult plants. In transverse section the leaf shows a normal bifacial structure. Under the epidermis of the adaxial side, are two to three rows of palisade cells above the spongy parenchyma. The vascular bundles are surrounded by a bundle sheath consisting of parenchyma cells. Stomata are restricted to the lower epidermis. The guard cells are reniform. The leaves of Chaenomeles are serrate, each serration and alternating indentation ending with a gland. In cross section it can be seen that both phloem and xylem elements terminate just below the tip of the gland. These tracheids are enclosed by bundle-sheath cells. The epithem, a thin- walled parenchyma without chloroplasts, is little differentiated from the spongy tissue. Glands are known to differentiate relatively early in leaf ontogeny and to function for a short time. In Chaenomeles they are nearly full size in unrolling leaf-buds. When functioning, the glands are trans- parent or tinted red by anthocyanins. The secretion can often be observed 178 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Fic. 4. Leaf shape (natural size ) ae ater: types (X 6) in Chaenomeles. japonica ‘Sargentii’; b. C. spec has Itzii’; c and d. C. catha ayensis, ci o 8 ee) le n old shrub, d. juvenile te af: C x superba ‘Superba’; f. — ‘Minerva’; g. C.X Saeed ‘Attetslow: h. C.X californica ‘Rose- mary’. 1964 | WEBER, CHAENOMELES 179 as dried transparent droplets on herbarium specimens. The foliar nectaries in Chaenomeles may be considered intermediate between hydathodes and glands by their position, their structure, and their relatively fluid secre- tion. As soon as the leaves mature and become coriaceous, the glands cease functioning and dry up according to Ono (1907). They appear dark-red or black on adult leaves. Leaves on short and long shoots were compared in order to find if there were any morphological differences. The leaves have the same shape, type of serration, abundance of glands, and petiole length. Short shoot leaves may be larger than long shoot leaves appearing on the crown, but long shoot leaves of adventitious shoots or sprouts are the largest of all. The leaf size is so variable that it may vary ten-fold in one short shoot. Size and shape could not be used to distinguish isolated leaves as to origin on either long or short shoots. The only difference is the presence on leaves of long shoots of large paired stipules, and their absence on the leaves of short shoots. The stipules are free, reniform, and show the type of serration and pubescence characteristic of the species. They vary in direct relation to the leaves, tending to be very sharply serrate when accompanying juvenile sharply serrate leaves, or nearly entire in some old shrubs of C. cathayensis. Occasionally, stipules have been found to be slightly lobed in three culti- vars of C. speciosa, ‘Alba Cincta’, ‘Shirataum’, and ‘Versicolor’. In Chaenomeles the stipules are foliaceous and have palmate venation. A branch of the lateral trace of the petiole goes into each stipule. This divides rapidly into three or five bundles of uniform size. The epidermis is not as resistant as in the leaves and tears off easily. The reniform stipules have a bifacial structure with two rows of palisade cells on the adaxial side. The minor vascular bundles forming a reticulate pattern are located between the palisade cells and the spongy parenchyma, close to the abaxial surface. The bundle sheath consists of parenchyma cells, but only three or five veins possess a bundle sheath extension. The glands, stomata and trichomes are comparable to those of the leaves. Critchfield’s theory (1960) on leaf dimorphism in Populus finds appli- cation in Chaenomeles where dimorphism is expressed in the absence or presence of stipules. Critchfield found in Populus that a first type called ‘‘early leaves” laid down in the leaf buds in the fall and existing during the winter as embryonic leaves will be produced on short shoots and possibly at the base of long ones. The second type which spent the winter as leaf primordia will appear on adventitious shoots and upper parts of long shoots of the crown. He observed three types of shoots, one with only early leaves, corresponding to short shoots; the second with early and late leaves corresponding to long shoots; and a third type with only late leaves, found in adventitious shoots. The same situation is encountered in Japanese Quinces. If there is an interruption between the development of short shoots and the start of the long shoots, the number of leaves (4 to 8) observed as embryonic in the bud will all be found fasciculate at the same level, and without stipules. 180 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV The production of stipules will start abruptly with the first alternate leaves of the young long shoot. When the long shoot develops without interruption, the first leaves, exstipulate, are alternate along the shoot. It is only when the supply of leaves (maximum 8) which spent the winter as embryonic leaves in the buds has been exhausted, that the leaves with stipules start to be produced. The production of all exstipulate leaves in a fascicle, or their successive production along an axis, is certainly dependent on the weather. It is not known why stipules are produced only in long shoots, but this may be observed in Pyrus, Malus, and Cydonia which have less showy stipules than Chaenomeles, and also in Platanus, a member of the same order, the Rosales. Stipules are also always present in long shoots develop- ing from adventitious buds, and in seedlings In C. cathayensis the juvenile leaves differ markedly from the leaves produced on adult shrubs. The juvenile leaves and their stipules are smaller, more sharply serrate, always glabrous, and so narrow as to look nearly linear. Their length can be as much as seven times greater than their width, compared to three to four times in the leaves of adult shrubs. It is not known how many years juvenile leaves persist in a plant which originated from seed. A shrub of C. cathayensis which germinated in 1936 at the Arnold Arboretum, and has been repeatedly killed back by frost every year, is still producing juvenile leaves on the sprouts it sends up each spring. In contrast, mature leaves can be obtained by grafting a scion of an old plant on a young understock. In C. cathayensis, mature or senile leaves often appear subentire and show a thick tomentum on the under surface. The juvenile leaves of C. japonica and C. speciosa do not differ markedly in shape from their mature leaves, but are only slightly smaller. In C. japonica, the teeth of the crenate leaves are evident at a later stage than the fine serrations of the other two species. It is probably a case of allometric growth, since crenate and serrate leaves look alike during the embryonic stage in a bud or in a young seedling. The foliage, as recorded in the last three years, appears during the last week of April, or more often the first week of May, in the climate of Boston. The first flush of growth during the blooming period produces fasciculate leaves on short shoots. These are nearly always green. By the middle of May to the middle of June the long shoots elongate. Their leaves are very often brightly colored pink or brown from an abundance of anthocyanin pigments. The coloration extends from the midrib toward the extremities of the veins, and the chlorophyll appears as green patches isolated in the reticulate red pattern. One cultivar at least, C. & superba ‘Foliis Rubris’ has been named for its red leaves. In spite of its name, the foliage coloration does not last any longer than in other cultivars. The cultivar which probably retains the red pigmentation longest is C. speciosa ‘Kermesina Semiplena’. By the middle of July the foliage of all the Japanese Quinces appears glossy green and remains so for the rest of the season, except at the growing tips of the shoots. 1964 | WEBER, CHAENOMELES 181 Leaves of C. japonica and C. & superba are often a yellow-green, while C. speciosa, C. cathayensis and the other hybrid groups usually possess leaves dark green on the upper surface and lighter below. Two cultivars have been named for their variegated foliage, C. speciosa ‘Foliis Variegatis’ and C. japonica ‘Tricolor’. Chaenomeles japonica ‘Tricolor’ which is still in cultivation shows leaves of three colors, pink, green, and white, on the new long shoots. Mature leaves are green or variegated green-and-white. An albino mutation was observed by Simirenko (1888) on a branch of C. speciosa in which all the leaves were white, without chlorophyll. Simirenko named it ‘Simirenkiana’ and grafted it on normal understock in order to produce Japanese Quinces with white leaves. On this branch the flowers were also of a lighter shade than on the rest of the shrub. The fate of this is unknown. In areas with dry hot summers Japanese Quinces may lose some of their leaves in midsummer. Some nurserymen suggest that by removing the fruits, or growing cultivars with weak fruit production, defoliation will be prevented. This fact is contested by other horticulturists, and has not been proved one way or another. Chlorosis has been observed on CAaenomeles growing in lime soils. In the autumn the leaves fall gradually, being mostly gone by November 10. There is no fall coloration of any kind, unlike most species of Cydonia, Pyrus, and Malus which color brightly in red, or at least tint with yellow, before falling. In warmer climates Chaenomeles is reported to have nearly persistent leaves. As observed, the leaves, especially when young, are less resistant to cold than the flowers. With a late frost in the spring leaves can be damaged, appearing crinkled when unfolding, although the flowers are not touched. Inflorescences. Along with the tendency to flower at any time of the year, weather permitting, Chaenomeles presents a great variety of inflor- escence types. A few cultivars have been named for their long peduncles, such as C. speciosa ‘Pedunculata’. These pedunculate inflorescences appear on any shrub if heat, light, and humidity are sufficient. They are not a varietal character, but rather a seasonal phase. In Chaenomeles, the inflorescences vary from one to six subsessile flowers appearing together at one point, to a very elaborate raceme or panicle, sometimes including leaves, stipules, and bracts. A correlation with the short and long shoots may explain what kinds of responses are brought about by the environmental conditions. The different inflores- cences, modifications of cluster, spike, raceme or panicle, which can be found on any one shrub are of two main types, one being analogous to a short shoot, the other to a long shoot. An intermediate type is sometimes observed. The short shoot type is found in the spring and in the late fall (Fic. 5 a; FIG. 6 a, 6e). Its flowers have spent months in an advanced embryonic state in an open bud. They are clustered, (1) 2 to 6, on old wood or on the long shoots of the preceding year. The flowers have extremely short 182 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Fic. 5. Inflorescence types (natural size). a. Spring inflorescence, with short pedicels, of C. X californica ‘Arthur Colby’ (May 4); b. fall inflorescence of C. speciosa ‘Phylis Moore’ (October 2); c. summer inflorescence of C. speciosa ‘Phylis Moore’ (May 28). or non-existent peduncles, and no bracts. This is the normal type of in- florescence at blooming time in early spring, before the leaves appear. If branches are lying on the ground partially buried in mulch, the flower buds will develop in the same manner, but the peduncles will expand in order to reach the surface. In any case, each flower peduncle will be distinct from the others. 1964] WEBER, CHAENOMELES 183 Flower buds are formed very early. If the cold weather does not set in, flowers which were prepared for the following spring start to open in late fall. It is difficult to tell whether the inflorescence type consisting of a few sessile flowers, fasciculate at one node, is brought about by a return of the early spring conditions of temperature, humidity, short days, and absence of leaves, or whether it depends, as in short leafy shoots, on the advanced embryonic condition of the flowers in the buds. An intermediate situation between short and long shoots has been ob- served several times in living plants as well as on herbarium specimens. The bud produces a short shoot during the spring like any other short shoot. It is continued, sometimes without interruption of growth, by a slightly branched cluster of three to five pedunculate flowers. No leaves, stipules or bracts are produced and the whole inflorescence is not more than two to five centimeters long. Normally, a short shoot does not elongate and forms a bud at the end, often destined to produce flowers the next spring. For an unknown reason these flowers develop the same sea- son. This intermediate condition may be found in late spring and in early fall (Fic. 6 b). The long shoot type of inflorescence occurs from late spring to late fall. So far, it has not been possible to predict where and when summer in- florescences are going to appear. At this time, the shrubs are covered with leaves. Summer inflorescences, like sprouts, develop very fast if the temperature is favorable. When they can be recognized, it is too late to find out in what kind of embryonic condition they spent the winter. This can be deduced by their location, often on old wood, even on the main branches, and also by the fact that at this time of the year there are only rudimentary dormant buds. Summer inflorescences, like the long leafy shoots, probably spend the winter in a primordial stage. The peduncle length appears to be dependent on the temperature, a few degrees change at night probably being enough for change from a spike to a raceme or a panicle. The variation in long shoot type of inflorescence is even more striking, involving the presence or absence of leaves, stipules, bracts, or intermediate organs, as well as the transformation of the sepals into leaves. This type of inflorescence is extremely variable and may be subdivided into four categories. 1. A few exstipulate leaves are found at the base or along the axis. They are succeeded by leaves with small stipules. Axillary to each leaf is a flower or a ramified branchlet with two or more flowers associated with bracts. The shoot is terminated by a flower. Unbranched, it may be a spike or a raceme, when branched, a panicle. It may reach a length of 20 centimeters in a few days. This is the commonest inflorescence type in the summer and occurs from late spring to fall (Fic. 5 c) In a variant of the preceding the leaves, usually not more than six. lack stipules. The flowers may occur in the axils of the leaves or above them. Bracts are usually absent. This type is common in the fall. The inflorescence may reach 20 centimeters or more (Fic. 5 b 3. A third and infrequent type of inflorescence may be formed in the fall. A 184 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV nell yi ih _. | v 2: RANG p as e Fic. 6. Inflorescence types (natural size). a. Spring inflorescence, with long pedicels, of C. x superba ‘Perfecta’ (May 10); b. intermediate inflorescence of C. speciosa ‘Simonii’ (June 25); c. fall inflorescence, with heart-shaped bracts and foliaceous sepals, of C. japonica ‘Sargentii’ (November 10); d. inflorescence appearing through the ground, of C. japonica ‘Sargentii’ (June 6); e. late fall solitary flower of C. speciosa ‘Simonii’ (November 17). few exstipulate leaves are laid down at the base. The whole shoot ramifies and heart-shaped, serrate organs, intermediate between leaf and stipule, are found in place of the usual inconspicuous bracts. This modification is carried out to the 1964 | WEBER, CHAENOMELES 185 calyx which becomes foliaceous. These leafy bracts and sepals bear glands on each tooth. It is not known if they ever secrete any liquid (Fic. 6 c). 4. In two instances, in summer and early fall, inflorescences were found com- ing up through the ground. One was formed on a very old shrub of C. japonica ‘Sargentii’, and the other, on a young plant of C. X superba ‘Elly Mossel’. The shoots elongated through the ground, without bearing any leaves. Once in the open they developed bracts, some normal, others heart-shaped, then flowers with foliaceous glandular sepals (Fic. 6 d). Sometimes there are other abnormalities accompanying foliaceous bracts and sepals such as glands appearing away from the edges (on the adaxial side only, and always connected with veins), stamens transformed into petals or sterile styles with extra branches, and so forth. At least some of these different types of inflorescences are common and usually appear at the same season year after year, probably as a response to similar environmental conditions. They cannot be associated with any particular species or cultivar. On one shrub of C. japonica ‘Sargentii’ which originated from seeds brought back by Sargent from the mountains of Japan, six of the seven types and subtypes have been observed during the last two years. Summer inflorescences are so frequent that not only cultivars, but also botanical species or varieties have been named for these seasonal phases. The type of C. speciosa (Sweet) Nakai, an illustration, plate 692 in the Botanical Magazine, represents an inflorescence interme- diate between a short and a long shoot. The syntypes of C. japonica var. pygmaea Maximowicz are specimens, collected in flowering condition in December, of the short shoot type with long peduncles, appearing excep- tionally in late fall. Other characteristics which vary according to the temperature were included in the original description. The different types of inflorescences found in Chaenomeles at different times of the year encompass the range of inflorescences found in the closely related genera. Since these genera have a limited blooming period, only one type of inflorescence occurs which is considered characteristic of the genus. Docynia possesses a few clustered subsessile flowers. They are analogous to the spring inflorescences of Chaenomeles, and comparable to those of short shoots. Malus has the short shoot type of inflorescence with long peduncles. Pyrus, also some species of Malus, present the inter- mediate situation with fasciculate leaves and a slightly branched, but short and naked inflorescence in the center. This has been called “pyrostele” by Bailey (1949a) who thought that this type was peculiar to Pyrus. This term is unfortunate for the confusion it can cause with the anatomical meaning given to the term stele. Cydonia possesses solitary flowers occur- ring at the tips of leafy shoots. This inflorescence is a typical modification of a long shoot, which in Chaenomeles ends up with a spine, or is modified into a branched inflorescence. The different types found in these four genera develop at the same time as the corresponding types in Chaenome- les. Consequently, they must not be considered as genetically determined, but rather as arising in physiological response to the environment. 186 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLV Flowers. The variation in inflorescences found on any one shrub is matched by an equal or greater variation found in the flowers. Normally, flowers are produced from special buds, early in the fall. In the spring, flowers open within a few days in good weather. They usually open before the leaves, but the leaves develop and start to unfold when the petals are still present. The aestivation of the sepals and petals is imbricate, as in Pyrus and Malus, and in contrast to the contorted condition in Cydonta. The hypanthium forming the base of the flower varies from cup-shaped to gourd-shaped. There is a tendency for the hypanthium of C. japonica to be more nearly cup-shaped; for C. speciosa, more gourd-shaped; and for C. cathayensis, to have an elongated hypanthium. Moreover, the shape of the hypanthium is complicated by the presence of numerous unisexual flowers in every species, which are more nearly cup-shaped when entirely male, more nearly gourd-shaped if entirely female. This does not always hold true, for flowers with long hypanthia sometimes have no ovules, and flowers with short hypanthia may have ovules which appear superficially to be borne in the peduncle. The receptacle, formed of the hypanthium and prolonged by the 5 sepals is always glabrous. It is woolly or at least pubescent in Cydonia, Docynia, and Pyrus. In Chaenomeles the recep- tacle is often brightly colored purple or brown on the sunny side, by anthocyanin pigments. The sepals appear as five short truncate or round, ciliate lobes, without glands, but pubescent inside at the place where the petals are attached. The lobes, often unequal in size, are always erect at anthesis. In contrast, the sepals are reflexed in Cydonia and Pyrus, and are glandular in Cydonia and Docynia. The calyx lobes, as well as the hypanthium of which they are a prolongation, are very much vascularized by anastomosing traces. The sepals are ciliate with single unicellular trichomes. Sepals can show two abnormalities. They may either become leaf-like, a relatively frequent condition, or more rarely, petal-like. Foliaceous sepals occur normally on at least two types of inflorescences, and in this case, they are serrate and glabrous on the edge, each serration being terminated by a gland. They are pubescent inside, as usual. If only par- tially leafy, the entire edge will be ciliate, and the trichomes are abruptly replaced by glands where the serrations begin. A partial development of petal-like sepals is sometimes found. This latter case is of common oc- currence in C. * superba ‘Texas Scarlet’. The genus Chaenomeles normally has five regular, free, slightly ungui- culate round petals. They are borne on the hypanthium, opposite the sepals. They may be carinate, especially in C. japonica and C. « superba, undulate, or flat at the end of anthesis. The petals are short pubescent on the claw, glabrous otherwise, and have exceptionally been found to be ciliate and bordered with small capitate glands in C. & superba ‘Crimson and Gold’. The petals possess slightly larger cells than the sepals, and are less vascularized. At least three traces enter each petal. They divide dichotomously and the veins anastomose to form a reticulate pattern often reaching the edge of the petal. 1964 | WEBER, CHAENOMELES 187 Petals are never less than five, but when broad and overlapping the flowers appear semidouble, even with a regular number. Often, there are more petals from a partial modification of stamens and the number may reach 15 which are quite variable in size and shape. Numerous cultivars with semi-double or double flowers have been named. This character may vary widely, flowers being double on a young shrub, or when the number of flower buds has been reduced by frost, single when the shrub is older and loaded with flowers. Isolated summer flowers often have a greater number of petals than spring flowers. The genera Cydonia and Chaenomeles are cited as having petals which drop before wilting. They do so after four or six days, even before the outer stamens have shed their pollen, if the weather is hot. A longitudinal section of the flower reveals an abscission layer formed between the hypanthium and the petals. On both sides of this constriction the cells are smaller and more tightly packed. The abscission of the petals which are shed when perfectly turgid occurs at this constriction. The separation seems to start from the center and proceed toward the outside. Since stamens do not possess an abscission layer, petals on double flowers are not shed, and can often be seen to persist on young fruits. The double flowers last considerably longer than the single ones. The petals of Chaenomeles cultivars can show a great array of colors, from white to darkest red, through all shades of pink, orange, and scarlet. Bicolored petals are also frequent. The color intensity and petal size are in direct relation to light, and probably to heat. In rainy spring sea- sons, flowers appear pale and dull. This relationship is also obvious from an observation of the flowers borne on one shrub from early spring to late fall. The color becomes more intense until August and September, revert- ing later to lighter shades. Very few cultivars are pure white during the summer months. The effect of light may also be observed by comparison of flowers opening at the same time on the outside and inside of a leafy shrub, or on double flowers in which the parts where the petals overlap remain uncolored. Most cultivars have been selected and named on the basis of petal color. A classification of cultivars according to their color can only be approxi- mate, due to the changes in intensity throughout the year. In addition to the weather, the soil may have an influence. Nakai (1916) is the only botanist who attempted to distinguish species by the color of the flowers alone. He raised C. speciosa var. eburnea to the rank of species and placed in it all the cultivated varieties with white petals. As a reverse case from petaloid sepals, sepaloid petals may also be found. C. speciosa ‘Simonii’, with semidouble flowers, has outer petals often marked with a green line, the red colored parts appearing as two wings. The sepals of this cultivar are normal. The stamens, 40 to 60 in number, are borne on the hypanthium below the petals, more or less in two rows. They are free and shorter than the petals. The filaments are glabrous, white or pink or dark rosy-red when developing in summer conditions. The anthers are introrse, basifixed at 188 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV first, becoming slightly versatile when old. They are yellow, turning to brown after the pollen has been shed. The extra petals in semidouble and double flowers are modified stamens which may retain one or more anther sacs on the edge. All transitional stages between stamen and petal can be found. There is no constriction at the base of the stamens, rather a slight en- largement. The tissue of the hypanthium with its large cells arranged vertically continues without interruption into the filament up to the anther. There is one trace only entering the filament. The few protoxylem ele- ments which form the vascular bundle show reinforcement by spiral thickenings. The center of the vascular bundle is often hollow. The vascular bundle stops at the base of the anther sacs, but diffused proto- xylem elements may be found in the connective. The epidermis of the filaments, formed of very small cells, is cutinized. Stomata have been observed, but there are no trichomes present. Under the epidermis is another row of slightly larger cells around the ground tissue. The anthers, in an early stage, show two lobes and four locules. Close to maturity the anthers have two locules only, by loss of the partition within each anther half. Beneath the epidermis is the endothecium which consists of large cells, tangentially arranged, their walls, except the one in contact with the epidermis, being reinforced by conspicuous secondary wall thickening strips oriented perpendicularly to the epidermal layer. There is an interruption of the epidermis and of the endothecium at the place where dehiscence will occur. The stomium is a longitudinal slit located between the locules of each half of the anthers. Before opening it is plugged by a chain of very small thin-walled cells. Just before dehiscence starts, the parietal layers inside the endothecium and the tapetum disintegrate and appear as collapsed or crushed cells. In the double flowers, some of the stamens are petaloid. In longitudinal section, it may be seen that in spite of their change of function and color, they have a single vascular trace, and lack trichomes completely at the base. This sole trace divides much higher up in the claw and eventually produces a normal reticulate vein pattern. A fragment of anther is often retained, with or without pollen production. There is neither constriction at the base, nor small cells arranged tangentially to interrupt the tissue of the hypanthium. This explains why “‘petals’’ of double flowers do not shed as in single flowers. The outer or normal petals of double flowers do possess a constriction and are shed unnoticed, while petaloid stamens persist until the abscission of the top of the hypanthium occurs a few weeks after pollination. Functionally female flowers are formed by the abortion of stamens and are less common than functionally male flowers. Although they are not recorded in the horticultural literature of Chaenomeles, functionally fe- male flowers occur occasionally in many shrubs, or very frequently in such cultivars as C. & superba ‘Alba Semiplena’, ‘Columbia’, ‘Early Apple Blossom’, and ‘Sunset’. Two shrubs of ‘Columbia’, grown side by side, have been observed for a few years to produce female flowers only. In the 1964 | WEBER, CHAENOMELES 189 spring of 1963, they both produced a few male flowers and no female ones. In these female flowers, stamens in reduced number were present but did not contain any pollen and often i sn sacs were transformed into pink appendages in the shape of hor The pollen grains of see ae are tricolpate, without ornamenta- tion; re three germ pores are evident at the rounded corners. The exine between the pores is thick and smooth. In herbarium specimens of wild and cultivated plants, the pollen grains measure from 25 to 55 micra. There is a great variation in size of pollen grains coming from the same anther. The three species and their hybrids have been found to have pollen grains with the same range in size. The greatest variation in pollen size has been observed in a specimen of C. japonica collected in Japan. The grains measured from 15 to 32 micra. The pollen of the cultivars representing the interspecific hybrid groups has been found to be on the lowest part of the range, while the pollen of C. californica ‘Clarke's Giant Red’, a tetraploid, was in the upper part of the same range. These measurements do not seem to be significant since the variation is so great in the pollen contained in one anther, and is no larger in cultivated plants or in hybrids than in wild plants. It is impossible to recognize the species or the genus by the pollen. The two species of Cydonia have similar pollen, also within the size range of Chaenomeles. The proportion of bad or empty pollen grains which do not stain with aceto-carmine can be from 4, to 4 of the total. The proportion of empty pollen has not been found to be significantly different in hybrids and in species, or in cultivated specimens and in wild specimens. It varies a great deal on the same shrub at different dates, and seems to be dependent on weather conditions. Between the inner row of stamens and the column formed by the fused styles, is a glabrous, nectariferous interval called a disc. The disc is larger in male flowers when the column formed by the styles is reduced in size or absent, than in predominantly female flowers. The disc covers the part of the carpels fused to the hypanthium and is the place where nectar is excreted. The epidermis of the nectary appears no different from that of very small round cells, closely packed and staining differently from the ground tissue. When the cell content persists, it shows numerous small refringent granules. The floral nectaries differ anatomically from the leaf glands in lacking vascular bundles. The column formed by the styles is characteristic of Chaenomeles, al- though it may vary widely within one species. The styles may be com- pletely fused, in which case the column is solid, or partially fused, forming a hollow column. At the base are some straight unicellular trichomes which may vary from abundant to entirely absent. Trichomes are occa- sionally found within the hollow columns. Each style is vascularized by two to four traces. The ground tissue consists of large cells longitudinally arranged, and bordered by an epidermis. The styles are always free above. 190 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV The decurrent stigmata have a glandular papillate epidermis. The stigma- toid tissue of large and longitudinally arranged cells forms anastomosing strands connecting the stigmata to the ovules. These cells stain in the same way as the nectariferous cells, and also possess refringent granules. In longitudinal sections, the stigmatoid tissue may be observed very close to the reticulum formed by the nectariferous tissue. The two kinds of cells may be distinguished by their size. The five styles in Chaenomeles are fused for 44 to % of their length. This is a generic character not present in the related genera which have free styles coalescent by the hairy covering. The column is commonl glabrous or glabrescent, less often pubescent, rarely woolly. It is always glabrous in C. japonica, pubescent to woolly in C. cathayensis, and varies from glabrous to woolly in C. speciosa, with the greatest number of culti- vars being glabrous or glabrescent. Japanese authors put much emphasis on this character and still recognize Nakai’s (1916, 1918, 1923, 1929) four “species,” separated according to the pubescence of the stylar column: C. eburnea, glabrous; C. speciosa, glabrescent; C. trichogyna, villose; C. extus-coccinea, woolly. Nakai himself (1929), gave as synonym of C. extus-coccinea with a woolly column, the cultivar ‘Alba Cincta’ (as C. japonica var. fl. roseo, albo-cincta) which has a glabrous column. Such variation in the pubescence of the column is not correlated with other morphological characters. The part where the styles are free is always glabrous. The styles are white, exceptionally red (observed in summer flowers only). They terminate in green decurrent stigmata, comparable to those of Cydonia and Docynia, but contrasting with the discoid stigmata of Pyrus and Malus. Many aberrations may take place in the styles. Often more than five styles, and up to nine, have been observed in the same flower. They usually correspond in number to the carpels, but extra styles may also occur by branching just under the stigmata. These extra branches are often found in flowers appearing in the fall. In some cultivated shrubs of C. speciosa, the styles are abnormal year after year. In a few flowers, they are fused into a hairy cone protruding above the petals. These flowers appear to have a superior ovary, in contrast to the normal inferior ovary of the Maloideae In other and common instances styles may vary from normal size, about '% longer than the stamens, to being completely absent. Every transition has been observed in the flowers of a single shrub, which can produce normal hermaphrodite flowers with long styles, others with styles of the same length as the stamens, some shorter, as well as flowers which are completely male by abortion of the female organs. Flowers are found with ovules and no styles or with normal styles and stigmata, and no ovules. In contrast to functionally female flowers which always contain sterile stamens, male flowers usually do not have styles or ovules. Contorted styles of normal size, but without stigmatic surface are extremely rare. Male flowers are very common, even in wild specimens. The syntypes of Ch.ienomeles japonica var. pygmaea Maxim. show male flowers only. So 1964 | WEBER, CHAENOMELES 191 does plate 692 of the Botanical Magazine which is the type of C. speciosa (Sweet) Nakai. Under the disc are found the five carpels. They are completely fused to the hypanthium on the abaxial side, and fused together on the adaxial side. They form an inferior syncarpous ovary with five locules, but num- bers from three to seven have been observed. Carriére (1876b) found a fruit in which the five carpels gave rise to ten locules by the production of false partitions perpendicular to the normal ones. Exceptionally, carpels are incompletely fused in the center (as in Cydonia and Docynia). transverse section through the ovarian region shows a cutinized epidermis formed by small cells without trichomes. Directly beneath the epidermis are two to three radially oriented rows of collenchymatous tissue of larger tightly packed cells. The ground tissue forming the bulk of the hypanthium and of the carpels, consists of large cells loosely arranged. Immediately internal to the collenchymatous tissue are numerous small vascular bundles forming an irregular circle. Near the inner side of the hypanthium are ten vascular bundles which vascularize the sepals and the petals. At the tip of each of the five carpels is another vascular bundle. Between the two circles is the boundary of the ovary and the hypanthium. These two organs are completely fused in Chaenomeles and are anatomi- cally indistinguishable. The ground tissue of each is a loose parenchyma. At the time of anthesis there are no sclereids in either of them. However, only the ovary tissue has the strands of stigmatoid tissue. Each locule contains about 20 anatropous ovules with two integuments, arranged hori- zontally in two rows. Fruits. The fruits of Chaenomeles are pomes as in the other Maloideae. There are two conflicting theories as to the interpretation of the inferior ovary in the Maloideae. The usually adopted one is the Axial Theory of the nature of the inferior ovary, also called the Receptacular Theory. In this theory, the hypanthium is considered basically axial, consisting, ac- cording to Eames (1961) of the rim of the receptacle. Under the Appen- dicular Theory, however, the hypanthium is considered to consist of the adnate bases of sepals, petals and stamens. Most authors commonly cite the Maloideae as showing inferior ovaries in which the receptacle has a prominent part. Eames, through comparative studies of the floral anatomy of the Rosaceae, and comparisons of proliferated rose and apple fruits, arrives at the opposite conclusion. For Eames, the fleshy outer part of the apple and the pear consists morphologically of the fused bases of appendages. The fruits of Chaenomeles in which the calyx lobes often persist and become accrescent, forming the upper part of the fruits, seem to accord with the Appendicular Theory. The transformation of the ovary and the hypanthium into a pome occurs in the same manner as in the apple, which has been studied by several authors. The ripe fruit is brightly colored yellow by chromoplasts. The cuticle on the outer side has thickened; it may become sticky in some cultivars. The stomata have been transformed into lenticels consist- 192 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV ing of patches of suberized cells. In some fruits, patches of suberized cells occur also at random on the skin, especially near the apex. This corresponds to a phenomenon called ‘“‘russeting” in apples where the outer layers of the fruits are replaced by cork (Tetley, 1930). The subepidermal parenchyma consists of several layers of tangentially elongated cells, which contain carotenoids. These cells were radially oriented in a cross section of the flower. They have divided and enlarged and by a difference in allometric growth, have changed their orientation. Their walls also have thickened, and they are as tightly packed as earlier, in contrast to an apple whose tissues become relatively soft at maturity. The fragrance as well as the coloration of the fruit appear to be localized in the epidermis and in the subepidermal parenchyma. The ground tissue is of loosely arranged parenchyma, well vascularized by a network formed by the anastomoses of the main traces. In a young fruit, the parenchyma cells contain chloro- and chromoplasts. Chloroplasts degenerate at ripening time, and carotenoids are present, but in lesser quantity than in the epidermis and subepidermal layer. In spite of the fact that the fruits of Chaenomeles do not soften in ripening, there are many intercellular spaces between the large more or less radially oriented cells of the parenchyma. The fleshy part of the fruit which corresponds to the hypanthium tissue is the exocarp. In a young ovary, the limit between the hypanthium and the carpels was undiscernable between the circle of ten traces going to the sepals and petals and the circle of five traces on the dorsal side of the carpels. In a mature fruit, patches of sclereids have developed around the median and lateral vascular bundles at the limit of the floral tube and the carpels. These patches become confluent and form an irregular ring of sclereids. There are no sclereids in the flesh of Chaenomeles nor are there any in Malus. In contrast, there are patches of sclereids or stone cells in the ground parenchyma of the fruit of Pyrus, Cydonia, and Docynia. Inside of the circle of sclereids are the carpels which became cartilaginous early in the development of the fruit. They also are strongly vascularized and possess sclereids. This is considered to be the endocarp lining the locules (McDaniels, 1940). In Chaenomeles, the carpels are usually completely fused in the center. The fruits present a great variation in their shape. More commonly, they are apple-, orange-shaped, or ovoid, when coming from spring flowers; pear- or fig-shaped, when formed on summer or fall inflorescences. Morren (1851) has also observed fruits of C. speciosa in the shape of a falciform and ribbed zucchini. As a rule, fruits of C. japonica are small and apple-shaped; fruits of C. cathayensis are large and ovoid; fruits of C. speciosa and of the interspecific hybrids vary in shape and size. A few cultivars (e.g. ‘Citri-pomma’, lemon-shaped; ‘Pyriformis’, pear-shaped) have been named for the shape of their fruits. In continental Europe, summer flowers on Chaenomeles are infrequent, and have attracted the attention of botanists when they appeared. In New England, where they are of common occurrence, pear-shaped or fig-shaped pedunculate fruits 1964] WEBER, CHAENOMELES 193 are found every year on many shrubs. Instead of appearing isolated on old wood, they are clustered on a ramified young shoot. They usually sara fewer ovules than spring fruits, and never mature, being killed by fro feieeadeal of their shape, fruits are smooth or ribbed in a few culti- vars. Usually there are ten ribs in C. speciosa bodied Rosea’, ‘Kermesina Semiplena’, ‘Phylis Moore’, ‘Simonii’, ‘Spitfire’, a : x superba ‘Knap Hill Scarlet’; and only five ribs in C. X ee ‘Ulidia’. This character of having ribs is constant in a cultivar and occurs in fruits formed from spring flowers as well as in fruits formed in the summer and fall. The fruits of Chaenomeles have another peculiarity not found in the closely related genera. They very often terminate in a prominent umbilicus. Other times the calyx becomes accrescent and protrudes as a fleshy rim above the fruit, or it dries up and may persist with the stamens (as in Malus, Pyrus, and Cydonia). The three different kinds of apices may occasionally be found in the fruits of the same shrub. However, the fruits usually offer good characteristics for recognizing some of the cultivars. As the flowers of Chaenomeles are similar in all three species, it is inter- esting to determine the development of these different shapes of fruits. A swelling of the ovary is evident 10 to 12 days after anthesis. When cut open longitudinally it may be observed that the center where the carpels are fused has increased considerably, and that the ovules have started to enlarge, while the hypanthium is of about the same width as in the flower. At this-time there is a natural thinning of the fruits. The flowers which are not pollinated, and certainly many others, fall off. In these, the ovules, instead of increasing, have by now dried up. Soon after, the calyx is shed in one piece by an abscission produced below the inser- tion of the stamens at the constriction occurring in a more or less gourd- shaped hypanthium. The styles are present and the separated calyx is sometimes retained for a while as a loose ring surmounting the young fruit. The fruit continues to increase in girth, more by enlargement of the center and of the locules than by a thickening of the flesh, which will take place after the locules are nearly full size. The shape of the apex of the fruits of Japanese Quinces is determined in the first few weeks after pollination, by the abscission or persistence of the calyx, and of the stylar column. The place where the abscission occurs on the column determines the presence or not of an umbilicus. If nothing is left of the column, and if the calyx was shed, the fruit will be terminated by a depression; or if the calyx persists and becomes accres- cent, by a hollow protuberance. If a few millimeters of the column re- main in the fruit this narrow base will enlarge and form an umbilicus. It may remain sunken in the fruit, be hidden under the accrescent calyx, or be evident by protruding in the cavity left in the hypanthium after its transformation into a pome. The umbilicus is usually hard and fibrous. The cultivar C. speciosa ‘Umbilicata’ has been named for having an umbilicate fruit. Fruits with a very large umbilicus are also found in 194 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV C. speciosa ‘Fireball’ and ‘Doctor Bang’s Pink’, and in C. % superba ‘Incendie’. Umbilicate fruits are common in other cultivars, but the umbilicus is not apparent unless the fruit is sectioned longitudinally. Since the column formed by the fused styles is a characteristic of the genus Chaenomeles, it is not surprising that umbilicate fruits do not occur in other Maloideae. Occasionally, a shrub shows fruits of different shapes, a condition first observed by Morren (1861). Irregular or malformed fruits may be caused by abortion of the ovules on one side. Ovules after pollination must produce auxins which govern the growth of the flesh. The grooves always correspond to partially or totally empty locules. The first part of the pome to develop is the core containing the carpels and the ovules. For flowers which were pollinated during the first week of May the core and the ovules have reached approximately their mature size by the beginning of August. Fruits formed in the summer develop much faster. When the core has become cartilaginous the flesh will start to increase. The stage of cell division described by Clements (1935) is replaced by a stage in which the cells enlarge slowly up to the maturity of the fruit. The skin of CAaenomeles fruit is always very thin and completely glabrous. Some of the interspecific hybrids of C. japonica possess char- acteristically sticky skin, which is also a sign of maturity. The skin remains dry in the other two species. The color of the ripe fruits is yellow. Sometimes the showiness of the fruits is the main merit of a cultivar, as in C. < superba ‘George Landis’ with orange fruits. This stage of ripening is not reached by all cultivars in the climate of Boston, for the earlier ones mature usually in October. Fruits of C. cathayensis need more heat to show color. If fruits do not ripen they remain green, but often are pink or brown on the sunny side. The ripening is also evident in the fragrance of the fruits. The mature seeds are brown. Chaenomeles fruits often fall off all at once, especially after a white frost. The fruits may be ripened indoors and formerly, were kept in closets for their pleasant odor. They remain hard and acid until they decompose. Spring inflorescences contain as many as six subsessile flowers clustered at one point. By natural thinning the number of fruits is usually reduced, but as many as four mature fruits have been observed at one node. The peduncles did not elongate, and they developed a special shape due to the crowding. In C. *& superba ‘Boule de Feu’ where the peduncles are com- pletely absent, the fruits often surround the supporting spiny twig. This abundance of sessile fruits in the fall is a special aspect of Chaenomeles shrubs. The fruits appear on old wood to the ground level. Fruit produc- tion is sometimes very heavy, especially in C. cathayensis and in C. japonica ‘Maulei’. Chaenomeles japonica is usually the species maturing first, which may be due to its low stature, as is its hardiness. Seeds. The numerous seeds are arranged horizontally in two rows, in each of the five locules. There may be up to 120 in C. cathayensis, about 80 in C. japonica, and an intermediate number in C. speciosa if the ovules 1964 | WEBER, CHAENOMELES 195 are all fertilized and develop. This is usually not the case, and seeds in full sized fruits number not more than 70 to 80. There appear to be fewer seeds in fruits of interspecific hybrids. The seeds of C. japonica and C. speciosa are ovoid and pointed at one end, while those of C. cathayensis are wedge-shaped. The testa is coria- ceous, dark or reddish brown in the first two species; coriaceous or very slightly mucilaginous, light brown and dull in C. cathayensis. In contrast, seeds of both species of Cydonia are highly mucilaginous. The raphe and chalaza are apical. The endosperm is lacking. Seeds help in recognizing interspecific hybrids, especially those involving C. cathayensis. The seeds of C. japonica and C. speciosa, in spite of being very numerous in each locule, are more or less ovoid (as in Pyrus and Malus). In C. cathayensis, they take a planoconvex shape by compression with each other (as in Cydonia). Seedlings. In Chaenomeles the percentage of seed germination is high, and does not appear to be decreased in interspecific hybrids. After a cold period the germination of the seeds takes place immediately. The root ramifies very fast. In five or six days the seedlings, still surmounted by the split testa, begin to emerge above the ground. After shedding the seed coat, the two, rarely three, cotyledons expand and the plumule becomes visible. The young leaves, as well as their stipules, are immediately serrate (also in Cydonia oblonga), each serration and interval being marked by a reddish gland, secreting liquid. The glands are sometimes found at a distance of a few millimeters from the edge, on the upper surface of the leaves. The hypocotyl is glabrous while the epicotyl is pubescent. One or two weeks after germination the seedlings of different species and hybrid groups are very similar. There are, however, a few specific characteristics. C. cathayensis and C. speciosa appear dull dark green, while C. japonica is a shiny yellow-green. The leaves of C. japonica and of C. cathayensis are glabrous, while C. speciosa shows a few trichomes on the midribs of the under surface. The leaves of C. cathayensis are more finely and sharply serrate, a characteristic also evident on the young stipules which are more leaf-like on the first few leaves and become reni- form on later leaves. Seedlings of the three species of Chaenomeles often show such ab- normalities as lobed leaves, cotyledons colored with anthocyanins, or albino leaves. Relatively often, when still in the fruit, seeds rupture and the testa and the cotyledons start to swell, and these seeds continue to grow without requiring a cold period. The seedlings obtained do not show juvenile leaves, but form a rosette of mature looking leaves. However, they do not elongate further without being submitted to a cold period. Influence of the environment. The genus CAaenomeles, as shown by its morphology, is extremely plastic. Nearly all organs may present abnormalities, and some of them are so common that they can scarcely be 196 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV called aberrations. The plant may be a shrub, or, excluding C. japonica, be trained as a small tree. The shoots are of three types, short and long shoots on the crown, and sprouts at the base. The leaves vary a great deal in size, pubescence, and serration. The inflorescences may be of eight different kinds, according to the season. The flowers show all stages from a normal hermaphrodite condition to plants which are predominantly or completely male or female. Unisexual flowers appear usually in the spring and late fall, and a change in sex is possible from year to year in the same shrub. Fruits have two basic shapes, one developed from spring flowers, another on summer inflorescences. Most of these differences can and do, in fact, appear on the same individual at different seasons. It is striking to observe two or more shrubs of the same clone react in the same manner under the same environmental conditions. For example, a change may occur at the same time in the sex of the flowers, or in the type of inflorescences on adjacent plants. A similar reaction to the climate is not limited to plants belonging to the same clone or even to the same species. Independent of their parental species or hybrid groups, all cultivars from Avery Island, Louisiana, are more spiny than is usual in the genus. In the specimens from Washington, D-C., the leaves are always smaller than on identical cultivars grown in other localities. Fruits are larger when coming from Georgia, but they show abnormally corky and prominent lenticels when grown in the Nether- lands. Shrubs in England, usually in late spring, bear the type of inflores- cence intermediate between a short and a long shoot. This type is rare elsewhere. These examples show that the genetic make-up of Chaenomeles allows the genus a large range of possible responses to the climate, and in part accounts for the selection of cultivars as ornamental plants on six continents. CYTOGENETICS Morphological and taxonomic studies show that the genus Chaenomeles comprises three species, namely C. japonica, C. speciosa, and C. cathayen- sis. These three species cross with each other in every possible way, forming three interspecific bihybrid groups: C. « superba (C. japonica « speciosa), C. clarkiana (C. cathayensis japonica), and C. x vil- moriniana (C. cathayensis X speciosa). A fourth interspecific hybrid, C. X californica (C. cathayensis < superba) is a synthesis of the three species. Due largely to confusion of nomenclature but also to the lack of her- barium vouchers, little of the published cytogenetic information on Chae- nomeles is reliable. Available information, however, does suggest the following: 1. The breeding behavior of Chaenomeles is pe as to that of its close spon = the Maloideae, Pyrus, Malus, and Cydon v, brids within Chaenomeles are formed coin under cultivation and are easily produced by artificial cross-pollination. As variations can be pro pa- 1964 | WEBER, CHAENOMELES 197 gated asexually these have been maintained in cultivation for long periods of time. 3. Although polyploids have been suspected, no chromosome counts support this assumption. There was need, therefore, to repeat the previous work for confirmation; to make additional observations in an attempt to understand or explain the multitude of cultivars described for Chaenomeles; and to determine the relationships of Chaenomeles to its related taxa. Chromosome Studies. The basic chromosome number of x = 17, common to all the Maloideae, was reported by Moffett (1931) for the three species of Chaenomeles which have 2n = 34. He found the chromo- somes of this genus to have the same morphology as those in the other genera of the Maloideae. They are small, measuring between one to three micra and, according to Moffett, possess median and submedian constric- tions. Moffett’s figures of meiosis in Chaenomeles japonica (as Cydonia maulei) and C. cathayensis show secondary pairing of the chromosomes into sexivalents, quadrivalents and bivalents. Sax (1932) thought that the apparent polyvalents might be due to Moffett’s technique of section- ing, for they do not appear in smears where, at anaphase, 17 pairs pass to each pole. In attempting to verify Moffett’s work, it was found that meiosis occurs over a period of a few days, usually at the beginning of April, when the buds are 2—3 mm. in diameter on plants grown out of doors, but there is some variation due to the temperature and the cultivar used. Forced flowers proved unsatisfactory for often the stamens fail to develop properly or dry out. The young flower buds were fixed in a solution of one part glacial acetic acid and three parts 95% ethyl alcohol, and kept in the refrigerator until used. To separate the pollen mother cells, the dissected anthers were put into HCl 1 Normal for ten minutes; deposited on a slide and dried; then squashed and mounted in a drop of aceto- orcein and heated slightly. Aceto-carmine was not a satisfactory stain. Because of the small size, the chromosomes were hard to see even in good stamens at the right stage. The preparation had to be examined with the immersion objective. The pollen cells also contained oil droplets about the same size and shape and easily confused with the chromosomes. All the species and cultivars examined were orthoploid, as in the other Maloideae, which means that they all have the same basic number. Aneuploid series have not been observed in this subfamily. The number of x = 17, and 2n = 34 was confirmed in all the species. It was possible to make counts in seventeen cultivars. As diploid with » = 17 were: C. japonica ‘Pigmani’; C. speciosa ‘Contorta’, ‘Nivalis’, ‘Simonii’ and ‘Umbili- cata’; C. superba ‘Cameo’, ‘Corallina’, ‘Glowing-Ember’, ‘Mount Shasta’, ‘Red Chief’, ‘Roxana Foster’, ‘Rowallane’, and “Texas Scarlet’; C. X californica ‘California’, ‘Flamingo’ and ‘Rosy Morn’. Only one tetraploid with n = 34 was found, C. x californica ‘Clarke’s Giant Red’, which has the largest flowers of all cultivars. 198 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV DOCUMENTATION OF CHROMOSOME COUNTS IN CHAENOMELES Diploid cultivars with n = 17. C. JAPO Pigmani Arnold Arb. No. 498-59, Weston, Mass., Weber, May 11, 1962 (AAH), C. SPECIOSA ‘Contorta’, Arnold re No. 126-42, Jamaica Plain, Mass., Weber, May 6 & May 23, 1960 (A ‘Nivalis’, ee Arb No. 13022, Jamaica Plain, Mass., Weber, May 6 & May “Simonii’, Avnold ee No. 178-40, Jamaica Plain, Mass., Weber, May 6 & May 23, 1960 (A ‘Umbilicata’, Arnold “At. No. 865-53, Weston, Mass., Weber, May 17, 1961 (AAH ) C. X SUPERBA ‘Cameo’, Arnold Arb. No. 179-58, Weston, Mass., Weber, May 31, 1961 (AAH). ‘Corallina’, — Arb. No. 198-42, Jamaica Plain, Mass., Weber, May 6, 1960 (AAH). ‘Glowing Ember’ Arnold Arb, No. 766-57, Jamaica Plain, Mass., Weber, May 14, 1961 & May 9, 1962 (Aan). ‘Mount Shasta’, Arnold Arb. No. 176-58, Weston, Mass., Weber. May 13, 1961 (AAH). ‘Red Chief’, Arnold Arb. No. 486-58, Weston, Mass., Weber. May 17, 1961 (AAH). ‘Rowallane’, Arnold Arb. No. 156-52, Weston. Mass., Weber, May 9, 1962 (AAH ) ‘Roxana Foster’, Arnold Arb. No. 483-58, Weston, Mass., Weber, May 9, 1962 AAH ) ‘Texas Scarlet Arnold Arb. No. 488-58, Weston, Mass., Weber, May 14, 1961 (AA C. X CALIFORNICA ssi ea oo Arb. No. 176-39, Jamaica Plain, Mass., Weber, May 6, 1960 ‘ ‘Flamingo. “Arnold ae ‘i 512-55, Weston, Mass., Weber, May 25, 1960 & May 10, 1962 (AA ‘Rosy Morn’, Arnold ree No, 181-58, Weston, Mass., Weber. May 31, 1961 Tetraploid cultivar with n = 34. C. X CALIFORNICA ‘Clarke’s Giant Red’, Arnold Arb. No. 751-60, Weston, Mass., Weber. May 31, 1961 (AAH). The only known tetraploid is usually sterile, as are a few other cultivars. Occasionally, it produces fruits containing viable seeds. Since the species of Chaenomeles have been shown to be self incompatible, and the shrubs of ‘Clarke’s Giant Red’ were surrounded by cultivars known to be diploid, a few seedlings were grown from one such fruit. As root tip smears are not 1964 | WEBER, CHAENOMELES 199 dependable, it is necessary to wait for the flower to determine a possible triploid condition. Cytological investigations do not seem promising in the genus Chae- nomeles. The pairing cannot be studied well in the genus because of the small size of the chromosomes. Meiosis was found to be normal in the species and in the hybrids, 17 chromosomes passing to each pole at anaphase. This was suspected, as the fertility appears to be normal in hybrids. Breeding Behavior. Flowers of Chaenomeles are formed very early in the fall, and open in the spring before the leaves develop, or when the buds are starting to expand. The regular flowers are cup-shaped or flat, and form an easy landing platform for insects. The stigmata seem to be receptive at once and the flowers are proterogynous. The disc is nectari- ferous although this is not evident out of doors for the flowers are eagerly visited by ants, and nearly all of them are inhabited by thrips which must consume the nectar as soon as it is produced. Nevertheless, if flowers are kept in the refrigerator for a few days in a closed jar, nectar appears as droplets on the disc. Its taste is sweet and its peculiar odor resembles that of apples, or of maple sap. Otherwise, the flowers have no scent. The young unfolding leaves also produce liquid through the glands at about the same time. The stamens in two rows are incurved in the buds. The stamens of the outer row have straightened up by anthesis and are ready to shed pollen after a few hours or within one or two days after the flowers open. The inner row of stamens will, in turn, straighten and start to shed pollen one or more days after the stamens of the outer row. At this time, flowers are visited by honey bees, and occasionally by bumble bees. Japanese Quinces are listed by some nurseries as melliferous plants. If the weather is hot and sunny, the flowers may drop the still turgid petals while the outer row of stamens is shedding pollen. By the end of anthesis, anther sacs have tipped over and stamens become slightly versatile. When empty. they change from yellow to brown. In the spring, each flower lasts about a week, but the blooming period is longer, for flowers start to open close to the ground (which warms up earlier) and follow later along the higher branches. In spring inflorescences containing several flowers, the one in the center usually opens first, eradually followed by the others. The terminal flower is the youngest in the racemes or panicles formed in the summer, and usually is the last one to open. When ripe, the fruits do not persist long on the shrubs. They often drop all at once, probably due to frost. When the fruits are on the ground (but sometimes while still on the shrubs) they are opened by birds seeking the } J ) seeds. The flesh is very acid and does not seem to be eaten. The seeds are also eagerly collected by rodents. A burrow (of rats?) was observed under a shrub of Japanese Quince and split open fruits were carefully piled up as refuse on both sides of the entrance. No seeds were left on the rejected cores, but were probably stored underground. 200 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLV Most likely, the seeds of Chaenomeles are dispersed by birds and small animals, when gathering food for winter preserves. A shrub, as a garden escape, was observed by Fernald in a hedge at Barnstable, Mass. Japanese Quinces are also known to become naturalized in Japan and Korea. In cultivation, no seedlings have been observed close to mature shrubs. This may be due to the numerous animals eating the seeds. Flocks of pigeons have been observed repeatedly near Japanese Quinces when the fruits were on the ground. It is difficult to believe that large seeds, often already breaking the seed coat inside the fruits, can pass undigested through the intestinal tract of a bird, but enough must be lost by the animals collecting the seeds to ensure reproduction in the wild. From old observations, Chaenomeles was suspected to be self incom- patible but the possibility of apomixis had to be investigated since it is of frequent occurrence in Malus and other genera of the Maloideae. It is recorded that C. speciosa, when introduced into Europe at the end of the eighteenth century, did not bear fruits for many years. At that time, it was believed that the temperature was not adequate for fruit production. Shrubs often were kept in greenhouses but still did not form fruit. When Japanese Quinces became common, and many varieties were grown in adjacent areas, fruits appeared in any climate. The early reports of no fruit development were certainly due to self incompatibility as subsequent experiments seem to have demonstrated. Apomixis includes reproduction in which embryos (and seeds) are produced asexually (agamospermy). Simple experiments involving the bagging of flowers or the isolation of plants may give the first indication of the presence or absence of this phenomenon. Colby (1929) bagged various cultivars of Chaenomeles and reported that except for C. speciosa ‘Grandiflora’, no fruits were produced. Similar experiments were per- formed for this study on C. speciosa ‘Early Apple Blossom’; C. superba ‘Columbia’, ‘Sunset’, and ‘Superba’; as well as on C. x californica ‘Cali- fornia’. No fruits were produced. The explanation of fruit production in one cultivar is not clear. The flowers of Chaenomeles are visited by ants and extremely small insects attracted by the nectar present inside the floral cup at anthesis and along the unfolding leaves. It is possible that these small insects penetrated the bag placed around a branch of C. speciosa ‘Grandiflora’. A word of caution must be expressed. The position of the flower clusters of Chaenomeles on old wood, requires the shoot system to be enclosed in plastic or paper bags. Artificial conditions of heat and moisture are evident in the bag and are detrimental to the young foliage and perhaps to the flowers. However, in most cases sufficient normal shoot development was observed to eliminate this factor in ac- counting for the lack of fruit production. A second test of apomixis and self incompatibility was made by the observation of isolated plants without bagging, and insect pollination limited to natural selfing. A shrub of C. speciosa located near the Harvard Law School was observed for two years during which it did not produce iruits. The closest Japanese Quinces, members of the Superba group, 1964 | WEBER, CHAENOMELES 201 are in the Cambridge Common, a distance of about 200 meters. In the spring of 1962, flowers of C. x superba were used to pollinate, by hand, a few flowers of the isolated Aoi of C. speciosa at the Law School. The artificially pollinated flowers were tagged and all of them produced fruits. No fruits were formed from flowers not cross pollinated by hand. The seeds were sown, and all the seedlings showed traces of hybridization between C. speciosa and C. superba. In another experiment, to serve as a control, a shrub of C. & superba ‘Knap Hill Scarlet’ known for its fruit production was transplanted from the shrub collection of the Arnold Arboretum in Jamaica Plain to an isolated spot in Cambridge. It bloomed abundantly in its new location during the spring of 1962, but in spite of numerous bees that visited the flowers, not one fruit was formed. The same condition occurred in a plant of C. speciosa “Tani-no-Yuki’ and one of C. X superba ‘Otto Froebel’ moved into the greenhouse. In the absence of pollinators to effect cross pollination, no fruits were pro- duced. A few flowers were self-pollinated by hand and these also failed to develop fruits. However, a few flowers were cross-pollinated between the two varieties and all produced fruits with fully developed ovules. In addi- tion to the fruit formation, the pistils of these self- and cross-pollinated flowers were also observed. The styles were collected from one to three days after hand pollination; dipped for two days in Gentian violet and then mounted in lactophenol. Under a microscope, it was possible to examine the growth of the pollen tubes. Self-pollinated flowers of Chaenomeles have been observed to have aberrant pollen development. When pollen is placed on the stigmata there may be complete failure of pollen tube development or the tube may develop and not penetrate the stigmatoid or stylar tissues. In such flowers there may be a slight swelling of the hypanthium but the resulting small fruits turn yellow and fall within a few weeks. Sections of such fruits show no ovule development. In the collections of the Arnold Arboretum many cultivars of Chae- nomeles are grown side by side in rows. With this abundance of plants fruit production is usually heavy, although some cultivars remains sterile year after year. In contrast, in the Boston suburbs isolated plants may be found as ornamental shrubs in gardens, which rarely fruit or produce malformed fruits containing a reduced number of seeds, but never less than four. There seems to be no parthenocarpy as in Malus, where fruits of normal appearance may be formed without seeds. Hall (1931) studied the problem of fruit failure in the apple and recorded that the fruit production of a given variety may be inconsistent. A variety which is self sterile under one set of conditions may be self fertile under another. Self incompatibility may be the result of genetic factors but environmental factors may also influence fruit production. Hybrids. Intergeneric hybrids are of common occurrence in the Maloideae, but none are recorded with Chaenomeles as one of the parents in spite of the common chromosome number 2 = 34, and the same 202 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV chromosome morphology. Crosses between the two species of Cydonia and species of Chaenomeles were unsuccessful in several attempts. Crosses with other genera should be tried. The three species of Chaenomeles have been shown to be completely interfertile under cultivation and there seems to be no reduction of fertility in subsequent generations of hybrids. This should, however, be oe statistically. Genetically, Chaenomeles appears to be a distinct taxo Self incompatibility in seamen plants favors the production of hybrids by natural and artificial means. There was very little variation in Chaenomeles during the first twenty years after the introduction of C. speciosa from Japan into Europe, in 1796. During this period, all plants were probably propagated asexually from the first shrub introduced at Kew Gardens, as no fruits were produced. One variety with blush flowers, called white, existed very early, and may have been a separate introduction. It was only after 1830 when von Siebold returned from Japan with many color forms of C. speciosa that varieties were developed in Japanese Quinces by the selection of seedlings. The colored forms, grown side by side, certainly crossed naturally. Before the introduction of any other species, numerous selections had been made in C. speciosa on the basis of variation in flower color, in number of petals, in the shape of the fruits, and in the size of the plants. The selections made over a century ago were the start of clonal lines which, through asexual propaga- tion, are still grown in many gardens. In 1869, a second species, C. japonica, was introduced into Europe. This species was distributed to gardens where C. speciosa was planted, and from the beginning set fruits. C. japonica was reproduced by seeds, and was thought to breed true. Very soon, numerous color forms ap- peared in the progeny of this species. They were considered varieties, but later all proved to be natural hybrids with C. speciosa. Such was the origin of C. superba. In recent years, numerous cultivars have been added to this group by selection, or by deliberate crossing of the two parental species. C. cathayensis, introduced into Europe around 1880, appears to be very uniform, but more observations must be made on this species before its breeding behavior can be ascertained. This species is closest to Cydonia which contains two species known to be self compatible. The hybrid groups C,. x clarkiana, C. & vilmoriniana, and C. californica were produced by hand pollinating flowers of two different species, or, the latter by pollinating flowers of an interspecific hybrid by a third species. The progeny of the interspecific hybrids proved to be fertile with other hybrids, or with any of the parents. As a result, Chaenomeles has an amount of variation unequalled in the other genera of the Maloideae partially due to backcrossing and introgression. The taxonomic treatment of several hundred cultivars of Chaenomeles presents a significant problem. The horticulturist would accept a group- ing based on the color of the flower, the form of the flower, single, semi- deable or double, or perhaps a classification based on hardiness, fruit 1964 | WEBER, CHAENOMELES 203 shape, or habit. The botanical approach requires recognition of the parent species and suggests that the cultivars be treated as a hybrid swarm. The fundamental information on the species was obtained by a study of herbarium specimens collected in the localities where the species are not found together. C. japonica is restricted to Japan where no other species is native. C. speciosa and C. cathayensis are from China. The sparse collections and the lack of data on their distribution and ecology do not indicate whether they grow together within their sympatric range, but no hybrids of these species have been recorded in China, and none were found among the herbarium specimens. Information on individual variation within a pure species was also available from shrubs grown from seeds of wild plants. The morphological variation of the species is extreme and it was, there- fore, impractical to apply the usual techniques of scatter diagrams, ideo- graphs or hybrid indices for this reason, and because of the numerous cultivars of the Californica group which are hybrids of not two but three species. A different method had to be devised to analyse each of the more than 300 named cultivars produced by variation within three species, by simple crosses in four combinations, and by backcrosses with subse- quent asexual propagation. Anderson’s techniques inspired the method used in this study. Herbarium specimens of cultivars were gathered from different localities to observe the differences due to the environment. These specimens were supplemented by colored drawings of fresh flowers. The drawings, more reliable than colored photographs, allowed a comparison of characters important to the horticulturist, but not easily preserved in dried specimens. ince numerical evaluation could not be used with trihybrids, the usual scoring was replaced by attributing each morphological character of the cultivar, such as habit, twig, leaf, flower, and fruit to the species it re- sembled most. It was found that the different organs usually tended to be closer to the type encountered in one of the species, perhaps because of the presence of linked genes, or at least of several genes on a single chromo- some affecting the same organs. When the characters were intermediate, this was indicated. From the results of the listing, each cultivar was attributed to a species, or to a hybrid group, when possessing charac- teristics of more than one species. The records of the parentage of crosses made by a few nurserymen were generously supplied and in most cases substantiated the recording technique. On several occasions cultivars, supposedly authentic, could not have been produced by the parentage suggested, and indicated an error in the identification of the cultivar, or one of its parents. Many cultivars in Chaenomeles are known to be backcrosses. Others suggest that introgression has advanced far in the Japanese Quinces. As Anderson stated ‘In the second generation of hybrids and in backcrosses there will be various and multitudinous re- combinations of flower color, shapes and heights and no two plants will look very much alike.” This explains why so many cultivars can be produced from three basic species. 204 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV The relative ease with which characteristics of habit, twig, leaf, flower, and fruit could be attributed to the parents of one species or another, permits identification of the parents of a hybrid, and also prediction as to the progeny of the cross. After the intercorrelated characters were sorted, it was possible to list those characters tending to appear together. For example, a cultivar with orange petals always has crenate leaves, warted twigs, and is of medium size. These characters come from its parent, C. japonica. Leaves with serrations terminated by an awn-like tip usually show some pubescence on the under surface, both characters coming from C. cathayensis. Having so many hybrids of known parentage, it was possible to recognize the characteristics which tend to be expressed. The members of the Superba group, C. japonica XX speciosa, usually are shrubs of intermediate size and branching. The twigs have the warting of C. japonica. The leaves are intermediate between the leaves of the two species, or more japonica-like as far as the color (dark green in C. speciosa, yellow-green in C. japonica), pubescence, and serration are con- cerned. The flowers have the wide range of color of C. speciosa, with the addition of pure orange found only in this group as a combination of the salmon pink of C. japonica with the red of the other species. Fruits are intermediate in size and perhaps in shape, but are closer to those of C. japonica. The sticky skin in many hybrids is also derived from C. japonica. The seeds, which are very similar in shape and size, may be differentiated only by the shiny testa of C. japonica, often present in seeds of the hybrids, versus the dull testa of C. speciosa. The three cultivars of the Vilmoriniana group, C. cathayensis x speciosa, are intermediate as far as characteristics of the habit and twigs are concerned, but show the influence of C. cathayensis in the other characters. These two hybrid groups show that in a cross of C. japonica X speciosa, the characters of C. japonica tend to be expressed over those of C. speciosa; in a cross of C. cathayensis & speciosa the progeny will look more like C. cathayensis. It is interesting to compare these observations with the characters of the cultivars of the Clarkiana group, C. cathayensis japonica. Although two cultivars only are in this group, they are both intermediate in all their parts and neither parent dominates the other. An analysis of the numerous cultivars of the Californica group, C. cathayensis superba reveals mostly intermediate characters, those of C. cathayensis tending to be expressed in the shape and size of the shrubs, in the spines, and often in the fruits. The flowers encompass the range of color found in C. speciosa, but are often of two tones of pink, brighter but very much like the flowers of C. cathayensis. These results, obtained from a study of the existing cultivars, may serve as a guide for the plant breeder wanting to produce a certain type of shrub. Size is certainly due to multiple genes, being in the range of the size of the parental species. Branching and spininess of C. japonica tend to be expressed when crossed with C. spec‘osa. In a cross including C. cathayensis as one parent, the shrub will look more like C. cathayensis than any other. The warting of C. japonica is always present in its pro- 1964 | WEBER, CHAENOMELES 205 geny, and is an important character in detecting hybridization with this species. C. japonica leaves, characterized by glabrescence and crenate teeth, may be contrasted with the leaves of C. cathayensis, showing serra- tion, terminated by an awn-like tip and abundant fulvous tomentum on the under surface at maturity. The leaf characters of C. japonica and of C. cathayensis again, tend to be expressed over those of C. speciosa. Interspecific hybrids, including the two dominant species, possess leaves intermediate in size, serration, and pubescence between those of the par- ents. The flowers are usually intermediate, taking up the range of color of the species with the largest range. A new color, orange, has appeared in hybrids by combination of pigments separately present in C. japonica and C. speciosa. In hybrids including C. cathayensis, flowers are often bicolored as in this species. The fruits tend to be intermediate in shape and size between those of the two parents; the ovoid shape of C. cathayen- sis appearing often in its hybrids. In the seeds, the wedge-shape of C. cathayensis is expressed over the unspecified shape of C. speciosa (found also in C. japonica), while seed characteristics of C. japonica dominate those of the others in shape, size, and the nature of the testa. To these morphological characters may be added the physiological one of hardiness. In a cross between a hardy species and one which is not, the lack of hardiness, probably due to multiple genes (as present in C. cathayensis) is usually expressed. This survey of the morphological characters and behavior of the different cultivars belonging to the three species, and to the four hybrid groups shows that C. cathayensis and C. japonica are two well defined species. Their morphological characters are easily recognized, and are evident in their progeny. This is not the case with C. speciosa, which is less specialized and does not possess characters not present in some degree in the other two species. This was evident when preparing the key to distinguish species and hybrid groups. The separation was found to be difficult between the cultivars belonging to C. speciosa = those of C. clarkiana, a hybrid between C. cathayensis and C. japonica. Evreinoff (1937) quotes Pachkevitch (1930) ¢ as eet that he believed C. lagenaria to be a hybrid between the other two species. The nomen- clature is so confused in Evreinoff’s paper that it is not possible to ascertain whether C. lagenaria in his sense means C. speciosa or C. cathayensis. Unfortunately, Pachkevitch’s treatise on pomology, in Russian, could not be obtained. His idea of the origin of one of the species of Chaenomeles by hybridization of the other two is appealing. It would explain why C. speciosa has no morphological character and range of its own, and why it is by far the most variable species in cultivation as well as in a native state. It may also suggest an explanation of the fact that the three species are still so closely related that they cross freely when grown in gardens. C. japonica is native only in Japan, C. cathayensis and C. speciosa probably occurring in different habitats in the region of Yunnan, China, did not develop reproductive isolation because other isolation mechanisms are present to keep them apart in the wild. [To be concluded | 206 JOURNAL OF THE ARNOLD ARBORETUM [| VOL. XLV THE GENERA OF CELASTRALES IN THE SOUTHEASTERN UNITED STATES! GrorGE K. BrIzICKY THE ORDER CELASTRALES, as delimited here, includes the families Celastraceae, Hippocrateaceae, and Aquifoliaceae, as well as several other allied families (e.g., Siphonodontaceae, Stackhousiaceae) not represented in the southeastern United States. The Staphyleaceae, usually regarded as members of either Celastrales or Sapindales, are excluded from both these orders, and will be treated within Cunoniales (or Saxifragales sensu lato), where they seem to belong (cf. C. G. G. J. van Steenis, Fl. Males. I. 6: 49. 1960). CELASTRACEAE R. Brown in Flinders, Voy. Terra Austr. 2: 554. 1814, “Celastrinae,’” nom. cons (STAFF-TREE FAMILY) Usually glabrous trees or shrubs, rarely woody vines, sometimes gutta- percha [or rubber | bearing; nodes unilacunar. Leaves simple, mem- branaceous to coriaceous, opposite or alternate [rarely minute or rudi- mentary], petiolate, de ciduous or persistent; stipules minute, usually caducous. Inflorescences axillary and/or terminal, cymose (usually dicho- tomously branched) or racemose, or the flowers solitary and/or fascicled in the leaf axils. Flowers small, regular, hypogynous, perigynous, or semi- epigynous, bisexual or unisexual by abortion (the plants then monoecious or dioecious), pediceled (except Gyminda). Sepals 4 or 5, small, connate about half of their length or more, or nearly distinct, usually imbricate. ‘Prepared for a generic flora of the southeastern United States, a joint project of the Arnold Arboretum and the Gray Herbarium of Harvard University made possible through the support of George R. Cooley ei the Scat Science Foundation and under the direction of Carroll E. Wood, Jr., and Reed C. Rollins. This treatment follows the pattern established in the first paper in the series (Jour. Arnold Arb. 39: 296-346. 1958) and continued through those in volumes 40-44 (1959-1963). The area covered is bounded by and includes North Carolina, Tennessee, Arkansas, and Loui- siana. The descriptions are based primarily on the plants of this area, with any supplementary material in brackets. References which the author has not seen are marked by an asterisk. The author is indebted to Carroll E. Wood, Jr., for his many valuable suggestions ; to Harry Ahles, George R. Cooley, Wilbur H. Duncan, and James W. Hardin, for e was drawn by Dorothy H. Marsh, in part from fresh fruiting material sent by R. B. Channell. 1964] BRIZICKY, GENERA OF CELASTRALES 207 Petals 4 or 5, distinct, inserted under the disc, usually imbricate. Stamens 4 or 5 [very rarely 8-10], distinct, alternipetalous, inserted under or on the disc at its exterior margin, reduced or wanting in @ flowers; filaments short; anthers basi- or dorsifixed, introrse, longitudinally dehiscent. Nec- tariferous disc intrastaminal, usually conspicuous, variously shaped, rarely wanting. Gynoecium syncarpous, 2—5-carpellate, rudimentary in ¢ flow- ers; stigma single, + lobed, or stigmata 2—5; style single or wanting; ovary superior, distinct or + immersed in and sometimes fused with the disc, or semi-inferior, 2—5-locular; placentation axile; ovules usually 2, more rarely 1 or 4-6 [to many] in a locule, pendulous from the top or ascendent from the base of the placenta, anatropous, usually apotropous, 2—-integumented. Fruit a loculicidal [rarely septicidal] capsule or a drupe [berry or samara], (1)2—5-locular, the locules 1— or 2—6-seeded. Seeds often arillate; endosperm present [rarely absent]; embryo straight, axile; cotyledons large, foliaceous, often green; radicle short, superior or inferior. TYPE GENUS: Celastrus L. A primarily pantropical family of 50-55 genera and over 800 species, extending with several genera into the temperate zones, to which a few are restricted. Six genera occur in our area. The genus Paxistima Raf. (also incorrectly spelled Pachystima or Pachistima; see Wheeler, p. 293) is excluded from the flora of the southeastern United States. Pollination by insects seems to be the rule. Data on floral biology and anatomy, as well as on embryology are sparse, pertaining to a few genera and species. Chromosome counts made for four genera (eight species) are 2n = 32 and 64, 46, and 80. Though sparse, these numbers may suggest that both aneuploidy and polyploidy have been of importance in the evolutionary development of Celastraceae. The family is closely related to Hippocrateaceae and Aquifoliaceae. A relationship with Siphonodontaceae and Stackhousiaceae has also been assumed (Takhtajan). * The occurrence = Paxistima Canbyi Gray in North pars mag first recorded . Cu by Chapman (FI. So. U.S. ed. 2. 613. 1889) on the authority of M. rtis. Ap- ee on the same a ee the species was included by Small (FI. SE. is S, 735. 1903 ; Man. SE Fl. 818. 1933), who repeatedly mentioned the occurrence of the species in introduced, apparently from West Virginia, into the Green & lor nursery, now abandoned; the plants are growing well and the population is expanding, but still without competition from native vegetation; it rem e seen whether the plants survive the invasion of the forest in the s to com I e meantime, and without constant cultivation, we still cannot call it a definite member of the native southeastern flora 208 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV REFERENCES: ANpeErsSoN, A. Studien iiber die Embryologie der Familien Celastraceae, Oleaceae und Apocynaceae. Lunds Univ. Arsskr. II. Sect. 2. 27(7): 1-110 pls. 1-4. 1931. [Celastraceae, 1-40, pls. 1, 2.| BarLton, H. Monographie des Célastracées et des Rhamnacées. Hist. Pl. 6: 1-92. 1875. [Celastraceae, including Hippocrateaceae, 1-50. Bentuam, G., & J. D. Hooker. Celastrineae. Gen. Pl. 1: 357-371. 1862. dns. Hippocrateaceae. | BERKELEY, E. Morphological studies in the caine Jour. Elisha Mitchell Sci. Soc. 69: 185-206. pls. 3, 4 [pp. 207, 208]. 1953. [Includes floral anatomy of Celastrus (2 spp.), Euonymus (4 spp.), and Paxistima (1 sp.).] BLAKELOocK, R. A. Celastraceae. Jn: Hutchinson, J., & J. M. Dalziel, Fl. W. Trop. Afr. ed. 2. 1(2): 623-634. 1958. [Hippocrateaceae included in Celastraceae; “the genus Campylostemon with 4-5 stamens and a Hippo- cratea-like fruit connects the two into one family.” | Boote, J. A. Studies in the anatomy of the family Celastraceae. Ph. D. thesis (unpublished). Univ. North Carolina. Chapel Hill. 1955.* CaANDOLLE, A. P. pe. Celastrineae. Prodr. 2: 2-18. 1825. [Includes Aqui- foliaceae. | CroizaT, L. A study in the Celastraceae. Siphonodonoideae subf. nov. Lilloa 13: 31-43. 1947. [Includes a hypothesis on the derivation of the typical celastraceous flower and the flower of Siphonodon from an ancestral flower (or inflorescence) composed of superposed rows of carpels and stamens. ] sabi oeoreag C. Uber die Nervation der Blatter bei den Celastrineen. schr. Akad. Wiss. Wien Math. Naturw. 13: 43-83. pls. 1-10. 1857. Harun, "|. W. Pachystima Cambyi in North Carolina. Castanea 28: 177, 178. 1964. Herr, J. M., Jk. Embryological evidence for the Sarees of Aquifoliaceae to Celastraceae. (Abstr.) Va. Jour. Sci. II. 10: Hou, D. Celastraceae. In: C. G. G. J. VAN eae Fl. Males. I. 6: 227-291. 962 1 . LoESENER, T. Celastraceae. Nat. Pflanzenfam. III. 5: 189-222. 1892. . Celastraceae. Jbid. ed. 2. 20b: 87-197. Uber die geographische Verbreitung einiger Celastraceen. Bot. Jahrb. 24: 197-201. 1897. Massey, A. B. Discovery and distribution of Pachystima Canbyi Gray. Castanea 5: 8-11. 1940. Mavrirzon, J. Zur Embryologie und systematischen Abgrenzung der Reihen Terebinthales und Celastrales. Bot. Not. 1936: 161-212. 1936. Metz, A. Anatomie der Laubblatter der Celastrineen mit besonderer Beriick- sichtigung des Vorkommens von Kautschuk. Beih. Bot. Centralbl. 15: 309-386. 1903. [Thirty ate 72 spp. | J. B. Traité d’organogénie comparée de la fleur. 748 pp. Ty pls. 1-154. Paris. 1857. [Celastraceae, 167 170, pl. 36. figs. 1-13, PiLouvier, V. Nouvelles recherches sur le quebrachitol des ae et Hippocastanacées, le dulcitol des Celastracées et la saccharose de quelques autres familles. Compt. Rend. Acad. Sci. Paris 228: 1886-1888. 1949 Recorp, S. J. The American woods of the orders Celastrales, Olacales, and Santalales. Trop. Woods 53: 11-38. 1938. [Celastraceae, 14—21.] & R. W. Hess. Timbers of the New World. xv + 640 pp. pls. 1-58. New Haven. 1943. [Celastraceae, 119-124.] Renurous, L. Les stomates des Celastraceae. Bull. Soc. Bot. Genéve II. 6: 1964 | BRIZICKY, GENERA OF CELASTRALES 209 13-18. 1914. [Euonymus (5 spp.), Celastrus (1 sp.), Catha (1 sp.).] Etude sur les stomates. /bid. 9: 245-350. 1917. [Celastraceae, 310- 323. SARGENT, C. S. Manual of the trees of North America (exclusive of Mexico). ed. 2. xxvi + 910 pp. Boston & New York. 1922. [Celastraceae, 674-680. ] SMITH, A. C., & I. W. BatLey. Brassiantha, a new genus of Hippocrateaceae from New Guinea. Jour. Arnold Arb. 22: 389-394. pl. 1. 1941. [Includes notes on gross morphology of the flowers and fruits and wood anatomy of Celastraceae. | STENZEL, G. Anatomie der Laubblatter und Stamme der Celastraceae und Hippocrateaceae. Inaug.-diss. 91 pp. Breslau. 1892(?). TAKHTAJAN, A. Die Evolution der Angiospermen. viii + 344 pp. Jena. 1959. [Celastraceae, 241-243. ] TIEGHEM, P. vAN. Structure de quelques ovules et parti qu’on eu peut tirer pour améliorer la classification. Jour. Bot. Morot 12: 197-220. 1898. [Celastraceae, 210. TRELEASE, W. Revision of North American Illicineae and Celastraceae. Trans. Acad. Sci. St. Louis 5: 343-357. 1889. Ls eames 349-357. ] Ursan, I. Celastraceae. Symb. Antill. 5: 48-94. 190 WEST, E. .. & L. E. ArNoip. The native trees of cae xx + 212 pp. Gaines- ville. 1946. [Celastraceae, 126-128 WHEELER, L. C. History and orthography of the celastraceous genus “Pachy- stima” Rafinesque. Am. Midl. Nat. 29: 792-795. 1943. Witczexk, R. Celastraceae. Fl. Congo Belge 9: 113-125. 1960. KEy TO THE GENERA OF CELASTRACEAE General characters: woody plants with opposite or alternate iad bai minute, usually caducous stipules; flowers small, bisexual or unisexual merous, usually with a conspicuous nectariferous disc, in axillary ee cymes or few-flowered fascicles, more rarely in terminal Hishses- stamens alterni- petalous ; ovary usually superior, 2-5-locular, sometimes + immersed in the disc, the placentae axile, the ovules usually 2 or 1; fruit either a loculicidal capsule with arillate seeds or a drupe. A. Plant scandent, woody, dextrorsely twining; leaves alternate; flowers in terminal or axillary raceme-like thyrses; capsule subglobular, ca. 1 cm. in diameter, orange or yellow; seed covered with a scarlet aril. 2. Celastrus. A. Plant nonscandent, shrubby, sometimes prostrate or creeping, or arbores- cent; seiae in axillary dichasial cymes or fascicles. B, Lea es opposite; flowers in axillary dichasially branched cym Cc. ey deciduous, leaves membranaceous (rarely eee anther-halves diverging; ovary locules 2-6(10?)-ovulate; fruit a 3—5-lobed, -valved, sometimes tuberculate capsule; seed covered with a scarlet aril; occurring north of central Florida. 1. Euonymus. Plant evergreen, leaves thin-coriaceous; anther-halves parallel; ovary locules l-ovulate; fruit a small drupe; seed exarillate; sub- tropical Florida. Flowers slender pediceled, 2-bracteolate at base of pedicel, bisexual; sepals connate nearly half their length; stigmata 4, narrow, spreading, raised on a short style; ovary 4-locular; ovule ascendent from the base of the placenta; drupe red, O 210 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV ies) — ° = oOo ca n w i¢") n nw =: oO 5 : ee a oO Lone} jo) = ~ wm jo) Loney = a im” p> wm —-s p ? ion 4 2 Qa fe Oo 2. p> ad oO et c na < =) Q. oO tad s fleshy-coriaceous, with indistinct venation on the upper surface; flowers 5-merous, with a 3-carpellate gynoecium; fruit a red ellipsoidal to obovoid capsule, 8-12 mm. long, dehiscing by 3 valves; seed arillate. ......... oo eee... 3, Maytenus. Leaves thin-coriaceous, conspicuously veined on the upper surface; flowers 4-merous, with a 2-carpellate gynoecium; fruit a scarlet subglobular drupe 5-6 mm. in diameter, containing 2 stones; seed exarilaté,. ....4¢94063004eade83 SENG sea eee ass 6. Schaefferia. res) Subfamily CELASTROIDEAE Tribe CELASTREAE 1. Euonymus Linnaeus, Sp. Pl. 1: 197. 1753; Gen. Pl. ed. 5. 91. 1754. Deciduous [or evergreen] shrubs [sometimes climbing by aérial root- lets] or trees, with green, usually quadrangular branchlets [sometimes corky ridged or warted]. Leaves opposite [rarely whorled or alternate], membranaceous [to coriaceous], crenate-serrate or serrulate, with gland- tipped teeth [rarely prickly or entire], usually petioled; stipules caducous [rarely persistent]. Inflorescences axillary {and terminal], simple or com- pound dichasially branched (1)3-15(31)-flowered peduncled cymes with minute caducous bracts. Flowers 4- or 5-merous, hypogynous (or semi- epigynous?), bisexual (rarely also unisexual by abortion, the plants then polygamo-monoecious or -dioecious). Se als connate about half their length. Petals ovate-orbicular to suborbicular [to linear], sometimes shortly clawed, greenish white to purple [or greenish yellow|. Stamens inserted on the margin of a conspicuous fleshy usually 4- or 5-lobed nec- tariferous disc; filaments subulate, very short [more rarely + elongate, or wanting]; anthers basifixed, broad, the anther-halves connivent [or con- fluent |at apex and much divergent to subhorizontal toward the base, 2|1]|- locular at anthesis, dehiscing by 2 lateral [or 1 continuous subapical | slit. Gynoecium 3-5-carpellate; stigma small, indistinctly 3-—5-lobed; style short, stoutish; ovary superior (or semi-inferior?), | 2 |3—5-locular, + immersed in, except the top, and fused at base with the disc; ovules 2-6(10)[-12] in a locule, + ascendent or rarely pendulous, superposed or collateral [or in 2 rows], apotropous, the nucellus thin. Fruit a leathery capsule, usually pink to red or purple, subglobular or oblate |or obovoid, obconical or obpyramidal], smooth or “tuberculate” |or echinate], usually [2]3—5-locular, -lobed |-angled or -winged, or rarely round in cross- section], and -valved; locules 2—6(10?)-seeded. Seed partially or com- 1964 | BRIZICKY, GENERA OF CELASTRALES 211 Pic, J, nT aaa E. americanus: a, flowering branchlet, from above, * 1 b, flower, X 10; c, flower in diagrammatic vertical section, to show i ees petal and stamen Y aitice removed) and ovary immersed (except papillate a in disc, X 16; d, fruiting branchlet bearing incompletely developed capsule a seeds from a single locule, arils squeezed together , three low er seeds of same cluster separated to show varying ee of aril, g, h, two views of seed, showing raphe, the micropyle below, X 6; i, seed in diagrammatic vertical section in plan of raphe, showing on see coats and three of four embryos embedded in endosperm (stippled), oe Fp pletely covered by a scarlet or orange aril; seed coat leathery, orange- or reddish-brown [to black, or whitish]; endosperm abundant, fleshy; embryo large, the cotyledons roundish to oblong, usually green; the radicle inferior or superior. LECTOTYPE SPECIES: E. europaeus L.; see N. L. Britton, N. Am. Trees 630. 1908. (Name from Greek, eu, good, and 212 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. XLV onoma, name, i.e., of good name, famous, apparently applied ironically to the genus, the plants having had the bad reputation of poisoning cattle.) — SPINDLE-TREE. A genus of nearly 180 species of the North Temperate Zone and the Tropics, concentrated in eastern Asia, a few in Europe, Africa, Madagas- car, and Australia, completely absent in South America and the West Indies; five to seven unsatisfactorily known species in Central America (south to Costa Rica), four in the United States, and three in our area. Blakelock’s classification, which recognized subgenera EvonyMus and Katonymus Beck emend. Blakel. with six sections and 14 series, is fol- lowed here. All American species belong to subg. EUONYMOUS, character- ized by small, ovoid, acute winter buds, filamented (vs. subsessile ) stamens, anthers two- locular at anthesis and ‘dehiscing by two distinct slits, and unwinged capsules. The sections are based primarily on the shape and degree of lobation and the character of the surface of the capsules. Section Evonymus (§ Biloculares Rouy & Fouc. emend. Blakel.) in- cludes about 70 Old World and two (or three) American species, all with two-ovulate locules and smooth fruits. The primarily northeastern Ameri- can Euonymus atropurpureus Jacq., wahoo or burning bush, a large shrub or small tree with four-merous purple flowers, occurs in thickets along streams and in rich damp woods from southern Ontario to southern Min- nesota, southeastern South Dakota (and locally in Montana), south to Nebraska, Kansas, Oklahoma, northeastern Texas, Arkansas, and Ten- nessee, with scattered stations in North Carolina, Georgia, Alabama, Mississippi, and northernmost Florida (Gadsden County). Euonymus occidentalis Nutt. ex Torr., differing in part in its five-merous flowers, occurs from Washington to California. Euonymus europaeus, 2n = 64, with greenish-white petals and 3—5-flowered inflorescences, often grown as an ornamental, has been recorded as naturalized in at least the north- eastern United States. Primarily of eastern Asia, sect. EcHtNococcus Nakai emend. Blakel., characterized by five-merous flowers (in our species), 4—6(—10?)-ovulate locules, and tuberculate fruits, includes five North American species, two in our area. Euonymus americanus L. (including E. angustifolius Pursh), strawberry-bush, 2m = 64, an upright or straggling (very rarely creeping) + stoloniferous shrub with rather variable, rarely obovate thickish leaves, densely minutely pustulate branchlets, flowers about 1 cm. in diameter, distinctly clawed petals, and (3-)5-locular capsule is common throughout much of our area.’ It occurs in rich lowland woods and thickets, in ravines ® Two poorly known variants a uncertain taxonomic status have been described as varieties of Euonymus americanus. The var. angustifolius (Pursh) A. Wood, with narrowly lanceolate to sublinear leaves (as against lanceolate or ovate, rarely obovate, ones in var. americanus) was described from Georgia and is also known from at least North Carolina and/or Tennessee (French Broad River), and northernmost Florida (Apalachicola River). The var. sarmentosus Nutt., a prostrate, usually creeping shrub, rarely seen with flowers (Meehan), seems to be of rare occurrence in some parts of our area (at least in Load and Arkansas). Perhaps it represents a distinct 1964] BRIZICKY, GENERA OF CELASTRALES 213 and bottoms, and on river bluffs from New York west to southern Illinois and southeastern Missouri, south to eastern Texas and Highlands County, Florida. The primarily northeastern American E. obovatus Nutt., running strawberry-bush, a creeping shrub with upright flowering shoots 3-6 dm. high, thin, usually obovate leaves, smooth branchlets, flowers 6—7 mm. in diameter, nonclawed petals, and usually three-locular capsules, reaches its southern limits in Tennessee and southwestern North Carolina, where it occurs in rich, moist or dry woods, on moist shaded banks, or on rocky bluffs. The Asiatic Euonymus alatus (Thunb.) Sieb., of sect. MELANOCARYA (Turcz.) Nakai emend. Blakel., an ornamental shrub with spreading, usually corky-winged branches and smooth four-parted capsules, has be- come established locally in at least Connecticut, West Virginia, and Kentuck Flowers of the genus usually are bisextial and proterandrous, but the occasional occurrence of unisexual flowers (the plants consequently monoecious or dioecious) has been recorded at least in European species. Hymenoptera and Diptera, especially short-tongued bees and flies, seem to be the main pollinators. The attractive arillate seeds are dispersed by birds. Seeds in many (perhaps most) species have a long dormant period, requiring stratification before germination. Apomixis (adventi- tious polyembryony from the inner integument of the ovule, perhaps rarely from the endosperm) is known in Euonymus latifolius (L.) Mill., E. ameri- canus, and E. alatus, as well as in several eastern Asiatic species. Fertiliza- tion is necessary for the development of adventitious embryos. rkeley, on the basis of four species, interpreted the “broad flat pier surrounding the ovary” [the disc] as a floral tube and considered the ovary to be semi-inferior, the flowers semi-epigynous. The occurrence of fibriform vessel elements in the wood of the stem and of secretory cells (sacs) containing gutta-percha in the phloem of the roots (and sparingly of the axis) is nota Reported chromosome numbers (four species) are 2n = 32 and 64. The lower number was found in Euonymus japonicus Thunb. and FE. Fortunei (Turez.) Hand.-Mazz., representing the presumably primitive sect. In1c1- FOLIA, while the higher pertains to E. europaeus and EF. americanus, of two presumably advanced but not closely related sections. The genus is closely related to the West Indian Torralbasia Krug & Urb. and the New Guinean Xyvlonymus Kalkm. Many species are grown as ornamentals. At least Euonymus verrucosus Scop. and E. europaeus are cultivated in the Soviet Union as sources of gutta-percha. The seeds (containing the glucoside evonosid, a_ heart poison) and the vegetative organs of at least the European species and E. atropurpureus are said to be poisonous. population of the species which seems to be more pronouncedly stoloniferous than the typical variety. In the study of the variants the common occurrence of apomixis (adventitious embryony) in E. americanus should be taken into consideration. 214 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV REFERENCES: The large number of references has been greatly reduced here. Under family references see ANDERSSON, BERKELEY, Hou (pp. 243-254), LoEsENER (1942, pp. 115-124), Mez (pp. 333, 334), REHFOUS (1914; 1917, pp. 310-323), and SARGENT. BARTHOLOMEW, E. A. Euonymus alatus established in West Virginia. Castanea ye hae IK 1957 BLAKELOCK, R. A. A synopsis of the genus Euonymus L. Kew Bull. 1951: 210-290. 1951. Braun, A. Uber Polyembryonie und Keimung von Caelebogyne. Abh. Akad. Wiss. Berlin Physik. 1859: 109-263. pls. 1-6. 1860. | Includes a survey of polyembryony; Euonymus latifolius and E. americanus, 156-159. pl. 4.] Brizicky, G. K. Polyembryony in Euonymus (Celastraceae). Jour. Arnold Arb. 45: 251-259. 1964. Copacciont, M. Etude phyllotaxique d’Evonymus japonicus Th. Revue Gén. Bot. 61: 740-784. pl. 8. 1954. FREEMAN, O. M. New or noteworthy plants from Polk Co., North Carolina or vicinity. Castanea 21: 41-43. 1956. [E. atropurpureus, 42.] GRAENICHER, S. Flowers adapted to flesh-flies. Bull. Wis. Nat. Hist. Soc. II. 2: 29-38. 1902. [E. atropurpureus, 36-38 Institut LesA AKADEMII NAUK S. S. S. R. Cex Institute, Academy of Sciences of the U. S. S. R.) Euonymus as a gutta-percha-bearer and the scientific foundations of its cultivation and utilization. (Collected articles by various authors; in Russian.) Trudy Inst. Lesa Akad. Nauk SSSR 1: 1-196. 1947 | Russian bibliography for 1931-1944, pp. 190-192]; ibid. 11: [paging?| 1953 * [Russian bibliography for 1945-1950, pp. 5-13]; ibid. 46: 1-153. 1958 [Russian age for 1951-1955, pp. 148-153]. [See also RyCHNOvsKA-SoupKova, Lesn. Prace 34: 436-441. 1955. ] Kniazeva, L. A. The formation of gutta-percha in young roots of European Euonymus. - Russian.) Dokl. Akad. Nauk SSSR 119: 602-605. 1958. [See Am. Inst. Biol. Sci., A translation of Doklady (Bot. Sci. Sect.) 119: 86-90. poy KRASILNIKOV, P. K. Some — features of the root systems of five species of Euonymus at the age of two years under conditions of the Leningrad Region. (In Ru a ee Bot. Zhur. 45: 394-397. 1960. LAWRENCE, G. H. M. Euonymus europaea, E. hamiltoniana, and relatives. Baileya 3: 113, 114. 1955 Leonova, T. G. In regard to the time of starting of gutta-sacs in Euonymus europaeus L., and E. maackii Rupr. (In Russian.) Bot. Zhur. 43: 430- 433. 1958. ———. De speciebus generis Euonymus L. seriei Lophocarpi (Loes.) Blakel. Not. Syst. Leningrad 19: 315-329. 1959. [Sect. HESPERIDINYMUS Leon. established for E. occidentalis Nutt., and E. melanantha Fr. et Savat.] . A contribution to the knowledge of the genus Euonymus L. Bot. Zhur. 45: 750-758. 1960. | Proposes a new classification somewhat different from Blakelock’s; sketch of history and evolution. | McNAaIrR, a _T. Comparative anatomy within the genus Euonymus. Univ. Kan . Bull. 19: 221-260. pls. 22-27. 1930. [Eight spp.. including F. puny eT eve and E. americanus. | Marskevicu, N. V., & A. L. KosHcHEEv. Rooting and gutta accumulation in 1964 | BRIZICKY, GENERA OF CELASTRALES 215 the layered stems of Euonymus. (In Russian.) Dokl. Akad. Nauk SSSR 92: 1069-1072. 1953.* Mazak1, T., & M. Aritomi. On the triterpenoids in the leaves of Euonymus radicans Sieb. and E. japonica Thunb. oe Japanese; English summary.) Jour. Pharm. Soc. Japan 79: 980-985. 1959. MEEHAN, T. Contributions to the life-histories i“ plants, IV. On parallel habits in allied species from widely separated localities. Proc. Acad. Nat. Sci. Phila. 1889: 64-66. 1890. [E. americanus, E. japonicus, habit dimorphism, 65 Ls Nakal, T. Subdivisions of the genus Euonymus. (In sewers and Latin.) Jour. Jap. Bot. 17: 615-619. 1941. [Eight sects. established. ] Genitia, gn. [genus] novum Celastracearum. (Appendix.) Systema novum generis Euonymi Nipponensis. (In Japanese and Latin.) Acta Phytotax. Geobot. 13: 20-32. 1943. [A segregate from Euonymus; Euonymus subdivided into 4 subgenera and 10 sects.; see also Jour. Jap. Bot. 24: 8-14. 1949. ] NIKOLAEVA, M. G. Biology of seed germination of Euonymus in connection with its specific characteristics and geographic origin. (In Russian.) Bot. Zhur. 41: 393-403. 1956. [See also Acta Inst. Bot. Acad. Sci. URSS 4. Bot. Exp. 10: 267-295. 1955; 11: 331-350. 1956; and V. M. LruscHENKOo, Bot. Zhur. 45: 144-148. 1960. ] PesinA, K. Influence of arillus and time of planting on germination of European Evonymus (Evonymus europaea L.). (In Czech; English summary.) Preslia 29: 186-192. 1957. PROKHANOV, Y. I. Celastraceae Lindl. Fl. URSS 14: 546-573, 744-746. 1949. | Kalonymus (Beck) Prokh., a segregate genus, 566-573 Conspectus systematis Celastracearum URSS. Addenda et corrigenda. (In Russian and Latin.) Not. Syst. Leningrad 20: 409-412. 1960. SARGENT, C. S. Seni Silva N. Am. 2: 9-12. pl. 53. 1891. [E. atropur- pureus. Evonymus obovatus. Garden Forest 9: 384, 385. 1896. SHAMEL, A. D. A bud variation of Euonymus. Jour. Hered. 8: 218-220. 1917. [|E. japonicus argenteo-variegatus. Soper, J. H. Some genera of restricted range in the Carolinian flora of Canada. Trans. Roy. Canad. Inst. 34: 1-56. 1962. [E. atropurpureus, 26; E. obovatus, 26-29. | SPRAGUE, T. A. The ana fruited species of Euonymus. Bull. Misc. Inf. Kew 1908: 29-36. 19 . The correct ar of certain generic names: 6. Euonymus or Evony- mus. Ibid. 1928: 294-296. 1928. [See also Int. Code Bot. Nomencl. 1961. 56 (Art. 74).] STRASBURGER, E. Uber Polyembryonie. Jena. Zeitschr. Naturw. 12: 647-670. pls. 15-19. see [E. latifolius, adventive nucellar polyembryony, 658, 659. pl. 19. fig. 44. | ZABEL. H. oo obovatus, Nutt. Gartenflora 38: 638-640. 1889. ZECHMEISTER, L., & R. B. ESCuE. Isolation of prolycopene and progamma- carotene from Evonymus fortunei. Jour. Biol. Chem. 144: 321-323. 1942.* 2. Celastrus Linnaeus, Sp. Pl. 1: 196. 1753; Gen. Pl. ed. 5. 91. 1754. Scandent, usually dextrorsely twining shrubs with alternate, elliptic to suborbicular or obovate, serrulate to serrate [or subentire], deciduous 216 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLV [rarely persistent] leaves. Inflorescences few- or many-flowered terminal and/or axillary, sometimes raceme-like, thyrses [or simple to compound dichasia] with minute caducous bracts. Flowers small, greenish [white], hypogynous or perigynous, [bisexual or | unisexual (the plants dioecious), with articulate pedicels. Calyx campanulate, 5-lobed, the lobes imbricate, persistent in fruit. Petals 5, inserted under the margin of the disc. Stamens 5, inserted on [or under] the margin of the disc, abortive and sterile in ? flowers; filaments subulate [or linear]; anthers introrse, ovate [oblong] in outline, apiculate, cordate at base, dorsifixed and usually versatile, laterally or introrsely dehiscent by 2 longitudinal slits. Nectariferous disc mem- branaceous and cupuliform [or fleshy and flat], shallowly 5-lobed [or entire]. Gynoecium usually 3(4)-carpellate, rudimentary in 2 flowers; stigma usually 3(4)-lobed [the lobes sometimes bifurcate|; style short; ovary superior, usually incompletely [completely] 3-locular, with axile placentae at the very base; ovules 2 [1] in each locule, ascendent, usually with a cup-shaped aril at base, the nucellus thin. Fruit a small orange to yellow subglobular [to cylindric] leathery capsule, about 1 cm. in diameter, tipped by the persistent style, usually incompletely 3-locular, 3-valved; locules 2(1)-seeded. Seed ascendent, ellipsoidal, usually + plano-convex, obtuse at base, attenuate toward the apex, inclosed in a fleshy crimson aril open at apex; seed coat thin, brown, leathery; endo- sperm copious; embryo broad-spatulate; radicle inferior. LECTOTYPE species: C. scandens L.; see Britton & Brown, Illus. Fl. No. U.S. ed. 2. 2: 492, 493. 1913. (Name from Greek, kelastros, the ancient name of an evergreen tree, presumably Phillyrea latifolia L., Oleaceae, applied by Linnaeus to this genus.) A primarily tropical genus of about 30 (or more) species, widely dis- tributed in eastern Asia, Malaysia, Australia, Oceania, Madagascar, and continental America. Hou (1955) established subg. CeLastrus (flowers unisexual, ovary locules 2-ovulate, fruits usually 3—6-seeded), with 23 species in the Old World and one in eastern North America, and subg. RACEMOCELASTRUS Hou (flowers bisexual, ovary locules 1-ovulate, fruits usually 1-seeded) ranging from central Mexico to South America (seven species ). The primarily northeastern American Celastrus scandens, climbing or false or American bittersweet, waxwork, 2m = 46, with ovate to elliptic leaves and terminal inflorescences, occurs in thickets, woods (most fre- quently in stands of young trees), fence-rows, and along streams, usually in rich soil, from southern Quebec west to southern Manitoba, south to Oklahoma and central Texas, Arkansas, Tennessee, northern Alabama, and western North Carolina, although it is scattered and rare in the states of our area.t This species “is related to the eastern Asiatic center of “Fernald (Gray’s Man. Bot. ed. 8. 984. 1950), Small (Man. SE. Fl. 818. 1933), and some other authors, include also Georgia and Louisiana, or Mississippi (Small), in the range of Celastrus scandens. The occurrence of C. scandens in Louisiana is doubtful, since C. A. Brown (Louisiana trees and shrubs, 1945, p. 169) says, “Reported 1964] BRIZICKY, GENERA OF CELASTRALES 217 dispersion rather than to the one in Central America” (Hou, 1955, p. 224). The Asiatic C. orbiculatus Thunb., 2n = 46, with suborbicular to obovate leaves and axillary inflorescences, often grown as an ornamental, has be- come naturalized in at least the Northeastern States, south to Virginia. Recently it has also been noted by Hoffman as naturalized in Tennessee in the Great Smoky Mountains National Park. Little is recorded regarding floral biology in the genus. Hymenopterous insects, especially bees, seem to be the main pollinators, although wind also be involved (cf. Wyman, pp. 84, Apomixis (adventitious embryony from the inner integument), and perhaps “false polyembryony” (supposed on the basis of the occasional occurrence of two nucelli in an ovule), and parthenocarpy have been recorded in Celastrus scandens (Andersson). Birds are considered to be the seed dispersing agents. Seeds in C. scandens which mature in the fall seem to have a long dormancy period and should be stratified to accelerate germination. Vegetative reproduction by suckers seems to be common. Chromosome numbers are known only for the two species mentioned here. The artificial hybrid C. scandens x orbiculatus is less vigorous than the parents, sparingly fertile, and smaller fruited than either parent. The occurrence of natural hybrids between some species in eastern Asia seems probable (Hou, 1955 p. 225). The flowers are regarded by Berkeley as perigynous, the cup-shaped disc (free from the ovary) being interpreted as a floral tube. The genus is closely related to the pantropical Maytenus Molina. Celastrus scandens, C. orbiculatus, and a few other species are grown as ornamentals. All are strangling vines, and our species may cause distinc- tive injuries to the stem and branches of young trees (Lutz). Leaves of C. scandens are said to be poisonous to horses. REFERENCES: Under family references see ANDERSSON (pp. 1-40), BERKELEY, Hou (pp. 233-238), LOESENER (1942, pp. 131-134), and Metz (pp. 340, 341), Basu, N. K., & P. R. Paprar. A chemical investigation of Celastrus paniculata Willd. Jour, Am. Pharm. Assoc. Sci. Ed. 35: 272, 273. 1946. Croizat, L. The concept of inflorescence. Bull. Torrey Bot. Club 70: 496-509. 1943. [Celastrus, intercalary inflorescence, 500-506. ] DILLINGHAM, F. T. The staff-tree, Celastrus scandens, as a former food supply from West Feliciana Parish by R. S. Cocks. There are no specimens of this in the Tulane herbarium. The material so labeled is the star vine, Schizandra coccinea Michx. The aut thor has searched for this plant in the state but never found it.” No herbarium Hitchcock’s specimen is located and checked and/or the species is re-collected in northern Florida. 218 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV of starving Indians. Am. Nat. 41: 391-393. 1907. [Bark contains mannose and mannan Harris, J. A. Correlation in the inflorescence of Celastrus scandens. Missouri Bot. Gard. Rep. 20: 116-122. 1909. Hart, H. T. Delayed germination in seeds of Peltandra virginica and Celastrus scandens. Publ. Puget Sound Biol. Sta. 6: 255-261. 1928 Horrman, H. L. Check list of vascular plants of the Great Smoky Mountains. 44 pp. mimeogr. 1962. [C. orbiculatus, 27.] Hou, D. A revision of the genus Celastrus. Ann. Missouri Bot. Gard. 42: 215-302. 1955. Ketter, I. A. The coloring matter of the aril of Celastrus scandens. Proc. Acad. Nat. Sci. Phila. 1896: 212-218. 1897.* Lawson, G. Contributions to microscopical analysis. No. 2. Celastrus scandens Linn., with remarks on the colouring matters of plants. Trans. Bot. Soc. Edinb. 6: 362-368. 1860. LeRosen, A. L., & L. ZECHMEISTER. The carotenoid pigments of the fruit of Celastrus scandens L. Arch. Biochem. Biophys. 1: 17-26. 1942.* [Isolation and Pees of een Lutz, H. J. Injuries to trees caused by Celastrus and Vitis. Bull. Torrey Bot. Club 70: 436-439. 1943. Panisset, M., & A. NANTEL. Etude des propriétés antibactériennes des extraits des racines de Celastrus scandens. (Abstr.) Ann. Assoc. Canad.-Frang. Avance. Sci. 15: 83, 84. 1949.* ane ae F. Seedlings of certain woody plants. Minn. Bot. Stud. 2: 69-85. pls. 1-4. 1899. [C. scandens, 76, 77. pl. 2.] SHAH, M. 8 N. L. PHALNIKAR, & B. V. Buipe. A note on the chemical inves- tigation of the fruits of Calaxirs paniculata, Willd. (n. 0. Celastraceae). Curr. - Bangalore 16: 57, 58. 1947.* Wuirte, O. E., & W. M. Bowben. Oriental and American bittersweet hybrids. Jour. ae 38: 125-127. 1947. [C. scandens < orbiculatus. | Wyman, D. Fruiting habits of certain ornamental plants. Arnoldia 10: 81-85. 1950. | Celastrus, 84, 85.| 3. Maytenus Molina, Sagg. Stor. Nat. Chili 177. 1782; emend. Bosc, Nouv. Dict. Hist. Nat. 14: 211. Evergreen, usually glabrous shrubs or small trees. Leaves alternate, fleshy-coriaceous [or membranaceous to coriaceous], entire [or toothed], short-petiolate; stipules minute, caducous [or wanting]. Inflorescences compact fascicles or flowers solitary [or in dichasial cymes, short thyrses, or racemes| in leaf- or more rarely in cataphyll-axils. Flowers short- pediceled, small, usually hypogynous, | bisexual or| unisexual by abortion, the plants monoecious or dioecious. Sepals (4) 5, small, connate at least at base, persistent. Petals (4) 5, greenish white or white [or yellow, or red], longer than the sepals, spreading. Stamens (4) 5, inserted on or under the margin of the disc, shorter than the petals, reduced and sterile in ? flowers: filaments subulate, distinct; anthers ovate-suborbicular in outline, half as long as the filaments, introrse. Nectariferous disc conspicuous, fleshy, flattish and (4)5-angular [to -lobed] in ¢ , annular and fused with the base of the ovary in @ flowers. Gynoecium |2]|3(—5)-carpellate, rudi- 1964 | BRIZICKY, GENERA OF CELASTRALES 219 mentary in ¢ flowers; stigma [2]3(—5)-lobed; style very short or want- ing; ovary superior, + immersed in and fused at least at base with the disc, short-conical, slightly 3(—5)-angular, incompletely [or completely | 3(—5)-locular, with axile placentae at base: ovules 1 [2] in each locule, ascendent, apotropous. Fruit an ellipsoidal or obovoid, usually slightly 3(—5)-angular, coriaceous, red (our species) capsule 8-12 mm. long, 1[-3]-locular, 3(—5)-valvate, (1)3(-5)[-6]-seeded. Seed ascendent, ovoid-ellipsoidal, 3-5 mm. long, covered entirely [or partly] by a fleshy scarlet [white] aril; seed coat crustaceous [coriaceous]; endosperm fleshy [or wanting]; embryo with the radicle inferior. Type species: M. boaria Molina. (Derived from maytén, the vernacular Chilean name of the type species. ) A pantropical genus of over 200 species. In accordance with a broad concept of the genus (cf. Hou), Maytenus is here delimited to include Gymnosporia (Wight & Arn.) Benth. & Hook. and Moya Griseb. The subgeneric classification of the inclusive genus has not yet been established, but all the American species (except four formerly included in Moya) belong to Maytenus proper, i.e., to subg. MAyTENus. The primarily South American sect. TRICERMA (Liebm.) Loes., including seven species, is represented in our area by Maytenus phyllanthoides Benth., a shrub or small tree with rather small fleshy-leathery obovate to oblanceolate leaves, elsewhere known from Mexico (Baja California and Sonora to Yucatan and Puebla), Central America (?), coastal southwestern Texas, and Cuba. It occurs in the hammocks and coastal sand dunes on the Florida Keys, northward along the coasts of peninsular Florida to Levy County (Sea- horse Key) on the west, and Palm Beach County on the east. Little is known concerning the floral biology, pollination, seed germina- tion, and other biological features of either our species or the genus in general. Birds seem to be dispersal agents. The only chromosome number reported is 2n = 80 for Maytenus Vitis-Idaea Griseb. (Argentina, Uru- guay, Paraguay, and Bolivia), a close relative of M. phyllanthoides. The genus differs from Celastrus mainly in the nonscandent habit, the flowers largely in compact axillary fascicles or solitary, and the ovary + immersed in and fused at least at base with the disc REFERENCES: Under family references see Hou (pp. 238-243), LoEsENER (1942, pp. 134- 146), Merz (pp. 341, 342), SarcENT (pp. 676, 677), and West & ARNOLD (p. 126). HorrMan, A.. & J. KuMMERow. Anatomical, morphological and physiological aspects of the germination of seeds of Maytenus boaria. (In Spanish; English summary.) Phyton Buenos Aires 18: 51-56. 62 Larsste, A. M., & C. H. WHarton. Northern extensions in the recorded ranges of plants on Seahorse and associated keys, Levy County, Florida. Quart. Jour. Fla. Acad. Sci. 22: 105-113. 1959. [M. phyllanthoides, 110.] SMALL. J. K. Maytenus phyllanthoides. Addisonia 13: 39, 40. pl. 436. 1928. 220 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XLV Subfam. CASSINOIDEAE Loes. Tribe CASSINEAE 4. re aaa P. Browne, Civ. Nat. Hist. Jamaica 145. pl. 17. fig. 1. 756. feats evergreen shrubs or small trees. Leaves opposite, rarely some [or all] alternate or whorled, chartaceous to thin-coriaceous, entire, or more often serrulate to serrulate: crenate or spinose-dentate [rarely rudi- mentary and caducous], short-petiolate; stipules minute, subulate or rudimentary (punctiform). Inflorescences short [rarely rather ample | axillary, simple or compound (once to four times forked) dichasial cymes. Flowers very small, hypogynous, bisexual, 2-bracteolate at the base of the slender pedicels. Sepals 4, connate into a 4-lobed calyx, persistent. Petals red or purplish | ereeniah or white], spreading to reflexed. Stamens 4, distinct, inserted between the lobes of the disc; filaments subulate, short; anthers small, suborbicular in outline, introrsely or laterally dehiscent. Nectariferous disc + annular to low- cupuliform, usually + 4-lobed. Gynoecium [2]|4-carpellate; stigmas [2] 4, narrow, spreading; style usually short [rarely wanting]; ovary superior, [2]4-locular; ovule | in each locule, ascendent from the base of the axile placenta, apotropous. Fruit a red drupe, usually obliquely obovoid, 3-6 mm. long, excentrically apiculate by the persistent style, 1(2)-seeded; mesocarp fleshy, thin; endocarp bony or crustaceous. Seed small, obovoid, exarillate; seed coat membranaceous, pale brown, obsoletely minutely reticulate, with a branched raphe; endosperm fleshy, usually copious [sparse to wanting] ; embryo spatulate; cotyledons ovate or suborbicular-ovate; radicle inferior. (Rhacoma L. Syst. Nat. ed. 10. 2: 896. 1759, nom. illeg.; a proposal to conserve this almost universally used name has been rejected. Including Myginda Jacq.) Type sprectrs: Rhacoma Crossopetalum L. = Crosso- petalum Rhacoma Crantz. (Name from Greek, crossos, fringe, and petalon, petal, i.e., referring to the fimbriate petals of the type species. ) A tropical American genus of nearly 25 species, best represented in the West Indies and Central America, a few species in northern South America, and two in the southernmost part of our area. The West Indian Crossopetalum Rhacoma (C. austrinum J. R. Gardn., Rhacoma Crossopetalum L.), a shrub or small tree with relatively small obovate or elliptic, usually serrulate-crenate leaves, occurs in hammocks and on coastal sand dunes on the Florida Keys and in southernmost peninsular Florida (Dade County). Crossopetalum ilicifolium (Poir.) Kuntze (C. floridanum J. R. Gardn., Myginda ilicifolia Poir., Rhacoma ilicifolia (Poir.) Trel.), a prostrate shrub with coriaceous spiny-toothed leaves 10-15(20) mm. long, occurs in pinelands in the same areas of subtropical Florida ° and beyond in the Bahamas and Hispaniola. A recent record (Hayes) of this species from a locality in the Coastal oo of eastern North Carolina is apparently based either on an introduced plant on a 1964 | BRIZICKY, GENERA OF CELASTRALES 221 The genus appears to be closely related to Cassine L. (Elaeodendron Jacq. f.) and more remotely allied to Gyminda Sargent. REFERENCES: Under family references see LOESENER (1942, pp. 183-185), METz (pp. 380, 381), URBAN (pp. 69-79), and West & ARNOLD (p. 127). Hayes, D. W. Two remarkable range extensions (Rhabdadenia corallicola and Rhacoma ilicifolia). Castanea 11: 61, 62. 1946. [Both collected by W. A. Dayton in acid soil at Blackland Branch Station near Wenona, Washington C0563 5. Gyminda Sargent, Garden Forest 4: 4. 1891. Dioecious glabrous evergreen shrubs or small trees. Leaves opposite, thin-coriaceous, entire or crenulate [serrate], short-petiolate; stipules minute, membranaceous, lanceolate to linear-lanceolate, caducous. Inflo- rescences short, simple to compound (1—4 times forked), peduncled axillary cymes. Flowers very small, hypogynous, unisexual by abortion, minutely 2-bracteolate just below the calyx, sessile. Sepals 4, nearly distinct, per- sistent. Petals 4, white, longer than the sepals, spreading or reflexed. Stamens 4, alternipetalous, distinct, inserted between the lobes of the disc, wanting in ? flowers; filaments subulate; anthers suborbicular in outline, laterally-introrsely dehiscent. Disc fleshy, 4-lobed in ¢, crenu- late, adnate to the base of ovary in 9 flowers. Gynoecium 2(3)-carpellate, rudimentary in ¢ flowers; stigmata 2, ample, commissural, peltate, nearly semiorbicular, sessile, persistent on fruit; ovary superior, 2(3)-locular; ovule 1 in each locule, pendulous from the top of axile placenta, anatro- pous, apotropous. Fruit an ellipsoidal to obovoid drupe, 6-8 mm. long, black or dark blue, crowned with persistent stigmata, 2(3)-locular (1 locule often rudimentary), (1)2(3)-seeded; mesocarp fleshy, thin; stone bony. Seed ellipsoidal or cylindric-ellipsoidal, exarillate; seed coat mem- branaceous, finely foveolate-reticulate; endosperm fleshy, thin; embryo spatulate, the cotyledons ovate or oblong, the radicle subcylindrical, superior. Type SPECIES: G. latifolia (Sw.) Urb. (Name an anagram of Myginda, to which this plant had been referred.) A genus of two or three (?) species of subtropical North America, the West Indies, and Central America south to Costa Rica. The West Indian Gyminda latifolia (G. Grisebachii Sarg.), false-boxwood, occurs in ham- mocks on the Florida Keys. Records of this species from Mexico may perhaps refer to the closely related Central American G. Tonduzii Loes. The latter seems to be rather variable in the shape and size of leaves, and some specimens (forms?) with smaller and more rounded leaves (than those usually seen) can easily be confused with G. /atifolia, as often also are sterile (and rarely flowering) specimens of Crossopetalum Rhacoma. The genus is most closely related to and perhaps congeneric with the mistake in labeling. At present there hardly is a reason to regard C. ilicifoltum as either native to or naturalized in North Carolina 222 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLV monotypic Jamaican Tetrasiphon Urb., which differs from Gyminda only in its four-carpellate gynoecium, four oblong-linear stigmata, and fibrous innermost layers of mesocarp. The genus also shows a relationship to Maurocenia Mill., Cassine L. (Elaeodendron Jacq. f.), and Crossopetalum. It differs from Cassine and Crossopetalum in its pendulous ovules and, if combined with either of these, would be an anomalous group with regard to this characteristic. REFERENCES: Under family references see LOESENER ee p. 182), Metz (pp. 377, 378), SARGENT (pp. 678, 679), URBAN (pp. 80-83), and West & ARNOLD (p. 127). SARGENT, C. S., Gyminda. Silva N. Am. 2: 13, 14. pl. 54. 1891. 6. Schaefferia Jacquin, Enum. Syst. Pl. Ins. Carib. 10. 1760. Dioecious, evergreen, usually glabrous shrubs or trees. Leaves alternate, + coriaceous, entire [rarely toothed], short-petiolate; stipules subulate or rudimentary. Flowers small, hypogynous, unisexual by abortion, mani- festly pediceled, in axillary few-flowered fascicles or solitary. Sepals 4, nearly distinct, persistent. Petals 4, much longer than the sepals, yellow- ish or whitish, spreading or reflexed, cochleate-imbricate. Stamens 4, hypogynous, shorter than the petals, wanting in @ flowers; filaments subulate; anthers broad-elliptic-subquadrangular, dorsifixed near base, hardly versatile, introrsely dehiscent. Nectariferous disc indistinct and adnate to the base of ovary in @, wanting in ¢ flowers. Gynoecium 2-carpellate, rudimentary in 4 aowenn stigmata 2, commissural, large, usually 2-lobed to -fid, spreading or reflexed; style very short, stoutish, or wanting; ovary superior, sessile, 2-locular; ovule 1 (2) in each locule, ascendent from the base of the axile placenta, apotropous. Fruit a small, scarlet, subglobular, thin-fleshed drupe, nearly 5-6 mm. in diameter, indistinctly 2-lobed, crowned by the persistent style and/or bases of stigmata; stones (pyrenes, “nutlets”) 2, or 1 by abortion, somewhat obovoid, plano-convex, with rough + tuberculate cartilaginous endocarp. Seed ellipsoidal or ovoid, plano-convex, exarillate; seed coat membra- naceous, smooth, brownish; endosperm copious; embryo spatulate, the radicle inferior. Typr species: S. frutescens Jacq. (Named for Jacob Christian Schaeffer, 1718-1790, distinguished German naturalist and mycologist. ) A tropical American genus of nine species, ranging from Peru, northern Argentina, and Uruguay, northward to the West Indies, southernmost Florida, western Texas, and northern Mexico; one species in the sub- tropical part of our area. Schaefferia frutescens, Florida boxwood, yellow-wood, distributed in the West Indies, Venezuela and Colombia, and northward in Central America to southern Mexico, occurs in hammocks on the Florida Keys and in extreme southeastern peninsular Florida. According to Sargent (1891), 1964 | BRIZICKY, GENERA OF CELASTRALES 223 the ripe fruits possess an acrid disagreeable flavor but are greedily eaten by many birds. Schaefferia cuneifolia Gray, a small shrub, occurs in the arid region of western Texas and northern Mexico. Loesener (1942) assumed Schaefferia to be closely related to Crosso- petalum but to approach Aquifoliaceae in its discless staminate flowers. The genus also resembles the latter family in the general structure of the fruits (convergent evolution?). REFERENCES: Under family references see LoesENER (1942, p. 189), Merz (pp. 383, 384), SARGENT (pp. 689, 690), URBAN (pp. 84-87), and West & ARNOLD (p. 128). Recorp, S. J., & G. A. Garrat. Boxwoods. VIII. Florida boxwood of the family aeupod a School Forestry Bull. 14: 75-77. 1925. [S. frutescens, wood an | SARGENT, C. S. daar Silva N. Am. 2: 15-18. pl. 55. 1891. Be gc ica . L. de Jussieu, Ann. ea ae Nat. Paris 18: 486, . “Hippocraticeae,” nom (H1ppocrRATEA FAMILY) Woody vines with slender, scandent branches |or shrubs, or slender trees], often with + anomalous wood, frequently with latex ducts in all organs [or only in the axis], differing from Celastraceae mainly in the almost exclusively opposite leaves; extrastaminal nectariferous disc; usually 3-merous androecium; usually extrorse anthers often dehiscing by transversally confluent slits; occurrence of pollen tetrads in some taxa; usually 6—8[2-20]-ovulate locules; pleurotropous ovules; and frequent occurrence of peculiar, usually 3-locular, deeply 3-parted capsules with large, dorsiventrally flattened, radially divergent segments containing endospermless seeds winged from the base. Type GENuS: Hippocratea L. A pantropical family of about 21 genera with nearly 400 species. It is centered and almost equally represented (with regard to the genera) in America and Africa, with a few taxa in Asia and Australasia. Only Hippo- cratea is represented in the subtropical part of our area. No data on floral biology or anatomy are available; those on embryology pertain to six genera. Pollen has been studied in 44 species of 14 genera (mostly African, three Afro-American, one pantropical). Chromosome counts for ten species in seven genera (two Afro-American, one pantropi- cal) invariably have been 2m = 28 in the genera with berry- or drupelike fruits (e.g., Salacia L., Salacighia Loes., of subfam. Salacioideae Hallé), and 2m = 56 in the genera with capsular fruits (e.g., Campylostemona Welw.; Hippocratea, sensu stricto; and Loeseneriella A. C. Sm., of subfam. Hippocrateoideae). Polyploidy apparently has been of importance in the evolutionary development of the family The peculiar shape of the capsules of the members of subfam. Hippo- crateoideae is due to a disproportionate growth of the ovary after fertiliza- 224 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLV tion. While the size of the axis of the ovary and the lower placentiferous parts of the locules remains almost unchanged during the process of fruit development, the tips of the locules expand greatly, forming segments [or lobes] up to 50-100 times longer than the length of the central placenti- ferous part of the mature fruit. Since the basal placentiferous parts o the locules (including the funicles) remain permanently united with the persistent fruit axis after dehiscence of the expanded locular tips, such a fruit is to be classified as a somewhat modified capsule, rather than as an apocarpous or schizocarpous fruit. The taxonomy of the Hippocrateaceae has been based primarily on the gross morphology of the reproductive organs, on some characters of the inflorescence, and, in some cases, on palynological data. Differences in evaluation of these characters have resulted in varying numbers of recog- nized genera, e.g., nine (including Campylostemon and Cheiloclinium Miers [placed in Celastraceae] in Loesener, 1942) or 21 (Smith, 1940, 1941, 1945; Wilczek, 1960; Hallé, 1962). The generic concepts of the last three authors are accepted here, but it is likely that additional data on floral anatomy, embryology, cytology, etc., will create a basis for a con- siderable reduction in the number of genera. This very natural group is closely related to and sometimes included in (most recently by Blakelock) the Celastraceae, which name is conserved over Hippocrateaceae when the two are united. A relationship to Siphono- dontaceae and Stackhousiaceae is also presumed. REFERENCES: BLAKELocK, R. A. Celastraceae. Jn: Hutcuinson, J., & J. M. Dauziet, Fl. W. Trop. Afr. ed. 2. 1: 623-634. 1958. [Hippocrateaceae included in Celas- traceae, 626-634; Hippocratea treated in a broad sens CANDOLLE, A. P. pE. Hippocrateaceae. Prodr. 1: 567-572. 1824. Davip. E. Embryologische Untersuchungen an Myoporaceen, Salvadoraceen, Sapindaceen und Hippocrateaceen. Planta 28: 680-703. 1938. |Hippo- crateaceae, 700-702. Fritscu, F. E. Untersuchungen uber das Vorkommen von Kautschuk bei den Hippocrateaceen, verbunden mit einer anatomisch-systematischen Unter- suchung von Blatt und Axe bei derselben Familie. Beih. Bot. Centralbl. 11: 283-358. 1 pl. 1902. [Campylostemon, 1 sp.; “Hippocratea,” 23 spp.; Salacia, 44 spp. | HALLE, N. Monographie des Hippocratéacées d’Afrique Occidentale. Mém. Inst. Franc. Afr. Noire 64: 1-245. 1962. [Fifteen genera: 9 African, 3 Afro-American, 2 Afro-Asiatic, 1 pantropical. | LoEsENER. T. Hippocrateaceae. Nat. Pflanzenfam. III. 5: 222-230. 1896. —. Hippocrateaceae. /bid. ed. 2. 20b: 198-231. 1942. MANGENoT. S.. & G. MANGENOT. Nombres chromosomiques nouveaux chez diverses dicotylédones et monocotylédones d’Afrique Occidentale. Bull. Jard. Bot. Bruxelles 27: 639-654. 1957. [Hippocrateaceae, 645.] Mavritzon,. J. Embryologische Angaben liber Stackhousiaceae, Hippocrateaceae und lcacinaceae. Sv. Bot. Tidskr. 30: 541-550. 1936. [Six genera of Hip- pocrateaceae, 544-547. | 1964] BRIZICKY, GENERA OF CELASTRALES 225 Miers, J. On the Hippocrateaceae of South America. Trans. Linn. Soc. 28: 319-432. pls. 16-32. 1872 OxaTon, M. Les lianes ligneuses 4 structure anormale des foréts denses d’Afrique Occidentale. Thése. 220 pp. Paris. 1960. [Hippocrateaceae, 107-119.] Paver, J. B. Traité d’organogénie comparée de la fleur. 2 vols. Masson, Paris. 1857. decreas 161-166, pl. 35.) Recorp, S. J. The American woods of the orders Celastrales, sere and Santalales. ey Woods 53: 11-38. 1938. [Hippocrateaceae, 22, _ W. Hess. Timbers of the New World. xv + 640 pp. pls. oa Smitu, A. C. The American species of Hippocrateaceae. Brittonia 3: 341-555. 1940. [Twelve genera. | Notes on Old World Hippocrateaceae. Am. Jour. Bot. 28: 438-443. 1941. panes -Pacific and Papuan genera and spp. ; n Hippocrateaceae in southeastern Asia. Jour. Arnold Arb. 26: 169-179. "4045. & I. W. Battey. Brassiantha, a new genus of Hippocrateaceae from New Guinea. Jour. Arnold Arb. 22: 389-394. pl. 1. 1941. [Includes a general survey of gross morphology and wood anatomy of Hippocrateaceae and Celastraceae. | STENZEL, G. Anatomie der Laubblatter und Stémme der Celastraceae und Hip- pocrateaceae. Inaug.-diss. 91 pp. Breslau. 1892(?). [Hippocrateaceae, 84-88. ] Takutajan, A. Die Evolution der Angiospermen. viii + 344 pp. Jena. 1959. [ Hippocrateaceae, 241, 242. Van Campo, M., & N. HAtieé. Les pollens des Hippocratéacées d’Afrique de rail aa Spores 1: 191-272 [193-272, pls. 11-50]. 1959. [Pollen in 14 gen WILCZEK, R. teasers Fl. Congo Belge 9: 133-232. 1960. 1. Hippocratea Linnaeus, Sp. Pl. 2: 1191. 1753; Gen. Pl. ed. 5. 498. 1754. Latex-bearing vines with scandent, usually opposite branches. Leaves opposite, membranaceous to subcoriaceous, petiolate, persistent; stipules minute, subulate, caducous. Inflorescences axillary, corymbose-paniculate, dichotomously branched, peduncled cymes, usually with 1 or 2 short, few- flowered branchlets in each dichotomy, with minute bracts and bractlets. Flowers small, regular, hypogynous, bisexual, pediceled. Sepals 5, connate at least at base, imbricate [or valvate]. Petals 5, distinct, yellowish [or whitish], thickish, much longer than the sepals, spreading, with a transverse tomentulose band below the apex within [or tomentulose except at the base], narrowly imbricate [or valvate]. Disc extrastaminal, conspicuous, fleshy, annular-pulvinate, subcylindrical or truncate-conical [or + cup shaped], puberulent [or glabrous] without. Stamens 3, suberect, spread- guous at base, often adnate to the upper margin of the disc; anthers basifixed, oblate-spheroidal, dehiscing by horizontal (transverse) extrorse- apical slits; pollen 3- or 4-porate, in compound tetrads (polyads) of 16 grains. Gynoecium syncarpous, 3-carpellate; stigma small, punctiform or 226 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV inconspicuously 3-lobed; style short, subulate; ovary 3-locular, 3-lobed, sessile, completely immersed in and fused with the disc, at least at the base; ovules usually 6[—8] in a locule, 2-ranked, + ascendent, anatropous. Fruit a 3-locular, deeply 3-parted capsule, with divergent, dorsiventrally (vertically) flattened segments (the greatly expanded tips of the locules), each dehiscing loculicidally along an inconspicuous median suture; peri- carp membranaceous to thin-coriaceous. Seeds 5 or 6 in each capsule segment, each with a basal wing jointed at the base with a fleshy, nearly obdeltoid funicle, the embryoniferous apical portion of seed coriaceous, flattened, the wing membranaceous, with a marginal and a submedian longitudinal nerve; endosperm wanting; embryo straight; cotyledons large, distinct, fleshy, flat; radicle minute, inferior. Type species: H. volubilis L. (Named for Hippocrates, 460-377 B.C., eminent Greek physician, “Father of Medicine.’’) A genus of one American and two African species. The tropical American Hippocratea volubilis (H. scandens Jacq., H. ovata Lam. occurs in hammocks and mangrove swamps in southernmost peninsular Florida (Dade, Monroe, Collier, and Lee counties) and on the Florida Keys, in the West Indies, and from central Mexico, south to Peru, Bolivia, northern Argentina (Misiones), Paraguay, and southern Brazil. Within its extensive range the species exhibits numerous variations on which many supposedly distinct species have been based (see A.C. Smith, 1940, for a complete synonymy). The genus has often been delimited in a broad sense (cf. Loesener, 1942: Blakelock, 1958) to include all the species (over 100) of Hippocrateaceae with very similar capsular fruits and winged seeds (except Campvylostemon Welw. and Tristemonanthus Loes., both with introrse stamens). In 1940, however, A. C. Smith restricted Hippocratea to the type species. Recently two African species (H. myriantha Oliver and H. Vignei Hoyle) have been included in Hippocratea, sensu stricto (see Hallé, 1962). Thus delimited, Hippocratea differs from closely related segregate genera (e.g., Pristimera Miers, Prionistemma Miers, Elachyptera : Sm., Reissantia Hallé, Loeseneriella A. C. Sm.) both in some floral characters (e.g., petals tomen- tulose within in Hippocratea, configuration of the disc) and in the com- pound pollen tetrads and the presence of latex in all parts of the plant (a questionable generic character in this family). In general, the characters distinguishing Hippocratea, sensu stricto, from the closely related genera are of such a nature that they may eventually be regarded as subgeneric, in view of the rather uniform anatomy of vegetative organs and the gross morphology of fruits and seeds (cf. Aceraceae). REFERENCES: Under family references see BLAKELOcCK (pp. 626-629), FritscH (pp. 288- 300, 302-310, 336-344), HALLE (pp. 123-131), LoESENER (1942, pp. 206-216), OBATON (pp. 108, 109), Recorp (pp. 22, 23), Recorp & Hess (p. 191). SMITH (1940, pp. 356-367), and VAN Campo & HALLE, (pls. 35, 36 [pp. 242-245] ). 1964] BRIZICKY, GENERA OF CELASTRALES 227 AQUIFOLIACEAE Bartling, Ord. Nat. Pl. 228, 376. 1830, nom. cons. (HoLty FAMILy) Shrubs or trees, differing from Celastraceae mainly in the almost ex- clusively alternate leaves, the usually sympetalous rotate corolla, the absence of a nectariferous disc (but cf. Schaefferia) ; the almost exclusively one-ovulate locules, the always pendulous one-integumented ovules, and the minute “rudimentary” embryo in the vicinity of the micropyle. TYPE cENus: Aquifolium Mill., nom. illeg. = Ilex L. A family of three genera: the primarily pantropical polytypic /lex; the monotypic Nemopanthus Raf., endemic in northeastern North America; and the tropical Phelline Labill. (ten species), endemic in New Caledonia. Only Jlex is represented in our area; Nemopanthus mucronatus (L.) Trel. occurs from Newfoundland to Minnesota, southward to northern Illinois, Indiana, Ohio, and upland West Virginia and Virginia. Aquifoliaceae exhibit a number of primitive characters (e.g., the minute “rudimentary” embryo; the exclusively scalariform perforation plates in vessels), as well as specialized features (e.g., the often sympetalous corolla; the absence of a nectariferous disc, apparently due to abortion; the solitary, one-integumented ovules). A relationship to Celastraceae (espe- cially to subfam. Cassinoideae) has generally been assumed, and the evidence from gross morphology of flowers and fruits, wood anatomy, and embryology seems to support this view. Pollen morphology does not furnish positive indications in favor of such a relationship (Erdtman, 55), but it should be noted that only two genera of subfam. Cassinoideae, the group apparently closest to Aquifoliaceae, have been investigated. REFERENCES: Battton, H. Monographie des Labiées, Verbénacées, Ericacées et Tlicacées. Hist. Pl. 11: 1-220. 1891. [Aquifoliaceae, 211-220. Bentuam, G., & J. D. Hooker. Ilicineae. Gen. Pl. 1: 355-357. 1862. CANDOLLE, A. P. pe. Celastrineae. Prodr. 2: 2-18. 1825. [ Aquifoliaceae, 13- 18.] ErptMaNn, G. Pollen morphology and plant taxonomy. Angiosperms. xii + 539 pp. Stockholm. 1952. [Aquifoliaceae, 54, 55. Herr, J. M., Jr. Embryological evidence for the relationship of Aquifoliaceae to Celastraceae. (Abstr.) Va. Jour. Sci. II. 10; 259. 1959. [See also under Ilex. | LoresENER, T. Vorstudien zu einer Monographie der Aquifoliaceen. Inaug. Diss. 45 pp. pl. 1. Berlin. 1890 _ Monographia Aquifoliacearum. Pars I. Nova Acta Acad. Leop.-Carol. 78: 1-598. pls. 1-15. 1901; Pars II. Jbid. 89: 1-313. maps 1-3. 1908. ———. Uber die Aquifoliaceen, besonders iiber //ex. Mitt. Deutsch. Dendrol. Ges. 28: 1-66. pls. 1-4. 1919. . Aquifoliaceae. Nat. Pflanzenfam. ed. 2. 20b: 36-86. 1942. PENNINGTON, M. J. A comparative study of wood anatomy of fifty-four species of the family Aquifoliaceae. Thesis (unpublished). Univ. Va., Charlottes- ville. 1953.* 228 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLV Recorp, S. J. The American woods of the orders Celastrales, Olacales, and Santalales. Trop. Woods 53: 11-38. 1938. [Aquifoliaceae, 12-14.] SARGENT, C. S. Manual of the trees of North America (exclusive of Mexico). ed. 2. xxvi + 910 pp. map. Boston & New York. 1922. [Aquifoliaceae, 668-674. | TRELEASE, W. Revision of North American Ilicineae and Celastraceae. Trans. Acad. Sci. St. Louis 5: 343-357. 1892. [Aquifoliaceae, 343-349. ] 1. Ilex Linnaeus, Sp. Pl. 1: 125. 1753; Gen. Pl. ed. 5. 60. 1754. Dioecious evergreen, or more rarely deciduous, shrubs or trees of acid, often wet soils. Leaves alternate [exceptionally opposite], simple, thick- coriaceous to membranaceous, entire, crenate, serrate, or spiny toothed, petioled; stipules minute, usually caducous, more rarely persistent. In- florescences axillary to leaves or bracts on either long- or short-shoots, single or fascicled 1- to many-flowered dichasial cymes, sometimes umbel- like and/or panicled [or racemes], or flowers solitary or in fascicles. Flowers small, hypogynous, unisexual by abortion. Sepals 4-9, connate in a 4—9-lobed (to -parted) calyx, imbricate, persistent. Petals 4-9, [rarely distinct] usually connate into a sympetalous rotate corolla, usually white, greenish-white or -yellow [rarely pink, red, or lavender], imbricate. Stamens 4~9, alternipetalous, usually adnate at base to the short corolla tube, more rarely distinct, nearly as long as the petals in ¢ , abortive and sterile in ¢ flowers; filaments subulate or straplike; anthers ovoid to ellipsoidal, introrse, dorsifixed, not versatile, dehiscing by 2 lateral longi- tudinal slits, 2-locular at anthesis; pollen 3-colp(oroid)ate, spheroidal to subprolate (or prolate?), medium sized, coarsely and densely granular on the surface. Gynoecium syncarpous, 4—9|—22]-carpellate, rudimentary in ¢ flowers; stigma mostly sessile or subsessile, large, capitate or discoid [or columnar], mostly + lobed, persistent in fruit; style wanting or very short; ovary superior, 4-9[—22]-locular; ovules 1 (occasionally 2) in each locule, pendulous from the top of the axile placenta, anatropous, apotropous, 1l-integumented, with a thickish nucellus; funiculus often with a prominent protuberance on the entire adaxial surface. Fruit a subglobular to ellipsoidal berry-like drupe, usually red, more rarely dark purple to black, occasionally yellow to whitish, containing [2]4-9[-22] unilocular stones (pyrenes, “nutlets”) shaped like citrus-fruit segments [exceptionally a single 4-locular stone]; exocarp chartaceous; mesocarp fleshy or sometimes + coriaceous because of stone cells interspersed in its outer layers; endocarp bony or woody, rarely coriaceous, the back (and sides) of stones either completely smooth and even, or more often longitu- dinally striate and/or ridged and grooved, sometimes sparsely irregularly reticulate. Seed solitary (occasionally 2) in a stone, pendulous; seed coat membranaceous; endosperm copious, fleshy: embryo minute, obversely heart shaped, situated near the micropyle. (Including Prinos L.) Lrcto- TYPE SPECIES: /. Aquifolium L.; see N. L. Britton, N. Am. Trees. 621. 1908. (Classical Latin name of the Mediterranean holly oak, Quercus Ilex L., applied by Linnaeus to this genus.) — Ho tty. 1964] BRIZICKY, GENERA OF CELASTRALES 229 A primarily pantropical genus of 300-400 species, centered in eastern Asia (southern and southwestern China and Indochina) and South America (Brazil), with one species in tropical Africa, one in Australia, and one in Europe; about 14 species (all represented with us) occur in eastern North America. Since the genus is currently being studied by several workers and its classification is still in flux, the relatively simple but at least partly unnatural one of Asa Gray is followed here (Gray, Man. eds. 1-5. 1848-1869; Fernald, Gray’s Man. Bot. ed. 8. 980-982. 1950).° Subgenus ILEx (subg. Euilex Loes., subg. Aquifolium (DC.) Gray). Leaves persistent, thin- to thick-coriaceous, sometimes spiny toothed; short-shoots lacking; carpellate flowers usually 4-merous; drupes red; stones (pyrenes) longitudinally (2)3—5-striate (sometimes faintly so) and/or ridged and grooved or irregularly thinly reticulated on the back. The largest and most widely distributed subgenus, it is represented with us by Jlex opaca Ait. var. opaca, American holly, 2m = 36, and var. arenicola (Ashe) Ashe (J. arenicola te I. cumulicola Small), scrub holly; J. vomitoria Ait. var. vomitor on or cassena, 2m = 40; I. Cassine L., dahoon holly; and /. ees Walt. The principal ieee nomic problems revolve around the status of J. opaca var. arenicola, of the sand scrub of peninsular Florida, which has been treated as a distinct species (correctly named /. arenicola), and that of the relationships of I. Cassine and I. myrtifolia, which have sometimes been regarded as varieties of a single species. //ex vomitoria is notable for the caffeine con- tent of the leaves, a concentrated infusion of which was used by the Indians of the Coastal Plain in a stimulating and emetic “black drink.” The species occurs disjunctly in Chiapas and Vera Cruz, Mexico, as the pubescent var. chiapensis Sharp. Ilex %< attenuata Ashe, a presumed natural hybrid between 7. Cassine and J. opaca, has been recorded from northwestern Florida and both Carolinas. Subgenus Prinos (L.) Gray, Man. Bot. No. U. S. ed. 2. 264. 1856 (Prinos L.; including subg. Euilex Loes. sect. Cassinoides Loes. Nat. Pflanzenfam. Nachtr. 218. 1897, in part). LrcTOTYPE SPECIES: Prinos verticillata L. = Ilex verticillata (L.) Gray; see Hitchcock & Greene, Int. Bot. Congr. Cambridge 1930. Nomencl. Propos. Brit. Bot. 147. 1929. Leaves deciduous or persistent; carpellate flowers (4)6—8(9)-merous; stones smooth and even, rarely indistinctly 1-striate on the back. Largely a temperate eastern Asiatic-North American group of about 12 species, extending with one or two species into the West Indies and Central America. Five species, comprising three morphological groups, occur in our area. Jlex verticillata (L.) Gray, winterberry, black alder, or fever bush, 2” = 36, and /. laevigata (Pursh) Gray, smooth winterberry, are deciduous shrubs or small trees with red (rarely yellow) drupes. By contrast, J. glabra (L.) Gray, inkberry, gallberry, or bitter gallberry; ° Following Gray’s classification, Fernald aire his subgeneric names for those of sections (cf. Brizicky, Jour. Arnold Arb. 44: , footnote 3. 1963). For much more elaborate classifications a Ilex, see Loesener ie. 1942) and Hu (1949-1950; and in Dengler, 1957, pp. 4). 230 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV I. coriacea (Pursh) Chapm., large or sweet gallberry; and the West Indian I. Krugiana Loes., which reaches Dade County, Florida, are all black- fruited evergreen species placed by Loesener in subg. ILEx. The last species differs, however, from the former two in its epunctate leaves, four-merous flowers, and fascicled inflorescences and seems to represent a distinct, primarily South American group (sect. Microdontae Loes.). Although the three groups have many similarities in the morphology of flowers and fruits, their relationships are highly questionable. Jlex glabra and /. coriacea are similar in many characteristics and are often confused, although they appear to be distinct species. /lex glabra is regarded by foresters as an undesirable plant in pinewoods, but it seems to be an important source of honey in the Gulf region. Subgenus Prinoipes (DC.) Gray, Man. Bot. No. U. S. 276. 1848 (subg. Prinos sect. Prinoides (DC.) Gray ex Loes. Nat. Pflanzenfam. Nachtr. 221. 1897). Lectotype species: /lex prinoides Ait. = I. decidua Walt.). Leaves deciduous, membranaceous to subcoriaceous, often crowded on short spurs (short-shoots); carpellate flowers 4—5(6)-merous, often crowded on short-shoots; drupes red (occasionally yellow); stones as in subg. ILex. A temperate eastern Asiatic-North American group of about 12 species, five or six in our area. The subgenus seems to be closely re- lated to and is often included in subg. Prinos (cf. deciduous, red-fruited species). Our species include /lex montana Torr. & Gray var. montana (J. monticola Gray), mountain winterberry or mountain holly, 2” = 40, and the pubescent-leaved var. mollis (Gray) Britton (including J. Beadlei Ashe ex Kearney); /. decidua Walt. var. decidua, possum-haw, 2n = 40, and a small-leaved Florida variant, var. Curtissii Fern. (J. Curtissii (Fern.) Small); J. ambigua (Michx.) Torr., Carolina holly; J. longipes Chapm. (including /. collina E. J. Alex., fide Edwin), Georgia holly; and the very rare 1, Amelanchier M. A. Curtis (J. dubia (G. Don) BSP., 1888; not Weber, 1851, pl. fossil.), sarvis holly, of sandy swamps from Louisiana to Georgia and southeastern Virginia. Jlex Cuthbertii Small apparently should be included in /. decidua, while 7. Buswellii Small, found in ham- mocks along the Caloosahatchee River in southwestern Florida, may per- haps be a small-leaved form of J. ambigua. Ilex longipes is remarkable for its fruits with pedicels 1.2—2.2 cm. long. It is notable that several eastern North American species have counter- parts in eastern Asia, e.g., 7. montana—I. Tsoi Merr. & Chun; J. decidua— I, aculeolata Nakai; I. longipes—I. asprella (Hook. & Arn.) Chapm.; /. glabra—I. yunnanensis Franch.; /. verticillata—/. serrata Thunb. var. Sie- boldii (Miq.) Rehder (Hu, 1949, p. 255). It seems likely that diversification of inflorescence types in J/ex has generally been due to the reduction and simplification of an original ample compound dichasial cyme (or cymes) along several divergent or convergent lines, rather than to the elaboration of a simple dichasium. Solitary or fascicled single-flowered inflorescences and solitary or fascicled flowers seem to be the final products of lines of reduction. Within a species, the carpellate inflorescences usually are somewhat more reduced than the 1964 | BRIZICKY, GENERA OF CELASTRALES 231 staminate ones (sexual dimorphism), e.g., fascicled simple staminate dichasia, but fascicled single-flowered carpellate inflorescences in J. Krugiana; solitary 3—7-flowered dichasial staminate cymes, but solitary single-flowered carpellate inflorescences in J. glabra; fascicled staminate flowers, but solitary carpellate flowers in J. coriacea. No correlation be- tween the types of inflorescence and the fruit (pyrene) characters has yet been demonstrated. Since the species of the genus usually are dioecious, although the occa- sional occurrence of bisexual flowers has been recorded in J. Aquifolium, cross-pollination and -fertilization are the rule. Insects, especially bees, have been recorded as pollinators, but wind apparently also is a factor in pollination in some species. The nectar is said to be excreted from small papillose swellings found at the base or near the middle of the petals on their adaxial (upper) surface (Loesener, 1942, p. 45). Although a single axile ovule in each locule is typical, the occasional occurrence of two-ovulate locules has been recorded. Herr (1959) demon- strated that the two-ovulate condition is due to the presence of a second parietal rudimentary ovule (“parietal protuberance”) in the ovary locule which maintains a potentiality to develop. Also, the “funicular pro- tuberance,”’ considered by Herr to be a partially suppressed nonfunctional ovule, can sometimes develop into a regular ovule. Chromosome counts made in ten species (2” = 36, 40) possibly indicate polyploidy, with the basic numbers supposedly « = 9, 10. Ilex * attenuata is the only wild interspecific hybrid known in eastern North America; the occurrence of many forms intermediate between the species of Jlex in eastern Asia, especially in south and southwest China (Hu, 1949, p. 254) may suggest more extensive hybridization in that region. Many artificial and/or natural hybrids of garden and nursery origin are known at present. Most are within subg. ILEx and involve J. Aquifolium as one of the par- ents, but a few are in subg. Prrnos (sensu Gray), e.g., 7. glabra X serrata, I. crenata < glabra. The established hybrid 7. Aquifolium yunnanensis, and the suspected J. chinensis < glabra are intersubgeneric (according to the classification used here) or intersectional (according to that of Hu). Birds are considered to be the seed dispersing agents. Parthenocarpy seems to be common in /. laevigata and J. cornuta Lindl., less so in J. vomitoria, but apparently is of wider distribution, since a natural tendency to parthenocarpy has also been found in several other species (e.g., J. Aguifolium, I. Cassine, I. myrtifolia). The rudimentary embryo of Jlex undergoes maturation after dispersal but prior to germination (Herr. 1962, p. 32), which usually occurs in nature from one to three years after dispersal. The percentage of germination is low, and as a rule, the seeds are very difficult to germinate. Many foreign and indigenous species are grown as ornamentals (e.g.. the European /. Aquifolium with its numerous varieties and hybrids, the eastern Asiatic /. cornuta and J. crenata, the domestic J. opaca). The berried branches of J. Aquifolium and I. opaca are in great demand for Christmas decoration, and those of J. laevigata and J. verticillata are 232 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLV sometimes sold for this purpose. Leaves of the South American J. para- guartensis St. Hil. (and possibly of a few other closely related species) yield the maté or Paraguay tea which is much used in South America. The wood of some arborescent species is valued for lathe work, engraving, and cabinet-making. REFERENCES: The large number of references has been reduced here primarily to those either of general interest or dealing specifically with the southeastern United States. For an extensive list of references see B. F. THomson, Bibliography on holly. Bull. Holly Soc. Am. 8: 1-36. 1955. Under family references see ERDTMAN (p. 54), LoESENER (1890; 1901 & 1908; 1919; and 1942, pp. 53-83), PENNINGTON, REcorD (pp. 12-14), and SARGENT (pp. 668 8-674). ALEXANDER, E. J. Jlex decidua, Addisonia 11: 47, 48. pl. 376. 1926. : 0 new species from the southern Appalachians. Castanea 6: 30-32. 1941. [/. collina, 30, 31. See also F. Woops, Castanea 16: 126, 127. 1951, and G. Epwin, Rhodora 59: 20-23. 1957, regarding the status of this sp.] Arston, A. H. G., & R. E. Scuuttes. Studies of early specimens and reports of Ilex vomitoria, Rhodora 53: 273-279. 1951. [See also R. E. ScHuLTEs, The correct name of the yaupon. Bot. Mus. Leafl. 14: 97-105. pls. 24-26. 1950. | Ammons, N., & E. L. Core. The hollies of West Virginia. Castanea 10: 57- AsHwortH, R. P. Investigations into midvein anatomy and ontogeny of certain species of the genus /lex L. Jour. Elisha Mitchell Sci. Soc. 79: 126-138. 1963. [Fourteen (?) spp.; unilacunar nodes prevalent, trilacunar in J. opaca; see also Diss. Abstr. 21: 1721. 1961.* Barrett, R. E. Germinating seeds of the American holly. Holly Soc. Am. Proc. 28: 7, 8. 1960.* Barton, L. V., & N. C. THornToN. Germination and sex population studies of Ilex opaca Ait. Contr. Boyce Thompson Inst. 14: 405-410. 1947. BICKNELL, E. P. Jlex verticillata. Addisonia 3: 71, 72. pl. 116. 1918. | See also W. J. Bean, Bot. Mag. 146: pl. 8832. 1920.] Box, C. O. A study of the time of flower bud initiation in Jlex cornuta burfordi. (Abstr.) Assoc. So. Agr. Workers Proc. 52: 130, 131. 1955.* [See also W. F. Kosar, Natl. Hort. Mag. 37: 110, 111, 1958; * J. L. Frierson, Am. Nurseryman 107(9): 9, 56. 1958.* Brooks, A. B. A new holly for West Virginia. Castanea 1: 83-85. 1936. [T. longipes.] new form of /lex longipes Chapman. /bid. 5: 15, 16. 1940. [Forma Vantrompii Brooks, with bright golden-yellow fruits. ] Burton, G. W., & R. H. Hucues. Effects of burning and 2,4,5-T on gallberry Ilex glabra| and saw-palmetto [Serenoa repens|. Jour. Forestry 59: 497- 500. 1. Capor, L. Anatomische Untersuchung der Mateblatter unter Berucksichtigung ihres Gehaltes an Thein. Bot. Centralbl. 84: 241-251, 275-283, 309-315, 340-345, 369-374. 1900. [Leaf anatomy of 19 spp. of Nex) CuiLpers, J. T., & W. E. Snyper. The effect of time of taking cuttings on the rooting of three cultivars of American holly (Jlex opaca Ait.). Proc. Am Soc. Hort. Sci. 70: 445-450. 1957. Connors, C. H. Parthenocarpy in holly. Proc. Am. Soc. Hort. Sci. 63: 453-456. 1964 | BRIZICKY, GENERA OF CELASTRALES 233 1954. [See also F. E. Garpner & P. C. Martu, Bot. Gaz. 99: 184-192. 1937.] DENGLER, H. ed. Handbook of hollies. Natl. Hort. Mag. 36(1): 1-193. 1957. oh especially F. C. Gatte, North American hollies, 11-31; S. Y. Hv, Oriental Hollies, 31-64; W. F. Kosar, Hybridizing hollies, 121-129.] Focc, J. M., Jr. The deciduous fioiies Morris Arb. Bull. 11: 59-63. 1960. Foret, J. A., & S. L. Sotymosy. A new variant of Jlex vomitoria. Baileya 8: 83. 1960. [Forma pendula Foret & Solymosy. | FrIERSON, J. L. Some chromosome counts in the hollies. Holly Soc. Am. Proc. 30: 5. 1961.* [See also Diss. Abstr. 21: 288. 1960.*] Gusa, E. F., & J. A. STEVENSON. Fungus and nematode inhabitants and diseases of holly (Ilex). Exp. Sta. Coll. Agr. Univ. Mass. Bull. 530. 43 pp. 1963. Herr, J. M. The development of the ovule and megagametophyte in the genus Ilex L. Jour. Elisha Mitchell Sci. Soc. 75: 107-128. 1959. [Includes dis- cussion of the relationship of Aquifoliaceae to Celastraceae. | . Endosperm development and associated ovule modification in the genus /lex L. Ibid. 77: 26-32. 1961. [See also ASB Bull. 7: 29. 1960. | ————. Maturation of the embryo in //ex crenata Thunb. (Abstr.) ASB Bull. 9: 32, 33. 1962. Hotty Society or America. Preliminary holly check list. Bull. Holly Soc. Am. 6: 1-56. 1953 Hom, T. Anatomy of certain species of Jlex of the sections Aguifolium and Prinos. Am. Jour. Sci. V. 18: 497-504. 1929. [J. opaca, I. glabra, I. verticillata, I. laevigata. | Hv, S. Y. The genus //ex in China. Jour. Arnold Arb. 30: 233-344, 348-387. 1949; 31: 39-80, 214-263. 1950. [Monograph. ] Hume, H. H. Sern oe native in the United States. Natl. Hort. Mag. 26(3): 143-179. . Holly pyrenes ean, Bull. Holly Soc. Am. 10: 1-16. 1959. Ives, S. A. Maturation and germination of seeds of /lex opaca. Bot. Gaz, 76: 60-77. 1923. Jensen, H. W. Heterochromosome formation in the genus /lex. Am. Nat. 78: 375-379. 1944 Littte, E. L., Jr. Miscellaneous notes on nomenclature of United States trees. Am. Midl. Nat. 33: 495-513. 1945. [Jlex, 497, 498; see also his check list, U.S. Dep. Agr. Handb. 41: 207-212. 1953.] LUNDELL, C. L. ig Cay DC. Fl. Tex. 3: 112-122. 1943. McFartin, J. B. Two new evergreen hollies from central Florida. Rhodora 34: 16-18. 1932. “TSee also ibid. 233-236. pls. 224-231.] McMenamin, J. P. Observations on the stomatal structure of //ex opaca. Proc. Indiana Acad. Sci. 52: 58-61. 1943. MELL, C. D. The early uses of the yaupon. Am. Forestry 28: 531. 1922. [I. aceon REcorp, S. J. me Jlex woods without spiral elements. Trop. Woods 8: 9. 1926. [See Ee ibid, 3: 14. 1925. Roserts, A. N., & C. A. BoLLer. eae ae requirements of ui holly, Ilex aquifolium. Proc. Am. Soc. Hort. Sci. 52: 501-509. 19 SarGENT, C. S. Jlex Amelanchier. Garden Forest 2: 40. fig. 88 a 41]. 1889. . Llex laevigata. Ibid. 4: 220, 221. 1891. . Ilex. Silva N. Am. 1: 103-116. pls. 45-50. 1891. 234 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLY ScHUruHorrF, P. N. Die oe von J/lex aquifolium. Ber. Deutsch. Bot. Ges. 39: 377-379. | Embryology SHARP, A. J. A new variety of Jlex ae from southern Mexico. Bot. Mus. Leafl. 14: 107, 108. 1950. SMALL, J. K. flee myrtifolia. Addisonia 17: 39, 40. pl. 564. 1932 Tryon, E. H., & R. W. Pease. Shading effects of natural canopies on holly characteristics. Castanea 18: 70-83. 1953. West, E., & L. E. Arnotp. The native trees of Florida. xx + 212 pp. Gaines- ville. 1946. [Jlex, 118-125.] Woops, F. W. The genus //ex in Tennessee. Rhodora 53: 229-240. 1951. [Five spp. Wyman, D. Ilex crenata and its varieties. Arnoldia 20: 41-46. 1960. 1964] GRAHAM, GENERA OF LYTHRACEAE 2a5 THE GENERA OF LYTHRACEAE IN THE SOUTHEASTERN UNITED STATES ! SHIRLEY A. GRAHAM LYTHRACEAE Jaume St.-Hilaire, Expos. Fam. 2: 175. 1805, ‘“Lythrariae,” nom. cons (LOOSESTRIFE FAMILY) Herbs or shrubs [rarely small trees], with quadrangulate or terete stems. Leaves opposite, seldom alternate or whorled, simple, entire, exstipulate. Flowers regular or irregular, axillary or in terminal racemes stamen length, bisexual, [3]4—6-merous perigynous, the flowers rarely cleistogamous; bracteoles 2, opposite on the pedicel, rarely none. Floral tube campanulate to tubular, persistent, often conspicuously nerved; calyx lobes 4—6, alternating with 3—5 deltoid appendages, or the appendages wanting. Petals 0-6, distinct, crumpled, deciduous [rarely persistent], inserted on the inner surface of the floral tube between the calyx lobes. Stamens as many as or twice as many as the petals [or more numerous], often alternately unequal, inserted on the inner surface of the floral tube below the petals, included to exserted; anthers versatile [rarely basifixed], introrse, 2-locular, longitudinally dehiscent. Gynoecium syncarpous, often surrounded at the base by a hypogynous disc, or in irregular flowers the disc on the upper (adaxial) side only; stigma capitate, rarely bilobed; style filiform; ovary superior, free in the floral tube, 2—4[-6]-locular, the septa complete or reduced; placentation axile [rarely parietal]; ovules anatropous and epitropous, with 2 integuments. Fruit a membranaceous [or leathery] capsule in- closed by the persistent floral tube, septicidally or loculicidally dehiscent or indehiscent, splitting irregularly. Seeds 3 to many, minute or up to Ca. 4 mm. long, pyramidal, ovoid, or discoid, sometimes slightly winged, with ‘Prepared for a 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 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). The area covered in this, as in former treatments, is bounded by and includes North Carolina, Tennessee, Arkansas, and Louisiana. Material included in the descriptions in brackets applies to species outside this area, and references marked by an asterisk have not been seen by the author. The aid of Dr. Wood and the observations and suggestions of Drs. H. E. Ahles, R. K. Godfrey, D. B. Ward, and C. A. Brown are gratefully acknowledged. Dr. R. B. Channell kindly supplied living material of Cuphea for study. 236 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLV little or no endosperm; embryo straight with a short radicle, the cotyle- dons flat [rarely convolute|, often auriculate-cordate. Embryo sac de- velopment of the normal (Polygonum) type (rarely intermediate between Polygonum and Oenothera type); endosperm development of the nuclear type. (Salicariae Juss., 1789.) Typr Genus: Lythrum L. A family of about 24 genera and 450 species, most abundant in the New World tropics and subtropics, but present in temperate to tropical areas of Europe, Asia, Africa, and Australia; six genera represented in our area. Koehne (1903) divided the family into two tribes based on the condi- tion of the septa in the ovary. In tribe Lythreae the septa are interrupted or split above the placenta, which is then discontinuous with the style; in Nesaeeae Koehne the septa are complete and the placenta is continuous with the style. This character is variable and difficult to ascertain in many specimens. If tribal divisions are to be used at all, a firmer basis for their delimitation is needed. Similarities in anatomy and embryology closely link the Lythraceae with the Myrtaceae, Onagraceae, Punicaceae, and Sonneratiaceae (see Mauritzon, 1939). The Lythraceae are distinguished from these families by the superior ovary which is free in the floral tube. The family is char- acterized anatomically by intraxylary phloem which forms bicollateral bundles in the leaves. Unique, spirally grooved, inverted hairs are found in the epidermal layer of the seed coat in several genera (e.g., Cuphea, Peplis, and some species of Lythrum). Upon contact with water, the hairs emerge through the outer wall of an epidermal cell and slowly uncoil to a length of two or three millimeters. The hairs are slightly mucilaginous and may serve in dispersal by attaching the seed to passing objects. Heteromorphism is widespread in the family. Lythrum, Decodon, and Nesaea have species with trimorphic flowers, and species with dimorphic flowers are known in Lythrum, Rotala, Pemphis, and Nesaea. Cleisto- gamous flowers have been reported in Ammannia and are thought to occur in the apetalous species of Rotala, Peplis, and Nesaea. Fossil remains of the Lythraceae are known from the Eocene through the Pleistocene. At least six genera of fossil seeds belonging to the Lythraceae have been described. All are well suited to aquatic dispersal by means of winglike expansions of the seed coat and/or buoyant spongy tissue on the outer seed coat. The aquatic or semiaquatic habit of many Lythraceae is undoubtedly an ancient characteristic, still present, but less prevalent in modern species. Two genera, Lawsonia and Lagerstroemia, are cultivated in warm- temperate and tropical regions of the world. Lawsonia inermis L., native to northeastern Africa, is the henna of commerce, yielding an orange-red dye which has been used for centuries in the Near and Far East for color- ing the hair, fingernails, and soles of the feet. The leaves contain hydroxynaphthaquinone which reacts directly with the keratin of human hair and skin to form the bright pigment. 1964 | GRAHAM, GENERA OF LYTHRACEAE 237 The crape-myrtle, Lagerstroemia indica L., 2n = 48, 50, a native of Asia, is a very popular shrub or tree widely cultivated throughout our southern states. It is showy in bloom, producing great numbers of magenta flowers. A number of horticultural varieties bearing purple, pink, or white flowers are planted. The trees are long lived and may persist where planted long after dwellings have disappeared. For this reason, they may appear to be naturalized but are usually only remnants of cultivation and not true escapes. One possible exception is the presence of several young plants in the sparsely inhabited sand-hill country of Berkeley County, South Carolina (H. E. Ahles 52982, Ncu). Collectors of this species should note the circumstances under which the plants are growing and record any evidence of reproduction from seed. REFERENCES: BaILLon, H. Lythrariacées. Hist. Pl. 6: 426-457. 1877. BENTHAM, G., & J. D. Hooker. Lythrarieae. Gen. Pl. 1: 773-785. 1867. CaNDOLLE, A. P. pe. Lythrarieae. Prodr. 3: 75-94. 1828. eric E. B. Daily growth movements of Lagerstroemia. Philip. Jour. Sci. ot. 8: 287-298. 1913. BO C. B. The hundred days’ flower. re history of the crepe myrtle. La. Conserv. Rev. 6(1): 21-23, 25. GIN, ee clas sur les Lythracées: his ia. Mat. Méd. Paris 6: 1-166. Prete W. ... Beitrage zur anatomie der Myrtifloren mit besonderer Beriicksichtigung der Lythraceae .. . Inaug.-diss. 39 pp. Universitat Breslau. 1905.* Harris, J. A. Variation and correlation in the flowers of Lagerstroemia indica. Missouri Bot. Gard. Rep. 20: 97-104. 1909. . On a chemical peculiarity of the dimorphic anthers of Lagerstroemia indica, with a suggestion as to its ecological significance. Ann. 499-507. 1914. [Suggests presence of a chemical substance in the yellow, but not the red, anthers which keeps the pollen moist, so it can be readily gathered by insect visitors. ] Josu1, A. C. Embryological evidence for the relationships of the Lythraceae and related families. Curr. Sci. Bangalore 8: 112, 113. 1939.* [Lythraceae and allied families not closely related to Caryophyllaceae. & J. VENKATESWARLU. Embryological studies in the Lythraceae. I-III. Proc. Indian Acad. Sci. B. 2: 481-493, 523, 524. 1935. (See also ibid. 3: 377-400. 1936.*) [Lawsonia and Lagerstroemia investigated. KOEHNE, Zz Pgs Lythraceae of the United States. Bot. Gaz. 10: 269-277. pl. 6 Oe Ge Pflanzenreich IV. 216(Heft 17): 1-326. 1903. ee oe Bot. Jahrb. 41(2): 74-110. 1907; ibid. Mavritz0N, i Zur Pa ater einiger Lythraceen. Acta Horti Gothob. 9: 7a ee, Cetin to the embryology of the orders Rosales and Myrtales. Lunds Univ. Arsskr. II. Sect. 2. 35: 1-121. 1939. [Lythraceae, 64-68.] Repcrove, H. S. The cultivation and uses of henna. Gard. Chron. III. 85: 13, 14. 1929 238 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLV SAHNI, B. Indian silicified plants. 2. Enigmocarpon parijai, a silicified fruit from the Deccan, with a review of the fossil history of the Lythraceae. Proc. Indian Acad. Sci. B. 17: 59-96. 1943 ScHocH-BopMer, H., F. Buxpaum, & W. WANGERIN. Lythraceae. Jn: O KIRCHNER, E. Loew, & C. SCHROTER, Lebensgeschichte der Bliitenpflanzen Mitteleuropas. 3(5): 1-128. 1937. [Includes extensive bibliography. | SpracuE, T. A., & C. R. METCALFE. The taxonomic position of Rhynchocalyx. Kew Bull. 1937: 392-394. 1937. [Closely related to Lawsonia and wrongly excluded from the Lythraceae by Koehne. | KrEY TO THE GENERA OF LYTHRACEAE General characters: Leaves entire, mostly opposite; floral tube campanulate to cylindrical, persistent; ovary superior, free in the floral tube ; petals crumpled; fruit capsular; seeds without endosperm, the embryo straight. A. Floral tube campanulate to globose, about as long as wide. B. Plant woody, perennial, over 1 m. ta C. Flowers in axillary dichasia; stamens 8-10; shrubs of shallow water. be Reg yew tebvemetes Gaganes sen eenaepwesnan be went ecodon. C. Flowers in terminal panicles: stamens numerous, more than 10; terrestrial trees or shrubs. .......... | Lagerstroemuia. | B. Plant herbaceous, annual, generally less than 50 cm. ‘tall. D. Appendages in the sinuses of the calyx lobes present; capsule dehis- cent; flowers 1 to many in the axils of leaves; petals 4 (rarely 0). E. Middle and upper leaves attenuate at base; capsule dehiscing septicidally, the outer wall of eines finely and densely striate; flowers solitary in the axils of leaves. ............ 2. Rotala. E. Middle and upper leaves cordate auriculate at base; capsule dehiscing irregularly, the outer wall of the capsule smooth, not striate; flowers (1—)3 to many in the axils of leaves. ........ hed RSA ate Rd dP ae Lae a BS a ea ee 3. Ammannia. D. ig ela in the sinuses of the calyx lobes none; capsule seri flowers solitary in the axils of leaves; petals . 4. Peplis. A. Floral ae ginwienty elongate, about twice as long a as wide. Flowers regular; floral tube entire in fruit; capsule dehiscing septicidally from the apex, the placenta included; seeds numerous, mostly more 20. 5. Lythrum. F. Flowers irregular; floral tube and capsule splitting longitudinally along the upper (adaxial) side in fruit, the placenta then exserted; — 3-20. LeGbyne et edge eba pune cea euanaaeaheeediag head eeewaws Cuphea. 1. Decodon J. F. Gmelin, Linn. Syst. Nat. ed. 13. 2(1): 677. 1791. Short-lived glabrous or pubescent perennial shrubs or tall herbs of marshy habitats, spreading by means of arching branches rooting at the tips; stems terete to 6-angled, the submerged portions thickened by spongy aerenchyma. Leaves membranaceous to leathery, decussate or verticillate, lanceolate, shortly petiolate, the base and apex acute. Inflorescence of 1-3 shortly pedunculate axillary dichasia at a node, the peduncles partly fused with the petiole. Flowers regular, 4- or 5-merous, trimorphic; 1964 | GRAHAM, GENERA OF LYTHRACEAE 239 bracteoles 2, ovate to linear, opposite, at the base of the pedicel, early deciduous. Floral tube campanulate, greenish, inconspicuously nerved; calyx lobes 4 or 5(—7), the apex acuminate; appendages narrowly es angular, thickened, mostly twice the length of (rarely equal to) the calyx lobes. Petals 4 or 5(—7), rose-purple, deciduous, about twice the length of the floral tube. Stamens 8—10, deciduous; filaments of 3 possible lengths, two of the three lengths occurring in any one flower and alternating in a single whorl; anthers orbicular in outline. Gynoecium sessile, the disc wanting; stigma capitate; style thin, included to exserted, of 3 possible lengths; ovary 3(4)-locular, the septa complete. Fruit a loculicidally dehiscent capsule. Seeds ca. 20-30, small, inverted-pyramidal with rounded edges, an oval germination valve covered by a spongy epidermal layer present on the adaxial side. Typr species: D. verticillatus (L.) Ell. (Name from Greek, deca, ten, and odous, tooth, in reference to the five sepals and five appendages of the floral tube.) — SWAMP LOOSESTRIFE, WATER-OLEANDER, WATER-WILLOW. A monotypic genus now occurring only in the eastern United States from Louisiana to Florida and north to Minnesota and central Maine. The glabrous var. /aevigatus Torr. & Gray has been recognized in the northern part of the range. In the Mio-Pliocene the genus was repre- sented by at least two other species from western Europe and central Russia. Fossil seeds of these suggest they were very closely related, but not identical, to the extant Decodon verticillatus. Silicified fruit remains from the Eocene of India may represent either Decodon or an extinct genus similar to it. Extensive aerenchyma is produced in concentric layers on submerged stems. The tissue is composed of large thin-walled parenchyma supported by radial columns of cells and by additional strands of elongated cells which obliquely cross several layers to strengthen the tissue. The meri- stem which produces aerenchyma develops in a narrow zone of paren- chyma between the phloem fibers and sieve tube area. Some species of Ammannia and Lythrum in aquatic situations also develop aerenchyma, but to a lesser degree than in Decodon. Endlicher subordinated Decodon to a section of Nesaea, but most sub- sequent authors have regarded it as distinct. REFERENCES: Under family references see KoEHNE (1885, 1903). KoeHNE, E. Ueber Inflorescenz und Trimorphismus von Decodon (Nesaea) verticillatus, einer nordamerikanischen Lythracee. Verh. Bot. Ver. Brandenb. 16: 42, 43. 1874. NIkiTIN, P. A. The systematic position of the fossil genus Diclidocarya E. M. Reid. (With a note by Mrs. E. M. Reid.) Jour. Bot. 67: 33-38. pl. 589. 1929. ia of three spp. = Decodon.] SCHRENK, J. the floating tissue of Nesaea verticillata (L.) H.B.K. Bull. Torrey a Club 16: 315-323. pls. 95-97. 1889. 240 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV 2. Rotala Linnaeus, Mant. Pl. Alt. 175. 1771. Annual [rarely perennial], glabrous herbs of aquatic or marshy habitats. Leaves membranaceous, decussate, verticillate, or alternate, linear to oblanceolate [to ovate], attenuate to the base, sessile to short petioled. Flowers regular, solitary and axillary [or in terminal spikes], [3]4—6- merous, [seldom dimorphic, | sessile, with 2 opposite bracteoles at the base of the floral tube. Floral tube campanulate to globose or urceolate, 2.5—5 mm. long, greenish; calyx lobes 4[—6], the apex acute to acuminate; appendages shorter than to exceeding the lobes [or wanting]. Petals [0 or| 4[—6], white or pink, small, scarcely exceeding the calyx lobes, deci- duous [rarely persistent]. Stamens 4[—6], included [to exserted]. Gy- noecium without a disc at the proximal end; stigma capitate; style very short or wanting; ovary incompletely 3- or 4[-6]-locular, the upper portion of each septum + incomplete. Fruit a membranaceous, septicidally dehiscent capsule with dense, transverse striations on the outer wall, | 2-] 3- or 4-valved. Seeds many, minute, ovoid in outline, plano-convex. TYPE SPECIES: R. verticillaris L. (Name from Latin, rota, wheel, in allusion to the whorled leaves of the type species.) —- TooTH-cuP A genus of about 45 species, best represented in Asia and Africa. Of the three North American species, only Rotala ramosior (L.) Koehne, 2n = 32, is present in our area, occurring in all the Southeastern States. The more robust plants occurring inland from the Atlantic Coastal Plain have been recognized as var. interior Fern. & Grisc. The species closely resembles Ammannia in general habit and was originally described in that genus. Bentham and Hooker considered Rotala synonymous with Am- mannia, but the close resemblance to Ammannia is found only in R. ramosior;, the genera can be distinguished easily on the basis of at least three characters. Rotala has solitary axillary flowers, a septicidally dehiscent capsule, and the outer wall of the capsule densely striate, a character present in no other genus of the Lythraceae. In Ammannia the flowers occur in axillary cymes, the capsule dehisces irregularly, and the outer wall of the capsule is smooth. REFERENCES: Under family references see KOEHNE (1885, 1903). FERNALD, M. & L. Griscom. Three days of botanizing in southeastern Virginia. Rhodora 37: 129-157, 167-189. pls. 332-351. 1935. [R. ramosior var. interior, 169.] Lewis, W. H., H. L. Striptinc, & R. G. Ross. Chromosome numbers for some angiosperms of the southern United States and Mexico. Rhodora 64: 147- 161. 1962. [R. ramosior, 152.] 3. Ammannia Linnaeus, Sp. Pl. 1: 119. 1753; Gen. Pl. ed. 5. 55. 1754. Annual glabrous herbs of aquatic or marshy habitats. Leaves mem- branaceous, decussate, linear to lanceolate or oblanceolate, sessile, the base cordate to auriculate, rarely attenuate. Flowers regular, 4(5)-merous, 1964 | GRAHAM, GENERA OF LYTHRACEAE 241 never heteromorphic, in sessile or pedunculate axillary cymes, (1—)3-15 flowers at a node; bracteoles 2, linear, opposite, at the base of the floral tube. Floral tube campanulate to urceolate, becoming globose in fruit, 1.5—6 mm. long, greenish to rose colored, 8-nerved with 4 nerves especially prominent at anthesis; calyx lobes 4 (5), short and broad; appendages thick, shorter than to equaling the calyx lobes [or wanting]. Petals 0-4, small, pink [white], early deciduous. Stamens 4(-8), included to exserted. Gynoecium without a disc at the base; stigma capitate; style thin, longer than the ovary and exserted, or thick, shorter than the ovary and included: ovary incompletely 2—4(5)-locular, the upper portion of the septa + in- complete. Fruit a membranaceous, irregularly dehiscent capsule, with the outer wall smooth, not striate. Seeds many, minute, ovoid. aera SPECIES: A. latifolia L.; see Britton & Brown, Illus. Fl. No. U:S. ed. 2. 2: 577. 1913. (Named e honor of Paul Acraiann: 1634-1691, a on botanist from Leipzig, not Johan Ammann; see Linnaeus, Crit. Bot. 91. 1737.) A genus of world-wide occurrence, with about 20 species; represented in the southeastern United States by two to four species. Ammannia coccinea Rottb. is found throughout the eastern United States, in the western coastal states, and in Mexico, Central America, the Caribbean region, and northern South America. In Mississippi and elsewhere outside our area (e.g., Oklahoma, Missouri, and Indiana), intermediates occur between this and A, auriculata Willd., which is generally found west of the Mississippi River. Ammannia auriculata differs from A. coccinea mainly in having more densely clustered, smaller, long-pedicellate flowers. A third species, A, teres Raf., of swamps or fresh to brackish tidal marshes, ranges from New Jersey to Florida, Mississippi, and Texas. In southern Florida it is replaced by A. latifolia L., which extends southward through the Antilles into northern South America. The two are distinguished only by the lack of petals in A. latifolia and should probably be considered variants of a single species. Cleistogamous flowers have been reported in A. latifolia. A re-evaluation of specific limits in the genus is needed. REFERENCES: Under family references see KoEHNE (1903) and Mauritzon (1934, 1939). Jones, F. B. Ammannia teres Raf. (Lythraceae) in coastal Texas. Field Lab. 5. 1958. Josut, A. C., & J. VENKATESWARLU. Structure and development of the synergids in Ammannia baccifera Linn. New Phytol. 34: 144-150. pl. 3. 1935. [See also correction, ibid. 35: 92. 1936; developing endosperm mrongly inter- preted as syn-synergids. | a Peplis Linnaeus, Sp. Pl. 1: 332. 1753; Gen. Pl. ed. 5. 154. 1754. Delicate annual herbs growing along lake margins or submerged in shallow water. Leaves membranaceous, opposite or alternate, sessile, nar- rowly linear or oblanceolate, the base acute to obtuse. Flowers regular, 242 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV 4{—6]-merous, solitary, axillary on short pedicels or sessile; bracteoles 0 [or 2, opposite on the pedicel]. Floral tube broadly campanulate [to globose], 2-3 mm. long; calyx lobes 4|—6], the appendages wanting |or long-subulate]. Petals 0 [or 6, but early deciduous |. Stamens (2?) 4|-6], included. Gynoecium without a disc at the base; stigma capitate or slightly bilobed; style very short or wanting; ovary incompletely 2-locular, appearing 1-locular by reduction of the septum. Fruit a membranaceous, indehiscent capsule, splitting irregularly. Seeds many, small, dorsiventrally flattened, spathulate, the distal end slightly enlarged and curved. (Includ- ing Didiplis Raf.) Typr species: P. Portula L. (Ancient Greek name applied by Dioscorides to Euphorbia Peplis and by Pliny to “porcillaca or purslain,”’ Portulaca oleracea.) —WATER-PURSLANE. A European genus of eight species; represented in the United States by the endemic Peplis diandra Nutt. ex DC., which ranges from Florida to eastern Texas, northward to Minnesota, and to North Carolina. This species displays a terrestrial form having cuneate-lanceolate leaves and a short style, and an aquatic form which generally has longer, narrower, linear leaves, shorter internodes, and no style. The specific epithet diandra is perhaps misleading, being based on an atypical plant with only two, rather than the normal four, stamens. The separation of Peplis diandra by some authors as Didiplis diandra (DC.) Wood has been made on the basis of its four-merous, rather than six-merous, flowers and the lack of appendages on the floral tube. In view of the numerous other shared characters, e.g., irregularly splitting capsules with many seeds, campanulate floral tube, general plant habit, there seems to be no reason for retaining Didiplis as distinct from Peplis. Koehne considered Didiplis synonymous with Peplis but circumscribed subg. Dip1ptis for this one species. No insect visitors have been reported for the genus. The most wide- spread European species, Peplis Portula L., 2n = 10, is known to be self- pollinated, and the floral morphology suggests that this is probably the rule for the other species. The stigma is short and in a position to receive pollen from the introrse anthers lying above it in the floral tube, while insect-attracting petals and nectaries are wanting. Cleistogamy is suspected in the apetalous species. REFERENCES: Under family references see KozEHNE (1885, 1903). Korune, E. Genus-Recht der Gattung Peplis. Verh. Bot. Ver. Brandenb. 19: 47-53. 1877. Wixus, J. C., & I. H. BurKILy. Flowers and insects in Great Britain. Ann. Bot. 9: 227-273. 1895. [P. Portula, 266. | 5. Lythrum Linnaeus, Sp. Pl. 1: 446. 1753; Gen. Pl. ed. 5. 205. 1754. Annual or perennial herbs or shrubs of moist, often brackish habitats. Leaves membranaceous, decussate, alternate |or verticillate], ovate to 1964 | GRAHAM, GENERA OF LYTHRACEAE 243 linear [or obovate], sessile or shortly petiolate, the base attenuate to cordate. Flowers regular [or slightly irregular], (4—)6-merous, sometimes heteromorphic with 2 [3] floral forms, the flowers axillary [or in terminal spikes or racemes], 1 or 2 [to several] sessile or shortly pedicellate flowers at a node; bracteoles [0] 2, on the pedicel, linear, opposite. Floral tube cylindrical, greenish, 8-12-nerved, 4-8 mm. long; calyx lobes (4—)6, the apex acute to acuminate, the appendages narrowly triangular, shorter than to exceeding the calyx lobes. Petals (4—)6 [rarely 0], rose-purple or white, deciduous. Stamens (4—)6[-12] in 1 [2] whorls, inserted deep in the floral tube, included to exserted; filaments in heteromorphic forms of 2 [3] lengths. Gynoecium sessile or stipitate, with or without a hypogy- nous disc at the proximal end; stigma capitate; style thin, included to exserted, in heteromorphic forms of 2 [3] lengths; ovary 2-locular, the septum + incomplete. Fruit a membranaceous [rarely leathery] capsule, dehiscing septicidally or septifragally. Seeds [8 to] many, small, ovoid in outline. Lectotyre species: L. Salicaria L.; see Britton & Brown, Illus. Fl. No. U.S. ed. 2. 2: 580. 1913. (Name from Greek, lythron, gore or blood; thought to refer to the purple color of the petals.) — LoosEsTRIFE. A genus of about 30 species, with its greatest development in North America (ca. 16 species), but also occurring in Europe, Asia, Africa, and Australia. The North American species are badly in need of revision. Past treatments have depended heavily on variable characters of the leaf for delimitation of species. Hybridization among species in different parts of their range is suspected but has not been studied. Understanding of the breeding patterns and morphological variation seems necessary for a more satisfactory treatment of this group. The genus is represented in our area by six weakly defined species of sect. EUHYSSOPIFOLIA Koehne (plants grayish or bluish-green, calyces mostly tubular, flowers 4—6-merous, petals present). The most common and widespread indigenous Lythrum in the eastern United States is L. lanceolatum Ell., characterized by appendages exceeding the sepals and alternate, linear to elliptic leaves with tapering bases. Within its range is L. Curtissii Fern., reportedly endemic to southwestern Georgia and adjacent Gadsden County, Florida. This species, with slightly smaller flowers, larger stem leaves, and smaller branch leaves than L. lanceolatum, probably represents a local variant of L. lanceolatum. Lythrum lineare L.., named for its linear, strictly opposite leaves, is restricted to brackish marshes on or near the coast from New Jersey to Florida, and west along the gulf coast to eastern Texas. Specific lines separating L. alatum Pursh, 2n = 10, L. dacotanum Nieuw., and L. flagellare Shuttl. ex Chapm. are very tenuous. Misinterpretation of Z. alatum has increased difficulties in recognizing species within this group (see Shinners). Several species in North America (including Lythrum lanceolatum, L. lineare, and L. alatum) have two floral forms, an individual having either flowers with long styles and short stamens or short styles with long 244 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLV stamens. The plants are generally self-sterile and outbreeding. The most successful crosses occur between stamens and styles of corresponding lengths. Pollination is accomplished by insects. The most detailed studies on dimorphic flowers have been conducted on the genus Primula (Prim- ulaceae). The dimorphic Lythra have not been investigated. The occurrence of plants having three floral forms was first noted by Vaucher (1841) in Lythrum Salicaria L., 2n = 30, 50, 60, a Eurasian species long naturalized in Canada and the northeastern United States and now occurring southward at least to Virginia. The first extensive account of this phenomenon was given by Darwin (1865). In this species there are three style lengths and three sets of stamens of lengths corresponding to those of the styles. Each floral form has one style length and two whorls of stamens, one whorl for each of the other two possible lengths. The forms are termed long-, mid-, or short-styled, depending on whether the style exceeds, lies between, or is shorter than the two whorls of stamens. An individual plant bears only one of the three floral forms. Genetically, the long-styled form has been shown to be homozygous and recessive for long styles, while the short-styled form differs from long- and mid-styled by a single epistatic gene. The mid-styled flowers carry duplicate factors for “mid-” in the same linkage group, with the homozy- gous condition of either or both factors producing a lethal effect. Pollen differs in color, size, and amount of stored starch in each of the three stamen lengths. The longest stamens have the largest grains, the anthers are green, and the pollen grain is filled with starch. The two shorter stamen lengths have yellow anthers and correspondingly smaller pollen containing less starch. The two kinds of pollen in each floral form carry identical hereditary factors for the inheritance of trimorphism. In like manner, the stigmatic papillae vary in length with the length of the style, being longest on the longest style. The flowers are slightly irregular, with the stigma and anthers bent upward toward the open side of the flower. Insects of several different orders visit the species. They are able to pass only along the open side of the flower to reach the nectar at its base, and, in moving toward the nectaries, their underside brushes both the upturned stigma and the anthers. The pollen is transported on the ventral side of the insect mainly in three places, corresponding to the position of the three stamen lengths, and is consequently transferred to styles of those lengths, effecting cross- pollination. Crosses occurring between style and stamens of corresponding length are very fertile and are termed “legitimate” unions; conversely, those between style and stamens of unequal length (including self-pollinated flowers) are generally unsuccessful and are termed “illegitimate” unions. Exceptions are the successful crosses obtained between mid-styled plants pollinated by long stamens of the short-styled form and short stamens of the long-styled form. It is probable that pollen from more than one stamen length is dusted onto a style, but studies indicate that unidentified chemical 1964 | GRAHAM, GENERA OF LYTHRACEAE 245 growth inhibitors present in the style prevent the pollen of illegitimate unions from competing with that of legitimate unions. Products of most illegitimate crosses are highly sterile, often dwarfed, and rarely persist. Seedlings of mid-styled illegitimate crosses are, how: ever, more hardy and fertile than other illegitimate unions. Darwin found that although the mid-styled form produced seed by illegitimate as well as legitimate unions, the stamens of that form were less fertile than were those corresponding stamens of the other two forms, suggesting that in the mid-styled form there is a tendency toward functional dioecism. All three forms are believed to have been derived from a single homomorphic, self- compatible form. A semihomomorphic form has been reported in which the style and one whorl of stamens are the same length, being intermediate in position between long- and mid-styled forms. It has been suggested that heterostyly is an unstable state in Lythrum Salicaria, the presence of an intermediate form and the variable degrees of self-compatibility evi- denced by the three floral types tending to support this conclusion. Tri- morphism has been reported for only four other genera: Nesaea and Decodon, of the Lythraceae; Oxalis; and a single monocotyledonous genus, Pontederia. Fossil flowers of Lythrum have been found in the Lower Pliocene of Japan (see Miki). This determination is based on the flower’s being four- sepalate, a rare condition in modern species. Miki also found fossils of Trapa and Hemitrapa Miki (an extinct genus) and, on the basis of a comparative study, concluded that Trapa was derived from Lythrum through Hemitrapa. Other writers consider Trapaceae closely related to, and perhaps derived from, Onagraceae. Chromosome numbers of 2” = 10, 20, 30, 50, and 60 have been reported for the genus, but only one count has been published for the indigenous American species. REFERENCES: Under family references see KOEHNE (1885, 1903). ARARATIAN, A. G. Observations on heterogeny in Lythrum Salicaria. (In Rus- sian.) Akad. Nauk Armian. SSR. Dokl. 23(4): 187-192. 1956.* BarLow, N. Inheritance of the three forms in cae species. Jour. Genet. 13: 133-146. 1923. [Oxalis valdiviana and L. Salicaria.| BopMer, H. Beitrage zur Anatomie und Physiologie von Lythrum Salicaria.] Beih. Bot. Centralbl. 45: 1-58. 1928 Darwin, C. On the sexual relations of he three forms of Lythrum Salicaria. Jour. Linn. Soc. Bot. 8: 169-196. 1865 . The different forms of flowers on plants of the same species. xx + 352 p. London. 1892. East, E. M. The inheritance of heterostyly in Lythrum Salicaria. Genetics 12: 393-414. 1927. [See also Proc. Natl. Acad. Sci. 13: 122-124. 1927.] Further observations on Lythrum Salicaria. Genetics 17: 327-334. 1932 FERNALD, M. L. Some little-known plants from Florida and Georgia. Bot. Gaz. 33: 154-157. 1902. [L. Curtissii, sp. nov.] 246 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV FisHer, R. A. On the selective consequences of East’s (1927) theory of heterostylism in Lythrum. Jour. Genet. 30: 369-382. 1935 V. C. Martin. Spontaneous occurrence in Lythrum Salicaria of plants duplex for the short-style gene. Nature 160: 541. & K. Matuer. Polyploid inheritance in Lythrum Salicaria. Nature 150: 430. 1942 Fyre, V. C. Double reduction at the mid locus in Lythrum Salicaria. Heredity 7: 285-292. 1953. ad D. Pollen-tube growth in Lythrum Salicaria. Proc. Natl. Acad. Sci. 13: 253-255. 1927. LEHMANN, E. Keimungsversuch mit Samen von Lythrum Salicaria. Ber. Deutsch. Bot, Ges. 42: 55-60. 1925. [Effects of varying light and tempera- ture on seed germination. | Mrx1, S. Evolution of Trapa from ancestral Lythrum through Hemitrapa. Proc. Japan Acad. 35(6): 289-294. 9. Miter, H. Fertilisation of flowers. (Transl. by D. W. Thompson.) xii + 669 pp. London. 1883. [List of insect visitors to L. Salicaria, 255, 256. | Rao, L. sarin Untersuchungen iiber die Wirkung des Lichtes auf die se keimung von Lythrum Salicaria. Jahrb. Wiss. Bot. 64: 249-280. 19 ae a H. The influence of nutrition upon pollen grain size in Lythrum Salicaria, Jour. Genet. 40: 393-402 ———_.. Pollenbeschaffenheit ar elias bei Lythrum ‘Solicaria L. Bull. Soc. Bot. Suisse 52: 317-35 . ScnouTe, J. C. Ueber die Se oraisde der Heterostylie, insbesondere bei Lythrum Salicaria. Rec. Trav. Bot. Néerl. 25A: 271-340. ; SHinners, L. H. Synopsis of the United States species of Lythrum (Lythra- ceae). Field Lab. 21: 80-89. 1953 SovEces, R. Embryogénie des Lythracées. Développement de l’embryon chez le Lythrum Salicaria L. Compt. Rend. Acad. Sci. Paris 180: 1417, 1418. 1925. Stevens, N. E. Observations on heterostylous plants. Bot. Gaz. 53: 277-308. pls. 21-23, 1912. [Extensive bibliography. | StirLinc, J. Studies of flowering in heterostyled and allied species. IT. The Lythraceae: Lythrum Salicaria Linn. Publ. Hartley Bot Lab. Liverpool 10: 1-24. 1935; III. Gentianaceae, Lythraceae, Oxalidaceae. Ibid, 15: 1-24. 1936. Strout, A. B. Studies of Lythrum Salicaria. I. The efficiency of sel f-pollination. Am. Jour. Bot. 10: 440-449. 1923; II. A new form of flower in this species. Bull. Torrey Bot. Club 52: 81-85. 1925. Tatese, T. Physiological studies on the fertilization in Lythrum Salicaria Linn. I. Presence of pollen-germination inhibitors in the ovary. Bot. Mag. Tokyo 74: 291-295. 1961. TIscHLeR, G. Uber die Entwicklung und phylogenetische Bedeutung des Embryosackes von Lythrum Salicaria. Ber. Deutsch. Bot. Ges. 35: 233-246. 1917. imei related embrvologically to Onagraceae 2 suchungen iiber den anatomischen Bau der Staubund Frucht- blitter bei Lythrum Salicaria mit Beziehung auf das “ Tie emi Flora 111/112: 162-193. 1918. Vaucuer, J. P. Histoire physiologique des plantes d’Europe 2: 371. 1841.* [First published reference to tristyly; observed in L. Salicarta. | 1964 | GRAHAM, GENERA OF LYTHRACEAE 247 6. Cuphea P. Browne, Civ. Nat. Hist. Jamaica 216. 1756. Herbaceous or woody annuals or short-lived perennials, mostly with viscid, glandular hairs on stem, leaves, and flowers. Leaves membranaceous [or leathery], decussate or verticillate [rarely alternate], ovate to lan- ceolate, elliptic or linear, sessile or petiolate. Flowers irregular, 6-merous, homomorphic, sessile or pedicellate, 1-3 at a node: when alternate on the stem one flower at each node interaxillary, the others, if present, axillary: when opposite or verticillate on the stem all flowers interaxillary and often internodal; bracteoles 2, opposite, on the pedicel. Floral tube cylindrical, 4-10[-25] mm. long, green or purple, distinctly 12-nerved, the base gibbous or distinctly spurred, the spur curving downward toward |or up- ward away from] the pedicel; calyx lobes 6, deltoid, acute to acuminate [rarely apiculate] at apex, equal in size or the upper (adaxial) one larger than the others; appendages shorter than [to exceeding| the length of the calyx lobes. Petals 6 [4, 2, or 0], pale to deep purple, equal in size or the upper (adaxial) 2 or the lower (abaxial) 4 largest, early deciduous [rarely persistent]. Stamens (5—)1!1(—12), included to exserted, alter- nately unequal, the filaments often covered with dense hairs, the 2 upper (adaxial) stamens inserted deeper in the calyx tube than the others; pollen triangular or orbicular in outline in polar view, [2]3-colporate, the exine striate. Gynoecium with a disc at the base on the upper side only and free from the side of the ovary; stigma capitate, seldom bilobed; style included or exserted; ovary incompletely 2-locular, appearing 1-locular by reduction of the septum to 2 thin threads, the lower (abaxial) locule smaller, sometimes sterile. Fruit a membranaceous, loculicidally dehiscent capsule, splitting longitudinally the length of the adaxial wall of the ovary, the upper (adaxial) side of the persistent floral tube also splitting, the placenta projecting up and out of the capsule and floral tube. Seeds 3- 20|—numerous], orbicular to ovoid in outline, dorsiventrally flattened [rarely not flattened], the surface pebbled, dark brown [mottled green and brown, or light brown], often narrowly margined; cotyledons flattened, cordate to auriculate; radicle short. (Parsonsia P. Br.) Typr SPECIES: C. decandra Ait. (C. ciliata (Sw.) Koehne, 1881, not C. ciliata Ruiz & Pav., 1798). (Name from Greek, kuphos, hump, referring to the gibbous or spurred floral tube.) A genus of 200-250 species, in 13 sections, mostly restricted to the American tropics and subtropics; represented in our area by four species in three sections. The sections are very weakly delimited, and discovery of new species and accumulation of knowledge of variability in the species have further lessened their validity. The entire genus is in need of revision, with emphasis on a more natural arrangement of the species. The most common species of our area is the annual Cuphea viscosissima Jacq. (C. petiolata (L.) Koehne, 1881, not C. petiolata Pohl ex Koehne. 1877), tarweed or blue waxweed, an endemic to the United States, which ranges from northern Alabama and Georgia and western North and South 248 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLV Carolina, to central Arkansas, and northward to easternmost Kansas and Massachusetts. Reportedly a bad weed in pastures, old fields and gardens, it is avoided by livestock because of its dense, sticky, glandular hairs. It is most closely related to species of western and central Mexico. A second indigenous species, C. aspera Chapm. (Parsonsia lythroides Small), with opposite flowers, whorled leaves, pale-purple petals, and tuberous roots, is known only from the immediate vicinity of Port St. Joe, Florida. It is most closely related to C. Ayssopoides St.-Hil., of Brazil. The South American Cuphea glutinosa Cham. & Schlecht. was first collected in the United States in 1884, in Vermilion Parish, Louisiana, a general area to which it has remained restricted. The petals are charac- teristically very early deciduous. The presence of C. carthagenensis (Jacq.) Macbr. (C. balsamona Cham. & Schlecht.) along our coast from Louisiana to North Carolina often has been overlooked. It has been confused with the common C. viscosissima, in spite of its distinct geographical range and very different morphology (floral tube green vs. purple and green, 4.5-6 mm, vs. 8-10 mm. long, petals equal in size vs. upper (adaxial) petals largest, and stamens deeply vs. shallowly inserted in the tube). The species is native to South America but has spread as far west as the Fiji and Philippine islands and is undoubtedly a relatively recent introduction to the United States. Most early collections are misidentified as C. vis- cosissima. A showy-flowered Mexican species, C. procumbens Cav., has been reported from the Blue Ridge, North Carolina, and Andover and Springfield, Massachusetts. These plants were most likely garden escapes which would not have persisted. There is no evidence that the species is established in the eastern United States today. The genus is morphologically the most highly specialized in the family. It is well adapted to insect and hummingbird pollination by virtue of the long, irregular floral tube, the loss of a stamen in the honey entrance, and the introrsely directed anthers. Glandular-viscid bristles on the floral tube and stems discourage crawling insects from entering the flower, while a disc at the base of the flower narrows the passage to the nectar secreted by the walls of the spur, excluding all but long-tongued insects. Most species are apparently self-compatible but require an insect vector to accomplish pollination. The seeds emerge from the capsule and floral tube while green and remain on the placenta, exposed to the air, for a ripening period. The interaxillary position of the flowers of Cuphea is unique in the family. In the alternate-flowered species, a vegetative bud is present in the axil of one leaf and a floral bud in the axil of the other. The vegeta- tive bud may produce a flowering branch (sometimes highly reduced), or remain suppressed; the floral bud produces a single flower, the pedicel of which traverses and is fused with the internode above, the flower emerging just below or at the node above, between the leaves. In opposite-flowered species, both flowers at the node are interaxillary. Pollen morphology differs among some species and in these cases pro- vides a useful taxonomic character. One fossil pollen grain of Cuphea is known from the Miocene of Alabama. It is three-colporate but does not 1964] GRAHAM, GENERA OF LYTHRACEAE 249 belong to any of the species present in the United States today. Accord- ing to Erdtman, Cuphea is isolated from the other genera of the family on the basis of pollen morphology. Embryologically, Cuphea is quite distinct from Ammannia, Nesaea, Rotala, Peplis, and Lythrum. Its relationship to other genera in the family has not been investigated. Chromosome numbers of 2m = 12, 16, 18, 20, 24, 32, 36, 72, and ca. 84 have been found in the genus (Graham, unpub- lished). No counts have been reported for the species in our area. A few species are cultivated; the best known are Cuphea platycentra Lem. (C. ignea A. DC.) and C. procumbens Cav., commonly called cigar or firecracker plants, after the long, purple or red floral tube. REFERENCES: Under family references see KoEHNE (1885, 1903) and Mauritzon (1934, 1939). AHLES, H. E., C. R. BELL, & A. E. RApForp. Species new to the flora of North or South Carolina. ae 60: 10-32. 1958. [C. carthagenensis, 19.] ArpDAo, M. I. Contribucion al estudio de la Cuphea glutinosa Cham. & Schlechtd. y breve noticia de la Litraceas Uruguayas. I. Litraceas del Uruguay; II. Estudio botanico-quimico y farmacodinemico de la Cuphea glutinosa Cham. y Schlechtd. Arch. Soc. Biol. Montevideo 8: 173-201. 1938 BACIGALUPI, R. Taxonomic studies in Cuphea. Contr. Gray Herb. 95: 1-26. pls. 1-5. 1931 BarciaANnu, D. P. Ueber die Bliithenentwicklung der Cupheen. Mitt. Gesammt. Bot. 2: 179-193. pl. 11. 1875. Baum, H., & W. LEINFELLNER. Die Plazenta des dorsiventralen Cuphea- Gynozeums. Osterr. Bot. Zeitschr. 98: 187-205. 1951. [Correctly interpret the lower (abaxial) locule as reduced, contrary to Koehne’s view that it is the upper (adaxial) locule which is reduced. | Bousier, A. M. Recherches anatomiques sur l’inflorescence des Cuphea alterni- flores (Lythrariées). Bull. Herb. Boiss. 4: 328-335. 1896. CorreNS, C. Ueber die Epidermis der Samen von Cuphea viscosissima. Ber. Deutsch. Bot. Ges. 10: 143-152. pl. 8. 1892. Duncan, W. H. Stamen-numbers in Cuphea. Rhodora 52: 185-188. 1950. [C. viscosissima. | Focc, J. M. Weeds of lawn and garden. vii+ 215 pp. Philadelphia. 1945. [C. viscosissima, 126. GRAHAM, S. A. Systematic studies in the ore oe (Lythraceae). vi + 235 pp. Ph.D. diss. (unpublished). Univ. GuicNarp, M. L. Recherches sur le sac ae onnaire a phanérogames angio- spermes. Ann. Sci. Nat. Bot. VI. 13: 136-199. pls. 3-7. 1882. GutTTerson, M. E. Cuphea procumbens at Andover, Massachusetts. Rhodora 4: 2 1902 KERNER, A. Die Schutamittel der Bliithen gegen unberufene Gaste. Festschr. Zool.-Bot. Ges. Wien. 189-261. pls. 1-3. 76. Kiess, G. Beitrage zur Morphologie und Biologie der Keimung. Unters. Bot. Inst. Tubingen 1: 536-635. 1885. [Discussion of hairs in seed coat of Cuphea, 583-585. ] KoEHNE, E. Berichtigung der von D. P. Barcianu gemachten Angaben iiber 250 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Bliithenentwickelung bei den Cupheen. Bot. Zeit. 33: 291-296, 302-307. 75. Merrit, E. D. The generic name Parsonsia and the status of Parsonsia heli- candra Hooker and Arnott. Brittonia 1: 233-237. 1933 [Parsoma R. Br., 1810, conserved for a genus of Apocynaceae. | Los nombres Parsonsia y Cuphea. Revista Sudam. Bot. 1: 97-99. 1934. Sarto, K. Studies on inducing polyploid flower plants and their utilization. III. On several polyploid plants of Cacalia, Cosmos, Cuphea, and others. (In Japanese; English summary.) Jour. Hort. Assoc. Japan 19: 195-199. 1950.* [Colchicine treatment. | 1964 | BRIZICKY, POLYEMBRYONY IN EUONYMUS 251 POLYEMBRYONY IN EUONYMUS (CELASTRACEAE) 1 GrEoRGE K. BriziIcKy POLYEMBRYONY, the production of two or more embryos within a seed, is rather widely, although sporadically, distributed among the sperma- tophytes. Although less common than in the gymnosperms, polyembryony in the angiosperms is much more diverse in respect to the origin and mode of development of embryos. In the angiosperms, embryos may be derived from an egg cell (rarely from two or more “egg cells”) and/or synergids and/or antipodals within a single reduced or unreduced embryo sac of an ovule, either following or without fertilization; from a fertilized egg cell through the cleavage of the zygote or the proembryo (cleavage polyem- bryony) ; or directly from the cells of the nucellus or the inner integument of an ovule (adventitious embryony) (Ernst, pp. 436-438; Lebégue, pp. 333-336). Although polyembryonate species or populations usually show a predominance of one type of polyembryony, there also are those (e.g., Allium odorum L.) which may produce seeds with embryos derived in several different ways. Of all the types of polyembryony those involving apomixis (either gametophytic [diplo- or apospory, parthenogenesis, or apogamety] or somatic [adventitious embryony]) are of special interest because of the effects of apomixis on the groups in which it occurs. Apomixis makes possible the survival of well-adapted but sexually sterile genotypes, per- mits the building up of large populations of genetically similar individuals for rapid colonization of newly available habitats, and limits the genetic variability of those plants which have adopted this mode of reproduction (Stebbins, 1950, pp. 414-416). Although a rather restrictive factor in evolution, apomixis is important in the increase of polymorphism and geographic distribution of the genera in which it occurs. The study of agamic complexes, i.e., groups of closely related apomictic and sexual species, may yield valuable information on the present and past distribu- tion of plants and may help in the determination both of centers of origin and routes of dispersal of genera and/or their sections and of the relative ages of different floras (Stebbins, 1941, pp. 533-536). In view of the role of apomixis in evolution and distribution, and of its bearing on taxonomy, the records of polyembryony in three species of Euonymus L., coupled with the great polymorphism and wide distribution "Continuing a series of miscellaneous notes and papers on the flora of the south- eastern United States made possible through the interest and support of George R. ndatio expresses his sincere gratitude to Dr. Carroll Wood, Jr., for his critical reading of the manuscript, valuable suggestions, and advic 252 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV of this genus, suggested that it would be desirable to verify Braun’s record (1860) of polyembryony in E. americanus and to investigate the distribu- tion of polyembryony within the genus. Since only a few special seed collections were available, the necessity of using herbarium specimens for this purpose has limited investigations to those species in which the speci- mens have numerous loose seeds or fruits. Consequently, only a few seeds of each species could be examined, and no strictly statistical evaluation of the results can be made. Despite this limitation, the occurrence of polyem- bryony in several species has been detected, and the possibility of its absence or rare occurrence in other species has been indicated. POLYEMBRYONATE SPECIES OF EUONYMUS It is expedient to begin with Evonymus latifolius and E. americanus, in which polyembryony has long been known and thoroughly investigated by several authors. Other species are surveyed in alphabetical order. In citations of the material investigated by the writer the abbreviations “P” and ‘““M” are used for polyembryonate and monoembryonate, respectively. In regard to size, embryos are described as full sized (3.53 mm.), sub- normal (3-2 mm.), small (2-1 mm.), very small (1—0.5 mm.), minute (0.5— 0.25 mm.), and very minute (under 0.25 mm.). The minute and very minute embryos usually are club shaped, subcylindrical, or subglobular, often notched at the apex; all the larger ones are distinctly differentiated into cotyledons and radicle and are similar in shape. Euonymus latifolius (L.) Mill. Du Petit-Thouars (1808, p. 199), the first to record polyembryony in the genus, wrote that in Euonymus latifolius, although sometimes two equally developed embryos were found in the endosperm (‘‘perispérme’’) of a seed, more commonly one was much smaller than the other. Later, Grebel (1820) and Treviranus (1838) also reported the occurrence of polyembryony in this species. The main points of Grebel’s observations (pp. 324, 325) were that not rarely (but neither in every capsule nor twice in a capsule) two completely developed embryos lying side by side in the endosperm cavity of a seed were observed, but more frequently one embryo of the pair was much smaller than the other, had spread cotyle- dons, and was leaning to the radicle of the larger embryo; and, once, three normally developed, completely mature and two minute embryos were found in the endosperm cavity. In all the cases the embryos were lying in the seed in the same position, their radicles being directed toward the hilum (micropyle). Grebel did not mention the number of seeds examined. Treviranus (p. 556, pl. 3, fig. 40) reported that a half of nearly a dozen seeds examined by him were two-embryonate, both embryos of each pair lying in the same position in a seed. In 1860, Braun (pp. 157, 158, pl. 4, figs. 1-12; pl. 5, figs. 1-4) exam- ined nearly fifty seeds of E. latifolius. Twenty-eight of these were polyem- 1964] BRIZICKY, POLYEMBRYONY IN EUONYMUS 253 bryonate (twenty-four with two embryos, three with three, and one with four). All embryos in a seed were lying in the same position in respect to, but often at different distances from, the micropyle. Strasburger (1878, pp. 658, 659, pl. 19, fig. 44) confirmed Braun’s observations and stated that most seeds examined by him were polyembryonate. He also investi- gated the embryology of E. latifolius and concluded that the egg cell, al- though fertilized, only rarely developed into an embryo, but the adjacent nucellar tissue penetrated into the embryo sac and gave rise to adventi- tious embryos. Usually, only a single cell of the nucellar tissue became the initial of an adventitious embryo, which appeared to be situated similarly to an embryo developed from a fertilized egg cell, but sometimes the development of adventitious embryos began from a large cell-complex. Most recently Andersson (1931), ages oes the embryology of the same species, reported (pp. 34, 35), “In den allermeisten Samen und Samenanlagen mit Stadien nach der eaeti von Evonymus latifolius, die ich untersucht habe, kamen zwei bis mehrere Embryonen vor. In einigen Fallen habe ich feststellen kénnen, dass die befruchtete Eizelle stirbt. Ob das aber immer der Fall ist, weiss ich nicht. Es ist aber nicht zu verneinen, dass sogar in Samen mit einem einzigen Embryo, dieser ein Adventiv- embryo sein kann. Auch die Eizelle wachst. Sie teilt sich erst ziemlich spat, und dann ist es sehr schwierig zu entscheiden ob sie noch da ist oder nicht. Denn zu dieser Zeit haben sich einige Zellen des inneren Integu- ments vorgewolbt . . . Die Zellen geben durch ihren Plasmareichtum ihren Charakter als Initialen der Adventivembryonen zu erkennen. . . . Nicht alle Adventivembryonen werden zu derselben Zeit anuclest. eas scheint mir aber, dass Embryonen auch von einem schon vorhandenen Embryo eon konnen, der dann als Embryotrager fungiert. Einmal habe ich eine embryoahniliche Bildung wahrgenommen, von der ich geneigt bin zu glauben, dass sie aus dem Endosperm stammt.” An- dersson also showed that in Euonymus latifolius, as in E. europaeus L. (q. v.), the nucellus disintegrates, except for a negligible residue in the chalazal region, and can not be regarded as a source of adventitious embryos. Thus, according to him, in E. latifolius the adventitious embryos originate from the inner integument, although he once observed an embryo possibly of endospermous origin. Euonymus americanus L. Braun (1860, p. 159) investigated seeds of this species and stated that polyembryony in Evonymus americanus is even more common than in E. latifolius. Only five of twenty seeds examined by him were monoembry- onate, six contained 2 embryos, four 3, three 4, and two 5. Two embryos in a seed often were subequal and full sized, the others gradually smaller, with one or two minute and easily overlooked. The relative position of the embryos in the endosperm was the same as in E. latifolius. One hundred twenty-six seeds of nine collections of Euonymus ameri- canus from various parts of its range were examined by me. 254 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLV a a Jackson County, Indiana (GH): 5 seeds; 1 M; 4 P ‘three seeds and one with 3 embryos; supernumerary embryos small, minute, or nute. Short pes pee (GH): 6 seeds; 2 M; 4 P (one with 2, two with 3, and ) with 4 embryos; in all cases only one rian in each seed full sized, e picecatcatcies small, minute to very minute). seed ed Long 7522, Nansemond rea Virani (GH): 25 seeds; 1 M; 24 e with 10 embryos, one with 9, se ‘ four with 6, four with 5, ae ee with 4, eight with 3, and one wit ‘th Di ne 6-embryonate seed the largest (full-sized) embryo reversed, with he a directed toward the chalaza; in one 6- and one 8-embryonate seed all the embryos small to very minute and lying in various positions both axially [with their radicles directed to opposite poles] and + transversely). This specimen shows not only the highest percentage (96%) of polyembryonate seeds, but also the highest number of embryos in a seed (10), the average number being 5. Demaree ele Montgomery County, Arkansas (GH): 5 seeds, all P (two with 2, on with 3, and two with 4 embryos; in all cases only one embryo in each seed full sized, the supernumerary small, minute to very minute). Demaree 16568, Geant wee Arkansas (cH): 5 seeds; 2 M; 3 P (two with 3, and one with 4 embry R. _ ‘Channel, 13 Sept. 1067, Davidson County, Tennessee (preserved fruits) : 50 seeds; 9 (18%) M (Fic. 1, a); 41 (82%) P (nine ine 2 oo nine with 3, fourteen with 4, four with 5, three with 6, and two with 7; and one 4-embryonate seed one embryo reversed FIG. a b, cl, ae the radicle pieree toward the chalaza; in a 4-embryonate seed three minute embryos situated in the micropylar region and one subnormal embryo lying across the seed | Fic. 1, au Four was the average number of embryos in a seed. (See also Fic. 1, Wiegand & Manning 1908, an County, Alabama (GH): 5 seeds, all P (two with 2, one with 3, and two with 5 embryos; in all the cases only one embryo in a seed full sized, the supernumerary mostly minute to very minute; in a 5-embryonate seed one minute embryo situated at the reverse pole of the endosperm cavity [opposite the micropyle] immersed in the endosperm and covered with an oil droplet). An endospermous origin of the last-mentioned mbryo seems to be possi Pie 6782, Polk County, Texas (A): 5 seeds, all P (two with 2, one with 3, and two with 4 embryos; in a 4-embryonate seed two embryos full sized). Kral es Godfrey 3570, Gadsden County, Florida (GH): 20 seeds; 19 M; 1 P (with one full- sized and one small embryo). My observations show that polyembryony in Evonymus americanus is rather frequent, as long ago stated by Braun (1860). Ten was the highest number of embryos found in a seed. Although all supernumerary embryos usually were found in the micropylar region in normal positions, in seven seeds one or more supernumerary embryos of a seed were situated in reverse axial and/or transverse positions (Fic. 1, b-e). In six of these seeds an integumentary origin of such embryos is assumed, but in one (Wiegand & Manning 1908) an endospermous origin of a reverse embryo is to be suspected. The specimen Kral & Godfrey 3570, perhaps represent- ing E. americanus var. angustifolius (Pursh) A. Wood or a form close to it seems to be an exception in which polyembryony is rare. 1964] BRIZICKY, POLYEMBRYONY IN EUONYMUS 255 Euonymus alatus (Thunb.) Regel var. apterus Regel Thirteen seeds of three herbarium collections, and 200 seeds of a living cultivated plant were examined. W. T. Tsang 23362, Kwangsi, China (a): 3 seeds; 1 M; 2 P (each with one full-sized and two somewhat smaller embryos). Wilson 354, Western Hupeh, China (A): 10 seeds; 8 M; 2 P (each with two ubequal, slightly undersized embryos). Sargent, 16 Sept. 1892, Sapporo, Japan (a): 3 seeds, all P (two with 2 subequal embryos in each; one 4-embryonate seed with one embryo in a reverse position at the chalazal end of endosperm [Fic. 1, g}). Arnold Arboretum; no. 14543-B (grown from seeds sent from Korea, by J. Jack): 200 seeds; 166 (83%) M; 34 (17%) P (twenty-eight 2-embryonate and six 3-embryonate; in all cases embryos located in the micropylar region [Fic. 1, f]). Euonymus dielsianus Loes. Stewart, Chiao, & Cheo 811, Kweichow, China (A): 4 seeds; 1 M; 3 P (all 2- mbryonate; in each of two seeds both embryos subequal and nearly full sized but lying in reverse positions with their radicles directed toward micropyle and chalaza respectively). 256 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Euonymus macropterus Rupr. Dorsett & Morse 1336, Odomari, Sakhalin (A): 10 seeds; 1 M; 9 P (six with 2, and three with 3 embryos) Euonymus oxyphyllus Miq. Wilson, 1914, Hondo, Japan (A): 5 seeds, all from the same capsule; 2 M; 3 P (one with 2, one with 3, and one with 4 embryos). Euonymus vagans Wall. ex Roxb. (E. bockti Loes.; E. hupehensis var. brevipedunculatus Loes.) Seven seeds of two herbarium specimens were investigated. C. S. Fan & Y. Y. Li 673, Hunan, China (a) [determined as E. bockii]: a single seed examined contained two subequal, full-sized and one minute Henry 9106B, Yunnan, China (A) [determined as E. hupehensis var. brevi- pedunculatus Loes.]: 6 seeds; 2 M; 4 P (all 2-embryonate). Euonymus verrucosoides Loes. Sixteen seeds of two herbarium specimens were examined. Wilson 3102, Western ua! China (A): 7 seeds; 1 M; 6 P (two with 2, ith 3, and one wi embryos; one 3- and the single 5- embryonate bia only minute nae os). Rock 14978, Southwestern Kansu, “China (A) [determined as E. verrucosoides var. viridiflorus Loes. & Re hd.]: 9 seeds; 2 M; 7 P (one with 2, five with 3, and one with 4 embryos; one 3-embryonate oe with two equal, full- sized and one minute embryo). MONOEMBRYONATE SPECIES OF EUONYMUS Since Euonymus europaeus has sometimes been mentioned as polyem- bryonate, the question of polyembryony in this species should be considered in some detail. Jager (1814, p. 202) listed this species among a few plants in seeds of which two embryos occur. In order to verify Jager’s record Braun (1860, p. 156) examined seeds of E. europaeus and remarked, “Wenn hier nicht eine Verwechselung mit der folgenden Art [£. atifolius | im Spiel ist, so ist ein solches Vorkommen jedenfalls sehr selten, denn mir gelang es nicht bei dieser Art mehr als einen Keimling zu finden, wiewohl ich mehr als ein halbes Hundert Samen zergliederte.”’ In 1916, however, Bally reported the occurrence of polyembryony in E. europaeus, According to his observations, sometimes the embryo sacs in this species degenerate in an earlier or later stage of development. The space which was occupied by the embryo sac becomes surrounded by the nucellar tissue. The cell walls of these “‘tapetum-cells” dissolve themselves, and their naked contents wander into the cavity and soon fill in the latter with a homogeneous tissue as has been observed in the pollen sacs of certain monocotyledons. When the cavity is almost completely filled with nucellar tissue, the exterior layers of the latter in the micropylar region begin to execute divisions which lead to the development of one or more 1964 | BRIZICKY, POLYEMBRYONY IN EUONYMUS 257 embryos growing into the endosperm tissue. However, Andersson, on the basis of his own study of embryology in E. europaeus, criticized Bally’s observations, saying (p. 30), “The nucellar cells can not be involved at all, because the nucellus disintegrates early and its small chalazal remain- der is of no significance. . . . I believe that Bally’s statement is to be attributed to defective staining” (translation supplied). Regarding Bally’s explanatory account of the occurrence of adventitious nucellar embryony in E. europaeus, Andersson (p. 32) further remarks, “The formation of embryos in the micropylar region from the cells of the nucellus is im- possible, since even this part of the nucellus disintegrates early. The origin [of embryos] from the cells of the inner integument would perhaps be possible [in this species], but I have never observed it. I also have not found polyembryony, and to my knowledge the latter has been observed [in this species] otherwise only once by Jager (1814).” The writer examined 30 seeds from each of two living cultivated speci- mens of Euonymus europaeus (Arnold Arboretum no. 18253, and Fairfield Garden Apts., Watertown, Mass., Brizicky, Oct. 1963) and found no evi- dence of polyembryony. Since Braun, Andersson, and the present writer have not found polyembryony in this species, one may assume either that Jager’s and Bally’s records were based on misidentified material or that polyembryony in £. europaeus is extremely rare. The following species investigated by the writer showed no polyem- bryony: Euonymus bungeanus Maxim.: 30 seeds of a living specimen, Arnold Arboretum 2128-D (grown from seeds received from J. Hers, China, 1919). Euonymus maackii Rupr.: 30 seeds of a living specimen, Arnold Arboretum no. 4563-1 Euonymus sachalinensis (Schmidt) Maxim.: 15 seeds of a living specimen, rnold Arboretum no. 13213-A. Euonymus semiexsertus Koehne: 50 seeds of a living specimen, Arnold Arbore- tum no. 99-35-B. Euonymus yedoénsis Koehne: 150 seeds of three living specimens, Arnold Arbo- retum no. 34340-A (cutting from Arnold Arboretum no. 14576-1-A), thirty seeds; Fairfield Garden Apts., Watertown, Mass., Brizicky, Oct. 1963, tree no. 1, ninety seeds, and tree no. 2, thirty seeds. Although a few seeds, taken from random herbarium specimens of E. atropurpureus Jacq., E. obovatus Nutt., and E. occidentalis Nutt., were all monoembryonate, no conclusion can be made about the nature of these species until adequate material is investigated. Seeds of E. fimbriatus Wall. and E. verrucosus Scop. were found to be monoembryonate by Braun (1860). Although the present investigations indicate that all the above-mentioned species may be monoembryonate, the monoembryony of at least some species, assumed on the basis of not very abundant material, should be reinvestigated. Furthermore, it seems to be of even greater importance to study the embryology of these species in order to determine the origin of the embryo, since even a single embryo within a seed could be apomictic. 258 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV SUMMARY AND CONCLUSIONS The occurrence of polyembryony in Euonymus latifolius and E. amert- canus, recorded in the nineteenth century, has recently been confirmed b Andersson (1931) for the former and by the writer for the latter species. On the contrary, polyembryony in EF. europaeus, recorded by Jager (1814) and Bally (1916) has not been substantiated (Braun, 1860; Andersson, 1931; the writer). If these records were not due to misidentification, polyembryony in E. europaeus may at least be extremely rare. I have found polyembryony to be of common occurrence in E. alatus, FE. dielsi- anus, E. macropterus, E. oxyphyllus, E. vagans, and E. verrucosoides, all of eastern Asia. In EF. bungeanus, E. maackii, E. sachalinensis, E. semiex- sertus, and E. yedoénsis only moncenbevonste seeds were observed. Al- though no polyembryonate seeds were found in E. atropurpureus, E. obovatus, and E. occidentalis, the material investigated was inadequate. Braun (1860) found £. verrucosus and E. fimbriatus also to be monoem- bryonate. In the polyembryonate species the number of supernumerary embryos in a seed varies from one to nine. Only rarely are two or three embryos of a seed + full sized and subequal; usually the supernumerary embryos are small to very minute. They are usually located with at least their radicles in the micropylar region and directed toward the micropyle (Fic. 1, a, f). In some cases, however (e.g., in E. americanus, E. alatus, E. dielsianus), one supernumerary embryo was lying in a reverse position with its radicle directed toward the chalaza (Fic. 1, b, c, g), and in a few cases one or more embryos were situated + transversely (Fic. 1, d, e). In a few instances (e.g., in E. verrucosoides), a full-sized embryo was absent, but there were several minute ones in the micropylar region. The cotyledons of all the embryos of a seed usually lie in nearly the same longitudinal (vertical) plane, the position of which with regard to the raphe varies from one seed to another. Although sometimes parallel to the plane of the raphe, the plane of the cotyledons is more often at an angle to it. Since Andersson showed that in Euonymus europaeus and F. latifolius the nucellar tissue disintegrates early and the adventitious embryos in the latter species arise from the inner integument of the ovule, one may assume that the early disintegration of the nucellus is a generic character and that in all polyembryonate species the adventive embryos originate from the inner integument. In fact, the integumental origin of the super- numerary embryos located in the micropylar region in seeds of the polvem- bryonate species of Euonymus can hardly be doubted. An identical origin of the embryos found in a reverse or + transversal position is most probable, although the embryos at the chalazal end might also be of an antipodal origin (through apogamety). In one instance (E£. americanus, Wiegand & Manning 1908), an endospermous origin of a reverse embryo is suspected Adventitious embryony from the inner integument seems to be rela- tively rare and has so far been recorded in only about ten species in nine 1964 | BRIZICKY, POLYEMBRYONY IN EUONYMUS 259 genera and eight families (Johansen, pp. 283, 284). In the family Celas- traceae this kind of polyembryony is also known in Celastrus scandens L. It is notable that “for some reason, adventitious embryony seems to be relatively frequent in species native to warm temperate or tropical cli- mates” (Stebbins, 1950, p. 384) The question of whether a full-sized, normal embryo in polyembryonate seeds or a single embryo in monoembryonate seeds of Euonymus de- velops from a fertilized egg cell, or whether it is at least sometimes apomictic, can not be answered without detailed embryological study of the species involved. Andersson’s statement (p. 34) regarding FE. latifolius reads in translation, “I can ascertain that the fertilized egg cell dies in some cases. Whether it is always the case, I do not know. However, it is not to be denied that even in seeds with a single embryo this can be an adventitious one.” LITERATURE CITED Anpersson, A. Studien iiber die Embryologie der Familien Celastraceae, Oleaceae und Apocynaceae. Lunds Univ. Arsskr. II. Sect. 2. 27(7): 1-110 pls. 1-4. 1931. Batty, W. Zwei Falle von Polyembryonie und Parthenokarpie. Verh. Schweiz. Naturf. Ges. 98(2): 169, 170. 1916. BLAKELocK, R. A. A synopsis of the genus Euonymus L. Kew Bull. 1951: 210-290. 1951. Braun, A. Uber Polyembryonie und Keimung von Caelebogyne. Abh. Akad. Wiss. Berlin 1859: 109-263. pls. 1-6. 1860. [Euonymus, polyembryony, 156-159. pl. 4.] Du Petit-THovars, A. Observations sur la germination de |’Allium fragrans et de quelques autres plantes dont les graines renferment plusieurs embryons distincts. Nouv. Bull. Soc. Philom. Paris 1: 198-201. Ernst, A. Bastardierung als Ursache der Apogamie im Pflanzenreich. 665 pp. pls. 1, 2. Jena. 1918. GREBEL, Dr. Ueber die Saamen des Evonymus latifolius. Flora 3: 321-332. Gustarsson, A. Apomixis in higher plants. Part I. The mechanism of apomixis. Lunds Univ. Arsskr. II. Sect. 2. 42(3): 1-68. 1945; Part II. The causal aspects of apomixis. /bid. 43(2): 69-180. 1946; Part III. Biotype and species formation. /bid. 43(12): 181-370. : JAcrer, G. F. Ueber die Missbildungen der Gewachse. 320 pp. pls. 1, 2. Stutt- gart. 1814. JoHaAnseNn, D. A. Plant embry aa Embryogeny of the spermatophyta. xvi 950. LEBEGUE, A. La polyembryonie ree les angiospermes. Bull. Soc. Bot. France 99: 329-367. 1952. SteBBINs, G. L. Apomixis in the angiosperms. Bot. Rev. 7: 507-542. 1941. _ Variation and evolution in plants. xix + 643 pp. New York. 1950. STRASBURGER, E. Ueber Polyembryonie. Jena. Zeitschr. Naturw. 12: 647-670. pls. 15-19. 1878. Treviranus, L. C. Physiologie der Gewachse. Vol. 2. 809 pp. pls. 1-3. 1838. 260 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV CONE MORPHOLOGY IN PINUS SABINIANA ! JAMES R. GRIFFIN THE DiIccER PINE (Pinus sabiniana Dougl.) of California’s foothill woodland, is unique in several ways. Few other pines can tolerate such sterile soils or xeric habitats, have such a diversity of irregularly shaped crowns, or show such impressive variation of cone forms within and be- tween populations. Cone morphology has been emphasized in all taxonomic accounts of Pinus sabiniana. After commenting on the branchy form of the crown, most authors mention the conspicuous ‘‘spurs’’ — as the combined umbo and apophysis of the basal cone scales will be termed here. A few authors have been impressed by extra long spurs in certain local populations. For example, Jepson (1909, p. 217) named one population on Mount Diablo having “strongly spur-hooked” scales and relatively long seed wings as variety explicata. Early workers had insufficient material at their disposal to study geographical patterns in cone variation. Still, a few California botanists with wide field experience suggested some regional trends. Thus, Lemmon (1888) thought that the northern forms of the cone were less strongly spurred than the southern forms, adding that trees in the Tehachapi had large cones with strong hooks “closely approaching’ Pinus coulteri D. Don. The most detailed reference to cone variation in Pinus sabiniana was given by Stockwell (1939), who stated: “. . . the degree of variation in cone size and morphology exhibited by this pine is approached by few others.”” He also felt that this variation often fell into two patterns within the species as follows: “In central and northern California the cones of 1 conformation. . . toward the southern end of its range, however, and near the coast, colonies of digger pine are known that produce an entirely different appearance. . . the general aspect of the cone is similar to that of Torrey pine. . A broad study of this species was made and reported by the author in an unpublished doctoral dissertation (1962). As part of that study, cone samples from populations throughout the range of Pinus sabiniana were analyzed for many characters. Preliminary sampling had suggested that s paper is based on parts of a thesis submitted to the University of California in areal satisfaction of the requirements for the Ph. D. degree in botany. Portions t and Range Dae Station, Forest Service, U. S. Department of Agriculture, B.rkeley, Californi - 1964 | GRIFFIN, CONE MORPHOLOGY IN PINUS 261 100 Miles -———_—_——4 Fic. 1. Natural distribution of Pinus sabiniana in California. Circled dots indicate po opulation oa aad locations; numbers offset from dots refer to sample descriptions in TABLE 1. intra-tree variation in both size and form of cones was small in contrast to intra- or inter-population variation. To minimize the expected high vari- ance within populations, trees of a “standard” size and age class were sampled. Resulting cone samples from the upper, lateral branches of standard trees were more comparable physiologically than samples from convenient ground reach would have been. Unfortunately, scarcity of standard trees in the field, as well as storage and handling problems in the laboratory, forced the major samples to be somewhat restricted — one cone from each of 20 different trees. In some localities only 10 cones were obtained. The natural distribution of the species and the relative location of samples appear in Fic. 1. In addition to population samples, the 262 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV complete cone crops on two trees in one population were collected for two consecutive years in order to estimate the nature and magnitude of intra-tree variation.” CONE SIZE Stockwell (1939) suggested that a small-cone population existed on Figueroa Mountain in the southwestern corner of the range. ‘These examples,” he emphasized, “are not of rare individual cones or of isolated trees, but of local races. . .””. This note prompted Gaussen (1960) to describe forma microcarpa. Vague reports of other small cone populations in the Tehachapi Mountains in the southeastern corner of the range have circulated, although Lemmon alluded to large cones in the same region. Stockwell (1939) and Mason ® had been impressed with large-cone popu- lations in the northwestern portions of the range. Preliminary sampling in this study suggested that a more detailed survey of cone size was desirable. Aside from the possibility of general cone size differences between populations, absolute cone size was investi- gated because of its influence on cone form, cone scale development, seed size, and other associated characters. The spurs on many cones so complicated direct measurement of length or width on either open or closed cones that air-dried weight was selected as the size criterion. The results of the population survey appear in Fic. 2. No simple correlation of latitude, temperature, precipitation, or other habitat factor with cone size was apparent. The ranges of the largest and smallest samples just overlapped. Clearly, the larger and smaller samples are distinct from one another in a statistical sense, although many of the intermediate samples do not differ significantly from each other. The complete intra-tree analyses revealed that mean cone-weight per tree could vary at least 15 per cent in different years. The standard errors of the mean in Fic. 2 su gest the variable nature of many populations. Despite the high variances involved, the magnitude of difference between the more extreme samples suggests that some mean- ingful population differences exist. The phenotypic differences expressed in Fic. 2 might be thought to reflect merely environmental differences — in part they must. But several factors suggest that genetic differences also are involved. One line of evidence supporting this view is found by comparing specific pairs of popu- lations. For example, the Plummer Creek (#1) sample with the largest cones grew in a moderately unproductive “serpentine” habitat. The Applegate (#8) sample, with rather small cones, grew under a comparable precipitation regime, but on a more productive soil. General crown vigor in the two samples as measured by annual branch-length increments and * Five cone subsamples from all collections are on deposit at the Institute of Forest Genetics, Placerville, California. *H. L. Mason. Botany Department, University of California, Berkeley, California. Personal communication. * 1000 — Y) > de < ad or O ef EEE Eee er 1710 9 8 7 4 19 16 21 13 1 14 3 2012 22 15 5 23 6 1% 2 1 POPULATION Fic. 2. Cone weight variation in Pinus sabiniana. Data are arranged in order of increasing mean weights. Horizontal cross- bars represent sample means, vertical bars represent sample St dig and rectangular boxes indicate + 2 standard error of the mean values. Population numbers refer to descriptions in TABLE 1. 264 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV TABLE 1. Description of population sample localities SAMPLE NO. CONES NUMBER IN SAMPLE Locality LATITUDE LONGITUDE ELEVATION M.A.P.* 1 20 Plummer Creek 41° 09’ 123° 13’ 2,400’ 45” 2 20 Big Bar 40 45 123. 13 1,300 47 3 20 Murken Bench 40 50 121 26 4,500 20 -f 20 Bear Creek 40 31 122 04 1,200 35 5 10 Bennett Creek 39 49 122 39 2,000 25 6 20 Summit Rock 39 09 122 44 3,500 40 7 10 Buckeye Creek 38 55 122 Ol 200 17 8 20 Applegate 39 00 120 59 2,000 47 9 10 ~Volcanville 38 58 120 48 3,200 55 10 10 Folsom Dam 38 41 121 10 250 24 11 10 Con 26 122 20 300 30 12 10 Mt. Hamilton-1 37 20 121 40 3,300 25 13 20 Mt. Hamilton-2 37) 21 121 37 3,900 30 14 10 Uvas Creek 37 06 121 45 800 25 15 20 San Benito River 36 23 120 57 1,800 15 16 20 Clear Creek 36 23 120 44 3,800 18 17 20 Sycamore Creek 36 «653 119 15 1,800 23 18 20 Redwood Gulch 35 50 121 23 1,800 40 19 10 = Las Pilitas 35 21 120 29 1,500 17 20 20 Nine Mile Canyon 35 52 118 00 6,500 10 21 10 +Figueroa Mt.-1 34 44 120 Ol 3,000 20 22 20 + Figueroa Mt.-2 34. 44 119 49 3,700 25 23 10 Cow Spring 34 44 118 39 3,800 18 * Mean annual precipitation, estimated for most stations. needle-lengths was similar. Tree height growth was superior in the Apple- gate stand. Small cones in this case seemed not to be a function of “‘poor”’ tree-growth, and conversely, large cones were not necessarily a function of “good” tree-growth. No evidence of unusually small cones was found in the Figueroa Moun- tain region (see #21 and #22 in Fic. 2). Two samples were taken in the area— one (#21) located on the most sterile site available. The cones of sample #21 were smaller than the species mean; those of sample $22 were larger than the species mean. Ironically, the heaviest individual cone of any sample was found in the general ‘“microcarpa” area in sample #22. CONE FORM The shape of Pinus sabiniana cones varies widely, but population patterns were not clearly suggested in the samples. In large part the relationship of length to width depends upon absolute size. Many of the ovoid cones were larger than normal (Fic. 3). Although large cones have more massive individual scales, they also tend to have a greater number o: scales per cone. The longer axis needed to accommodate the extra 1964 | GRIFFIN, CONE MORPHOLOGY IN PINUS 265 é| <—10 CM—> Fic. 3. pan cone form extremes in Pinus sabiniana. The elongated cone A has a grea mber of scales than “normal,” while the small ee cone B ha aires te scales. The cones in Fic. 4 are of more usual shap scales of big cones makes the cones appear more cylindrical. On the trees that were intensively studied for two successive years, average cone weight decreased and average number of scales per cone decreased. The length-width ratios also shifted slightly with the decrease in cone size and scale number. In this case as cone size decreased, number of cones per e No significant differences in phyllotaxis were noted in any of the hundreds of cones examined. The usual scale arrangement in the central portion of the cone consists of 8 obvious parastichies in one direction, and 266 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLV | Applegate (8) Fic. 4. Two examples of spur variation within Pinus sabiniana populations. The pairs of cones came from comparable trees growing a few feet 13 obvious parastichies in the other direction. This arrangement is essen- tially the 8/21 ratio in terms of classical phyllotaxy, and it is found in the cones of many other pines. There is a tendency for the “primary 1964 | GRIFFIN, CONE MORPHOLOGY IN PINUS 267 spiral” to wind in one direction, but this sometimes varies between cones from the same branch. SPUR FORM One of the most conspicuous features of cone variation in Pinus sabiniana is in relative spur length (Fic. 4). Although spur length is obvious on a given cone, the condition is rather difficult to quantify for a population. Absolute spur length is not a satisfactory measure, because it is affected by absolute cone size. In order to achieve a more independent criterion of spur development, a “spininess index” was assigned to each cone. This value was computed by measuring the length and width of one spur of constant relative position on the dorsal base of each cone. This length included both the umbo (the first year’s external scale development ) and the apophysis (the second year’s extension growth of the cone scale). Spur length was divided into the width of the spur base. Spurs with values of 1.0 or greater were rather “blunt” (Fic. 3; Fics. 4B, 4D). Spurs with ratios of 0.5 or less were very ‘“‘spiny” in aspect (Fics. 4A, 4C). The differential scale-growth causing spiny cones is evident early in ontogeny. During the first season’s growth of the cone, the spiny condi- tion is obvious if it is to develop at all. The proportion of total number of scales that become spiny in a cone is variable. Occasionally all the scales are markedly affected (Fic. 5D), but more frequently only the more basal scales are affected (Fics. 5A, 5C). No practical way to describe the proportion of spiny scales in a cone was found. Spininess values refer only to the spininess at the basal portion of the cone. One feature of cone spininess, its stability within a given tree, is quite constant. If cones on a tree were spiny, every cone on the tree was spiny to the same degree vear after year. In addition to relative length, spurs vary greatly in degree of curvature. Some spurs are strongly recurved (Fics. 4A, 4C); others are straight except for the extreme tip of the umbo (Fic. 3A). Jepson (1910) long ago commented on such straight and recurved spur forms. Curvature seemed to be independent of spur length in many cases. In this survey each cone was assigned an arbitrary spur “curvature index.’ Only spurs of constant relative position on the dorsal base of the cone were used in assigning curvature values. Cones with the straightest spurs were classed ‘‘1”’; cones with the most recurved spurs as “5.” Intermediate values were assigned, with some difficulty, to intermediate conditions. As with spur length the degree 2 curvature was rather constant within all cones of a tree vear after The eee of the spur survey are difficult to interpret. Geographic patterns in spininess or curvature are obscured by variation within a population. Very spiny cones were found infrequently in most of the populations sampled. In the area on Mount Diablo of Jepson’s var. explicata, for example, several trees with very spiny cones were observed. Small samples in the vicinity suggested that no spiny “population” was involved. 268 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV TABLE 2. Spininess variation of several cone samples, by populations FREQUENCY DISTRIBUTION WITHIN SPININESS INDEX CLASSES POPULATION 03 0.4 0.5 06 0.7 08 0.9 1.0 1.1 1.2 PLUMMER CREEK (#1) 1 2 3 5 7 1 =«1 Figueroa Mr. (#22) 2 4 6 3 2 2 41 Bic Bar (#2) 1 2 11 5 1 APPLEGATE (#8) 9 5 3 2 TABLE 3. Curvature variation of several cone samples FREQUENCY DISTRIBUTION WITHIN POPULATION CURVATURE INDEX CLASSES 4 = PLUMMER CREEK (#1) 12 6 2 Figueroa Mt. (#22) 2 4 11 2 1 Bic Bar (#2) 4 9 2 5 APPLEGATE (#8) The sampling intensity was not great enough for strong quantitative conclusions, but the frequency of some combinations of spur characters does suggest population differences (TaBLEs 2, 3). Two of the more extreme patterns in cone morphology are compared in Fic. 6 in which weight, spininess, and curvature of all cones in two samples are plotted. A common spur situation in the species seems to be wide variation around intermediate spininess and curvature values. In Fic. 6 the Plummer Creek cones represent a skew toward straight, blunt spurs, while the Applegate data is skewed somewhat in the direction of curved, attenuated spurs. Of all the populations sampled, Plummer Creek had the most distinctive spur pattern. This population is also one of the most effectively isolated stands in the scattered Klamath Mountain populations. The blunt spurs of the Plummer Creek cones contrasted sharply with the spiny cones of the Big Bar (#2) sample 35 miles away on the other side of the Trinity Alps (see modal spininess values in TABLE 2). One population which had an unusual display of cone size, form, and spur variation was Clear Creek (£16). Three facts about this region should be mentioned. First, the edaphic situation is extreme with an extensive area of highly serpentinized parent material occurring under a relatively low rainfall. Secondly, the mixture of pine species (Pinus sabiniana, P. coulteri, and P. jeffreyi) occurs nowhere else. Pinus sabiniana grows with P. coulteri in many other areas in the south and with P. jeffreyi in the north, but all three grow together only near Clear Creek. Thirdly, the morphological variation in these three species is quite high here. When Zobel (1951) studied introgression between Pinus coulteri and P. jeffreyi, he found the greatest array of cone variation in his “pure” 1964 | GRIFFIN, CONE MORPHOLOGY IN PINUS 269 — 10 CM———4 | ms D. . Umbo variation in four normal, healthy Pinus sabiniana cones at end of first season’s growth. Cones A and B came from trees growing a few feet apart. Coulter and Jeffrey samples in this area. This study also showed very large morphological diversity in Pinus sabiniana. Zobel speculated that Pinus sabiniana in some way may have been involved in the great Jeffrey- Coulter variation at Clear Creek. This possibility must be seriously con- sidered. However, the extreme nature of the habitat also may have con- tributed to the variation in cone form in all three species. The relative influence of the associated pines and the unique habitat on Pinus sabiniana variation must remain speculative. Many authors have remarked that Pinus sabiniana cones occasionally resemble P. coulteri cones. The attenuation and curvature of the basal spurs on cones of the two species are often identical. The basic difference is that obvious spur development in Pinus sabiniana is often restricted to 270 e, | Spininess Index O O | JOURNAL OF THE ARNOLD ARBORETUM [ VoL. XLV _.. Plummer wn ~... Creek / | ~ en ‘ ms : “hy rs. t 1 ‘ fa ; 1 1 2 1 1 1; ‘ ‘ ' e ‘i ‘ ' 41 a : ri 1 2 1 2 a ‘ 1 ra : , t 1 . y i 4 ‘ ‘ 2 Pa : \ \ 3 co t 4 9 : 1 Pe ; 4 \ . 2 rd ( \ \ -o ; 3 L iF \ Poe 1 \ . - a a . 2 4 rs) t * 5 t 15 5 5 5 5, , ~-—- aw’ Applegate | it LL 1 | rn 1 J 0.0 O Fic. samples. 500 1,000 Cone Weight (g) Relationship of Plummer Creek (#1) and Applegate (#8) cone ‘Each plotted figure represents the curvature value for one cone of given spininess and weight. the basal portion of the cone, while in P. coulteri hooked spurs commonly extend to the cone apex. No clear geographical pattern to Coulter-like cones was suggested by the limited sampling here. A less well known aspect of Pinus sabiniana morphology is the occasional resemblance of cones to those of Pinus torreyana Parry. Stockwell’s ob- servations of Torrey-like cones were confirmed in several southern Califor- nia localities. On one tree, cones with weakly developed spurs over the whole surface were essentially indistinguishable from cones of Pinus tor- reyana. Pinus torreyana has long been isolated from P. sabiniana, being restricted to local coastal situations in southern California. Interestingly, the particular Torrey-like cones mentioned above were from the only coastal population of Pinus sabiniana (Redwood Gulch, #18). One near- 1964 | GRIFFIN, CONE MORPHOLOGY IN PINUS 271 definitive character that separates Pinus sabiniana and P. coulteri cones — the relative length of seed and seedwings — is not helpful in separating cones of Pinus sabiniana and P. torreyana, because the characters of seed and wing in these species overlap. SPECIFIC GRAVITY Since four of the samples were from stands where Pinus sabiniana grew with P. coulteri, cone specific gravities were taken to compare with the values Zobel (1951) reported for P. coulteri. He had found specific gravity helpful in characterizing cones of Pinus coulteri, and its hybrids with P. jeffreyi. All the Pinus sabiniana sample means were significantly different from Zobel’s P. coulteri means. But the Pinus sabiniana values did over- lap into the P. coulteri range. Specific gravity might be helpful in separat- ing Pinus sabiniana and P. coulteri cones, but large samples and careful analyses would be required. The specific gravity survey was later expanded to include all cones collected, and it was found that this physical character of the cones was loosely associated with latitude. Northern samples had lower specific gravities than southern samples (TABLE 4). With only minor deviations the intervening populations had intermediate values. Although the experimental methods (by water displacement) were crude, they were reproducible. Successive annual sampling in the Apple- gate trees gave similar results despite changes in cone size. The two Figueroa Mountain samples had virtually identical specific gravity dis- tributions as did the two independent samples at Mount Hamilton. Why the latitudinal trend in values was found or even why different samples from the same local area were so similar is not clear. Critchfield (1957) found specific gravity helpful in characterizing regional cone variation within Pinus contorta. But his specific gravity data were related to such morphological features as serotinous habit. Here, no relation to obvious morphological characters or habitat was apparent. TABLE 4. Specific gravity of cones from several samples NORTHERN SAMPLES MEAN ANGES OF POPULATION SPECIFIC GRAVITY SPECIFIC GRAVITY MurkKEN BENCH (#3) 0.70 0.59-0.79 Bic Bar (#2) 0.71 0.65-0.81 PLUMMER CREEK (#1) 0.73 0.68-0.79 SOUTHERN SAMPLES FIGUEROA MrT. (#22) 0.82 0.71-0.88 CLEAR CREEK (#16) 0.84 0.76—-0.88 Repwoop GULCH (#18) 0.85 0.77-0.96 272 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV CONE COLOR The magnitude of cone color variation in this species has never been described adequately. Color extremes were quantified by using Munsell soil color notations (Soil Survey Staff, 1951). These standardized soil colors conveniently covered the whole range of colors encountered in mature, unweathered cones. Modal cone color varied from reddish brown (5 YR 4/4) to dark reddish brown (5 YR 3/3). Lightest extremes were: yellow (2.5 Y 7/6) and yellowish brown (10 YR 5/4). Darkest extremes were brown (7.5 YR 5/8), and dark red (2.5 YR 3/6). Cone color within a tree was relatively constant, and color appeared to be stable on a given tree year after year. The common color is clearly a reddish brown. Color variations are stressed because of the tendency one might have to attribute occasional yellowish Pinus sabiniana cones to introgression from the yellow coned Pinus coulteri. From the limited sampling here, the striking color deviations suggested no geographic pattern. Yellowish or reddish cones appeared in low frequency in all parts of the range. Some yellowish Pinus sabiniana cones appeared in mixed forests containing P. coulteri; others appeared far removed from P. coulteri. CONCLUSIONS As a species, Pinus sabiniana seemed quite distinct from other “related” California pines with which it grows. No clear suggestion of introgression between this species and Pinus coulteri or P. jeffreyi was encountered in the field. Despite the wide range of habitat conditions in which the species grows, there was a general impression of uniformity in the appearance of the tree. In this context of tree similarity, the variation in cone features was impressive. The irregular pattern of this variation made it difficult to characterize the cones of local populations or to detect geographic trends within the species. A few small samples could lead to very erroneous generalities about Pinus sabiniana cones. With this in mind the following generalities are cautiously given: (1) Large cones were aresuently found in the North Coast-Klamath ranges; they may be elongated i (2) Small cones were pau found in the Sierra Nevada; they may be more ovoid in fo (3) Cones with low specific gravity were frequent in the north; low specific gravity was not closely related to the larger size of northern cones. (4) Cones with high specific gravity (overlapping into the range of P. coulteri) were found in the south. (5) Spur form was quite constant within a tree but quite variable within and between population samples. Plummer Creek, a small iene Klamath Mountain population, had a high frequency of blunt, straight spur ( ones were occasionally quite sini in size, color, and spur form to those of either Pinus coulteri or P. torreyan 1964 | GRIFFIN, CONE MORPHOLOGY IN PINUS 273 This study of cone morphology when combined with other investigated features suggested that Jepson’s variety explicata and Gaussen’s forma microcarpa were based on characters derived from inadequate sampling and are not useful. Distinguishable geographic groupings of populations were not obvious within the variable species. LITERATURE CITED CRITCHFIELD, W. B. Geographic variation in Pinus contorta. Maria Moors Cabot Found. Publ. No. 3. 118 pp. Harvard Univ., Cambridge. 1957. GAussEN, H. Les Gymnospermes, actuelles et fossiles. Laboratoire Forestier Toulouse. Tome II. Vol. 1. Fasc. 6. Chapt. XI. 272 pp. 1960. GRIFFIN, J. R. Intraspecific variation in Pinus sabiniana Dougl. Ph. D. Thesis, 274 pp. Univ. of Calif. 1962 [unpubl.]. Jepson, W. L. Trees of California. 228 pp. San Francisco. 1909. —. The Silva of California. Mem. Univ. Calif. 2: 480 pp. Berkeley. 1910. Lemmon, J. G. Pines of the Pacific Slope. pp. 69-140. Jn: Second Biennial Report. Calif. State Board Forestry, Sacramento. 1888. Sort SurvEY StaFF. Soil Survey Manual. U.S. D. A. Handb. 18. 503 pp. 1951. STOCKWELL, W. P. Cone variation in digger pine. Madrofio 5: 72, 73. 1939. ZOBEL, B. ae The natural hybrid between Coulter and Jeffrey pines. Ph.D. Thesis, 114 pp. Univ. of Calif. 1951 [unpubl.]. PacIFIC SOUTHWEST FoREST AND RANGE Expt. STA., Forest Service, U. S. D. A., REDDING, CALIFORNIA 274 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV THE ELAEAGNACEAE IN THE SOUTHEASTERN UNITED STATES * SHIRLEY A. GRAHAM ELAEAGNACEAE A. L. de Jussieu, Gen. Pl. 74. 1789, “Elaeagni,”” nom. cons. (OLEASTER FAMILY) The family is easily recognized and homogeneous, consisting of three genera, Hippophaé L., Shepherdia Nutt., and Elaeagnus L., all character- ized by a dense covering of silvery or rusty, peltate or stellate hairs on the flowers, twigs, and leaves; 2—4-merous flowers lacking petals; and a fleshy or mealy berry- or drupelike accessory fruit derived from the accrescent floral tube. Only Elaeagnus is represented in our area. Nelson and Fos- berg considered the family to be unigeneric, recognizing only Elaeagnus. The genus Hippophaé, sea-buckthorn, of Europe and Asia, is distin- guished by alternate leaves, unisexual flowers with a two-lobed floral tube, four stamens, poorly developed nectaries, and a long stigma. Generally, only H. rhamnoides L., composed of three subspecies, is recognized. In subsp. rhamnoides two races have been described, the one diploid (2n = 12), the other tetraploid (2n = 24). A count of 2m = 20 has also been reported for the species. The orange flesh of the “fruit”’ is reported rich in vitamin C. Vegetative multiplication is common by means of underground stems. The buffalo-berries, Shepherdia, comprising three North American species, have opposite leaves, unisexual flowers, a four-lobed floral tube, eight stamens, a ring of eight nectariferous glands in the floral tube at the base of the calyx lobes, and a short stigma. Chromosome numbers are known for S. canadensis Nutt., 2n = 22, and for S. argentea Nutt., on. 26, On the basis of anatomy, Hippophaé and Shepherdia are considered by Servettaz to be more closely related to each other than to Elaeagnus. The family is most commonly placed in the Myrtales next to the Thyme- laeaceae. It also shares several characters with the Proteaceae and Penaeaceae. Hutchinson alone closely relates the Elaeagnaceae to the Rhamnaceae. ’Prepared for a generic flora of the southeastern United States, a joint project of the Arnold Arboretum and Gray Herbarium made possible through the support of George R. Cooley and the National Science Foundation and under the direction of Carroll E. Wood, Jr., and Reed C. Rollins. The scheme follows that outlined at the beginning of the series (Jour. Arnold Arb. 39: 296-346. 1958). The area covered in this, as in former treatments, is bounded by and includes North Carolina, Tennessee, Arkansas, and Louisiana. Material included in the descriptions in brackets applies to — this area, and references marked by an asterisk have not been vetified by 1964 | GRAHAM, ELAEAGNACEAE 275 Clusters of root nodules are commonly formed in all three genera through infection of developing lateral roots by an unidentified endophyte. The causal organisms have been variously interpreted as filamentous fungi, nitrogen-fixing bacteria, actinomycetes, and myxomycetes, but their identity is still unknown due to difficulties in isolating and culturing them. The endophyte is soil borne and must be reintroduced into the plants of each new generation. It is apparently capable of fixing nitrogen and in this way maintains a symbiotic relationship with its host, much as do bacteria associated with legumes. The only other nonleguminous angio- sperm genera known to bear root nodules are Coriaria, Myrica, Alnus, Casuarina, Ceanothus, and Discaria; in none of these has the endophyte been positively identified. REFERENCES: Barton, H. Elaeagnacées. Hist. Pl. 2: 487-497. 1870. [See English transla- tion by M. M. Hartog, Nat. Hist. Pl. 2: 481-491. 1872. BENTHAM, G., & J. D. Hooker. Elaeagnaceae. Gen. Pl. 3: 203, 204. 1880. Bonp, G. The root nodules of non-leguminous angiosperms. pp. 72-92. pls. 1-4. In: Symbiotic Associations. 13th symposium. Soc. Gen. Microbiol. x + 356 pp. Cambridge, England. 1963. , J. T. MacConneti, & A. H. McCattum. The nitrogen-nutrition of Hippophaé rhamnoides L. Ann. Bot. II. 20: 501-512. pls. 23, 24. 1 Cooper, D. C. The development of the peltate hairs of Shepherdia canadensis. Am. Jour. Bot. 19: 423-428. pl. 36. 1932. . The chromosomes of Shepherdia canadensis. Ibid. 429-431. DarMer, G. Rassenbildung bei Hippophaé rhamnoides (Sanddorn). Biol. Centralbl. 66: 166-170. 1947. ErcHHo.z, W. Ein Beitrag zur vegetativen Vermehrung des Sanddorns (Hip- pophaé rhamnoides L.). Deutsch. Baumschule 9(7): 186-190. 1957.* FosBerc, F. R. Elaeagnaceae of Nevada. Contr. Fl. Nev. 26: 1—4. 1941. GARDNER, I. C. Nitrogen fixation in Elaeagnus root nodules. Nature 181: 717, 718. : . Bono. Observations on the root nodules of Shepherdia. Canad. Jour. Bot. 35: 305-314. 1957.* [ Nitrogen fixation. ] Gite, E. Elaeagnaceae. Nat. Pflanzenfam. III. 6a: 246-251. 1894. Groves, E. W. Hippophaé rhamnoides in the British Isles. Proc. Bot. Soc. Brit. Isles 3: 1-21. 1958. Hawker, L. E., & J. FrayMoutH. A re-investigation of the root-nodules of species of Elaeagnus, Hippophaé, Alnus, and Myrica, with special reference to the morphology and life histories of the causative organisms. Jour. Gen. Microbiol. 5: 369-386. pls. 1, 2. 1951. [Extensive bibliography; summar- rizes previous work. ] Massacetov, P. S. Alkaloids of plants of the family Elaeagnaceae. (In Russian; English summary.) Zhur. Obshch. Khim. 16: 139, 140, 1946.* Netson, A. Rocky Mountain herbarium studies. III. The Elaeagnaceae — a mono-generic family. Am. Jour. Bot. 22: 681-683. 1935. PENHALLow, D. P. Notes on Shepherdia canadensis. Canad. Rec. Sci. 3: 360- 363. 1889. [Fruits used in making jam.] PLouvier, V. Sur la présence de québrachitol chez les Elaeagnacées. Sa re- cherche dans quelques autres Myrtiflorae. Compt. Rend. Acad. Sci. Paris 232: 1239-1241. 1951. 276 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV SCHLECHTENDAL, D. F. L. von. Elaeagnaceae. DC. Prodr. 14: 606-616. 1857. SEerveTTAz, G. Monographie des Eléagnacées. Beih. Bot. Centralbl. 25: 1-420. 1909. [Includes taxonomy, anatomy, developmental morphology; 117 bibliographic references. | SoBOLEWSKA, H. Cinése somatique et cinése de maturation dans les Eléagnacées. Acta Soc. Bot. Polon. 4: 64-76. pls. 7-9. 1926. Trorimov, T. T. Geographical variability of seeds of Hippophaé rhamnoides L. (In Russian.) Vestn. Mosk. Univ. Biol. Pochvov. 1961(1): 45-58. 1961. Vries, V. DE. Some observations on the ecology of the common seabuckthorn. (In Dutch; English summary.) Nederl. Dendrol. Ver. Jaarb. 16: 48-64. 2 pls. 1. Elaeagnus Linnaeus, Sp. Pl. 1: 121. 1753; Gen. Pl. ed. 5.57. 1754. Shrubs or rarely trees, with alternate, unarmed or spiny branches, densely covered with silvery or brownish stellate or peltate hairs (scales). Leaves alternate, deciduous or persistent, petiolate, lanceolate to oblong or ovate, simple, entire, exstipulate, both sides or only the lower (abaxial) side covered with peltate hairs. Flowers regular, bisexual or unisexual (the plants then polygamo-dioecious) , pedicellate, solitary or clustered in the axils of leaves or deciduous bracts, 4-merous. Floral tube cylindrical to globose at base, slightly contracted at apex of the ovary, then flaring and cylindric to campanulate above, the portion surrounding the ovary persistent, accrescent, the portion above deciduous in fruit; calyx lobes 4, silvery or rusty on the outer surface, yellow within; nectariferous disc cone- or cup-shaped, surrounding the base or diffuse at the base of the style. Petals none. Stamens 4, at the, base of and alternating with the locular, 1-ovulate, with basal placentation; ovule anatropous, 2-integu- mented. Fruit berry- or drupelike, the outer layers of the persistent floral tube fleshy or mealy, the inner layer hard and strongly 8-striate or leathery to membranaceous with very weak striations, ellipsoid, with or without hairs on the inner surface, surrounding a membranaceous achene. Seed 1, with little or no endosperm; embryo large, straight, the cotyledons fleshy. Embryo sac development of the normal (Polygonum) type. LECTOTYPE species: E. angustifolius L.; see Britton & Brown, Illus. Fl. No. U.S. ed. 2. 2: 575. 1913. (Name from Greek, elaia, olive, and agnos, the classical name for the chaste-tree, Vitex agnus-castus.*) — OLEASTER, RUSSIAN OLIVE, SILVERBERRY. 2 The Greek feminine word agnos was adopted by Linnaeus and the ending altered to the Latin masculine -us. The gender of Elaeagnus is, therefore, masculine and not feminine as regarded in most taxonomic literature. (See Art. 73, sect. 2(2), Int. the altered Greek word. According to Linnaeus (Critica Botanica, pp. 127, 128. 1737; transl. by A. Hort, 1938, pp. 100, 101), “He [Gaza] translated Greek names into Latin according to the meaning. For example, Vitex was called in Greek Agnos, 1964] GRAHAM, ELAEAGNACEAE 277 A taxonomically difficult genus of about 40 species, including many subspecies and varieties, placed by Servettaz in sect. ELAEAGNuS (§ Deci- duae Serv.), defined as spring-flowering with deciduous leaves, and sect. SEMPERVIRENTES Serv., comprising fall-flowering species having evergreen leaves. The genus is best represented in the steppe regions of Asia and in southern Europe; E. commutatus Bernh. (E. argenteus Pursh, not Moench), 2m = 28, is native to Canada and the northern United States. Three Asiatic species cultivated in the southeastern United States have become naturalized to varying degrees in scattered localities. Elaeagnus pungens Thunb. (sect. SEMPERVIRENTES), 2” = 28, is established in North and South Carolina and Tennessee. Elaeagnus umbellatus Thunb., 2n = 28, (flowers in axillary umbel-like clusters) and E. multiflorus Thunb. (flowers solitary or paired in the axils), both of sect. ELAEAGNUS, occur in isolated populations as escapes. Definition is difficult in many of the European and Asiatic species be- cause of the morphological variability displayed over the extensive and ecologically diverse parts of their ranges. The presence of juvenile and adult leaves and the change in form of the fruit with maturation add to the difficulty. Development of spiny branches is most pronounced under dry, poor soil conditions and in young plants. Several species are cultivated for their handsome silvery foliage and decorative red or silvery fruits; the Russian olive, Elaeagnus angustifolius, 2n = 12, 28, is one of the most widely planted. Forms with variegated leaves are available in a few cultivated species. REFERENCES: Under family references see Bond (1963), GARDNER (1958), G1Lc, HAWKER & FraymMouTtH, NELSON, and SERVETTAZ AHLES, H. E., C. R. L, & A. E. Raprorp. Species new to the flora of North or South a ee 60: 10-32. 1958. [E. pungens, 18.] Brown, N. E. Elaeagnus argentea. Bot. Mag. 137: pl. 8369. 1911 FREEMAN, O. M. Notes on the flora of Polk County, North Carolina. Castanea 0: 37-57. 1955. [E. umbellatus var. parvifolius an occasional escape. ] Fucus, A. Untersuchungen iiber den mannlichen Gametophyten von Elaeagnus angustifolia. Osterr. Bot. Zeitschr. 85: 1-16. 1936. Hooker, J. D. Elaeagnus multiflora. Bot. Mag. 120: pl. 7 1894. James, R. L. Introduced plants in northeast Tennessee. dee 21: 44-52. 1956. [E. pungens, 44. NasH, G. V. Elaeagnus multiflora. Addisonia 4: 69, 70. pl. 155. 1919. Pomeroy, C. S$. Bud variation in Elaeagnus. Jour. Hered. 12: 227-230. 1921. that ie in Greek signifying castus; hence Gaza called Vitex ‘Castus’ instead of ‘Agnu “The barbarians who followed him niga agnus to be actually a Latin word, sub a synonym for ‘sheep’: hence they took agnus as a stantive and castus as an adjective, as ovis casta or agnus castus; and this acacia specially pleased the Pharmacists, though none more bairats could have been found. From Agnus also is formed Elecagnus, as it were Olea-Vitex: hence the right spelling is Elaeagnus, not Elaeachnus.’ 278 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLV Recorp, S. J. Gum canals in Elaeagnus. Trop. Woods 12: 21. 1927. [Vertical canals of traumatic origin in E. pungens and E. umbellatus. | SaKISAKA, M., & Y. Sueniro. Notes on the development of the ‘‘star hair’ of Elaeagnus. (Abstr.) Jap. Jour. Bot. 4: 102. 1928/1929. TroirsKata, O. V. On certain morphological mechanisms in Elaeagnus angusti- folia. (In Russian.) Akad. Nauk Kazakh. SSR. Izv. Biol. 1955(10): 153- 162. 1955.* 1964 | HOWARD, NOTES ON ROSACEAE 279 NOTES ON ROSACEAE IN THE LESSER ANTILLES RiIcHARD A. HOWARD THE NOTES WHICH FOLLOW illustrate some of the problems encountered in the slow task of preparing a flora for the islands of the Lesser Antilles. Much of the field work on which the discussion below is based was sup- ported by grants from the National Science Foundation. I am grateful, as well, to the director and staff of the Muséum National d’Histoire Naturelle, Paris, for the opportunity of studying several historical collec- tions which have long been unidentified. Chrysobalanus cuspidatus Griseb. ex Duss, Fl. Phan. Antill. Fr. 258. 1897; Griseb., Fl. Brit. West Indies 711. 1864 (nomen); Urban, Symb. Antill. 5: 351. 1907. Licania oligantha A. C. Smith, Jour. Arnold Arb. 28: 333. 1947, syn. nov. In a small section of his Flora of the British West Indies, following the list of corrections and preceding the Index, Grisebach published a ‘‘List of species received since the impression of the first volume.” Included among these names are a dozen epithets, each followed by: “Gr. (n. sp.),” the country of origin, and the collector. Not one of these names is ac- companied by descriptive data and none can be considered validly pub- lished. All of them, however, have been cited in subsequent publications dealing with the West Indies or have been indexed in various works and attributed to Grisebach. Each offers difficult problems of citation and typification; Chrysobalanus cuspidatus will serve as an example. This name was published by Grisebach who indicated (by an exclama- tion mark) that it represented a new species he had seen collected by Imray in Dominica. In the January 27, 1897, issue of Botanisches Central- blatt, E. Kiister published an article on the anatomical characters of the Chrysobalaneae in which he referred to “Chr. cuspidatus Griseb.” He cited an Eggers collection (edit. Toepffer 618) and a description of the upper epidermis and of the hairs (p. 137), although earlier in the text he had considered additional anatomical details. Although a protologue in all senses, this is regarded as an inadequate description of the species. Also in 1897, and presumably later in the year, Duss published a treat- ment of the plants of Guadeloupe and Martinique and supplied a more usual description of CArysobalanus cuspidatus which he attributed to Grisebach. Duss cited two of his own collections from Guadeloupe, 3476 and 3633, and one from Martinique, 154. The identification of much material Duss collected and cited was made or verified by Urban, and the original collections were deposited in Berlin. In 1907 Urban included 280 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Chrysobalanus cuspidatus in a list of new genera and species, attributing the name to Grisebach and supplying a Latin description. He cited only two of the three specimens Duss had listed and the Eggers collection Kiister had cited. There are, therefore, four possible authors but the correct citation of the name should give credit to Duss for the first pub- lished description of taxonomic value and should be — CArysobalanus cuspidatus Griseb. ex Duss. The identification of a holotype or the selection of a lectotype of Chrysobalanus cuspidatus is more difficult. The Imray specimen from Dominica, cited by Grisebach, is in the herbarium at the Royal Botanic Gardens, Kew, but this collection was not cited by any of the three men who published a description of the species. Duss did not select a holotype from the three collections he cited, nor did Urban from among the col- lections he mentioned. Regrettably no material studied by Duss or Urban could be located at the herbarium in Berlin, so the eventual selection of a lectotype is left to a future worker who, having material on loan from various institutions, can select the collection and the most complete speci- men to represent the species. In addition to its previously reported distribution from Guadeloupe, Dominica, and Martinique the range of this species can be extended to St. Lucia on the basis of collections of three individuals. Material col- lected by John Beard in St. Lucia during the years of the second world war was identified by A. C. Smith. Two of Beard’s collections, 488 and 492, were identified as a new species, Licania oligantha. A study of the cited material indicates an older description has been overlooked and Licania oligantha may be considered a synonym of Chrysobalanus cus pi- datus. Subsequent to Beard’s work additional material was collected in St. Lucia by R. A. Howard (11638) in 1950 and G. R. Proctor (216217) in 1959. This species, however, is still not represented by fruiting specimens and until that time its correct generic assignment is questionable. The field notes supplied by Beard from his collection 492, the type of Licania oligantha, describe the fruit as blue and attractive to birds. An annota- tion on the type sheet, however, indicated the fruit was not received with the collection and was lacking for the description published by Smith. A comprehensive study of the genera Chrysobalanus, Licania and Mogquilea, at least as they occur in the western hemisphere, is needed. The majority of recent floras have considered Moguilea synonymous with Licania and have recognized Chrysobalanus as a distinct genus. Distinc- tions between these two genera are suggested to be in the length of the inflorescence, or the number of flowers, or in the habit of the plant. CAry- sobalanus is considered by some to be monotypic; other authors suggest three to seven species may comprise the genus. CArysobalanus icaco, the type species, is admittedly polymorphic and of wide distribution. The fleshy blue colored fruit of typical C. icaco is different from the fruit of species of Licania. Beard’s description suggests a typical Chrysobalanus fruit for Licania oligantha. Smith did not consider CArysobalanus in his discussion of the relationships of Licania oligantha. This species with the 1964 | HOWARD, NOTES ON ROSACEAE 281 missing fruit may prove to be intermediate between typical C hrysobalanus and typical Licania. Once again, more material is needed and a detailed study not only of the variation of Chrysobalanus icaco wherever it occurs, but of the relations of the three genera C hrysobalanus, Licania and Moquilea. Potentilla anglica Laicharding, Veg. Eur. 1: 475. 1790. Potentilla procumbens Sibth., Fl. Oxon. 162. 1794. I can find no modern published record of the genus Potentilla in the Lesser Antilles. Potentilla argentea was listed as a cultivated plant in the Botanic Garden of St. Pierre, Martinique, in 1829 (De VHorme, Ann. Marit. Col. 1: 142. 1830), but apparently has not survived in cultivation anywhere in the area. In an old collection sent to me for determination is a specimen collected by Belanger in 1857. Although no specific locality is given, Belanger was director of the St. Pierre Botanical Garden between 1853 and 1881, and the specimen which is Potentilla anglica was pre- sumably a weed in the garden. Subsequent collectors have not recorded this species from the French islands but another unidentified collection, W. H. & Barbara T. Hodge 2157, made on Dominica, is the same species. The Hodges record their collection as:“‘a weed in moss in which roses sent from England were packed; only a few plants seen in a rose bed at Fern Villa, Ridgefield Estate.” Recent visitors have checked this location and failed to find the species established as a weed. Elsewhere in the West Indies Potentilla reptans L. was collected by Brown and Britton in Bermuda, in 1905, and has not been recollected. Potentilla anglica (as P. procumbens) was also collected by Harris above New Castle in Jamaica, in 1915, but has not appeared in recent collections from that island. With such scattered records, one questions the necessity for considering species of Potentilla in a floristic treatment of the Antilles. Rubus ferrugineus Wikstrém, Sv. Vet.-Akad. Handl. 1827: 68. 1828. Two recent collections from the island of St. Kitts appear to be best referred to this poorly represented species. Wikstrém described Rubus ferrugineus basing it on a Forsstroém collection from Guadeloupe. All subsequent authors have failed to consider the species with the single exception of Rydberg who included Rubus ferrugineus as a species endemic to Guadeloupe in his treatment of the genus in the North American Flora (22: 449. 1913). Unfortunately the description given by Rydberg is puzzling for he cited no specimens, and yet included details not given in Wikstrom’s original diagnosis. Two years later Rydberg (Bull. Torrey Club 42: 136. 1915) assigned to this species a collection, Duss 2215, from Guadeloupe and a collection, Nelson 3305, from Mexico. The Duss collection was originally identified as Rubus jamaicensis Sw. and so cited by Duss (Fl. Phan. Fr. Ant. 260. 1897), who indicated the plant was rare, being known from but one location on Guadeloupe and not reported 282 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV from Martinique. Dr. and Mrs. Henri Stehlé, residents of Guadeloupe, who have collected extensively in the French West Indies, have been unable to relocate the species and, in a most recent reference to the Rosaceae, list only the original Duss collection under the same specific determination given by Duss. (H. & M. Stehlé and Quentin, FI. Guad. Depend. et Mart. 2: 53. 1949). Neither Urban in his Symbolae Antillanae nor L. H. Bailey in his series of papers on Rubus refer to the Wikstrom species or the type collection by Forsstrom. In 1950 I visited St. Kitts and had the pleasure of going intc the montane forests of that island with the late Malcolm Smith. Instead of taking the routine trail up Mt. Misery, that excellent naturalist suggested a new route previously unbotanized which led through a palm break across a knife-like ridge into a mossy forest and finally, by a short descent, to a small crater lake called Dodan’s Pond at 2750 feet on Verchild’s Moun- tain. Mr. Smith explained there was a conflict regarding the existence of that pond which is reflected even in its name. Either the pond had been difficult to locate in former years and its existence was doubted, or the pond was evanescent, perhaps appearing in alternate years, for an optional spell- ing of the location he suggested was Dos d’Ans. At the edge of the pond and in a thicket of vegetation was a colony of blackberry plants. These had long weak trailing sterile stems and arching fertile ones capped with panicles of many flowers and excellent flavored small fruits. The speci- mens collected (Howard 11962) could not be associated with any known species from the Antilles, native or introduced, and additional material was needed. In 1959 Mr. Smith and George Proctor again visited Dodan’s Pond and recollected the Rubus. At first it seemed the material repre- sented a new species. It was similar to Rubus florifolius Bailey of His- paniola but had a single terminal panicle, smaller prickles, and longer sepals. It lacked the coriaceous leaves and glabrous pistils of Rubus florulentus Focke of Puerto Rico. Rydberg’s modern description of Rubus ferrugineus was then checked against the more general original description published by Wikstrém. Although a Forsstrém type has not been seen, the recent collections from St. Kitts appear to correspond accurately to the original description of Wikstrém, and suggest that many parts of Rydberg’s description be discounted. Rubus ferrugineus should be described as follows: Sterile shoots (primo- canes?) delicate, long and trailing, sparsely strigose-pubescent becoming glabrate, cylindrical, purplish-brown in color. Leaves of sterile shoots with petioles 1.5—2 cm. long, leaflets 3, ovate-lanceolate, 4 & 1 to 5.5 % 3 cm. long and broad, the apex acuminate, the base obtuse or rounded, the mar- gins sharply and finely serrate. Flowering canes arching, scrambling, slender and terete, strigose and non-glandular, pubescent becoming glabrate. Prickles numerous, scattered, recurved, broad-based and about 3 mm. long. Leaves 3-foliolate, the petioles 3-4 cm. long, strigose, con- spicuously armed with short recurved prickles. Leaflet blades oblong- ovate, 5.5 X 3 to 8.5 & 3.4 cm. long and broad, the apex abruptly short acuminate, the base rounded or very slightly cordate, soft in texture 1964 | HOWARD, NOTES ON ROSACEAE 283 (neither coriaceous nor parchment like), glabrate above, variously pubescent below. Leaflets with 8-9 pairs of primary veins. Inflorescence a terminal panicle to 20 cm. long and 10 cm. in diameter, the rachis gray-tomentose and armed with short recurved prickles. Calyx lobes gray- tomentose on both surfaces, ovate in flower, becoming elongated to 1 cm. long in fruit. Stamens numerous with the outer ones abortive and con- spicuous. Pistils gray-tomentose, forming mature achenes 3 mm. long, rugose. JOURNAL OF THE ARNOLD ARBORETUM VoL. XLV Jury 1964 NUMBER 3 THE GENERA OF RHIZOPHORACEAE AND COMBRETACEAE IN THE SOUTHEASTERN UNITED STATES ? SHIRLEY A. GRAHAM RHIZOPHORACEAE R. Brown in Flinders, Voy. Terra Austral. 2: 549. 1814, “Rhizophoreae.” (RED MANGROVE FAMILY) Evergreen trees or shrubs of muddy tidal shores, brackish streams or lagoons [or tropical rain forests], often with conspicuous prop or knee roots. Leaves mostly opposite with large, leaflike, caducous stipules. Inflorescences cymose or racemose. Fruit [a dehiscent capsule or] an indehiscent, leathery berry, panel the persistent calyx. Seeds 1 to many, viviparous in the genera of coastal mangroves. TYPE GENUS: Rhizophora L. A family of about 16 genera and 120 species distributed throughout the tropical and subtropical regions of the world, generally between 25° N. and 25° S. Latitude. Rhizophora L., Bruguiera Lam., Ceriops Arn., and Kandelia W. & A., which are the primary somponenits of much of the world’s coastal mangrove swamps,” are usually thought of as representative ‘Prepared for a generic flora of the southeastern United States, a joint project of the Arnold Arboretum and the Gray Herbarium of Harvard University made possible through the support of George R. oa and the National Science Foundation and apply primarily to the plants of this area, with avspeaponed information in bracket References not seen by the author are marked with an asterisk. The author is grateful for the helpful commen ie and suggestions of Dr. Wood and for the preserved materials of Rhizophora, eres and Conocarpus supplied by Drs. Wood, R. B. Channell, H. H. Iltis, and ja Tomlinson, and Mr. David Seligson. The illustration, ie work of Dor othy H. eae was supervised by Dr. Wo od. Theophrastus (305 8B.c.) wrote of mangroves in the Persian Gulf which were away up to the middle by the sea and are held up by their roots so that they look like a cuttlefish” (Theophrastus, Enquiry Into Plants eal A. Hort] 1: 343. 1916). An 286 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV of the family, but most of the genera are inhabitants of inland tropical rain forests. A single species, Rhizophora Mangle L., occurs in the south- eastern United States. Bentham and Hooker separated the four genera of mangroves from the rest of the family as tribe Rhizophoreae (leaves opposite, style one, endosperm scanty, and embryo macropodous). Wood structure of this tribe, also distinct from that of the inland genera, is characterized by scalariform intervascular pitting, heavy-barred scalariform perforation plates, scanty vasicentric parenchyma, multiseriate rays, and libriform fibers. In addition, only the coastal genera have prop roots and vivipary, apparent adaptations to the littoral habitat. Schimper ignored the mor- phological similarities among the genera of Rhizophoreae, believing them to be the result of parallel adaptations to the habitat, and related each genus to a different inland genus, a concept not generally accepted. Embryologically, the genera of Rhizophoraceae differ in several respects from one another (see Mauritzon); each genus shares one or more characters with several different families in the Myrtales. The family is considered most closely related to the Combretaceae on the basis of the unilocular ovary and usually single-seeded fruit. Chromosome numbers of 2n = 36 and 64 have been reported for the family. REFERENCES: BarLton, H. Rhizophoracées. Hist. Pl. 6: 284-304. 1875. BENTHAM, G., & J. D. Hooker. Rhizophoreae. Gen. PI. 1: 7 683. 1865. CANDOLLE, A. P. DE. ng mae mas 3: 31-34. Hov, D. oe In: C. G. G. J. VAN STEENIS, a Males. I. 5: 429- 493 Marco, H. 7 ‘Systematic ee of the woods of the Rhizophoraceae. Trop. Woods 44: 1-20. 193 MAvRITZzon, J. Contributions to the embryology of the orders Rosales and Myrtales. Lunds Univ. Arsskr. II. Sect. 2. 35: 1-120. 1939. ‘Rhizophora- ceae, 75-80. | ScHiImPER, A. F. W. Rhizophoraceae. Nat. Pflanzenfam. HI. 7: 42-56. 1892- 1893. 1. Rhizophora Linnaeus, Sp. Pl. 1: 443. 1753; Gen. Pl. ed. 5. 212. 1754. Gregarious trees or shrubs inhabiting mud flats of coastal tidal marshes or shallow brackish streams and lagoons, bearing conspicuous, bowed prop roots at the base of the tree and other adventitious roots from the lower branches. Branching opposite, rarely alternate. Leaves petiolate, decussate, simple, entire margined, leathery, persistent, elliptic, oblong to obovate or lanceolate, punctate on the lower (abaxial) surface, stipulate; stipules interpetiolar, leaflike, convolute, surrounding the unexpanded leaf and falling as the leaf expands. Inflorescence axillary, cymose, dichoto- mously branched with 2—8|—16]| flowers, or the flowers solitary, the pe- extensive account of Rhizophora in early literature may be found in Bowman (1917, pp. 592-603). 1964] GRAHAM, RHIZOPHORACEAE AND COMBRETACEAE 287 duncles and pedicels subtended by connate, 2—4-lobed bracteoles, the flowers borne among [or below] leaves of ‘the current year. Flowers pedicellate [or sessile], regular, bisexual, 4-merous. Sepals 4 (5), distinct, valvate in aestivation, persistent, leathery, each with a prominent central longitudinal rib on the inner surface. Petals 4, distinct, alternate with the sepals and folded lengthwise in bud into the depression between the ribs of adjacent sepals, deciduous, leathery, white or yellowish white, lanceolate, equal to or shorter than the sepals, the margins densely villous [or glabrous]. Stamens 8[—12], in 1 whorl, the filaments short or wanting; anthers elongate-deltoid, apiculate, introrse, the inner (adaxial) sides filled with numerous areolae containing pollen and covered by a thin membrane, the apex and 2 longest sides of the membrane becoming free during pollen discharge; pollen oblate-spheroidal, tricolporate, the pores fused laterally forming a zone around the equator, the exine smooth to faintly scabrate, the greatest diameter of the grain about 18. Gynoecium syncarpous, bicarpellate; stigmata punctate; style filiform and forked at the apex [or nearly wanting]; ovary 1, the upper portion solid, the 2(—4) locules inferior, each with 2 axile, anatropous ovules pendulous from the apex of the locule and surrounded by spongy tissue, the micropyle up- ward and outward. Fruit a conical, indehiscent, leathery berry, attached at the base to the receptacle. Seed 1, rarely 2, the other ovules abortive, the apparent placentation parietal due to displacement of the placenta and abortive ovules by the developing seed; seed coat thick, fleshy. Embryo straight, green, in the micropylar end of the endosperm, developing (with- out a dormant period) into the seedling while still within the fruit on the tree; hypocotyl pendulous, clavate, protruding from the fruit and ultimately disarticulating at the junction with the cotyledons; cotyledons thick, fused into a green tube, protruding from the fruit and withering with it after the seedling has fallen. Embryo sac development of the normal (Polygonum) type; endosperm development of the nuclear type. LECTOTYPE SPECIES: R. Mangle L., typified by removal of the remaining Linnaean species of RAizophora to other genera of the Rhizophoraceae. (Name from Greek, rhiza, root, and phoros, bearing, in reference to the conspicuous prop roots.) — MANGROVE. A pantropical genus of six to nine species in two sections, RHIZOPHORA and AéRopE Endl.; represented in the Americas by Rhizophora racemosa G. F. W. Mey., R. Harrisonii Leechman, and R. Mangle of sect. Ru1zo- PHORA (flowers borne among leaves produced the same year, peduncles as long as or longer than the petioles, flowers pedicellate, and stamens eight). Only R. Mangle, the red mangrove, a pioneer colonist of silted tidal shores, extends northward into our area to Levy County, on the west coast of Florida, and to Volusia County, on the east, with its greatest development along the low southwestern shore in the region of the Ten Thousand Islands. North of Volusia County the red mangrove does not survive severe winter frosts, and the mangrove swamp is replaced by . Rhizophora. a-k, R. Mangle: a, fruiting ser Sage successive stages in Pasi of hypocotyl, x 1 : b, infloresc 4 wer, > 2; d, stamen, inner view, showing mode of dehiscence a ‘po ollen "areolae, x 16; oecium and calyx in diagrammatic vertical section, the pendulous i Ne Spinel by spongy tissue, x 4; f, two-locular ovary in diagrammatic cross section at vel of broken line in Ke,” x 4; g-j, enlarging fruit in semi- diagrammatic vertical section, showing progressive sah orca from embryo to seedling, X 2 (except j, X 1)—note endosperm (stipple ed), 1-seed coat, 2- aborted ‘ovule, 3-embryo, 4-cotyledons, 5-hypoco ie is “j” hypocotyl disarticu- lated from the co tyledonary tube revealing plumule) ; 7 fruit with seedling showing fully elongated cotyledonary tube and bopbeuial M172, 1964] GRAHAM, RHIZOPHORACEAE AND COMBRETACEAE 289 salt-marsh vegetation. The mangrove areas of Florida have been estimated to cover more than 1000 square miles (Davis). The same species, Rhizophora racemosa, R. Harrisonii, and R. Mangle, also occur along the West African coast from Senegal to Angola. Rhizo- phora racemosa, infrequent in the Americas, is the dominant species along the African shore, while R. Mangle, the most widespread American species, is less abundant and restricted to dryer ground inland. According to Savory, R. Mangle, having a high salt tolerance, is confined in Africa to land which, during the dry season, is flooded only at high tide, so that salt accumulates by evaporation, and which, during the wet season, is not flooded by fresh river water in the delta areas. Differences in dis- tribution of this species on the two continents are difficult to explain on the basis of present information. The species of Rhizophora on East African shores are of Indo-Malaysian affinity, the East and West African coasts sharing no species in common. Plants of R. Mangle from the Pacific coasts of North and South America and a few Pacific islands (e.g., Fiji, Samoa, Tonga, and New Caledonia) have slightly smaller flowers with shorter styles than those in the rest of the range and have been recognized by some authors as a distinct species, R. samoénsis (Hochr.) Salvoza. This distinction is questionable, with at least one character, length of styles, so variable as to make separation of the species on that character impossible (Hou, 1960). There are no positive records of hybrids within the genus. Environmental factors determining the distribution of the red mangrove include type of soil, salinity of water and soil solutions, tides, water levels, and temperature. Rhizophora grows best in shallow water (with a salt concentration at least one-third that of sea water) on deep peat or marl soils free from violent wave action. Recent experiments with R. Mangle seedlings in various concentrations of salt solutions show that best de- velopment occurs with a salt concentration equalling that of sea water, although seedlings are apparently facultative to a wide range of salinity and are even capable of surviving in fresh water to a limited extent (Stern & Voigt). Plants growing naturally in waters with relatively low salt concentrations, as in parts of the Everglades, tend to be stunted or dwarfed. The high salinity of the water and soil in which Rhizophora grows suggested to Schimper that mangroves were physiologically xero- phytic and, consequently, had a low rate of transpiration, a much-disputed conclusion. Recent studies indicate transpiration rates are indeed low, but more work is needed on this problem, as well as on all aspects of the physiology of Rhizophora. In addition to Rhizophora, the name “mangrove” has been applied to several different genera in at least four other families. Avicennia (Ver- benaceae), Pelliciera (Theaceae), Sonneratia (Sonneratiaceae), and La- guncularia and Conocarpus (Combretaceae) are all ‘‘mangroves,”’ charac- terized by high salt tolerances and the ability to persist along heavily silted shores. The extensive tidal forests which they form are known as mangrove swamps. The swamp formation, an exceptionally uniform one 290 JOURNAL OF THE ARNOLD ARBORETUM | VOL, XLV for the tropics, where diversity is the rule, is generally comprised of three or four genera of mangroves, either in mixed populations, or occupying particular zones in relation to the distance from the shore line. When zonation occurs, Rhizophora is in the outermost zone where the substrate is covered by water even at low tide. Other genera, such as Avicennia or Laguncularia, occur farther inland, where they are only partially flooded at high tide. The tidal action apparently promotes good aération and gas exchange in the prop roots. Once established, mangrove swamps may play a minor role in accumulation of debris and silt leading to new land formation, but it is doubtful that they are of value in land reclamation, since accumulation of silt usually is not initiated by, but must precede, their establishment. e red mangrove has both absorptive underground roots and aerial roots. Clusters of superficial, underground, secondary roots appear early in the seedling, taking over the function of the weakly developed taproot. These bear numerous absorptive rootlets, each with a small ‘“rootcap” composed of cells adhering to the rootlet as it grows laterally outward from the parent root. Aérial roots, known as prop roots for their sup- posed stabilizing action in anchoring the plant to the substrate and generally most abundant on the offshore side ot the tree, form along the main stem of the young mangrove about the second year of growth. In older plants additional adventitious aérial roots may form along the lower branches. The surface of the aérial roots is supplied with lenticels which function in maintaining a high oxygen tension in the underground roots submerged in anaérobic mud The genera of mangroves in the Rhizophoraceae are notable for vivipary, the germination of the embryo in the seed while the fruit is still attached to the parent plant. In RAizophora normally only one or rarely two ovules develop. Polyembryony is unknown. The growth of more than one hypocotyl from a fruit, implying the occurrence of polyembryony, is due to development of more than one ovule. As the embryo and endo- sperm grow, the nucellus and inner integument disintegrate, while the outer integument becomes very thick and fleshy, with much vascular tissue, purportedly for transport of nutrition to the embryo. The endo- sperm, composed of large, translucent cells, is abundant and during de- velopment of the embryo grows out of the micropyle, spreading over the outer integument to form an arillode. The presence of absorptive ope s ] The embryo develops directly into the seedling without a dormant period. The hypocotyl elongates, piercing through the seed coat and apex of the ovary wall, and is followed by the distal end of the cotyledons, which are fused into a green tube. (See Fic. 1, j.) The plumule lies in the hollow center of the cotyledonary tube. When the hypocotyl is fully elongated, an abscission layer forms at the junction of the hypocotyl and the cotyle- dons causing disarticulation of the two parts, the hypocotyl and plumule falling into the water or mud below, the cotyledons remaining attached 1964] GRAHAM, RHIZOPHORACEAE AND COMBRETACEAE 291 to the fruit and withering with it. From fertilization to the falling of the seedling is said to occupy nine to ten months. The hypocotyl at the time of fall averages about 25 cm. in length. The question of how the seedlings become planted has been a contro- versial one. Early writers (and even many recent ones) believed the hypo- cotyls plunged vertically downward from the tree, like falling darts, into the mud below where they took root, but more recent observations show that the hypocotyl generally is unable to land erect in the mud because of its slight curvature. It is suggested that the hypocotyls upon falling float, first horizontally, later vertically, until stranded, then gradually be- come erect through strong upward curvature at the base just above the radicle. The seedlings may float for several months and can retain their vitality up to five months in a dry state, making them capable of being dispersed great distances (Guppy). Regeneration of roots and buds from short pieces of hypocotyl has been reported. Stellate and H-shaped sclereids are present in practically all parts of the plant. The air-dried bark contains up to 40% tannin, but the tannin is of poor quality, imparting an undesirable red color to leather, and is not used extensively. Large mangrove trees are used in building, par- ticularly as underwater pilings. Small trees are cut in many parts of the world for firewood, and the dense, hard wood is used in making charcoal. Fossil leaves from the Middle Eocene of Georgia, and a fossil calyx from the Lower Eocene of Tennessee have been described as possibly belonging to Rhizophora. REFERENCES: See also the family references. Berry, E. W. The Middle and Upper Eocene floras of southeastern North America. U. S. Geol. Surv. Prof. Pap. 92: 1-206. 65 pls. 1924. [R. Mangle, 188, 189.] . Revisions of the Lower Eocene Wilcox flora of the Southeastern States. Ibid. 156: 1-196. 50 pls. 1930. [R. Mangle, 142, 143. Bowman. H. H. M. Physiological studies on Rhizophora. Proc. Natl. Acad. U. S. 2: 685-688. 1916. [Transpiraiion rates in various soils with various salt concentrations. | . Ecology and physiology of the red mangrove. Proc. Am. Philos. Soc. 56: 589-672. pls. 4-9. 1917. ——. Histological palvaer in Rhizophora Mangle. Rep. Mich. Acad. Sci. 22: 129-134. pls. 9 1921. CHapMan. V. J. 1939 Cee Ge University Expedition to Jamaica. I. A study of the botanical processes concerned in the development of the Jamaican shore-line. Jour. Linn. Soc. Bot. 52: 407-447. pls. 16-20. 1944. [Mangrove vegetation, 423-445.] II. A study of the environment of Avicennia nitida Jacq. in Jamaica. /bid. 448-486. [Deals also with R. Mangle. | Cook, M. T. The embry ew is Pai Mangle. Bull. Torrey Bot. Club 34: 271-277. pls. 22, Davis, J. H. The ee aa poe role of mangroves in Florida. Pap. Tortugas Lab. 32: 303-412. pls. 1-12. 1940. 292 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Ecter, F. E. The dispersal and establishment of red mangrove, Rhizophora, in Florida. Carib. Forester 9: 299-310. 1948. FaBer, F. C. von. Zur Physiologie der Mangroven. Ber. Deutsch. Bot. Ges. 41: 227-234. 1923. [Transpiration rates of Rhizophora.|. GENKEL’, P. A., & I. S. Fan. On the physiological significance of viviparity in mangrove plants. (In Chinese; Russian summary.) Acta Bot. Sinica 7: 51-70. 1958. [Increase in salt concentration in the developing embryo serves to precondition the seedling for subsequent growth in saline water. | GLEDHILL, D. The ecology of the Aberdeen Creek mangrove swamp. Jour. Ecol. 51: 693-703. 1963. [Discussion of distribution of spp. of Rhizophora in coastal and riverine mangrove swamps, Sierra Leone, Africa; suggestion of hybridization between R. capil and R. Harrisonii. | Gottey, F., H. T. Opum, & R. F. Witson. The structure and metabolism of a Puerto Rican red mangrove ee in May. Ecology 43: 9-19. 1962. Guppy, H. B. Plants, seeds, and currents in the West Indies and hie xl + 531 pp. front., 3 maps. London. 1917. [R. Mangle, 96-110. | Harris, J. A., & J. V. LawreENcE. The osmotic concentration of the sap of the leaves of mangrove trees. Biol. Bull. Woods Hole 32: 202-211. 1917. Hovpripce, L. R. Some notes on the mangrove swamps of Puerto Rico. Carib. Forester 1(4): 19-29. 1940. [Composition of swamps, descriptions of woods, economic uses of mangroves. | Hou, D. A review of the genus Rhizophora with special reference to the Pacific species. Blumea 10: 625-634. 1960. — F. P. The genus Rhizophora in Suriname. Acta Bot. Neerl. 8: 58-60. 959. ae R. W. J. Rhizophora in West Africa. Kew Bull. 8: 121-127. 1953. Kumar, L. S. S., & W. V. Josut. False polyembryony in viviparous Rhizophora mucronata ne Curr. Sci. Bangalore 11: 242. 1942.* LaRue, C. D., & T. J. Mustx. Does the mangrove really plant its seedlings? Science a 661. 662. 1951. ———. Growth, regeneration, and precocious rooting in Rhizophora Manegle. Pap. Mich. Acad. Sci. Arts Lett. I. 39: 9-29. 1953 LAWRENCE, D. B. Self-erecting habit of seedling red mangroves (Rhizophora Mangle L.). Am. Jour. Bot. 36: 426, 427. 1949. LIEBAU, O. any zur Anatomie und Morphologie der Mangrove-Pflanzen, insbesondere hres Na cil Beitr. Biol. Pfl. 12: 181-213. 1914. Rhizophora, 205, 206. Luspsock, J. A contribution to our knowledge of oo Popular ed. vi + 288 pp. London. 1907. [R. Mangle, 236-238. | REINDERS-GOUWENTAK, C. A. Sonneratiaceae nd: fs mangrove-swamp fami- lies, anatomical structure and water relations. Jn: C. G. G. J. VAN STEENIS, Fl. Males. I. 4: 513-515. : RipLey, H. N. The dispersal of plants throughout the world. xx + 744 pp. 22 pls. Kent, England. 1930. [Rhizophoraceae, 287-290, pl. 14. | SaLvoza, F. M. mance Nat. Appl. Sci. Bull. 5: 179-237. pls. 1-9. 1936. [Monographic treatment. | Savory, H. J. A note on ‘cc ecology of Rhizophora in Nigeria. Kew Bull. 8: 127, 128. 1953. SCHIMPER, A. F. W. Plant geography upon a physiological basis. (Transl. by W.R. FISHER.) xxx + 839 pp. Oxford. 1903. [ Mangroves, 395-410. | 1964] GRAHAM, RHIZOPHORACEAE AND COMBRETACEAE 293 ScHOLANDER, P. F., L. vAN Dam, & S. I. ScHOLANDER. Gas exchange in the roots of mangroves. Am. Jour. Bot. 42: 92-98. 1955 Stncu, T. C. N., & A. T. Natarajan. Morphology of the pollen grains of the constituents of the mangrove vegetation. (Abstr.) Proc. Indian Sci. Congr. 40(3): 99, 100. 1953 [1954].* STEARN, W. T. A key to West Indian mangroves. Kew Bull. 13: 33-37. 1958. [Includes Rhizophora. | Sreenis, C. G. G. J. van. The distribution of mangrove plant genera and its significance for palaeogeography. Proc. Nederl. Akad. Wetenschap. Amster- dam C. 65: 164-169 6 Srern, W. L., & G. K. Vorcr. Effect of salt concentration on growth of red mangrove in culture. Bot. Gaz. 121: 36-39. 9. Tuorne, R. F. Flowering plants of the waters and shores of the Gulf of Mexico. U. S, Fish Wildlife Serv. Fish. Bull. 89: 193-202. 1954. [Mangrove swamps, 194-196. ] TrimpLe, H. Mangrove tannin. Contr. Bot. Lab. Univ. Penn. 1: 50-55. pl. 7. 1892 VAUGHAN, T. W. The geologic work of mangroves in southern Florida. Smithson. Misc. Collect. 52: 461-464. pls. 46-52. 1909. Warminc, E. Tropische Fragmente. II. Rhizophora Mangle L. Bot. Jahrb. 4: 519- 548. pls. 7-10. 1883. [Descriptive account. WenzeEL, H. Die Mangroverinden als Gerbmaterialen. Anatomische Unter- suchungen der gerbstoffreichsten Mangroverinden. Bot. Arch, 12: 59-96. 1925. [R. Mangle, 67-72.] COMBRETACEAE R. Brown, Prodr. Fl. Nov. Holland. 351. 1810, nom. cons. (WuITE MANGROVE FAMILY) Shrubs, trees [or woody climbers], mainly of tropical coasts. Buds naked (at least in our genera). Leaves evergreen or deciduous, entire, opposite, alternate, or spirally whorled in dense clusters at the ends of short branches, exstipulate, petiolate or sessile, the blade lanceolate, elliptic, or obovate, often with 2 glands at the base and numerous small pits (domatia) on the undersurface. Inflorescence an axillary spike, terminal panicle, or compact, globose head. Flowers regular | rarely slightly irregular], sessile, bisexual or both bisexual and ¢ in the same inflorescence, usually proterogynous, each subtended by a bract [or the bract absent]. Floral tube glabrous, hairy [or scaly], of 2 distinct parts [of 1 part in the West African Strephonema]|; lower part adnate to the ovary, tubular or flattened laterally and + 2-winged; upper part, above the ovary, deeply or shallowly cupuliform, persistent or deciduous; calyx lobes (4)5(6)-merous, deltoid, valvate or imbricate in the bud, hairy [or glabrous| within, sometimes scarcely developed. Petals none or 5 and distinct, greenish white, shorter [longer] than the calyx lobes, deciduous. Stamens 5-10, in 2 whorls, deciduous; filaments scarcely to well exserted; anthers orbicular or cordiform, versatile. 2-locular, longitudinally de- hiscent. Gynoecium probably 1-carpellate; stigma capitate or punctiform, style filiform, free from [or adnate to] the upper part of the floral tube, generally surrounded at the base by a nectariferous disc; ovary inferior 294 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV [half-inferior in Strephonema], 1-locular; ovules 2(—6), anatropous, pendulous from the apex of the locule on slender, generally papillose funiculi. Fruits indehiscent, of various types (scalelike and imbricate in a cone-shaped head, or ellipsoid and laterally flattened, or fleshy and drupelike [or dry with 2-5 broad, papery wings]), composed _pre- dominantly of flotation tissue, sometimes also with a sclerenchymatous layer. Seed 1, the other ovules abortive; seed coat membranaceous to thick ; endosperm of the nuclear type, scanty or none. Embryo large, straight, often green; cotyledons thick, fleshy, spirally convolute [or plicate|; radicle weakly developed, superior. (Terminaliaceae Jaume St.-Hil. Expos. Fam. 1: 178. pl. 29. 1805.) Type Genus: Combretum Loefl., nom. cons. (Grislea L., 1753, nom. rejic.). A pantropical family of about 18 genera and 450 species; the largest genera, Combretum (250 species) and Terminalia (200 species), best de- veloped in the Old World Tropics. The classification of Engler & Diels, with two subfamilies and five tribes, is generally followed. Four genera in two of the four tribes of subfam. Combretoideae are represented in the Southeastern States in southern Florida or the Keys. The Combretaceae are related to Rhizophoraceae and Myrtaceae and distinguished from them by the unilocular ovary with two to six ovules suspended from the apex of the locule by slender funiculi. The family is characterized anatomically by the common occurrence of intraxylary phloem in the stem, bicollateral veins in the leaves, and the presence at the base of the leaf blade of two flask-shaped cavities containing a multi- cellular gland. Transparent dots sometimes apparent in the leaves are due to large stellate idioblasts in the mesophyll which contain calcium oxalate crystals. Curious unicellular hairs containing a conical or concave membrane which makes them appear bicellular are widely distributed in the family (cf. Cistaceae). The fruits of practically all genera are well adapted to dispersal by water by virtue of their thick aérenchymatous mesocarp and the ability to float several months in sea water without adverse effect on germination (Guppy). The seed is protected in many genera by a stony endocarp. Most species are believed to be insect pollinated. The main independent evolutionary tendencies in the family, according to Exell (1954), are elongation of the upper part of the floral tube, leading to pollination by long-tongued insects; congestion of small tlowers into spikes or racemes, accompanied by reduction in size of petals; and development of winged fruits, allowing for dispersal by air. Fossil leaves thought to represent Combretum, Conocarpus, Laguncularia, and Terminalia are known from the Eocene (Claiborne flora) of Alabama, Mississippi, and Georgia. In addition, a fossil pollen grain of an unknown genus of the Combretaceae has been recorded from the same flora, and a fossil flower related to Combretum has been collected from Eocene beds of Tennessee. Timbers of several species of Terminalia are of local importance in construction, and Bucida and Laguncularia are also used to a limited 1964] GRAHAM, RHIZOPHORACEAE AND COMBRETACEAE 295 extent in the West Indies for building. A tea made from the roasted leaves of Combretum sundaicum Miq., a jungle climber native to Malaya, is believed to combat craving for opium. Another climber, Quisqualis indica L., Rangoon creeper, 2n = 22, 24, is cultivated in warm regions for its attractive rose or red tubular flowers. Chromosome numbers of 2n = 22, 24, 26, and 36 are reported in the family. REFERENCES: BaILton, H. Combrétacées. Hist. Pl. 6: 260-283. 1875. BENTHAM, G., & J. D. Hooker. ee Gen. Pl. 1: 683-690. 1865. Berry, E. W. A fossil flower from the Eocene. Proc. U. S. Natl. Mus. 45: 261-263. pl. 21. 1913. [Combretanthitis eocenica. | . The Middle and Upper Eocene floras of southeastern North America. U. S. Geol. Surv. Prof. Pap. 92: 1-206. 65 pls. 1924. [Combretaceae, 85- 87, 189, 190. ] BraNnpis, D. Combretaceae. Nat. Pflanzenfam. III. . a 1898. CANDOLLE, A. P. bE. Combretaceae. Prodr. 3: 9-24. ENGLER, A., & L. Diets. Monographieen eee Saget gates und -gattungen. III. Combretaceae-Combretum. iv + 116 pp. pls. 1-30. Leipzig. 1899. IV. Combretaceae excl. Combretum. 44 pp. re 145. Leipzig. 1900. EXELL, A. W. The genera of Combretaceae. Jour. Bot. 69: 113-128. 1931. . Combretaceae. Jn: C. G. G. J. vAN STEENTS, Fl. Males. I. 4: 533-589. 1954, Combretaceae. Jn: R. E. Woopson & R. W. ScuHery, Fl. Panama. Ann. Mo. Bot. Gard. 45: 143-164. 1958 Gray, J. Temperate pollen genera in the Eocene (Claiborne) flora, Alabama. Science 132: 808-810. 1960. [Fossil pollen grain of Combretaceae, not identified to genus. | Guppy, H. B. Plants, seeds, and currents in the West Indies and Azores. xi + 531 pp. front., 3 maps. London. 1917. [Dispersal of West Indian genera of Combretaceae by water currents; Terminalia Catappa, 116, 117; Cono- carpus erectus, 201-204; Laguncularia racemosa, 221-223. Herwen, H. Anatomische Characteristik der Combretaceen. Bot. Centralbl. 55: 353-360, 385-391; 56: 1-12, 65-75, 129-136, 163-170, 193-200, 225-230. 1 pi. Hitiier, J. M. Anti-opium plants. Kew Bull. 1907: 198, 199. 1907. | Includes Combretum sundaicum e Malayan anti -opium plant. /bid. 1908: 235, 236. 1908. [Com- bretum ana ] Lerévre, G. R. Contribution a l'étude anatomique et pharmacologique des ombrétacées. 126 pp. Paris. 1905.* Mauvritzon, J. Contributions to the embryology of the orders Rosales and Myrtales. a Univ. Arsskr. II. Sect. 2. 35: 1-120. 1939. [Combreta- ceae, 85- RAJAGOPALAN, = R. ee studies in Bee lancet I. (Abstr.) Proc. Indian Sci. Congr. 36(3): 138. 1949 SARGENT, C. S. ee ere N. ‘ci 5: 19-29. pls. 201-203. 1903. | Terminalia, 19, 20; ee 21, 22, pl. 201; ne 23-25, pl. 202; Laguncularia, 27-295, pl. SoLEREDER. H. Zur a a Systematik der Combretaceen. Bot. Centralbl. 23: 161-166. 296 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Stace, C. A. Cuticular characters as an aid to the taxonomy of the South-West African species of Combretum. Mitt. Bot. Staatssam. Miinchen 4: 9-17. 1961. [Includes key to spp. based entirely on cuticular characters. | STEARN, W. T. A key to West Indian mangroves. Kew Bull. 13: 33-37. 1958. [Includes Conocarpus and Laguncularia. | THORNE, R. F. Flowering plants of the waters and shores of the Gulf of Mexico. U. S. Fish Wildlife Serv. Fish. Bull. 89: 193-202. 1954. [Discussion of mangrove swamps, including Conocarpus and Laguncularia, 194-196. | KEY TO THE GENERA OF COMBRETACEAE General characters: leaves with 2 inconspicuous glands at the base of the blade; flowers regular, the floral tube mostly of 2 parts, the lower part generally tubular and adnate to the ovary, the upper part cupuliform; ovary inferior, 1-locular; ovules Foner iad on Jone funiculi; fruit indehiscent, 1-seeded A. Leaves spirally whorled in terminal clusters on short branchlets; fruits ovoid or obliquely ovoid, not conspicuously flattened or winged; petals none; inflorescence racemose; bracteoles on floral tube lacking. B. Fruit fleshy, drupelike. ovoid, 20-70 mm. —- calyx lobes well de- veloped, deltoid, deciduous from apex of fruit. ........ . Terminalia. B. Fruit dry and leathery, obliquely ovoid, 5- 0 mm. long; calyx lobes scarcely developed, persistent at apex of fruit. .......... 2. Bucida. A. Leaves alternate or ey not clustered; sett scalelike or obovoid, laterally flattened, 2-w C. Leaves alternate; fruits waits: imbricate in a cone-shaped cluster; 3 bracteoles on the floral tube none; petals none. . 4 onocarpus. C. Leaves opposite; fruits obovoid, solitary: bracteoles o on the floral tube 2; PME Be 5 ee he A ok EE aE hb aE UES OR Oe oS 4, Laguncularia. Subfam. COMBRETOIDEAE Tribe TERMINALIEAE DC. 1. Terminalia Linnaeus, Syst. Nat. ed. 12. 2: 674 (err. 638). 1767; Mant. Pl. 21, 128. 1767, nom. cons. Trees [rarely shrubs], often with sympodial branching. Leaves de- ciduous, large, glabrous or with sparse reddish-brown hairs, the apex acute to acuminate, the blade minutely verruculose [rarely with canal-like muci- lage cavities] and with 2 inconspicuous glands at the base. Inflorescence a densely flowered axillary spike [rarely a terminal or axillary panicle]. Flowers bisexual or both bisexual and ¢ with the ¢ flowers on the terminal part of the sp:ke, the bisexual flowers on the basal part of the spike, [4]5-merous, subtended by an ovate bract [or bract wanting], both flowers and bracts with dense reddish-brown pubescence. Floral tube greenish white, the lower part narrowly tubular, the upper part shallow, broadly cupuliform, deciduous; calyx lobes [4] 5 [6], deltoid, valvate in bud, densely silky haired within at the base jor glabrous]. Petals wanting. Stamens [8] 10 [12]; filaments exserted; anthers orbicu- Jar. Stigma punctiform. Style surrounded at the base by a thickened 1964] GRAHAM, RHIZOPHORACEAE AND COMBRETACEAE 297 5[4 or 6]-lobed disc, the disc densely covered with silky reddish-brown hairs [or glabrous]; ovules 2 (3 or 4). Fruit fleshy and drupelike, ovoid, apiculate at the apex; pericarp of several layers: a fleshy outer one, a fibrous one, a spongy one often containing air cavities, and an irregularly shaped sclerenchymatous inner one, often with pockets of aérenchyma [or the fruit dry, membranaceous, 2—5-winged]. Seed coat thin. (Ad- amaram Adans., 1763, and Panel Adans., 1763, nom. rejic.) TYPE SPECIES: T. Catappa L. (Name from Latin, terminalis, terminal, referring to the leaves clustered at the ends of the branchlets.) A complex pantropical genus of about 200 species in 20 sections; represented in our area by Terminalia Catappa L., Indian almond, 2n = 24, of sect. TERMINALIA (§ Eucatappa Engl. & Diels), in which the fruit is compressed but lacks wings. Native to Polynesia, 7. Catappa has be- come naturalized in scattered areas of southern Florida and the Keys. It is widely planted in the tropics as a shade tree, and the young trees, particularly, are admired for their regular, tiered manner of branching. Finger-like growths at the base of the petioles have been interpreted as rudimentary stipules which are drawn inward to an axillary position as the blade of the leaf grows. The seed is edible and is said to have a filbert-like flavor. Its size varies widely, the cultivated races generally having larger fruits than the wild trees (Exell, 1954). The fruits, contain- ing 30-35% tannic acid, find limited use in tanning. REFERENCES: Under family lepesanee see BRANDIS, EXELL (1931, 1954, 1958), Guppy (pp. 116, 117), and Sar AsEnjo, C. F., & J. A. Goyco. El sear ree almendro tropical. Puerto Rico p. Agr. ‘one Bol. Mens. 1: 5-7. * [T. Catappa. | CHowpHury, K. A. The so- -called ea spe ai ok cells in the wood of nee sc tomentosa W. & A. Nature 133: 215. 1934. [| Parenchyma t formed as last tissue of the annual ring, but as the first cells at the be- pecie of the new growing season Hooker, W. J. Terminalia Catappa L. “Bot. Mag. 57: pl. 3004. 1830. PAL, i ri the embryology of Terminalia Catappa. Sci. Cult. 17: 178, 179. 951. Ee i: R. Tropical-almond Terminalia: Fruit, oil, and tanning. (In Portuguese.) Bol. Agr. Minas Gerais Dep. Prod. Veg. 8: 69-71. 1959.* [T. Catappa. | WEBERLING, F. Weitere Untersuchungen iiber das Vorkommen rudimentirer Stipeln bei den Myrtales (Combretaceae, Melastomataceae). Flora 149: 189-205. 1960. [T. Catappa and Bucida Buceras studied. | 2. Bucida Linnaeus, Syst. Nat. ed. 10. 1025, 1368. 1759, nom. cons. Trees of tropical coastal hammocks, the young growth clothed with silky reddish-brown hairs; branching sympodial. Leaves evergreen, petio- late [or sessile], obovate to elliptical, mostly without glands at the base of the blade, arranged in spiral whorls in dense clusters at the ends of 298 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV branchlets, the branchlets subtended by 2 short spines or without spines. Inflorescence a densely flowered axillary spike [rarely pseudo-umbellate |. Flowers bisexual, 5-merous, subtended by an ovate bract, both flowers and bracts with dense reddish-brown pubescence. Floral tube greenish, the lower part tubular, the upper part shallowly cupuliform, persistent; calyx lobes 5, scarcely developed, valvate in bud, hairy within. Petals wanting. Stamens 10; filaments exserted; anthers cordiform. Stigma punctiform. Style surrounded at the base by a thickened disc. Fruit obliquely obpyriform, not winged or ribbed, the apex bearing the per- sistent upper part of the floral tube; pericarp composed of a leathery outer layer and a central stony layer with spongy aérenchyma on both sides. Seed coat thin. (Buceras P. Br., 1756, nom. rejic.) TYPE SPECIES: B. Buceras L. (Name from Latin, bucida or bucaeda, one who is beaten by a whip made of ox hide; application of the name obscure.) A tropical American genus of three or four species, with one, Bucida Buceras, black-olive, widely distributed from Panama to southern Mexico and in the West Indies, extending northward to Elliott’s Key in the Florida Keys. The genus is very similar to, and perhaps derived from, Terminalia, with which it has been united by several authors. The two are easily dis- tinguished in the New World by the difference in the shape and internal structure of the fruit. Exell (1958), although maintaining both genera, has indicated that the distinctive distinguishing fruit characters are lack- ing in one or two Malayan-Polynesian species of Terminalia, which, how- ever, appear unrelated to Bucida in any other features. A re-evaluation of the taxonomic relationship of the two genera is desirable, but should be preceded by a comprehensive study of Terminalia, a large and mor- phologically diverse genus which has never been monographed. In Bucida Buceras the fruits are occasionally invaded by a gall-forming mite (Eriophyes sp.). The fruits become much elongated, up to 16 cm., and curved or twisted. The name Buceras, horn of an ox, is derived from the resemblance of the galls to such horns. Spines derived from lateral buds are present on some plants of this species. Rudimentary stipules, similar to those in Terminalia Catappa, have also been reported. REFERENCE Under family references see BRANDIS, EXELL (1931, 1958), and SARGENT. Britron, N. L. The generic name Bucida. Bull. Torrey Bot. Club 35: 303, 304. 1 |Comments incorrectly on the meaning of the generic name. | CooK, M. T. The hypertrophied fruit of Bucida Buceras. Bull. Torrey Bot. Club 35: 305, 306. 1908. [Hypertrophy due to a gall-forming mite. | 3. Conocarpus Linnaeus, Sp. Pl. 1: 176. 1753; Gen. Pl. ed. 5. 81. 1754. Shrubs or trees of sandy coasts and hammocks marginal to mangrove swamps |or of borders of tropical rivers], with erect to prostrate, glabrous to censely silver-haired stems. Leaves evergreen, alternate, shortly petio- 1964] GRAHAM, RHIZOPHORACEAE AND COMBRETACEAE 299 late or sessile, the blade elliptic, ovate [or narrowly lanceolate], the apex acute to acuminate or apiculate, the base with 2 raised glands, the lower (abaxial) side with several small pits (domatia), each at the juncture of the midvein with a secondary vein and occasionally also along the outer margin. Inflorescence a globose, many-flowered, conelike head, sessile or pedunculate. Flowers bisexual or both bisexual and ¢ in the same in- florescence, 5-merous, subtended by an ovate, concave bract, the apex of the bract acuminate. Floral tube greenish white, the lower part laterally flattened and 2-winged, the upper part cupuliform, deciduous; calyx lobes 5, valvate in bud, velutinous within. Petals wanting. Stamens 5—8(—10), well exserted, deciduous; anthers orbicular. Stigma punctiform. Style surrounded at the base by a fleshy, velutinous disc. Fruits leathery, laterally flattened, 2-winged, scalelike, imbricate in a conelike cluster; pericarp of thin, leathery outer and inner layers with spongy aérenchyma between. Seed coat thin. Lectotype species: C. erectus L.; typified by the union of the two original species under C. erectus by De Candolle, Prodr. 3: 16. 1828. (Name from Greek, konos, cone, and karpos, fruit, in reference to the conelike head of fruits.) A genus of two species, Conocarpus erectus, the buttonwood, of tropical America and West Africa, and C. lancifolius Engl., known only from the river valleys of the northern Somaliland highlands of East Africa. Cono- carpus erectus reaches northward into Florida as far as Hernando County, on the west coast, and Merritt Island (Brevard County), on the east, occur- ring inland from the mangrove swamps on ground generally free from inundation by high tides. The alternate, elliptical leaves and fruits clus- tered into globose, conelike heads are characteristic. It fruits abundantly, but only an estimated 3-10% of the fruits contain seeds (Guppy). The cause of low seed set has not been determined. The species shows a wide variation in stature and pubescence. Plants range in form from prostrate shrubs in rocky habitats to erect trees on sandy hammocks. The low-growing plants have been recognized by some authors as var. procumbens DC. A second, var. sericeus DC., has been recognized for plants with silver-haired pubescence which occur in the northern part of the species range. The common glabrous form and the densely pubescent extreme occur mixed in the same population. Both are without distinct geographical range and probably should be regarded only as minor variants. The wood of C. erectus is very hard, burning slowly like charcoal, and is prized as a fuel in the West Indies. REFERENCES: Under family references see BRANDIS, EXELL (1931, 1954), Guppy (pp. 201- 204), SARGENT, STEARN, and THORNE. CHIOVENDA, E. Flora Somala. xvi + 436 pp. 50 pls. Rome. 1929. [C. lanci- folius, 173.) HARSHBERGER, J. W. The comparative leaf structure of the sand dune plants of Bermuda. Proc. Am. Philos. Soc. 47: 97-110. pls. 1-3. 1908. | Includes C. erectus. | 300 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV LaessLe, A. M., & C. H. WuHarton. Northern extensions in recorded ranges of plants on Seahorse and associated keys, Levy County, Florida. Quart. Jour. Fla. Acad. Sci. 22: 105-113. 1959. [C. erectus, 110, 111; seed dispersal mainly by sea currents. | LittLte, E. L. Miscellaneous notes on nomenclature of United States trees. Am. Midl. Nat. 33: 495-513. 1945. [C. erectus, 507, 508 RussELL, A. Conocarpus erecta (buttonwood, Zaragoza-mangrove), a new domestic source of tannin. Chemurg. Dig. 2: 27-29. 1943.* Tribe LAGUNCULARIEAE Engl. & Diels 4. Laguncularia Gaertner f. Fruct. Sem. Pl. 3: 209. pl. 217. 1805.° Shrubs or trees of coastal mangrove swamps and hammocks, often with small, pencil-like pneumatophores. Leaves evergreen, opposite, leathery, glabrous, oblong to obovate with the apex generally obtuse or retuse, shortly petiolate with 2 glands on the petiole at the base of the blade, the blade with numerous submarginal pits (domatia?) at the ends of the secondary veins. Inflorescence of terminal panicles, or solitary spikes in the axils of leaves. Flowers greenish white, fragrant, bisexual, 5-merous, subtended by a deciduous bract. Floral tube cupuliform to tubular, fleshy, scarcely produced beyond the ovary, inconspicuously ribbed, accrescent, silvery pubescent to glabrous, bearing 2 ovate bracteoles at the apex beneath the calyx lobes; calyx lobes 5, broadly deltoid, imbricate in bud, persistent. Petals 5. Stamens 10; filaments scarcely exserted; anthers orbicular. Stigma capitate. Style surrounded at the base by a thick, flattened disc. Ovules suspended on very short funiculi. Fruit ellipsoid to mostly obovoid, laterally flattened, with 2 thick, spongy wings, densely silvery haired to glabrous, bearing the persistent calyx at the apex; peri- carp of leathery outer and inner layers with spongy aérenchyma between. Seed coat thick. Type species: L. racemosa (L.) Gaertn. f. (Name from Latin, laguncularis, flask- or bottle-shaped, in reference to the shape of the fruit.) — WHITE MANGROVE. A monotypic genus of the American and West African coastal mangrove swamps. Laguncularia racemosa, with opposite leaves, flowers with five small petals, and conspicuous sessile, gray, flask-shaped fruits has a range practically identical with that of the red mangrove, Rhizophora Manele. Where mangrove swamps occur in the Gulf of Mexico region, Laguncularia is found growing either mixed with R. Mangle and Avicennia germinans * Laguncularia has been equated, by some authors, with the earlier genus Horau Adans. (Fam. Pl. 2: 80. 1763), but, according to Adanson Adanson from Kaempfer, who applied it to an East Indian coastal plant with ainste leaves. On this evidence it appears that Laguncularia and Horau are not synonymou and that Laguncularia is in no danger of displacement. 1964] GRAHAM, RHIZOPHORACEAE AND COMBRETACEAE 301 (L.) L. (the black mangrove of the Verbenaceae) or in a distinct zone inland between these genera and Conocarpus erectus. It seems to grow best on land at least partly flooded by salt or brackish water at high tide. In Florida, Laguncularia occurs as far north as Hernando County, on the west, and Merritt Island (Brevard County), on the east coast. As with other genera of the family, few observations have been made on the biology of Laguncularia. It is known to be semiviviparous, the green embryo generally only piercing the seed coat while in the fruit on the tree, and germination generally not proceeding further until the fruit has fallen. No studies have been made on the origin of the floral tube in Combretaceae, but in Laguncularia (and the Old World Lumnitzera and Macropteranthes) bracteoles present on the tube suggest the possibility that it is partly cauline in origin. The wood of the white mangrove is used extensively in Puerto Rico for fence posts. Attempts are being made there to control cutting and production of Laguncularia, so it may become a profitable economic crop. REFERENCES: Under family references see BRANDIS, EXELL (1931, 1958), Guppy (pp. 221- 223), SARGENT, STEARN, and THORNE. ScHENCK, H. Ueber die Luftwurzeln von Avicennia tomentosa und Laguncu- laria racemosa. Flora 72: 83-88. 1 pl. 1889. [Anatomy and brief discussion of function of pneumatophores. | WapswortH, F. H. Growth and regeneration of white mangrove in Puerto Rico. Carib. Forester 20: 59-71. 1959. 302 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV THE GENUS CHAENOMELES (ROSACEAE) * CLAUDE WEBER TAXONOMY Chaenomeles Lindley, Trans. Linn. Soc. 13: 96. 1822, as “Chaenomeles.” Type species: Chaenomeles japonica (Thunb.) Lindl. ex Spach. Pseudochaenomeles Carriére, Revue Hort. 1882: 238. f. 52-55. 1882. Deciduous or half-evergreen shrubs with spiny branches and smooth bark. Buds small, glabrescent, with conduplicate vernation. Leaves alter- nate on voung shoots, alternate or fasciculate on old wood. subsessile, simple, ovate, serrate, or crenate, with glands (or foliar nectaries) terminat- ing and between the teeth, reticulate, glabrous or pubescent beneath when young, becoming glabrescent. Stipules reniform, free, occurring on young shoots only; leaves on old wood exstipulate. Flowers axillary, very numer- ous, solitary, or more often fasciculate, appearing on old wood and before the leaves in the spring, on leafy shoots in the summer and fall. Bracts caducous, subulate in spring inflorescences, subulate or reniform in sum- mer inflorescences. Flowers regular, hermaphrodite, or male by abortion of styles and/or ovules (more rarely female by abortion of stamens). Calyx tubular-campanulate, 5-lobed. Sepals 5, imbricate in aestivation, short, entire, obtuse, often irregular in length, glabrous on the outer sur- face, ciliate but not glandular, pubescent within, erect at blooming time, usually caducous. Petals 5, large, free and unguiculate, inserted in the throat of the hypanthium, pubescent at base, red or pink, less often white, with veins no darker than the rest of the petal, caducous before wilting. Stamens 40-60, arranged more or less in two rows, filaments free. red or white (depending on the season), anthers inserted below the petals, introrse, innate to slightly versatile when mature, bilocular, yellow. Pollen grains tricolpate, smooth. Disc small, nectariferous, covering the carpels which are fused on their abaxial side to the hypanthium. Carpels 5, completely fused in a 5-celled ovary containing numerous anatropous ovules arranged horizontally in 2 rows in each locule. Placentation axile. Styles 5, connate at the base or up to ™% their length, glabrous or pu- bescent at base, 14 longer than the stamens. The column formed b the fused styles sometimes persistent and becoming the umbilicus of the umbilicate fruits. Stigmata discoid or more often decurrent, extending on the styles following the junction of the edges. Fruit a short-stalked pome formed of the hypanthium, the flesh with grit cells, the carpels becoming cartilaginous. Seeds ovoid, numerous in each of the 5 locules, * Continued from volume XLV, p. 205. 1964 | WEBER, CHAENOMELES 303 pointed at one end, testa coriaceous and dark brown, raphe and chalaza apical, endosperm lacking. Germination epigeal. VERNACULAR NAMES. English: Japanese Quince, Flowering aie Japonica. French: Cognassier du Japon, Poirier du Japon. German Japanishe Quitte. Japanese: Boke. Uses: Ornamental shrubs cultivated for their abundant and early flowers which appear before the leaves. Numerous spines render the plants useful for hedges. The easily forced flowers are sold as cut flowers during the winter. The fruits are used in jelly and preserves. Leroi (1924) recommends double cooking to eliminate the strong and acid taste. Used in France in a liqueur called cydonade (Burvenich, 1900), and in the preparation of ratafia (Morren, 1851). In China the fruits (mostly in a dried state) are used medicinally in treating cramps, cholera, beriberi, stomach and intestinal pains, rheuma- tism, sunstroke, and as a sedative and excitant. The active principles are malic, tartaric and citric acids, and vitamin C (Loh, 1954). DISTRIBUTION. China, Tibet, Bhutan, Burma, Japan. Cultivated in Asia, New Zealand, Australia, Europe, Africa and North and South America Kaempfer (1712), the first European to have seen a Japanese Quince, described it under the name of “Buke,” a misspelling for the Japanese name ‘“Boke.”’ Thunberg (1784) described Pyrus japonica, based on a specimen he had collected which is still preserved at the Vaxtbiologiska Institutionen at Uppsala (upsv). This species became the basis for the genus de- scribed by Lindley (1822) as Chaenomeles (from the Greek chaino, to gape, and melon, apple). Although Thunberg’s type specimen consisted of flowering material only, he described fruit of his new species in a manner which has since been shown to be incorrect. He wrote ‘“‘Pomum subglobosum, glabrum, magnitudine iuglandis, 5-valve, 5-loculare.”’ Except for the expression ‘‘5-valve’ the description is accurate. There is no explanation why Thunberg thought the fruit was dehiscent; although the pome shrivels and dries, it never opens In describing the genus Chaenomeles Lindley wrote, aS fruit is only known from Thunberg’s map who says it splits into five valves.” His detailed description was, ‘““Choenomeles. Pyri species Soe Willd. Cal. campanulatus, 5 dentatus, carnosus. Stamina erecta, serie duplici inserta. Pomum quinquevalve, polyspermum. Frutex (Japoniae). Folia lucida, coriacea, crenata. Flores terminales, coccinei.” In 1825 DeCandolle treated Chaenomeles as a section of Cydonia and pointed out the error of Thunberg when he stated, ‘““Thunberg asserit po- mum 5-valve, sed vereor ne descr. sit erronea.”’ In spite of some confusion concerning the characters of its fruit, the genus Chaenomeles is distinct and homogeneous, and may be recognized 304 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV easily by a series of characters among which are the following: reniform stipules; short erect, entire, glabrous sepals; stamens in two rows; and numerous seeds. In 1875, Carriére wrote that on the basis of the figures given in the Gardeners’ Chronicle, Pvrus maulei, described by Masters in 1874, should be placed in Chaenomeles rather than in Pyrus; but he did not make the transfer. In 1876 he published an article with numerous figures illus- trating the variation found in the fruits of the other species of Chaeno- meles then commonly cultivated, C. speciosa. When later he saw fruits of the plant at that time called Pyrus maulei, he found them sufficiently different to justify the creation of a subgenus Pseudo-Chaenomeles. Ac- cordingly, although he considered the taxon a subgenus, he treated the combination simply as Pseudo-Chaenomeles maulei. Neither Carriére nor any later botanist used this combination again. It seems to me that the species of Chaenomeles are so very closely related that there is no need of subgeneric categories. It should be noted in considering specific characters that herbarium specimens of Chaenomeles are rarely complete. The flowers usually ap- pear before the leaves and are of no use in identifying the species, being on the one hand very much alike in the whole genus, and on the other hand offering a range of variation on the same shrub, according to the distinctive specific characters, found in the serration and the type of pubescence of the leaves, are of great use in determination of hybrids. The confusion between the three species of Chaenomeles and the two of Cydonia could have been completely avoided by using the distinguishing characters of the adult leaves. In this monograph the genus Chaenomeles is treated as including three species, C. japonica (Thunb.) Lindl. ex Spach, C. speciosa (Sweet) Nakai, and C. cathayensis (Hemsley) Schneider. As stated in my earlier paper (1963), hybrids have been developed between these species in every mathematical combination. The Superba group consists of hybrids be- tween C. japonica and C. speciosa; the first cultivars appeared after 1890 as chance seedlings in different nurseries. In 1921 a cross was made by Philippe de Vilmorin of Verriéres-le-Buisson, France, by pollinating flowers of C. speciosa with pollen of C. cathayensis, producing ‘Vedrariensis,’ the first cultivar of the Vilmoriniana group. Walter Clarke of San Jose, Cali- fornia, crossed C. cathayensis with C. japonica producing a group of seedlings from which he selected the first two cultivars of the Clarkiana group in 1945. Earlier, in 1938, he had crossed successfully C. cathayensis with C. & superba to produce cultivars which are a synthesis of the three species, and which form the Californica group. KEY TO THE SPECIES AND Hysrip GrRouPS OF CHAENOMELES A. Leaves crenate; shrubs not more than 1.5 m. high; branches spreading or 1964] WEBER, CHAENOMELES 305 erect- ee with slender spines; fruits small, to 4 cm., ripening early in the s B. ie regularly crenate, always glabrous; second year twigs ® heavily verruculose; flowers salmon-orange (rarely white); fruits apple- shaped. Seg ee geo 8 Satna char’. N 3 =a wm ° 5 5 5 a iF ce) p =a = rt) bs mn ana} ad o o >a Cc = fo) > o = o Loita! @ ! Las | iz. n yl o i=) = te are Lan | i= =p or 1964] DUDLEY, THE GENUS ALYSSUM 361 1. Section Meniocus (Desv.) Hook. in Benth. & Hook. Gen. Pl. 1: 74 1862. Type species: Alyssum linifolium Steph. ex Willd. — Boiss. Fl. Or. 1: 265, 286. 1867.— Schulz in Engler & Prantl, Nat. Pflanzen- fam. II. 17b: 492. 1936. — Nyar. Anal. Acad. Rep. Pop. Rom. Sect. Stiinte Geol. Geog. Biol. ser. A. mem. 3. 1: 7. 1949. Meniocus Desv. Jour. Bot. 3: 173. 1814 (!). DistrRIBUTION: European and Oriental, with maximum representation in Turkey. Plants annual. Long filaments unilaterally winged and toothed. Short filaments with free, or basally connate appendages. Nectaries globose. Silicles dehiscent, glabrous, or with indumentum or simple setae (rarely forked), frequently papillose, locules 2—4(—8)-ovulate with distinctly lateral placentation; valves always compressed. Seeds winged or wingless, mucilaginous. COMPONENT SPECIES: A. AUREUM (Fenzl) Boiss. Fl. Or. 1: 286. 1867 (!). Southern and eastern Turkey, Syria, and Israel. A. BLEPHAROCARPUM Dudl. & Hub.-Mor. Jour. Arnold Arb. 45: 61. pl. J, figs. 1-13. A. HETEROTRICHUM © Boiss. Diagn. 1(6): 15. 1845 (!). Iran and Kazakhstan. A. HUETII Boiss. Fl. Or. 1: 287. 1867 (!). Endemic to Turkey. A. LINIFOLIUM Steph. ex Willd. Linn. Sp. Pl. ed. 4. 3(1): 467. 1800 (!). Common in Europe, North Africa, the Levant, Caucasia, Siberia, and east to India. Naturalized in Australia. A. MENIOCOIDES Boiss. Ann. Sci. Nat. Paris II. 17: 158. 1842 (!). Widespread in the Levant, east to Afghanistan. A. STYLARE (Boiss. & Bal.) Boiss. Fl. Or. 1: 287. 1867 (!). Endemic to Turkey. 2. Section Psilonema (Meyer) Hook. in Benth. & Hook. Gen. Pl. 1: 74. 1862. Typr species: Alyssum dasycarpum Steph. ex Willd. — Boiss. Fl. Or. 1: 265, 285. 1867. — Schulz in Engler & Prantl, Nat. Pflan- zenfam., II. 17b: 491. 1936.— Nyar. Anal. Acad. Rep. Pop. Rom. Sect. Stiinte Geol. Geog. Biol. ser. A. mem. 3.1: 7. 1949. Psilonema Meyer in Ledeb. FI. Alt. 3: 50. 1831 (!). Alyssum subgen. Tetratrichia Gay in Cosson, Fl. Atl. 2: 236. 1885 (!). TypE SPECIES: Alyssum granatense Boiss. & Reut. Alyssum sect. Eualyssum subsect. Hygrochastica Zohary, Pal. Jour. Bot. Jeru- salem Ser. 2(2/3): 161. 1941 (!). Type species: Alyssum damascenum Boiss. & Gaill. DISTRIBUTION: Europe, North Africa, Egypt, Arabia, the Levant, and east to India. Plants annual. Filaments always slender, wingless, edentate, and un- 362 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV appendaged. Nectaries erect and peglike, or globose. Silicles dehiscent, glabrous, or with monomorphic or dimorphic indumentum, held in the same direction as pedicels, and with biovulate locules; valves equally inflated. Seeds winged or wingless, mucilaginous. COMPONENT SPECIES: A. Atyssomes (L.) L. Syst. ed. 10. 2: 1130. 1759 (!). Widespread in Europe, North Africa, Russia, and east to India. Doubtfully recorded from the Levant. roan in the United States and Canada, and in one locality in Argent A. DAMASCENUM Boiss. & Gail. Diagn. 3(6): 1859 (!). Syria, Jordan, and Israel. A. DASYCARPUM Steph. ex Willd. Linn. Sp. Pl. ed. 4. 3(1): 467. 1800 (!). Widespread in eastern Europe, the Levant, Caucasia, and east to India. A, GRANATENSE aoe. & Reut. Pugillus, 9. 1852 (!). Spain, Portugal, and North Afri A. HOMALOCARPUM (Fisch. & Meyer) Boiss. Fl. Or. 1: 285. 1867 (!). Egypt, Arabia, Iraq, Iran, and Pakistan. Rarely occurring in Syria, Jordan or Israel. Ww Section Alyssum. LEcTOTYPE sPEcIES: Alyssum montanum L.— Boiss. Fl. Or. 1: 264, 274. 1867.— Baumgartner, Beil. Jahresb. Nied.-Ost. Land.-Lehrers. Wien.-Neust. 34: xiv. 1907, ibid. 36: 1909; Beil. Jahresb. Kaiser Franz Josef-Land.-Gymn. Oberrealsch. Baden bei Wien 48: 1911.— Schulz in Engler & Prantl, Nat. Pflanzenfam. II. 17b: 491. 1936.—Nyar. Anal. Acad. Rep. Pop. Rom. Sect. Stiinte Geol. Geog. Biol. ser. A. mem. 3. 1: 7. 1949. Alyssum sect. Eualyssum Griseb. Spic. Fl. Rum. & Bith. 1: 274. 1843. Lecto- TYPE SPECIES: Alyssum montanum L. (!). DISTRIBUTION: Europe, North Africa, the Levant, Caucasia, and east to Siberia, Mongolia, and India Plants annual, biennial, or if perennial then with sterile shoots and winter rosettes. Sepals monomorphic, free, but often persistent in fruit; inner surfaces always glabrous. Petals predominantly yellow, occasionally pale, and rarely whitish. Filaments free, unilaterally or bilaterally winged, and/or dentate, and/or appendaged. Nectaries globose or triangular, occasionally lobed, rarely peglike. Silicles dehiscent, glabrous, or with monomorphic or dimorphic indumentum, held in the same direction as pedicels, with biovulate locules; valves equally or unequally inflated. Seeds winged or wingless, mucilaginous. COMPONENT SPECIES: A. a1zowes Boiss. Ann. Sci. Nat. Paris, II. 17: 153. 1842 (!). Endemic to Turke y. A. ARENARIUM Loisel. Fl. Gall. 2: 401. 1807 (!). Seacoasts of southern France and northern Spain. 1964 | DUDLEY, THE GENUS ALYSSUM 363 A. ARGYROPHYLLUM Schott & Ky. Oesterr. Bot. Wochenbl. 7(29): 229. 1857 (!). Endemic to Turkey A. ARMENUM Boiss. FI. Or 1: 278. 1867 (!). Eastern Turkey and the u sus. A. ARTWINENSE Busch in Kuznetsov, Busch & Fomin, Fl. Cauc. Crit. 3(4): 566. 1909 (!). Endemic to Lazistan (northeastern Turkey and the adjacent Caucasus). A. ATLANTICUM Desf. Fl. Atl. 2: 71. t. 149. 1798. Spain and North Africa. A. AURANTIACUM Boiss. Fl. Or. 1: 276. 1867 (!). Endemic to Turkey. A. BORNMUELLERI Hausskn. ex Degen, Osterr. Bot. Zeitschr, 48: 108. 1898 (!). Endemic to Turkey. A. BULBOTRICHUM Hausskn. & Bornm. Mitt. Thir. Bot. Ver. 20: 3. 1909- 1905 (!). Endemic to Turkey A. CAESPITOSUM Baumg. Beil. Jahresb. Nied. aes Land.-Lehrers. Wien.- Neust. 36: 26. 1909 (!). Endemic to Turkey A. CALYCOCARPUM Rupr. Mém. Acad. St. -Pétersb. Vil. 2527 )2. 103, 291. 1869 (!). The Crimea, Caucasia, northern Iran, and Turkish Kurdistan. A. CANESCENS DC. Syst. Nat. 2: 322. 1821 (!). Siberia, the Himalayas, and Mongolia. A. CEPHALOTES Boiss. Diagn. 3(1): 34. 1853 (!). Endemic to Turkey. A. conteMPTUM Schott & Ky. Oesterr. Bot. Wochenbl. 4(22): 177. 1854 (!). Widespread in southern and eastern Turkey, and in the Levant to northern Iran. A. CUNEIFOLIUM Ten. Cat. Pl. Hort. Neap. Append. 58. 1815 (!). Southern Europe A. DENSISTELLATUM Dudl. Notes Bot. Gard. Edinb. 24: 160. fig. 3, Aa—Ad. 1962 (!). Endemic to Greece. A. DESERTORUM Stapf, Denkschr. Akad. Wien. Math. Naturw. 51: 302. 1886 (!). Widespread in Europe and Asia. Naturalized in Idaho in the United States. A. pirFusUM Ten. Cat. Pl. Hort. Neap. Append. 58. 1815 (!). The Pyrenees, northern Spain, southwestern Alps, and Greece. A. DOERFLERI Degen, Denkschr. Akad. Wien. Math. Naturw. 64: 708. ¢. 2, figs. 5d-f. 1897 (!). Northern Greece and southern Yugoslavia. A. ERosuLUM Gennar & Pestal. Mem. Accad. Sci. Torino, II. 16: 248. 1855 (!). Endemic to Turkey. A. FASTIGIATUM Heywood, Bull. Brit. Mus. Bot. 1: 92. 1954 (1). Endemic to Spain A. FISCHERIANUM DC. Syst. Nat. 2: 311. 1821 (!). Southeast Russia, and Si A. rotrosum Bory & Chaub. Exped. Sci. Morée, 185. ¢. 23, fig. 1. 1832. Greece, Crete, Cyprus, Aegean Islands, and the Mediterranean coast of Turkey. A. FULVESCENS Sibth. & Smith, Prod. Fl. Gr. 2: 13. 1813 (!). Aegean Islands, Cyprus, and the Mediterranean coast of Turkey. Recorded doubtfully A. HANDELIT Hayek, Beih. Bot. Centr. 45(2): 1928 (!). Endemic to Greece. A. uirsutuM Bieb. Fl. Taur. Cauc. 2: 106. 1808 (!). Widespread in south- eastern Europe, western Russia, and northern and eastern Turkey. A. waeum Boiss. & Heldr. Diagn. 2(8): 35. 1849 (!). Endemic to Crete. A. rrANICUM Hausskn. ex Baumg. Jahresb. Kaiser Franz Josef-Land.-Gymn. 364 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Oberrealsch. Baden bei Wien 48: 9. 1911 (!), non Czerniak. (1924). Widespread in the Levant exclusive of Turke A. LANCEOLATUM Baumg. op. cit. 11 (!). Iran and Afghanistan. A. LassiticumM Hal. Suppl. Consp. Fl. Gr. 2: 10. 1908 (!). Endemic to rete. A. LENENSE Adams, Mém. Soc. Nat. Moscou 5: 110. 1817. Southeastern Russia and Mongolia A. LEPIDOTUM Boiss. Diagn. 1(1): 73. 1842 (!). Endemic to Turkey. A. MACROCALYx Cosson & Durand, Bull. Soc. Bot. Fr. 4: 12. 1857 (!). North Afri A. MAcRopopUM Boiss. & Bal. Diagn. 3(6): 18. 1859 (!). Endemic to Turkey. A. MARGINATUM Steud. ex Boiss. Ann. Sci. Nat. Paris, II. 17: 157. 1842 (!). Widespread in the Levant, exclusive of Turkey, and extending to south- eastern Russia and India. A. minus (L.) Rothm. oe Sp. Nov. 50: 77. 1941. Common in Europe, North pres and Asia A. MINUTUM Schlecht. ex DC. Syst. Nat. 2: 316. 1821 (!). Widespread in southeastern and eastern ie 7 Africa, Crete, Cyprus, the Aegean Islands, Turkey, and Syr A. ge seeat ren ex Beck, Ann. Naturh. Mus. Wien 2: 173. ¢. 2, figs. 5-7. 1887 (!). Yugoslavia. A. MONTANUM . Sp. Pl. 2: 650. 1753 (!). Widespread in Europe, mostly south of 50 degrees latitude. A. mourapicum Boiss. & Bal. Diagn. 3(6): 16. 1859 (!). Turkey and ebanon. A. MUELLERI Boiss. & Buhse, Nouv. Mém. Soc. Nat. Moscou 12: 6. 1860 (!). Russian Armenia, Caucasus, and northern Iran. A. MICROPHYLLUM (Meyer) Steud. Nomencl. Bot. ed. 2. 1: 68. 1841 (!). Siberia A. NEVADENSE Wilmott ex Ball & Dudley, sp. nov.2, Endemic to Spain. A. OCHROLEUCUM Boiss. & Huet, Diagn. 3(5): 36. 1856 (!). Endemic to urkey. A. OVIRENSE Kerner, Sched. Fl. Exsicc. — Hung. 2: 99. 1882 (!). Southeastern Alps and western Yugoslav A. PERSICUM Boiss. Ann. Sci. Nat. Paris, IL. “y. 152. 1842 (!). Iran. A. pRAECOX Boiss. & Bal. in Boiss. Fl. Or. 1: 275. 1867 (!). Endemic to Turkey. A. PROPINQUUM Baumg. Beil. Jahresb. Nied.-Ost. Land.-Lehrers. Wien.-Neust. 36: 22. 1909 (!). Endemic to Turkey. A. PSEUDO-MOURADICUM Hausskn. & Bornm. ex Baumg. op. cit. 6 (!). Endemic to Turkey. A. PULVINARE Velen. a -ber. Bohm. Ges. Wiss. Math. Naturw. 33: 30. 1889 (!). The Balkan * Alyssum nevadense Wilmott ex Ball & Dudley, sp. nov. Holotype, Spain, Sierra Nevada, Almeria, main ridge west of Cerro del Rayo, 25 June 1926, Wilmott (B Affinis A, montano sed habitu humilioris, caulibus misoribus decumbentibus, foliis sur- culorum sterilium lanceolatis et angustioribus, indumento ex toto valde dense albo- argenteo e pilis stellatis maioribus (0.5—-)0.7-1 mm. diametro 5-10 radiatis composito, floribus pallidis minoribus, racemi conferti differt.— Endemic to the Spanish Sierra evada. — Sierra Nevada above Minas de Bieres, between Minas de Bieres and Cerro del Rayo, 4 June 1924, Lofthouse (BM). 1964] DUDLEY, THE GENUS ALYSSUM 365 A. PURPUREUM Lag. & Rod. Anal. Ci. Nat. 5: 275. 1802. Endemic to Spain. A. REPENS Baumg. Enum. Stirp. Transsilv. 2: 237. 1816 (!). Widespread in southern and eastern Europe, the Levant, Caucasia, and rarely extending to Tran. A. ROSTRATUM Stev. Mém. Acad. St.-Pétersb. 3: 295. ¢. 15, fig. 1. 1809 (!). Southeastern Europe and the Caucasus A. SCARDICUM Wettst. Bibliot. Bot. 26: 24. 1892 (!). The Balkans A. SCUTIGERUM Durand, Bull. Soc. Bot. Fr. 4: 11. 1857. North Africa. A. SMYRNAEUM Meyer, Bull. nee Sci. St.-Pétersb. 7: 132. 1840 (!). Greece, the Aegean Islands, and Turk A. SPHACIOTICUM Boiss. & Heldr. Diagn. 2(8): 35. 1849 (!). Endemic to rete. A. staprit Vierh. Verh. Zool.-Bot. Ges. Wien 64: 261, t. 7, fig. 1. 1914 (!). Widespread in southern and eastern Turkey, the Levant, and east to Afghanistan. A. STRIBRNYI oe Fl. Bulg. 640. 1891 (!). The Balkans, Turkey, Syria, and Leban A. STRICTUM Willd. Linn. Sp. Pl. ed. 4. 3(1): 464. 1800 (!). Widespread in the Levant and Caucasia. A. strRIGosUM Banks & Sol. in Russell, Natural Hist. Aleppo 2: 257. 1794 (!). Common in southern and eastern Europe, North Africa, the Levant, and east to Afghanistan. A. SZOWITSIANUM Fisch. & Meyer, Ind. Sem. i Petrop. 4: 31. 1837 (!). Common throughout the Levant, east to In A. TAYGETEUM saa Verh. Zool.-Bot. Ges. "Wien 48: 703. 1898 (!). A. TENUIFOLIUM aa ex Willd. Linn. Sp. Pl. ed. 4. 3(1): 460. 1800 (!). Siberia and Mongolia. A. TRICHOCARPUM Dudl. & Hub.-Mor. Jour. Arnold Arb. 45: 68. pl. IJ, figs. 12-22. 1964 (!). Endemic to Turkey. A. TURKESTANICUM Regel & Schmalh. in oes Desc. Pl. Nov. Rar. 6. 1882 !). Transcaspia, Iran, and Afghanist A. UMBELLATUM Desv. Jour. Bot. 3: 173, “184, 1814 (!). Widespread in the Balkans, Crete, Cyprus, the Aegean Islands, western Turkey, and the Crimea. A. WIERZBICKII Heuff. Ral 18: 242. 1835 (!). The Banatus region of Yugoslavia and Rom A. WULFENIANUM Schlecht, in Willd. Suppl. Enum. Pl. Hort. Berol. 44. 1813. Southeastern Alps. A. XANTHOCARPUM Boiss. Ann. Sci. Nat. Paris, II. 17: 154. 1842 (!). Widespread in Turkey, the Caucasus, northern Syria, and rarely in Lebanon. 4. Section Gamosepalum (Hausskn.) Dudl. Jour. Arnold Arb. 45: 70. 1964. Lectotype species: Alyssum lepidoto-stellatum (Hausskn. & Bornm. ex Hausskn.) Dudl Gamosepalum Hausskn. Mitt. Thir. Bot. Ver. 11: 73. 1897 (!).— Schulz, Notizbl. Bot. Gart. Berlin 10: 109. 1927; in Engler & Prantl, Pflanzenfam. II. 17b: 494. 1936.— Bornm. Repert. Sp. Nov. Beih. 89: 58, 59. 1936. DIsTRIBUTION: The Levant, with maximum representation in Turkey. 366 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Plants perennial, with sterile shoots. Sepals distinctly dimorphic, always persistent, and often inflated in fruit, appearing connate because of inter- locking indumentum; inner surfaces with indumentum. Petals whitish with purple venation at limb bases, or yellow. Long filaments bilaterally winged, edentate and unappendaged, connate, or if free then contiguous with over- lapping wings. Short filaments with basal, connate appendages or teeth. Silicles dehiscent, always with indumentum, held in the same direction as pedicels, and with biovulate locules; valves equally or unequally inflated. Seeds narrowly winged or wingless, mucilaginous. A. Series Connata Dudl. Jour. Arnold Arb. 45: 70. 1964 (!). TYPE species: Alyssum lepidoto-stellatum (Hausskn. & Bornm. ex Hausskn.) Dudl. DisTRIBUTION: Endemic to Turkey. Long filaments connate. Indumentum overall composed of stellate hairs with few, branched rays. Petals whitish or pale yellow, with purple vena- tion at limb bases. COMPONENT SPECIES: A. LEPIDOTO-STELLATUM (Hausskn. & Bornm. ex Hausskn.) Dudl. loc. cit. (!). Endemic to Turkey. A. pAPHLAGONICUM (Hausskn.) Dudl. op. cit. 71 (!). Endemic to Turkey. A. TETRASTEMON Boiss. Ann. Sci. Nat. Paris, II. 17: 153. 1842 (!). Endemic to Turkey. A. tHymoprs (Hub.-Mor. & Reese) Dudl. Jour. Arnold Arb. 45: 71. 1964 (!). Endemic to Turkey. B. Series Libera Dudl. op. cit. 72 (!). Typr species: Alyssum baum- gartnerianum Bornm. Long filaments free, but contiguous with overlapping wings. Indumen- tum overall composed of lepidote hairs with many and branched rays. Petals sulphureous, not purple veined. COMPONENT SPECIES: A. BAUMGARTNERIANUM Bornm. in Beil. Baumg. Jahresb. Kaiser Franz Josef- Land.-Gymn. Oberrealsch. Baden bei Wien 48: 16. 1911 (‘!). Southern Turkey, Syria, Lebanon, Israel, and Iran. A. cornincit Dudl. Jour. Arnold Arb. 45: 72. pl. I//, figs. 1-11. 1964 (1). Endemic to Turkey. A. HaRPUTICUM Dud. op. cit. 76. pl. IV, figs. 1-12 (!). Endemic to Turkey. A. Lycaonicum (Schulz) Dudl. op. cit. 79 (!). Endemic to Turkey. A. n1veuM Dudl. op. cit. 78. pl. II, figs. 12-20 (!). Endemic to Turkey. A. SULPHUREUM Dudl. & Hub.-Mor. op. cit. 74. pl. IV, figs. 13-26 (1). Southeastern Turkey, and northern Syria and Iraq. 1964 | DUDLEY, THE GENUS ALYSSUM 367 wn Section Tetradenia (Spach) Dudley, comb. nov. LECTOTYPE SPECIES: Alyssum spinosum L. Sp. Pl. 2: 650. 1753 (!). Type: “Alyssum ramis senilibus spiniformibus nudis” (BM — Hort. Cliff. Alyssum no. 1). Konig Adans. sect. Tetradenia Spach, Hist. Nat. Vég. Phan. 6: 492. Plants perennial, often spinose, with sterile shoots and winter rosettes. Fruiting racemes candelabriform, Pedicels arcuate and recurved. Petals white, often pink or reddish. Long filaments free, narrowly bilaterally winged, edentate, or with small basal teeth. Short filaments with small basal teeth. Silicles dehiscent, always glabrous, conspicuously boat-shaped, erect or horizontal, held opposite to the direction of the pedicels, and with biovulate locules; valves strongly and unequally inflated. Seeds wingless, not mucilaginous. COMPONENT SPECIES: A. COCHLEATUM Cosson & Durand, Bull. Soc. Bot. Fr. 4: 11. 1857 (!). North Africa. A. LAPEYROUSIANUM Jord. Obs. Pl. Nouv. 1: 3. pl. 1, figs. Al-11. 1846. The Pyrenees. A. sprinosuM L. Sp. Pl. 2: 650. 1753 (!). Mediterranean Europe, and North Africa. 6. Section Odontarrhena (Meyer) Koch, Synop. Fl. Germ. Helv. 59. 1836. Lectotype species: Alyssum tortuosum Waldst. & Kit. ex Willd. — Hook. in Benth. & Hook. Gen. Pl. 1: 74. 1862.— Boiss. FI. Or. 1: 263, 267. 1867. — Nyar. Bul. Grad. Bot. Cluj 7: 1, 65. 1927, ibid. 8: 152. 1928, ibid. 9: 1. 1929; Repert. Sp. Nov. 27: 392. 1930; Bul. Grad. Bot. Cluj 9: 69. 1931; Bul. Soc. Stiinte Cluj 6: 446. 1932; Notizbl. Bot. Gart. Berlin 9: 631. 1932; Bul. Grad. Bot. Cluj 18: 82. 1938; Anal. Acad. Rep. Pop. Rom. Sect. Stiinte Geol. Geog. Biol. ser. A, mem. 3, 1: 1. 1949. — Schulz in Engler & Prantl, Nat. Pflan- zenfam. II. 17b: 491. 1936. Odontarrhena Meyer in Ledeb. FI. Alt. 3: 58. 1831 (!). DIsTRIBUTION: Widespread in Europe and the Levant, east to Siberia and Afghanistan. Plants perennial, with sterile shoots and winter rosettes, rarely biennial. Pedicels rigid, divergent and spreading, or slender, sigmoid and deflexed. Petals yellow, usually obovate, rarely with basal appendages. Long fila- ments free, always unilaterally or bilaterally winged, and variously dentate. Short filaments always with free or connate appendages, or teeth. Silicles dehiscent or indehiscent, glabrous or with indumentum, or rarely papillose, held in the same direction as pedicels or pendulous, and with uniovulate locules; valves equally or unequally inflated, or compressed and undulate 368 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV (samaroid and papyraceous in subsect. Samarifera). Seeds winged or wingless, rarely mucilaginous. I. Subsection Inflata Dudley, Jour. Arnold Arb. 45: 79. 1964 (!). TyPE species: Alyssum tortuosum Waldst. & Kit. ex Willd. Silicles dehiscent; valves equally or unequally inflated, with incon- spicuous venation, never undulate. Pedicels divergent and spreading. DISTRIBUTION: Widespread in Europe and the Levant, east to Siberia and Afghanistan. COMPONENT SPECIES: A. ALPESTRE L. Mant. 92. 1767 (!). Central and western Alps. A. ANATOLICUM Hausskn. ex Nyar. Bul. Grad. Bot. Cluj 9: 40. ¢. 5 fig. 111; t. 6, fig. 46; t. 7, fig. 56; t. 9, figs. 62 & 63; pl. 29. 1929 (!). Endemic to A. BAICALICUM Nyar. /bid. 18: 94. t. V, fig. 11; fig. 3, no. 3, 1939, Siberia. A. BERTOLONII iia Jour. Bot. 3: 172, 185. 1814 (!). Italy and the north- western Balkan A, BORZAEANUM Nyar. Bul. Grad. Bot. Cluj 6: 90. 1926 (!). The Balkans and Turkey. A. BRACTEATUM Boiss. & Buhse, Nouv. Mém. Soc. Nat. Moscou 12: 18. 1860 (!). Iran and Caucasia A. CALIACRAE Nyar. Bul. Grad. Bot. Cluj 6: 92. 1926 (!). The Balkans and the Crimea. A. CALLICHROUM Boiss. & Bal. Diagn. 3(5): 34. 1856 (!). Endemic to Turkey. A. CHONDROGYNUM Burtt, Kew Bull. 1949. 101. pl. 4. 1949 (!). Endemic to % ypru A. CONDENSATUM Boiss. & — in Boiss. Fl. Or. 1: 268. 1867 (!). Tur- key, Syria, Lebanon, and Iraq. A. CONSTELLATUM Boiss. rai Sci. Nat. Paris, IV. 2: 244. 1854 (!). Turkey and Iraq. A. corsicum Duby in DC. Bot. Gall. ed. 2. 1: 34. 1828 (!). Corsica and Turkey. A. CORYMBOSOIDES Form. Verh. Nat. Ver. Briinn 34: 329. 1895. The Balkans. A. cypricuM Nyar. Bul. Grad. “a — 7: 156. pl. 17; t. 5, fig. 81; t. 8, figs. 136-138. 1927. Cyprus and T A. DAVISIANUM Dudl. Jour. renee oe 45: 81. pl. I, figs. 14-22. 1964 (!). Endemic to Turkey A. piscoLtor Dudl. : Hub.-Mor. op. cit. 80. pl. V, figs. 1-11 (!). Endemic to Turkey. A. ERIOPHYLLUM Boiss. & Hausskn. in Boiss. Fl. Or. 1: 273. 1867 (!). Endemic to Turkey. A. EUBOEUM ae Consp. Fl. Gr. 1: 93. 1900 (!). Endemic to the Greek island of Euboe A. FALLACINUM fee: Mitt. Thiir. Bot. Ver. 3-4: 114. 1893 (!). Greece and Crete. A. FEDTSCHENKOANUM Busch, Not. Syst. Leningrad 4: 145. figs. p. 146. 1923 (!). Siberia. 1964 | DUDLEY, THE GENUS ALYSSUM 369 A. FILIFORME Nyar. Bul. Grad. Bot. Cluj 9: 35. t. 5, fig. 79; t. 6, fig. 40; t. 7, f. 49; t. 10, f. 53. 1929 (!). Endemic to Turkey A. FRAGILLIMUM (Bald.) Rech. f. Neue Beitrige zur Flora von Kreta, 77. 1943 (!). Endemic to Crete. A. GEHAMENSE Federov, Not. Syst. Geog. Inst. Bot. Tibilisi, fasc. 10: 65. fig. 4. 1941 (!). Turkish and Russian Armenia A. HAUSSKNECHTII Boiss. Fl. Or. 1: 269. 1867 ( (1), Endemic to Turkey. A. HUBER-MORATHII Dudl. Jour. Arnold Arb. 45: 83. pl. V, figs. 12-23. 1964 (!). Endemic to Turkey. A. INFLATUM Nyar. Bul. Grad. Bot. Cluj 9: 43. ¢. 5, fig. 113; t. 6, fig. 41; t. 7, figs. 50 & 51; t. 9, figs. 54-58. 1929 (!). Iran and the Transcaspian. A. LANIGERUM DC. Syst. Nat. 2: 308. 1821 (!). Iraq and east to Afghanistan. A. LIBANOTICUM Nyar. Bul. Grad. Bot. Cluj 18: 83. fig. 1. 1939 (!). Lebanon. A. LONGISTYLUM (Somm. & Lév.) Grossh. in Grossh. & Schischk. Scheda Herb. Pl. Or. Exsicc. 18. 1924 (!). Turkish Armenia, Caucasia, and the Crimea. A. MARKGRAFII Schulz, Ber. nig Bot. Ges. 44: 422. ¢t. VIJ-VIII, fig. 3. 1926 (!). Albania and Yugoslavi A. MASMENAEUM Boiss. Diagn. 3(5): 36. 1856 (!). Endemic : ar A. NEBRODENSE Tineo, Pl. Rar. Sic. Pug. 1: 12. 1817 (!). Sici A. OBOVATUM (Meyer) Turcz. Bull. Soc. Nat. Moscou in 57. 1837 (!). Southeast Russia, and Siberia. A. OBTUSIFOLIUM Stev. ex DC. Syst. Nat. 2: 305. 1821 (!). The Balkans and Russia A. OXYCARPUM Boiss. & Bal. Diagn. 3(5): 35. 1856 (!). Endemic to Turkey. A. PATERI os Bul. Grad. Bot. Cluj 9: 33. pl. 26. 1929 (!). Widespread in Turkey A. PEN J WINENSIS Dud]. Notes Bot. Gard. Edinb. 24: 162. pl. 7 & fig. 4. 1962 (!). Northern Iraq. A. POLYCLADUM Rech. f. Pi tee 32°55; 1951» Fran, A. ROBERTIANUM Bernard ex Gren. & Godr. Fl. Fr. 1: 117. 1847 (!). Endemic to Corsica. A. SERPYLLIFOLIUM Desf. Fl. Atl. 2: 70. 1798. Southwestern Europe and North Afri A. SIBIRICUM Willd. Linn. Sp. Pl. ed. 4. 3(1): 465. 1800 (!). Widespread in the Balkans, Turkey, the Crimea, and Caucasia. A. SINGARENSE Boiss. & Hausskn. in Buser, Suppl. Fl. Or. 49. 1888 (!). Iraq. A. SMOLIKANUM Nyar. Bul. Grad. Bot. Cluj 9: 43. t. 4, figs. 70, 71; t. 6, fig. 54; t. 7, figs. 09-71; t. 9, fig. 71; pl. 30, figs. 2 & 3. 1929 (!). Greece and Albania. A. syrtacuM Nyar. /bid. 18: 84. fig. 2. 1939 (!). Endemic to ees A. SZARABIACUM Nyar. op. cit. 97. fig. 3,7; t. 5, fig. 12. Siber A. TAVOLARAE Briquet, Prod. Fl. Corse 2: SS. ag. 1, £; eh e Be 1913 (1). Mediterranean islands of Tavolara and Sardinia A. TORTUOSUM Waldst. & Kit. ex Willd. Linn. Sp. Pl. ed. 4. 3(1): 466. 1800 (!). Widespread from eastern Europe to Siberia A. TRoopi Boiss. in Buser, Suppl. Fl. Or. 49. 1888 (!). Endemic to Cyprus. A. peg ae on Great Basin Naturalist 24: 7. pl. J, figs. A-E, G-K. 1964 (1). 370 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV II. Subsection Compressa Dudley, Jour. Arnold Arb. 45: 86. 1964. ype species: Alyssum murale Waldst. & Kit. DistrIBUTION: Southern and eastern Europe, and the Levant. Silicles dehiscent; valves compressed or rarely subinflated, + undulate, and always conspicuously nerved. A. Series Integra Dudley, Jour. Arnold Arb. 45: 86. 1964 (!). TYPE species: Alyssum murale Waldst. & Kit. DistTRIBUTION: Italy, the Balkans, Cyprus, Turkey, Spain, and Jordan. Silicles entire, without marginal wings. Pedicels rigid and spreading. COMPONENT SPECIES: A. aKamasicum Burtt, Kew Bull. 1949. 100. pl. 3. 1949 (!). Endemic to C a yprus. A. ARGENTEUM All. Miscell. Taur. 5: 73. 1774. Italy. A. casstum Boiss. Diagn. 2(8): 34. 1849 (!). Turkey and northern Syria. A. JANCHENII Nyar. Preslia 5: 109. fig. 1. 1927 (!). Albania. A. MURALE Waldst. & Kit. Pl. Rar. Hung. 1: 5. ¢. 6. 1799 (!). Europe and the Levant. A. suBspINosum Dudl, Notes Bot. Gard. Edinb. 24: 160. pl. 6. 1962 (!). Jordan. A. rentum Hal. Consp. Fl. Gr. 1: 93. 1900 (!). Greece. B. Series Crenulata Dudley, series nov. TyPE spEcIES: Alyssum crenula- tum Boiss. Diagn. 2(8): 33. 1849 (!). Holotype, Turkey, D6: Prov. Hatay, in collibus ad radices meridionales jugi Cassii (Akra dag) in via inter Latakieh et Suadieh, June (May-July) 1846, Boissier (c). DistrIBUTION: Greece, Turkey, and north Syria. A serie /ntegra siliculis marginibus alatis crenulatis pedicello capillari flexuoso deflexo suffultis differt. COMPONENT SPECIES: A. crticicum Boiss. & Bal. Diagn. 3(5): 34. 1856 (!). Endemic to Turkey. A. CRENULATUM Boiss. Diagn. 2(8): 33. 1849 (!). Turkey and north Syria. A. GiosnanuM Nyar. Bul. Grad. Bot. Cluj 7: 127. ¢. 3, figs. 52 & 53; t. 5, fig. 78, t. 8, fig. 102; t. 10, figs. 129-132. 1927 (!). Endemic to Turkey. A. HELDREICHU Hausskn. Mitt. Thir. Bot. Ver. 3-4: 113. 1893 (!). Greece. A. pterocarpum Dudley, nomen novum. Alyssum elatum Boiss. & Heldr. Diagn. 2(8): 33. 1849 (!) —non Kreutzer, Anthochron. Pl. Eur. Med. 99, 215. 1840. Lectotype, Turkey. C3: Prov. Antalya, ad rupes portus Tcherali (Cirali) (in campis sterilibus au pied des Montagnes, prés du portus), 13 May 1845, Heldreich 620 (cG); isolectotypes (BM, E, GH, HM, K, OXF, W). Endemic to Turkey. 1964 | DUDLEY, THE GENUS ALYSSUM Kar III. Subsection Samarifera Dudl. Jour. Arnold Arb. 45: 88. 1964. TYPE SPECIES: Alyssum samariferum Boiss. & Hausskn. Triplopetalum Nyar. Magyar Bot. Lap. 24: 97, fig. p. 98; t. 1. 1-14. 1925 (!).— Schulz in Engler & Prantl, Nat. Pflanzenfamt. Il. 17b: 492. 1936. DIsTRIBUTION: Lesbos, Turkey, and north Syria. Silicles indehiscent, samaroid, papyraceous, undulate, pendulous, always compressed. Pedicels slender, flexuose, and deflexed. Seeds mucilaginous. COMPONENT SPECIES: A. caricum Dudl. & Hub.-Mor. Jour. Arnold Arb. 45: 89. pl. IJ, figs. 1-11. 1964 (!). Endemic to Turkey A. DUBERTRETII Gombault, Mém. Soc. Bot. Fr. 33: 4. 1952. Endemic to Turkey. A. FLORIBUNDUM Boiss. & Bal. Diagn. 3(5): 33. 1856 (!). Endemic to Turkey A. LESBIACUM (Candargy) Rech. f. Fl. Aegaea, 226. 1943. Endemic to the Greek island of Lesbos A. oe Boiss. Ann. Sci. Nat. Paris II. 17: 158. 1842 (!). Endemic o Turkey. A. PINIFOLIUM (Nyar.) Dudl. Jour. Arnold Arb. 45: 92. 1964 (!). Endemic to Turkey A. SAMARIFERUM Boiss. & Hausskn. in Boiss. Fl. Or. 1: 272. 1867 (!). Turkey and north Syria A. TRAPEZIFORME ‘Baran: ex Nyar. Anal. noe Rep. ns Rom. Sect. Ay Geol. Geog. Biol. ser. A, mem. 3, 1: 83. ¢. 1, fig. 1; t. 6, figs. 4-6. 1949 (!). Endemic to Turkey. A. viRGATUM Nyar. Bul. Grad. Bot. Cluj 7: 115. pl. 10, t. 4, figs. 60-62, t. 6, fig. 61; t. 7, fig. 92. 1927 (!). Endemic to Turkey. SPECIES NON SATIS NOTAE A. americanum Greene, Pittonia 2: 224. 1892 (!). The Yukon in North America. Porsild (Rhodora, 41: 239. 1939 and Natl. Mus. Canada Bull. 121: 202. 1951) sae this species as the only indigenous American repre- sentative of the genus. Following Porsild, Anderson, in his Flora of Alaska. 267. 1959, lists Aly ee However, although Hultén (Lunds Univ. Arksskr. II. Sec 41(1): 886. 1945) maintains the binomial, he claims that it agrees aire with Siberian plants which have been known by a number of authors as “A. alpestre” or “A. sibiricum.” In 1939, Busch applied a nomen novum, A. biovulatum (see below), to refer to the Siberian plants; but if the plants from the Yukon and Siberia are conspecific, A. americanum is an earlier name and must be used. Busch (1939) states that A. biovulatum is a new name for that which he described, discussed. ne illustrated in detail as A. sibiricum in his Flora Siberiae et Orientis Extremit 6: 550-552. 1931. In this work Busch cites as a synonym o “A. sibiricum,” Odontarrhena obovata Meyer ( = Alyssum obovatum (Meyer) Turcz.). If in fact, A. americanum, A. biovulatum, and A. obovatum are the same species, A. obovatum must be used as the correct Sie JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV epithet. I have examined the holotype and isotypes of A. americanum, which agree in most characters with material of A. obovatum. I feel, however, that the question of whether the name 4. americanum should be maintained for an indigenous outlier of an entirely Old World genus should be left open until more specimens (as A. obovatum, or A. biovula- tum, or “A. sibiricum”) from the Kamchatka and Yakutsk provinces in Siberia have been carefully compared with the plants from the Yukon A. antiatlanticum Emb. & Maire, Bull. Soc. Hist. Nat. Afr. Nord 23: 165. 1932. North Africa. This plant is put in sect. PstLoneMa because the petals are only slightly longer than the sepals and the glands are erect and peglike. The fruits, however, are described as being cochleate and glabrous, and the plant is a suffrutescent perennial. Perennials are, to date, unknown in sect. PsILONEMA, as are fruits similar to those of A. spinosum or A, A. biovulatum Busch in Komarov, Fl. U.R.S.S. 8: 346. 1939. Siberia. See notes above under A. americanum. A. brughieri Colla, Herb. Pedem. 1: 142. 1833. Italy. No judgement can be made about this name without examination of the type in hb. Biroli (to). In the description, however, Colla allies it to A. argenteum, of which it may well be a synonym. A. costei Sennen & Pau, Bull. Acad. Geog. Bot. 18: 453. 1908. Spain. Accord- ing to the description, this plant is related to A. lapeyrousianum (sect. TETRADENIA), of which it may be a synonym. However, until the type material is studied, it is not possible to place this name in the correct section A, djurdjurae Chod. Bull. Soc. Bot. Fr. 36: 19. 1889. North Africa. This serpyllifolium (sect. ODONTARRHENA). The characters, noted in the descrip- tion appear to agree with A. atlanticum, but before examining the type material no judgement should be made. A, emarginatum Zahl ex Vis. Fl. Dalm. 3: 117. 1852. Yugoslavia. It is claimed that this species is intermediate between the perennial A. montanum and the annual “A. campestre” ( = A. minus). The fruits are described, however, as being glabrous. Type material must be examined before a decision is made concerning the application of this name. A, embergeri Quezél, Bull. Soc. Sci. Nat. Maroc 31: 254. 1951. North Africa, According to the description this plant is allied to A. montanum and A flahaultianum (see below), but it is impossible to tell from the description alone whether it should be maintained as a distinct species. The description appears to fit A. atlanticum. A. filifolium Wahlenb. in Berggren, Resor. i. Europa och Osterlanderna Bih. 1964 | DUDLEY, THE GENUS ALYSSUM 373 2: 54. 1826. ? Asia Minor. The original reference for this name has not been seen; however, an account of Jacob Berggren’s journey, by Wahlen- berg, appeared in /sis von Oken 21: 971-1006. 1828. There is on page 990 of this work a description of A. filifolium by Wahlenberg. Although in Index Kewensis 1: 97. 1895, this plant is cited from Asia Minor, Wahlen- berg states that it was collected by Bergren [one g]| in ‘“Morgenland.” Until the type is studied (spr) the name should not be used. The descrip- tion in /sis von Oken could apply to any annual with glabrous fruits in Alyssum. A. flahaultianum Emb. Bull. Soc. Sci. Natur. Maroc 15: 199. 1935, North frica. This plant is described as a perennial in sect. ALyssuM, allied to A. serpyllifolium (sect. ODONTARRHENA), having rose colored flowers, glabrous subcochleate fruits, and few-fruited racemes, characters nhich seem to apply better to A. cochleatum (sect. TETRADENTA). Judgement must be reserved, however, until the type material has been seen A. globosum Grossh. Izv. Azerbaidzh. Fil. Ak. Nauk. S.S.S.R. 5: 1939. Caucasus. The original publication of this species is apparently enavalabl in the United States or Great Britain, so application of the name must w A, hakaszkii Nyar. Anal. Acad. Rep. Pop. Rom. Sect. Stiinte Geol. i Biol. ser. A. mem. 3. 1: 32. 1949. Siberia. See A. ee below. This species is described as being “sp. trans. n. = micro phyllum— obovatum” and may be a synonym of either. A, heideri Kroch. Suppl. Fl. Silesia 4(2): 168. 1823. Germany. The informa- tion in the Senge of this plant is inadequate to place it in ne section of Alyssum A. “‘muricatum mihi vel tuberculatum” Kit. Linnaea 32: 498. 1863. Hungary. Kitaibel evidently could not make up his mind as to an appropriate specific epithet for this taxon, and supplied two names. Although he allies it to alyssoides. However, a decision should not be made until the type material has been examined. A, odoratum Colla, Herb. Pedem. 1: 141. 1833. Italy. There are no species of Alyssum with scented flowers as this specific epithet suggests. It is impossible to tell from the description, and without examining the type specimen (To, hb. Bellardi), whether the name applies to Alysswm or if it should be referred to another genus, for example, Lobularia (Lobularia maritima, the “Sweet Alyssum” having highly’. pear flowers ). A, ee ae _ Rae Rep. Pop. Rom. Sect. Stiinte Geol. Geog Biol. . A. mem. 3. 1: 1949, Siberia. The holotype of this ee, was aiilected i in ae in ne district Hakaszki. It is probably synonymous with A. hakaszkii, but the descriptions being inadequate, do not offer much assistance. Of interest is that the type of A. subbaicalicum was collected by Golubincev, and is deposited in the University of Tomsk herbarium (TK). One of the specimens cited as A. hakaszkii (see above) was also collected by Golubincev, deposited in the same herbarium, and may well be a du pli- cate. Alyssum subbaicalicum is described as being “SO. tYaRS.. f.-= 4: baicalicum—microphyllum” and may be a synonym of either taxon, both of which are apparently good species. 374 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV COMPARATIVE ANATOMY OF THE LEAF-BEARING CACTACEAE, XII PRELIMINARY OBSERVATIONS UPON THE STRUCTURE OF THE EPIDERMIS, STOMATA, AND CUTICLE I. W. Batrey ! THE TAXONOMIC SIGNIFICANCE OF structural variations in stomatal ap- paratuses 2 was strongly emphasized by Solereder (1899). Following the earlier work of Prantl (1872), Vesque (1889) and others, he distinguished a number of significant structural types, e.g., rubiaceous, false rubiaceous, cruciferous, caryophylleous. In the rubiaceous types (paracytic forms of Metcalfe and Chalk, 1950) the stomata are accompanied on either side by one or more accessory cells oriented parallel to the long axis of the pore and guard cells. In the true rubiaceous type the subsidiary cells are derived from the same mother cell as the guard cells; whereas in the false rubia- ceous forms the accessory cells arise by secondary divisions in neighboring epidermal cells. Since the publication of Strasburger’s (1866-67) paper, the Cactaceae have been regarded by later investigators as having stomatal apparatuses of the true rubiaceous form. In securing stems of varying sizes and ages preserved in formalin-acetic- alcohol (as a basis for investigations of the xylem reported upon in pre- ceding papers of this series), younger stems bearing leaves were usually included in collections of Pereskia, Pereskiopsis and Quiabentia. Such collections are difficult to obtain and have involved the active cooperation of many botanists. Young stems and leaves being thus available, it seemed advisable to initiate preliminary investigations, not only of stomata, but also of the epidermis and cuticle. The leaves of the three genera now available to me commonly have two, three, or four thin walled, crescent shaped subsidiary cells oriented parallel to the guard cells of the stomata (Fics. 1-4). At times there is but one subsidiary cell and at others secondary transverse divisions may occur in one or more of the subsidiary cells. Infrequently, in aberrant leaves some of the stomata appear to have no accessory cells derived from the same mother cell as the guard cells. The thin walled, crescent shaped accessory cells appear at least super- ficially to be derived from the same mother cell as the guard cells, in con- formity with Strasburger’s conclusion based upon detailed developmental ‘This investigation was financed by a grant from the National Science Foundation. I am aerreas to the American Philosophical Society for the loan of a Wild microscope. is paper stoma and stomata are used in referring to a pair of guard cells and the hesites between them, stomatal apparatus where subsidiary or accessory cells are involved in addition. 1964 | BAILEY, LEAF-BEARING CACTACEAE, XII 375 investigations of stomata in Pereskia aculeata Mill. and other cacti. Such an interpretation is strengthened by the not infrequent occurrence of groups of four or more thin walled, slender cells oriented parallel to each other, but not having completed differentiation of guard cells and an aperture between them As seen in surface views, the size and number of the stomata and the number and transverse diameter of the crescent shaped subsidiary cells varies more or less markedly in different leaves of the same taxon, in different leaves of the same plant, and in different parts of the same leaf. Similar forms of stomatal apparatuses occur in the epidermis of young stems. The size, form, and orientation of epidermal cells is highly variable in the three genera. At times the anticlinal walls of the cells are smooth (Fics. 2 and 4), whereas in other cases they are conspicuously undulating (Fic. 3). The stomatal apparatus may be jacketed by epidermal cells of but slightly modified size and form (Fic. 3). They may be completely (Fic. 1+), or partly (Fic. 4) jacketed by a pair of epidermal cells whose longer curved axis is oriented parallel to that of the guard cells and sub- sidiary cells. One or both of these curved epidermal cells may have sec- ondary transverse divisions (Fic. 2). Where the pair of epidermal cells does not completely jacket the stomatal apparatus, one, two or more con- spicuously small epidermal cells oriented at right angles to the major axis of the guard cells may be present (Fic. 4+). In some cases the stomatal apparatus tends to be jacketed by a ring of small epidermal cells. In other cases the stomatal apparatus is situated in a broadened complex of small epidermal cells (Fic. 12 Owing to variations in the breadth of subsidiary cells and their tendency to divide transversely at times, and to variations in the size, form, and orientation of epidermal cells which jacket the stomatal apparatuses, it is difficult in some cases — without developmental evidence — to determine with certainty which accessory cells arise from the same mother cell as the guard cells. Furthermore, owing to the wide ranges of structural variabil- ity of the epidermis and stomatal apparatuses within a taxon, morphologi- cal criteria for differentiating taxa are of uncertain reliability unless based upon very extensive collections of leaves and young stems from plants of different ages when grown under varying environmental influences. How- ever, the cuticle, epidermis, stomata, and sub-epidermal layers of available material, when studied in adequately stained periclinal and anticlinal sections, appear to provide potentially useful criteria in the comparison of different categories of Pereskia, and in distinguishing the genus from Pereskiopsis and Quiabentia. In addition, such investigations are par- ticularly desirable from a physiological point of view in regard to the structure and functioning of stomata, and in studying the conservation of moisture in leaves and in young stems prior to periderm formation. As shown in the second paper of this series (Bailey, 1961a), Pereskia sacharosa Griseb., P. grandifolia Haw., P. bleo DC., P. moorei Britt. & Rose, P. tam picana Web. and P. sorvrugata Cutak belong in a distinct 376 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV category of pereskias characterized by having consistently occurring forms of aggregated fiber-sclereids in their secondary phloem. Their xylem exhibits less divergent trends of specialization than that of other pereskias (Bailey, 1962, 1963c). In this category of taxa, having large, thin, pin- nately veined leaves, stomata are at times few in number or absent on the upper or adaxial surface of the foliar appendages. When epidermal peels are stained in sudan III or in sudan III and Haidenhein’s haematoxylin and mounted in glycerin, the cuticle and the outer periclinal wall of the epidermal cells either exhibit no conspicuous structural patterns (Fic. 1), or have variously oriented striations (Fic. 5) In view of the emphasis that has been placed upon cuticular striations in taxonomic investigations it is essential to determine whether such striations are present in unmodified living leaves or are produced by differential contractions of the cuticle and epidermal walls during fixation, dehydration, and other treatments. In the pereskias discussed in the preceding para- graph, cuticular striations, when present, are not due primarily to inequali- ties in thickness of the cuticle but to contractions of a cuticle of relatively uniform thickness forming folds or ridges. It is significant in this connec- tion that in unlignified, unsuberized and uncutinized meristematic and other tissues of plants the primary walls contract markedly during de- hydration. Furthermore, adjacent tissues in intact plants not infrequently are under differential tensions which may be released or modified in peels or sections of leaves producing morphological changes even prior to de- hydration. Where cuticular striations occur, the primary walls of the epidermal cells are smooth except for the outer surface of the outermost walls which have projections or ridges corresponding to the folds in the cuticle. In Pereskia sacharosa and allied taxa the cuticular striations, when present, vary markedly in form and orientation, not only in different leaves of the same taxon or plant, but also in different parts of a single leaf. At times, they may be linear and parallel to one another and may extend from cell to cell particularly in axially elongated cells overlying the mid- vein. They may be radially oriented in relation to the stomata or some particular cell of the epidermis (Fic. 5). In other cases they may be variously convoluted and oriented. The variations in pattern appear to be correlated with variations in the size, form, and orientation of the epider- mal cells and to the thickness and degree of contraction of primary walls during dehydration. The range of variability is so extensive as to render difficult and laborious any attempt to differentiate species upon the basis of such diagnostic criteria. It should not be inferred from this, however, that even where cuticular striations are induced by dehydration in other families, differences in the presence or absence of such structures may not be due to consistent differences in the cellular structure and behavior of the leaves in related species. In the leaves of the category of pereskias under discussion, the outer cells of the mesophyll are relatively large and thin walled (Fic. 7). The stomata are subtended by intercellular spaces of irregular ramifying form 1964 | BAILEY, LEAF-BEARING CACTACEAE, XII 377 in the chlorenchymatous tissue. As seen in median sectional view, the more or less circular guard cells have thickened walls on their upper and lower surfaces. The first pair of subsidiary cells are thin walled and their internal extensions curve about the guard cells exhibiting various degrees of approximation. The relatively thin part of the cuticle overlying the stomatal apparatus curves inwardly into the frontal cavity forming slender outer “ledges” but commonly more slender than those illustrated in Fic. 15. Very tenuous extensions of the cuticle pass inwardly lining surfaces of the aperture but without forming conspicuous internal ledges as in some dicoty- ledonous plants. Inner extensions of the tenuous cuticle may likewise occur on the exposed surfaces of cells in the subtending intercellular cavity of the mesophyll.* It appears likely that in the above category of stomatal apparatuses closure of the stomata may result (1) by contact between the outer cuticular ledges, (2) by contact between the guard cells, and (3) by approximation of the inner parts of the first-formed pair of subsidiary cells, as in (Fic. 7). In vigorous, rapidly elongating shoots, periderm formation is retarded: in the case of internodes, appearing first in parts at a considerable dis- tance below the apex of a stem. The lower internodes have a markedly thickened cuticle and massive, wedge shaped accumulations of cutin which extend inward between the epidermal cells (Fic. 16).4 Although the cells of the outer cortex tend to be smaller, no collenchymatous outer layers or crystalliferous hypodermis are formed, as noted in the third paper of this series (Bailey, 1961b). Stomata are of sporadic and relatively infrequent occurrence, particularly in P. sacharosa, being absent in some internodes or parts of them. Where stomatal apparatuses are present, the walls of their cells are very thin (Fics. 11 & 12), and the part of the cuticle which overlies them is tenuous. Furthermore, the cuticular ledges lining the en- trance to the aperture are slender as in the case of leaves The genera Pereskiopsis and Quiabentia differ from this and other categories of Pereskia in having no characteristic forms of sclereids in their secondary phloem. In addition, the genera differ from this category of pereskias, in having highly advanced trends of divergent specialization in their xylem (Bailey, 1964). When epidermal peels of the comparatively thick, palmately or pseudo-palmately veined leaves are obtained, and are subjected to the same treatments and staining as those from the thin, pinnately veined leaves of the preceding pereskias, epidermal cells (with the exception of those of the stomatal apparatus) exhibit more or less conspicuous punctate or reticulate patterns of varying degrees of coarse- ness (Fics. 2 & 4). At times cuticular striations are visible in addition. In anticlinal sections, these outer primary walls have a beaded appearance suggestive of inequalities in thickness. In surface views, the more intensive staining of parts of the reticulate patterns in haematoxylin suggests the “For clarification of nomenclature and a general discussion of the structure of Sia see Esau (1953). strictly developmental, rather than a purely descriptive, point of view the tniekening probably progresses largely from within outwardly. a ben j ° 378 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV possibility of differences in chemical composition. The structure of these walls obviously merits more detailed and comprehensive investigation since it may ultimately provide a significant criterion in differentiating leaves of Pereskiopsis and Quiabentia from those of Pereskia. All material available to me at present was preserved in F.A.A. It is essential to determine to what extent the reticulate patterns may have been induced or exaggerated by contractions due to dehydration as in the case of cuticular striations. The leaves of Pereskiopsis and Quiabentia usually have abundant sto- mata on their upper as well as their lower surface. The cuticle overlying the epidermal cells tends to be of relatively uniform thickness except where it forms more or less massive accumulations above the central part of the guard cells and at times between the guard cells and subsidiary cells (compare Fics. 17 & 18). These thicker parts of the cuticle deform the guard cells and depress them inwardly. In extreme cases, the central parts of the guard cells are greatly reduced in diameter (Fic. 18). This is in marked contrast to what occurs in the stomata of the pereskias discussed on preceding pages, where the cuticle merely forms slender ledges lining the entrance to the stomatal aperture. These structural differences are strikingly demonstrated in anticlinal sections of young stems (Fics. 17 & 18) which have numerous stomatal apparatuses, and in surface views of the epidermis when intensely stained in sudan III (Fic. 13). In the case of the pereskias, the thick cuticle forms massive intercellular intrusions between the cells of the epidermis except in the stomatal apparatuses (Fics. 11 & 12). On the contrary, in young stems of Pereskiopsis and Quiabentia massive intrusions of the cuticle are largely confined to the stomatal apparatuses (Fic. 13). It should be emphasized in these connections that it is essential in dealing with young stems to compare internodes in as nearly equivalent stages of development as possible, particularly those which have completed or nearly completed elongation prior to periderm formation. This is due to the fact that significant changes may occur in the cuticle, and in the cells of the epidermis and sub-epidermal layers during early stages of the elongation and increase in circumference of young stems. Similarly in dealing with leaves, changes in thickness of the cuticle, in the form of epidermal cells, and in the internal structure of stomata may occur during successive stages of the expansion and maturation of leaves. Furthermore, it should be kept in mind that unusually thin small leaves of shoots grown under certain environmental influences may vary more or less markedly in structure from larger thicker leaves of plants grown in different en- vironments, This may account at least in part for the structural dis- crepancy between Fic. 3 (leaf from a rapidly elongating shoot in a green- house) and Fic. 6 (leaf from the slower growing branch of an old tree in its native habitat). In sectional views of leaves and young stems of Pereskia sacharosa and allied taxa, the first pair of subsidiary cells commonly does not extend inwardly in excess of the general thickness of the epidermis (Fic. 7), whereas in the case of Pereskiopsis it is usually more extensive internally 1964 | BAILEY, LEAF-BEARING CACTACEAE, XII 379 (Fics. 17 & 18). However, in both cases the innermost parts of the first pair of subsidiary cells exhibit varying degrees of approximation when not actually in contact, as in Fic. 7. Young stems of Perekiopsis and QOuiabentia, in contrast to those of the pereskias, are characterized by having collenchymatous layers and a well developed crystalliferous hypodermis subtending the epidermis (Bailey 1961b). This taxonomically and physiologically significant difference in structure is illustrated in Fics. 14 & 19. Where collenchymatous layers are formed (Fic. 19) there are no intercellular spaces in the tissue except in clearly defined channels subtending the stomata. These channels com- municate internally with intercellular spaces in the subtending, large, thin walled, chlorenchymatous tissue. They are jacketed by unpitted parts of the walls of collenchymatous cells (Fic. 22), and in turn by very tenuous intrusions of the cuticle, so tenuous as not to be visible in Fic. 22. In young stems of Pereskiopsis and Quiabentia, the numerous, more or less closely aggregated druses of the hypodermal layer are compact and rotund (Fic. 20) and commonly differ from the diffusely distributed ones of the inner cortex which tend to be of stellate form (Fic. 21). During development of the crystalliferous hypodermis, enlargement of the crystals presses the epidermal cells outwardly and laterally, thus more or less drastically modifying their form and arrangement. Not infrequently the druses press laterally into the intercellular channels subtending the stomata and at least partly occlude their entrances. Anatomically more specialized species of Pereskia exhibit less consistent differences in distinguishing the genus from Pereskiopsis and Ouiabentia. In the case of P. aculeata Mill. and such Peruvian and Bolivian pereskias as P. humboldtiu Britt. & Rose, P. weberiana Schum., and P. diaz-romero- ana Card. (which have markedly divergent trends of specialization in their xylem, Bailey, 1962, 1963a) there is a tendency for the outer cortical layers in young stems to become conspicuously thicker-walled and to appear at least pseudo-collenchymatous. Furthermore, although no typical crystalliferous hypodermis is formed, there is a tendency for relatively widely spaced, often linearly arranged aggregations of rotund druses to occur in the outermost layer of the cortex. Such occurrences are at least suggestive of possible initial trends of specialization which become exag- gerated and dominant in Pereskiopsis and Quiabentia. The stomata in normally developing leaves of these taxa have cuticular ledges resembling those of Pereskia sacharosa and allied taxa. However. young stems have consistently more numerous stomata than do those of the primitive category of pereskias. The lowermost internodes of the Andean pereskias prior to periderm formation tend to have a thick cuticle and massive intercellular accumulations of cutin intruding between the epidermal cells as in P. sacharosa and allied taxa, but they have more extensive deposits of cutin overlying the guard cells and accessory cells as in Pereskiopsis and Quiabentia (Fic. 13). In the case of P. aculeata periderm formation is so precocious in young shoots that I have not suc- ceeded in obtaining internodes bearing stomata in stages of elongation and 380 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV increase in circumference comparable to those available in collections of the Andean pereskias. In younger developing internodes of P. aculeata the cuticle is thin and jackets a conspicuously papillose outer surface. As shown in the first paper of this series (Bailey, 1960), Pereskia pititache Karw. (P. conzattii Britt. & Rose), P. autumnalis (Eichlam) Rose and P. nicoyana Web. differ from the pinnately veined pereskias in having palmate or pseudo-palmate venation comparable to that which occurs in Pereskiopsis and Quiabentia. However, the divergent trends of specializa- tion in the xylem of these taxa are less advanced than in P. aculeata and the Andean pereskias (Bailey, 1963b). The cuticle overlying the epidermal cells may be smooth (Fic. 3) or it may have contraction folds, i.e., cuticu- lar striations (Fic. 6) comparable to those that occur in other pereskias. The stomatal apparatuses in surface views of leaves vary markedly in form in the three taxa. In thick, fully expanded leaves from trees growing in the wild, the stomata differ from those of previously discussed cate- gories of pereskias in having massive accumulations of cutin overlying the guard cells, as in Pereskiopsis and Quiabentia. Periderm formation is so precocious in young shoots of these pereskias that Dr. Boke and I have not succeeded as yet in obtaining material adequate for determining whether similar stomata occur in stems. However, since the cork cambium of cacti commonly arises by periclinal divisions in the epidermis, it is possible to observe after periderm formation is initiated that no collenchy- matous layers or crystalliferous hypodermis are formed in young stems of the three pereskias. In the case of Pereskia colombiana Britt. & Rose, P. guamacho Web., P. cubensis Britt. & Rose and P. portulacifolia Haw. adequate collections of roots and stems were obtained for studying divergent trends of special- ization in the xylem (Bailey 1963d). Unfortunately only a few collections of leaves and young stems have been obtainable thus far. Available evidence at present suggests, however, that the stomatal apparatuses of these taxa not only have slender cuticular ledges, but also accumulations of cutin penetrating between the guard cells and subsidiary cells. These deposits of cutin are much less massive than those which occur in the preceding category of pereskias and in Pereskiopsis and Quiabentia. Stomata are of infrequent occurrence in the epidermis of young stems, thus resembling the situation in P. sacharosa and allied taxa, but the cuticles are thin and devoid of massive intercellular intrusions of the epidermis. The stems are devoid of collenchymatous layers and a crystalliferous hypodermis. DISCUSSION AND CONCLUSIONS In preceding papers of this series I have shown that Pereskiopsis and Quiabentia may be differentiated anatomically from Pereskia by (1) the characteristic development in young stems of a highly specialized crystalli- ferous hypodermis in these genera (Bailey, 1961b), by (2) the absence of diagnostic forms of sclereids in their secondary phloem (Bailey, 1961a), 1964] BAILEY, LEAF-BEARING CACTACEAE, XII 381 and except in comparisons with Pereskia pititache, P. autumnalis and P. nicoyana, by (3) the characteristic palmate or pseudo-palmate vasculature of their leaves (Bailey, 1960). In addition, I have shown on preceding pages that, except in comparison with stems of Pereskia aculeata and the Andean pereskias, Pereskiopsis and Quiabentia are characterized by having typical collenchymatous outer cortical layers. A preliminary investigation of cuticles, epidermal cells, and stomatal apparatuses reveals such wide ranges of variability within taxa, particu- larly in superficial views of stomatal apparatuses, as to be of questionable taxonomic reliability unless based upon very extensive collections of ma- terial from plants of different ages when grown under different environ- mental influences. However, detailed investigation of periclinal and anti- clinal sections of available leaves and young stems reveals some structural differences of apparently potential taxonomic significance. The occur- rence of punctate and reticulate structures in the outer periclinal walls of epidermal cells in leaves of Pereskiopsis and Quiabentia may eventually provide a useful criterion in differentiating the foliar appendages of these genera from those of Pereskia. Furthermore differences in thickness of the cuticle, in intercellular accumulations of cutin between cells of the epider- mis, and in the deposition of cutin in the frontal cavity or vestibule of stomatal apparatuses have at least supplementary significance in differ- entiating certain categories of Pereskia, and in separating one or two of them from Pereskiopsis and Quiabentia. It should be emphasized again in this connection that investigations of the cuticle, epidermis and stomatal apparatuses, and of the subtending outer cortical layers of young stems prior to periderm formation are par- ticularly desirable from a physiological point of view, i.e., in regard to the functioning of stomata, and in retention of moisture in leaves and young stems of plants growing in semi-arid or arid environments. In the case of Pereskia sacharosa, P. grandifolia and allied taxa, stomata frequently are few in number or absent on the upper surface of the rela- tively large thin leaves, and on the surfaces of internodes of young stems prior to periderm formation which tends to be more or less precocious in its development. The stomatal apparatuses, even on the under surfaces of the leaves, exhibit no conspicuous structural modifications for the prevention of water loss such as occur in many dicotyledons growing in arid environments. This may be due at least in part to the abscission of leaves at the end of a rainy season. Although the internodes of young stems prior to periderm formation have very thick cuticles and massive intercellular accumulations of cutin between the epidermal cells (which may have some physiological significance during the dry season), the cuticle overlying the stomatal apparatuses is tenuous, and there are no con- spicuous structural adaptations for the prevention of loss of moisture. If the stomata are actually functional, as seems questionable in some cases, regulation of moisture probably is controlled merely by the opening and closing of the stomatal aperture. In the case of Pereskiopsis and Quiabentia and such pereskias as P. 382 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV pititache and allied taxa, the massive accumulations of cutin in the stoma- tal apparatuses of fully matured leaves prior to abscission, and in the fully elongated internodes of young stems shortly prior to periderm forma- tion raises important physiological questions regarding the functioning of stomata in these plants. In such stomatal apparatuses as those illustrated in Fics. 17 and 18 it appears unlikely that the stomata can be fully func- tional even where the guard cells contain vestiges of viable protoplasts. In certain of my collections the guard cells of some or all of the adjacent stomatal apparatuses have collapsed and are thoroughly impregnated with cutin (Fic. 9). It is of interest in this connection that some of the druses in fully matured cells of the crystalliferous hypodermis may be partly or completely jacketed by cutin. In general in the leaf-bearing Cactaceae the ordinary epidermal cells appear to be devoid of chloroplasts, but where the stomata appear to be potentially functional with partly or fully open apertures, the guard cells, and sometimes the subsidiary cells are packed with green plastids (Fic. 8). This raises the question whether stomata which are functional during earlier stages of the development of leaves and young stems subsequently become occluded and incapable of functional activity as in Fic. 9. In some taxa the epidermis is in direct contact with subtending thin walled chlor- enchymatous tissue, whereas in young stems of some taxa it is separated from such tissue by layers of thick walled collenchymatous cells. The col- lenchymatous tissue, like the epidermis, contains few if any chloroplasts and, since it is devoid of intercellular spaces, must retard the movement of gases and water vapor except in intercellular channels subtending the stomata. It should be emphasized in conclusion that a preliminary investigation of the cuticle, epidermis, stomatal apparatuses, and subepidermal layers of the most primitive surviving representatives of the Cactaceae reveal salient trends of structural specialization whose functional significance can be fully and reliably elucidated only by sustained experimental and other observations on plants growing in their native habitats. Such detailed experimental investigations, in adequate paca irae with anatomical and enviromental ones, are at present much needed for a clearer understanding of salient adaptations in the Cactaceae which are essential for survival in arid environments. LITERATURE CITED Bartey, I. W. 1960. Comparative anatomy of the leaf-bearing Cactaceae, I. Foliar _ of Pereskia, Pereskiopsis and Quiabentia. Jour. Arnold. Arb. 41: 196la Structure and distribution of sclerenchyma in the phloem of Pereskia, Pereskiopsis and Quiabentia. Ibid. 42: 144-156. ———. 1961b. III. Form and distribution of crystals in Pereskia, Pereskiopsis and aembes Ibid. 334 . VI. The xylem Pere skia sacharosa and Pereskia aculeata, Ibid, 43: ea 1964 | BAILEY, LEAF-BEARING CACTACEAE, XII 383 1963a. VII. The xylem of pereskias from Peru and Bolivia. /bid. 44: 127-137. . 1963b. VIII. The xylem of pereskias from southern Mexico and Cen- tral America. /bid. 211-221. . 1963c. IX. The xylem of Pereskia grandifolia and Pereskia bleo. Ibid. 222-231. . 1963d. X. The xylem of Pereskia colombiana, ao guamacho, mes cates and Pereskia portulacifolia. Ibid. 401. : The xylem of Pereskiopsis and Pana ee 45: 140-157. EsAu, K. non oe anatomy. John Wiley & Sons, Inc., New York. MetcaLre, C. R. & L. CHALK. 1950. Anatomy of the dicotyledons. Clarendon Press. Oxford. PRANTL, K. 1872. Die Ergebnisse der neueren Untersuchungen uber die Spaltoff- nungen. Flora. N.S. 30: 305-312, 321-328, 369-382 SOLEREDER, H. 1899, Systematische Anatomie der Dicotyledonen. F. Enke. uttgart STRASBURGER, E. 1866-67. Ein oe zur eee der Spalt- offnungen. Jahrb. Wiss. Bot. 5: VESQUE, M. J. 1889. De l'emploi ia mee anatomiques dans la classifica- tion der végétaux. Bull. Soc. Bot. France 36: XLI-LXXXIX. EXPLANATION OF PLATES PLATE I Fics. 1-4. Epidermal layers of leaves, pat in Haidenhein’s haematoxylin and sudan III and mounted in glycerin, L; ieee Matra hee [ Cas- tellanos |, epidermal layer from leaf preserved im FAA. 2, Pereskiopsis scandens N.Y. Bot. Gard.], epidermal layer from leaf preserved in F.A.A. 3, Pereskia pititache | Boke], epidermal peel from small thin leaf of young elongating shoot of a greenhouse plant. 4, Pereskiopsis [Boke B-20], epidermal layer from leaf preserved in F.A.A PLATE II 5-9. Cuticular striations and stomatal apparatuses. 5, Pereskia sacharosa IN.Y. eae Gard.], epidermal peel from leaf stained o ae an III showing radi- ally and irregularly oriented cuticular striations, & 420. 6, Pereskia agen [ Boke |, epidermal layer from large thick leaf of an ne mn growing in t Sei in sudan III, ae convoluted forms of cuticular striations, x 330. Pereskia sacharosa {N.Y. Bot. Gard.], anticlinal section of leaf stained in sian and haematoxylin, ‘showing stomatal apparatus and subtending inter- cellular cavity in the chlorenchymatous tissue, X 420. Pereskia grandifolia | Boke], epidermal peel from leaf, stained in sudan III, sho wing chloroplasts in guard cells of stomata, * 510. 9, Pereskia aff. humboldtii | Ferreyra|, epidermal layer from stem stained in sudan III, showing collapsed guard cells plugged with cutin, * 420 PLATE il Fics. 11-13. Epidermal layers of Pa oe in F.A.A ned in sudan III and mounted in glycerin, & 260. Pereskia ie Tet 12, Pereskia aff. bleo [ Rodriguez |. 13, ree scandens | N.Y. Bot. Gard. ]. 384 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV PLATE IV Fics. 14-18. Sectional views of young stems prior to periderm formation. 14, Pereskia tampicana [Boke], stained in haematoxylin and safranin, showing [Moran 7270], cuticle stained intensely with sudan III and mounted in glycerin, x 420. 17, Pereskiopsis aquosa |N.Y. Bot. Gard.], cuticle stained with sudan ITI, showing massive accumulation of cutin in frontal cavity, overlying the guard cells ~ penetrating between the guard cells and the subsidiary cells, 420. 18, Pereskiopsis porteri [N.Y. B . Gard. ], cuticle stained with sudan III show- ing massive intercellular intrusions of cutin, oo of guard cells inwardly, and marked reduction in their diameter, * 4 PLATE V Fics. 19-22. Details of collenchymatous and crystalliferous layers. 19, Pere- skiopsis scandens [N.Y. Bot. Gard.], section of young stem stained in haema- lin and safranin showing contrast in thickness of walls of the inner chlor- enchymatous cortex and the outer collenchymatous layers which are devoi intercellular spaces se ges _ channels subtending the stomata, 200. 20, bake ped ses Sire Y. Bot. Gard.], showing form of druse in crystalliferous s, X 510. 21, The me, showing form of — in chlorenchymatous ae x 31 O...2a; Bs tetirees chapistle [| Mo. Bot. rd.], tangential section of collenchymatous layers stained in haematoxylin, sfc intercellular channel and character of primary wall “pitting,” 200. PLaTE I Jour. ARNOLD Ars. VoL. XLV s BAILEY, LEAF-BEARING CACTACEAE, XII PLATE II Jour. ARNOLD Ars. VoL. XLV man eg a enna # # + € * t . ~~ BAILEY, LEAF-BEARING CACTACEAE, XII PuLaTE III] Jour. ARNOLD Ars. VoL. XLV % ey : ary eed Soe ea Ber 1 ea * Pe Mea Tage frees apne ee ri rai. cata x * “ge Q ne! li ie oY wt BAILEY, LEAF-BEARING CACTACE XII AE, PLATE IV Jour. ARNOLD Ars. VoL. XLV BAILEY, LEAF-BEARING CACTACEAE, XII Jour. ARNOLD Ars. VoL. XLV PLATE V 390 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV SYNOPSIS OF THE GENUS AURINIA IN TURKEY T. R. Duptey ! THE GENUS AuRINIA Desv. has long been considered a section of Alys- sum L., but, although both taxa are in the Cruciferae tribe Alysseae, affinity between them appears to be tenuous and remote. The major objec- tion to keeping Awrinia as a section of Alyssum is that, although composed of a relatively homogeneous group of species, Aurinia is anomalous in com- parison with the natural sections (i.e., MENIocus (Desv.) Hook., Ps1Lo- NEMA (Meyer) Hook., ALyssuM, GAMOSEPALUM (Hausskn.) Dudl., TETRA- DENIA (Spach) Dudl., and ODONTARRHENA (Meyer) Koch) of Alyssum. Lines of probable affinity may be traced between these sections, but Az- rinia does not appear to have any such clearcut relationships. Among the genera of the Alysseae, Aurinia seems to be most closely allied to Berteroa DC., or possibly to Alyssoides Mill. No taxon presently considered an Alyssum possesses the combination of characters which are evident in Aurinia and which permit it to be easily recognized, and distinguished feasibly from Alyssum. These characters are contrasted in the following table. AURINIA ALYSSUM ROSETTE LEAVES Usually repand-sinuate, den- Usually entire; (2—)5-15 tate, or pinnatifid, (20-) (-20) mm. long; petioles CAULINE LEAVES FERTILE STEMS FLOWER BUDS CALYX STIGMAS 40-60(-100) mm. long; pet- ioles deeply grooved with swollen bases which are per- sistent on caudices. Reduced, 1/2 (or less) the size of the rosette leaves. Sparsely foliate, often near- ly scapose. Globose, as long as broad. Cup-shaped, with © sepals widely spreading at matu- rity. (Immature ) bilobed. conspicuously flat and attenuate, bases never swollen or persistent on caudices. More or less equal in size to the rosette leaves. Always leafy, never scapose. Elliptic and oblong. Elongate, with sepals erect at maturity. Globose, rarely bilobed. “The assistance of Dr. James Cullen, Department of Botany, University of Edin- burgh, Scotland, is gratefully acknowledged, as well as that extended by Dr. Bernice G. Schubert of the Arnold Arboretum. 1964 | DUDLEY, THE GENUS AURINIA 391 Aurinia Desv. Jour. Bot. 3: 162. 1814 (!). Lecrorypr species: Alys- sum saxatile L. Sp. Pl. 2: 650. 1753 (= Aurinia saxatilis (L.) Desv. loc. cit.). — Meyer in Ledeb. FI. Ross. 1: 136. 1842. — Griseb. Spic. Fl. Rum. & Bith. 1: 271. 1843. — Koch, Hort. Dendrol. 23. 1853. — Schur, Envm. Pl. Transsilv. 61. 1866. Alyssum sect. Anodontea DC. Syst. Nat. 2: 317. 1821 (!). LecrotyPE Es: Alyssum edentulum Waldst. & Kit. ( = Aurinia petraea (Ard.) Schur). Anodontea (DC.) G. Don, Gen. Hist. Dichl. Pl. 1: 180. 1831 (!). Anodontea sect. Vesicaria sensu G. Don, loc. cit. (!), non Vesicaria Adans. Alyssum sect. Aurinia (Desv.) Koch, Syn. Fl. Germ. & Helv. 1: 58. 1837 ( Ch. — Boiss. Fl. Or. 1: 263 & 265. 1867.— Schulz in Engler & Prantl, Nat. Pflanzenfam. II. 17b: 491. 1936.— Busch in Komarov, Fl. U.R.S.S. 8: 348. 1939.— Nyar. Anal. a Rep. Pop. Rom. Sect. Stiinte Geol. Geog. Biol. ser. A. mem. 3. 1: 8 Aurinia sect. Corioceratium ce _ Fl. Rum. & Bith. 1: 272. 1843 (!). LECTOTYPE SPECIES: Aurinia ee (L. ) Lepidotrichum Velen. & Bornm. Oester a ae schr. 39: 324. 1889 (1). PE SPECIES: Lepidotrichum ee ein (Bornm.) Velen. & Bornm. Plants perennial, or biennial. Rosette leaves up to 100 mm. long, with repand or sinuate or dentate to pinnatifid margins and deeply grooved petioles with swollen and persistent bases (except Aw. halimifolia), usually distinctly rosulate and conferted on indurated caudices. Cauline leaves reduced, usually 14 (or less) the size of rosette leaves. Fertile stems sparsely foliate, often nearly scapose. Jndumentum of stellate or lepidote hairs with few or many branched or unbranched rays. /nflorescence un- branched and racemose, or multibranched, corymbose or paniculate. Pedi- cels divergent and spreading. Flower buds globose, as long as broad. Calyx cup-shaped. Sepals 4, widely spreading at maturity, free, never truly saccate. Petals 4, yellow or white, bifid or entire, spathulate and clawed. Long filaments 4, always free, equally bilaterally winged and dilated at bases. often with small basal teeth. Short filaments 2, winged with small basal teeth. Anthers short, obtuse, filament tips never prolonged. Vectaries 4, globose or triangular, one at each side of short filaments. Sfyes of vary- ing lengths, often widely dilated to bases, persistent. Stigmas (immature) conspicuously bilobed. Silicles latiseptate, always dehiscent, glabrous or rarely with indumentum (only Aw. rupestris), bilocular with conspicuous repla, locules 2—4(-6)-ovulate with placentation nearly apical, or rarely lateral (Au. halimifolia & Au. leucadaea); valves compressed or subin- flated, or equally inflated and turgid. Seeds winged, or rarely wingless (Au. uechtritziana), not mucilaginous. Embryo notorrhizal. DistrIBUTION: Central and southern Europe, east to the Ukraine and the Caucasus, north to Poland, and in Turkey. 392 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLV SPECIES NOT OCCURRING IN TURKEY: Au. corymbosa Griseb. Spic. Fl. Rum. & Bith. 1: 271. 1843 (!). The Balkans Au. halimifolia (Boiss.) Cullen & Dudley, comb. nov.? The maritime Alps of France and Italy. Au. leucadaea (Guss.) Koch, Hort. Dendrol. 23. 1853 (!). Yugoslavia and Italy Au. petraea (Ard.) Schur, Enum. Pl. Transsilv. 61. 1866 (!); Fuss, Fl. Transsilvy. 64. 1866, possibly the earlier combination, fide /ndex Kewensis. Central and southeastern Europe. Escaped from cultiva- tion in New York State. Key TO SPECIES OF AURINIA OCCURRING IN TURKEY “a Inflorescence compound, multibranched, corymbose; margins of rosette leaves always repand or sinuate, or dentate to pinnatifid; indumentum of stellate hairs, in or many rayed, never lepidote; valves of silicles mem- branous and sm Flowers ae seeds broadly winged; silicles (3-)4-8(-12) & (2-) 3-8(-15) mm., valves compressed or only slightly inflated at center; inflorescence 5-15(—20) cm. long; fertile stems greenish, with sparse POINTS - 5 og oc wee bh 28 8 v4 Abdio eat BE 1. Au. saxatilis. Flowers white; seeds wingless; silicles 2-4 2-2.5 mm., valves turgid, strongly and equally inflated; inflorescence 20-40 cm. long; fertile stems whitish with dense indumentum, especially towards base. |. 2. Au. uechtritziana. Inflorescence simple, unbranched, racemose; margins of rosette leaves always entire; indumentum of strongly lepidote hairs; valves of silicles cartilaginous and prominently ridged at apex. .......... 3. Au. rupestris. bd > Aurinia saxatilis (L.) Desv. ‘hae Bot. 3: 162. 1814 (!).— Kotov, Fl. Ukraine 5: 329. pl. 75. 19 Silicles elliptic or obovate, canine x 2-4(-5) mm., always longer than broad, obtuse or acute, rarely emarginate; styles 0.5-1 mm. long, not con- spicuously dilated to bases; margins of rosette leaves sinuate, or onl sparsely dentate, with teeth not more than 0.5 mm. long. a. Subsp. saxatilis. Silicles orbicular or oblate, 3.5-10(-12) & 4-12(-15) mm., usually wider than long, emarginate or truncate (or if obtuse or acute, then silicles 8-10 (-15) mm. long and wide, and styles 2-5.5 mm. long); styles 2.5 mm. long, strongly dilated at bases (bases 0.5-0.8 mm. in diameter; apices 0.1-0.2 — . — — * Aurinia halimifolia (Boiss.) Cullen & Dudley, comb. nov. Ptilotrichum halimi- folium Boiss. Voy. Bot. Mid. Esp. 1: 45. 1839, non Alyssum halimifolium Willd. Linn. Sp. Pl. ed. 4. 3(1): 460. 1800, nomen illeg. nec Alyssum halimifolium L. Sp. Pl. 2: 650. 1753 ( = Lobularia maritima (L.) Desv.). Boissier i name (cf. note to Article 72, Internat. Code of Bot. Nomencl. 52. 1961). Accordingly, the basionym of Aurinia halimifolie is Ptilotrichum halimifolium Boiss., not Alyssum halimifolium Willd. 1964] DUDLEY, THE GENUS AURINIA 393 mm. in diameter); margins of rosette leaves strongly dentate, often pin- natifid, with teeth (0.5—)1—5 mm. long (or if margins entire, then plants very reduced, only 5-10 cm. tall). 2. Silicles 3.5-6 mm. X 4-7 mm.; pedicels 4-10 mm. long; styles 1-1.5 mm. long; petals 3-4(-4.5) & 1.5-2 mm.; wings of seeds 0.4-0.5 mm. wide. tetas er ask kas es 2 ok b. Subsp. orientalis. 2. Silicles 6-12 X 8-15 mm.; pedicels (7—)10-20 mm. long; styles (1.5—-) 2-2.5 mm. long; petals (4s- 6-7 & (2-)2.5- ae mm.; wings of seeds (0.5-)0.7-1 mm. wide. ....................0.. Su bsp. megalocar pa. Subsp. saxatilis Alyssum saxatile L. Sp. Pl. 2: 650. 1753 (!). Holotype, Oe in the Leiden Botanic Garden (1, hb. Royen; hb. lugdb. Bat. no. 901050). Reichenb. Ic. Bot. 3: 30. ¢. 232. 1825; Ic. Fl. Germ. & Helv. 2: 9. ¢. 21, fig. 4280. 1837-1838. — Mansfeld, Repert. Sp. Nov. 46: 115.1939, — Nyar. Fl. Rep. Pop. Rom. 3: 327. pl. 56, fig. 2. 1955.— Moravec, Preslia 32: 360. 1960. Alyssum arduinii Fritsch, Exc. Fl. Ost. 253 & 268. 1897 (!); in Kerner, Sched. Fl. Aust.-Hung. 9: 24. 1902. Alyssum arduinii subvar. lamprocarpum Borb. Balaton FI. 391. 1900. Holo- type, Hungary. Tul-a-Dunan, Borbds (Bp, non vidi). — Jav. & Csap. Ic. Hung. 7: pl. 213, figs. 1600 & 1600a. 1930. Alyssum saxatile var. typicum Beck, Glasn. Zem. Muz. Bosn. Herceg. 28: 129. 1916 (1). Alyssum saxatile var. arduinii (Fritsch) Hayek, Prod. Fl. Pen. Balc. 1: 430. 1925 (!). Alyssum saxatile subsp. arduinii (Fritsch) Paw. Fl. Tatrorum. 1: 325. 1956 C) DISTRIBUTION AND HABITAT: Widespread on calcareous substrates in central and southeastern Europe, extending east to the Ukraine and the upper Caucasus, and north into Poland, alt. up to 1000 m. FI. Apr.—June. a FRANCE: Bei Muggendorf, frank. Jura, 410 m., Reinsch 157 (&, GH, . GERMANY: 1889, Dresler (GH); sua eae 1838, Souder (£); Aug. cr Koch (x); Koch 11 bis (em, GH, K, w). AustRIA: ad Danube, Kerner K, W). CzECHOSLOVAKIA: Thajortal, Obrony 160b (BM, E, G, GH, K, W); Burberg bei Kaaden, Stelzharmer 760 (BM, E, G, GH, K, W); oo iin ad Mikulov (Nikolsburg), 450 m., Sirjaev 219 (BM, GH, K, W); m, Oborny (BM, E, G, GH, K, W); Valle Radstnske: urb. Praha, May 1938, ie ee E, K, GH, W); Bohemia, Stekhemak 160 (£, Ww). Huncary: Buda, May-June 1876, Richter (£, w); Budors pr. Ofen, Richter 11 (BM, GH, K, Ww); Esckerberg, Tauscher s.n. (£, w); Mt. St. Gerade, Schiffer (cE, w); Comit. Zala, mont. Szent- Gyorgyhegy ad Kisapati, 250 m., Kiimmerle 189 (BM, E, G, GH, K, W); Comit. Pest., mont. Tordku-grato, pr. Buda6érs, 200 m., 6 June 1913, Jévorka (BM, GH, K, w); Buda, Ball (£, GH). Yucostavia: Macedonia, Sor. Planina, Popova Sapka, 30 miles west of Skopje, 210 m., Stainton 7962 (Kk, w); Serbia Suva Planina, pr. Babuschnitza, Rechinger 16036 (w); Djep, Jlié (£, w). BULGARIA: 394 JOURNAL OF THE ARNOLD ARBORETUM [| VOL. XLV Noé 4280 (£, K); Rhodope hills, gli ? (BM, K). ROMANIA: Matschan, Do- brudscha, 1873, Sintenis 576 (G, kK, ; Guravoie, 15 July 1895, Degen os E, K, G, W); Transsilvania, Koppandi-! ie pr. Torda, Wolff (8m, F£, G, G w). Russta: Konuva, 28 Mar. 1903 (£); Borythreum (Borysthenis), July i Steven (BM, G-DC); Podolia, 1820, Andrzejowski (G-vc). Fritsch, in Kerner (1902), who pointed out that Linnaeus (Species Plantarum, 1753) cited polynomials from Royen, and from Tournefort for Alyssum saxatile (the basionym of Aurinia saxatilis), correctly maintains that the polynomial from Tournefort, Corollarium Institutionum, 15. 1703, which reads: ‘“‘Alyssum creticum saxatile, foliis undulatis incanis” is syn- onymous with Alyssum orientale Ard. (subsp. orientalis (Ard.) Dudl.). Furthermore, because of Linnaeus’s discordant references, Fritsch applied a nomen novum, Alyssum arduinii, to refer to that part of Alvssum saxatile which was based on the polynomial from Royen, Florae Leydensis, 33. (1740) reading: ‘“‘Alyssum caulibus frutescentibus paniculatis, foliis lan- ceolatis undulatis integris.”” Examination of Royen’s original specimen in the Leiden herbarium proves without any doubt that it is subsp. saxatilis, and not subsp. ortentalis. The three well developed subspecies which are recognized in Aurinia saxatilis, subspecies saxatilis, orientalis, and megalocarpa, are morphologi- cally distinct, and have characteristic distributions. Subspecies saxatilis, the northern and most widespread in the species, is almost entirely re- placed in the Balkans and Aegean region by subsp. orientalis. Whereas subsp. orientalis is only partially sympatric in the northern Balkans with subsp. saxatilis, subsp. megalocarpa is completely sympatric, in terms of geography, with subsp. orientalis. Subsp. orientalis (Ard.) Dudley, comb. nov. Alyssum een ne Animad. Bot. Spec. Alt. 2: 32. ¢. 15, fig. 1. 1764 (!). pe, “ab Orie Bickel (PAD, non vidi).* — Boiss. Fl. Or. 1: 266. 1867. — Velen. FI. ci 37, 1891. Clypeola bpaanieke L. Mantissa, 92. 1767 (!). Holotype, “ab Oriente,” Arduino (LINN no, 828:3). Alyssum affine Ten. Syll. Fl. Neap. 315. 1831 (!). Holotype, Italy. In saxosis Japygiae: Lecce, Martina, 1819, Tenore 23716 (NAP, non vidi); isotypes (G-pc; K, no. 303 sent in 1824, and one sheet ex herb. Gay; GH, sent in 1846). ) “Seeds from the Orient, which were given to Arduino by Dr. Leonardo Seslerio were grown in the Padova Botanic Garden. Arduino states that the original descrip- herbarium, grown from these seeds. Evidently there were more than two specimens soarsd because in 1761 Arduino sent a specimen to Linnaeus. This specimen no. 8 n the Linnaean herbarium (LINN) was annotated-by Linnaeus ‘“Ard.”, indicating the source of the specimen. This Arduino duplicate, and the description and figure of Alyssum orientale, served as the basis of Clypeola tomentosa L. (1767), a synonym of Alyssum orientale, and accordingly of Aurinia saxatilis subsp. orientalis. There may be additional specimens, cultivated from the seeds given to Arduino by Seslerio, in the herbaria of the Botanical Museum in Copenhagen and the University of Firenze 1964 | DUDLEY, THE GENUS AURINIA 395 “4urinia orientalis (Ard.) Griseb. Spic. Fl. Rum. & Bith. 1: 272. 1843 (!). Alyssum orientale var. majus Hausskn. Mitt. Thir. Bot. Ver. 3-4: 112. 1893 (!). Holotype, Greece. In decliv. nemorosis infra Mon. Korona, Pindus Dolopicus, 1067-1097 m., 20-28 June 1885, Heldreich (ye, non vidi); isotypes (BM, E, G, K). Alyssum saxatile var. alpinum Hal. Denkschr. Akad. Wien Math. Naturw. 61: 496. 1894 (!). Holotype, Greece. In der oberen Region der Kyllene b Gura, Haldcsy (w). Alyssum denticulatum Form. Verh. Natur. Ver. Brinn 32: 177. lavia. In saxosis ad urbem Ochrida in oe Formdnek (BRNU, non vidi). — Vandas, Reliq. Formanek. 35. Alyssum saxatile var. majus (Hausskn.) Hal. Consp. os 1: 91. 1900 (!). pees oS var. albidum Tuzs. Bot. Kozlem. 8: 266. 1909 (!). Lecto- type, e. Cap. Maleca, Sieber (BM, E, GH, K, W Alyssum nil var. orientale (Ard.) Beck, Glasn. eu. Muz. Bosn. Herceg. 8: 129. 6 (!). — Hayek, Prod. Fl. Pen. Balc. 1: 430. 1925. Alyssum oo var. orientale f. maius (Hausskn.) Hayek, loc. cit. (!). Alyssum saxatile subsp. orientale (Ard.) Rech. f. Ann, Naturh. Mus. Wien 43: rem ent f. humilis Vierh. Oesterr. Bot. Zeitschr. 84: 1935 (!). Crete. Hagios Vasilis, alpine Stufe des Kedros, ‘o oe ok irfler 56 61 (WU); isotypes (BM, G, K). Alyssum saxatile auct. Levanticum — Sibth. & Smith, Ic. Fl. Gr. 7: 21. ¢. 624. 1830. — Hal. Consp. Fl. Gr. 1: 90. 1900.— Bornm. Mitt. Thur. Bot. Ver. 24: 11. fig. 2. 1909. DISTRIBUTION AND HABITAT: Saxatile on calcareous substrates in Cala- bria in southern Italy, the Balkans, the Aegean Islands, Crete, Turkey-in- Europe, and the western coast of Anatolia; alt. 50-1500 (—2286) m. Fl. Apr.—May. Iraty: Calabria, Catanzaro, near Tiriolo, Rigo 312 (A, K, w); Mt. al near Castrovillari & Mt. Pollina, 400-800 m., Huter, Porta & Rive 235 (BM, w); San Pietro in Devafane, May 1874, Groves (BM, K); near aren Lacaita 104/24 (Bm); 1820, Dierville (G-pc). Italia meridionalem, 1819, Moratti (G-pc). Yucostavia: Macedonia, Ohrid, Rechinger 15927 (kK, w); 20 miles south of Skopje, Katlanova, 305 m., Stainton 7819 (Kk). ALBANIA: Bertiscus, Pecska Bistrica near Pet (Ipek), 600-750 m., Rechinger & Scheffer 124 (kK, w); Sant. Graminta, 1897, Baldacci (pM). CorFu: San Deca, 7 May 1878, Spreitzenhofer (GH, K); pass to Paleocastuzza, Apr. 1855 (kK); H.T. (x); ex hb. Montbret (w); pass. Panteleimon in Skripero, 126-317 m., 10 May 1896, Baenitz (&, K. W). Butcarta: Montes Stara Planina, fluv. Isker near Lakatnik, Rechinger 1741 (w); in sax. ad Sveti Dimitri near Basarbova, Rechinger 569 (w); Dermende, May 1890, Pichler (Gc, w); Kricin, May 1895, Stribrny (£, K, w); May 1902, Stribrny (£, K, W); Kritchma, 27 June 1896, Stribrny (£, K, W); Tekir, 9 June 1895, St*ibrny (e, K, W); Eli Dere, May 1907, St?ibrny (E, K, W). Kleine eee Ruine Tallenschein bei Theben-Neudorf, June ROMANIA: 1902, Schneider (A, w); Danubii, inter Verciorova et Guravoie, et inter Svintza et Drenkova, Degen : 68 (BM, E, GH, K, W). CRETE: Tournefort (BM); Kis- samos, Polyrrhenen, Gandoger 8736 (BM, K); Cap Makea, Canée, Gundoger 82101 (BM, K). GREECE: Sibthorp 1057 (oxF); Attica, Mt. Parnes, Ilatoci, 396 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV 500 m., Samuelsson & Zander 120 & 476 (£, GH, K); 1000 m., Rechinger 5856 (w); Heldreich 2693 (c, w); Athens— ha Patten 196 (GH); Acropolis, Athens, 23 Mar. 1900, Patten D (GH); Spruner (G, GH, K, W); Mar. 1876, Pichler (G, GH, K, W); 1889, Sintenis 15 ae Feb. 1847, Heldreich (re, w); 1843-1844, Heldreich (£); Heldreich 412 (BM, kK); Athens, Lycabetho, Mar. 1862, Heldreich (G, GH, K); 60-457 m., Orphanidis 276 (BM, E, G, GH, K, W); Schrenk 27 (w); Davis 93 (BM, E); Heldreich 1112 (kK); Mar. 1842, Boissier (G, GH, K, w); Ardetto near Athens, Heldreich 1112b (pm, K); Thena ad Acropolis, Oct. 1867, Ball (£); Hieropolis, Spruner (kK); Attica, Guiol 915 (sm); Mt. Hymetto, Mar—Apr. 1847, Heldreich (£, w); Chelmos, 228 m., Davis 990 (£); Lemperg 1088 (K); Phocis near Delphi, Leonis 383 (a, E, G); Turko- vinis near Athens, 7 Apr. 1927, Rechinger (w); Kymelti versus Tankrati, Rechinger 482 (w); near Kalanistra, 500 m., 11 June 1893, Haldcsy (w); Thessalia, Agris prope Volo [Vélos] (Tersana), 1896, Sintenis 70 & 71 (BM, GH, K, Ww); Volos, Rechinger 22609 (w); Makiniata near Volo, Polunin 6552 (x); in valle Tempe fluv. Peneios, prope Rapsani, Rechinger 22572 (w); Kasabaliotiko pr. Kasabali, Rechinger 1190 (w); Kastri pr. lacus Karla, Rechinger 22842 (w); Pelion pr. Volos, 800-1500 m., Rechinger 22718 (w); Macedonia, Gjefgjeli, Rechinger 1525 (w); fluv. Angista pr. Phorolivos opp. Drama, Rechinger 6171 (kK, w); Salonika, Lembert hills, 1 May 1917, Wilson (e); Epirus, dist. Zagorion, fluv. Voidomatis, 17 km. versus Joannina, Rechinger 21428 (w); Mt. Tymphi, fluv. Vikos, 300-800 m., Rechinger 21149 (w); inter Vriskos et Joannina, Rechinger 20719 (w); pr. Klissoura, 47 km. south of Joannina, 400 m., Rechinger 23327 (w); 20 km. south of Joannina, Rechinger 23263 (W); Euboea, Aucher 239 (BM, K, OXF); pr. Steni, 350-500 m., Rechinger 1921 (K, W); in montis Xiron Oros pr. Hagia Anna, 750-950 m., Rechinger 17111 (w); Psachna et Achmet Aga (Prokopion) a Hagios, 300-500 m., Rechinger 16500 (Kk, w); Mt. Kandili, 10 km. ab Achmet Aga, 300 m., Rechinger 18166 (w); a Limni, Rechinger 16686 (w); Thrace, Derbend, 10 km. north of Alexan- driopolis, Rechinger 22258 (w); Lignitorichion et Potamos pr. Alexandriopolis, Rechinger 6048 (K, w); north of Arvas, 200 m., Ball 522 (Livu); Mt. Karlik- Dagh pr. Komotini (Gumuldschina), 1400 m., Rechinger 10497 (Kk, w); fluv. Nestos (Mesta), pr. Toxotai Spratt 60-100 m., Rechinger 9649, 9349, 9812 (w); Rechinger 10497, 10966 (w); Montes Rhodope supra pr. Iasmos (Jasi-Koj) Rechinger 9525 (w): Mesta valley, Tedd 547 (xk); above Yassikeuy, 366 m., Tedd 1805 (kK); Mt. Xanthi, 120 m., Tedd 2196 (kK); near Heraclitera, 90 m., Tedd 1964 (xk); Kizil Ada, 762 m., Tedd 1956 (xk); Peloponnesus, Laconia, Pen. Mani, supra Porto Kalion, Rechinger 20134 (w); Korinthia, lacum Stymphalia, Rechinger 10534 (w); Akropkari, Rechinger 573 (w); Messenia, Verga ad Kalamai (Kalamata) ad Mt. Taygetos, Rechinger 20400 (w); 950 m., Bornmiiller 119 (w). AEGEAN ISLANDS: Samothrake, Mt. Phen- gari, 500-1400 m., Rechinger 9812 (k, w); 28 June 1934, Tedd (Kk); Thasos, (BM, E, K); Bl; Prov. Canakkale, Giaow Hissar, valley of Rhodius, Apr. 1856, Kirk (£); Seidinlii in mont. Arablar-Depressi, 1883, Sintenis 226 (pm); Prov. Izmir, cei Sedcheni, 27 Apr. 1870, Peyronin (Gc); Koukoularoudja, Balansa (BM, E, G, K, W); W hittall ki (Ee); Mt. Sipylo (Manisa dag), 250 m., 1906, Bornmiller 9084 (BM, E, G, w); C2, Prov. Denizli, Denizli- -Acipayam, 25 m. from Denizli, 860 m. ” Dudley, D. 35340 (A, E, K). 1564 | DUDLEY, THE GENUS AURINIA 397 Specimens intermediate between subsp. saxatilis and subsp. orientalis. Yuco- sLAvIA: Macedonia, lacus Ohrid pr. Pistani, 17 km. a Sveti Naum, Rechinger 19717 (w); fluv. Radika inter Debar et aves: Hanovi, oo 15795 (w). Butcarta: Macedonia, Tartorly, 25 May 1917, Nixolaff (cH, w) Specimens intermediate in morphology between subsp. saxatilis and subsp. orientalis may be found in the area (primarily in the northern Balkans) where these two subspecies overlap. The occurrence of inter- mediates and the southern distribution of subsp. orientalis, replacing subsp. saxatilis, indicate that subspecific rank is the most appropriate. The density of indumentum, the degree of leaf lobation or dentation, and the stature of the plants of subsp. orientalis show considerable variation. Often the indumentum is denser, the leaf teeth fewer and less pronounced, and the plants dwarf at altitudes above 1500 meters. This does not, how- ever, appear to be a constant pattern. Recognition of these variations has accounted for the varieties listed in the synonymy. Subsp. megalocarpa (Hausskn.) Dudley, comb. nov. Alyssum orientalis var. megalocarpum Hausskn. Mitt. Thir. Bot. Ver. 3-4: 112. 1893. Holotype, the Aegean island of Chios. Bl: In saxosis ins. Chios. 1853, Pauli (JE, non vidi). — Bornm. Engl. Bot. Jahrb. 59: 449, 1925 Alyssum ephesium Bornm. Mitt. Thiir. Bot. Ver. 24: 11. fig. 1. 1909 (!). Holotype, Turkey, C1: Ad Ephesi ad moenia dirupta, 1 June 1906, Born- miller 9083 (B, non vidi) ; isotype (OXF). Alyssum saxatile subsp. megalocarpum (Hausskn.) eer f. Beih. Bot. Centr. 54(B): 611. 1936 (!).— Rech. f. Fl. Aegaea, 223. DISTRIBUTION AND HABITAT: Saxatile on calcareous substrates in the Cyclades, Sporades, western Aegean Islands, and the western coast of Anatolia; alt. 150-610(—1219) m. Fl. Apr.—May. GREECE: Peloponnesus, Laconia, Pen. Malea, Monemvasia, Rechinger 20093 (kK, w). AEGEAN IsLanps: B1, Chios, Scio, 1822, Oliver (G-pc) ; near Vrontados, Gathorne-Hardy 474 (£); Tal, valley of Livadi, 500 m., Rechinger 5311 (f, kK, w); Lesbos, Malea near Philia, Candargy (fide Rech. 1943); C1, Kos, Mt. Dikios Asphendiu near Pyli, Forsyth-Major 642 (A, BM, GH, K, W); 800 m., Rechinger 8023 (£, K, W); inter Pili & Kardamena, Rechinger 7981 (kK, w); C1, Samos, Mt. Kakis, 610-1219 m., Davis 1672 (BM, E, K); Mt. Ambelos, 700 m., Rechinger 3903 (£, G, K, Ww); Mt. Kerki, 1200 m., Rechinger 1966 (k, kK, ; 1000 m., Rechinger 4125 (£, K, Ww); ad monasterium Zoodochus Rigi, 300 m., Rechinger 3715 (©, K, W); Amorgos, ad Langada, Mt. Krikelas, Rechinger 2352 (w); Davis 955 (K); 500 ft., Davis 1433 (A, BM, E, K). TuRKEY: Cl, Prov. Aydin, Ruins of Priene, 150 m., Dudley, D. 34972a (A, E); 20 miles from Soke, Davis 25231 (BM, E, K). Subspecies megalocarpa is distinguished from subsp. orientalis, with which it is sympatric, entirely by the quantitative characters listed in the key to subspecies. The area of distribution of subsp. megalocarpa is rel- atively limited in comparison to that of subsp. saxatilis or subsp. orien- talis, and is completely contained within the range of subsp. orientalis. 398 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV There appears to be little morphological intergradation between subsp. orientalis and subsp. megalocarpa. The fruits of some gatherings of subsp. orientalis (determined as Alyssum affine), however, from Calabria in Italy (e.g., Rigo 312) approach in size some of the smaller fruited specimens of subsp. megalocarpa from the Aegean Islands (e.g. Forsyth-Major 642). 2. Aurinia uechtritziana (Bornm.) Cullen & Dudley, comb. nov. Ptilotrichum (Koniga) uechtritzianum Bornm. Osterr. Bot. Zeitschr. 38: 10. 1888 olotype, Bulgaria. Ost.-Bulgarien, massenhaft in der Bucht von Varna zwischen Pontus und Devno-See, bei Galata, 1886, Bornmiiller (B, non vidi); isotypes (G, K, W). Lepidotrichum uechtritzianum (Bornm.) Velen. & Bornm. Osterr. Bot. Zeitschr. 39: 324. 1889 (!).— Bornm. Bot. Centr. 41: 163. 1890. — Velen. Fl. Bulg. 42. 1891; Suppl. Fl. Bulg. 1: 27. 1898.— Aznavour, Bull. Soc. Bot. Fr. 64: 165. 1897. — Schulz in Engler & Prantl, Nat. Pflanzenfam. II. 17b: 495. fig. 271, m-o. 1936.— Hayek, Prod. Fl. Pen. Balc. 1: 428. 1925. — Stoj. & Steff. Fl. Bulg. ed. 2. 532. pl., p. 585. 1948. DISTRIBUTION AND HABITAT: Maritime sands along the coast of the Black Sea in Bulgaria; one coastal record from Asiatic Turkey. Fl. July— Aug. (—Sept.). Butcaria: Burgas, Jablonowski (prc, fide Velen. 1898). Turkey: A2(A), Prov. Istanbul, Bournov (Jum Burnu near Kartal), Nemetz 178 (c, hb. B.V.D. Post ) The long, linear rosette leaves, with dentate or repand margins, of Aurinia uechtritziana are conferted on indurated caudices. The bases of these leaves are grooved, swollen, and fleshy during the growing season, and persistent for a number of years. In addition to these characters, Aw. uechtritziana shares with the other species of Aurinia spherical buds, cup- shaped calyces, reduced cauline leaves, and very sparsely foliate, nearly scapose fertile stems. The recognition and correlation of these characters for Au. uechtritziana, show that there are no apparent reasons, especially as there are no differences i in floral morphology, to maintain the monotypic genus Lepidotrichum distinct from Aurinia. Although the silicles of Au. saxatilis, Au. leucadaea, and Au. rupestris are compressed or rarely sub- inflated, the strongly inflated and turgid silicles of Au. wechtritziana are not unique in the genus. Awrinia corymbosa and Au. petraea also have turgid silicles whose valves are strongly and equally inflated. In the majority of species of Awrinia the petals are yellow; however, in Au. rupestris and Au. uechtritziana they are white. The only Turkish record of Au. uechtritziana to date was considered previously (cf. Velen. 1898; Hayek, 1925) as coming from Turkey-in- Europe (Thrace). In the original publication of this record, however, Aznavour (1897) clearly states that the plant was collected near Kartal, a village on the Marmara Sea on the Asiatic side of the Bosphorus. This collection, unfortunately, is only a portion of an inflorescence, but the 1964 | DUDLEY, THE GENUS AURINIA 399 floral and indumentum characters indicate that it may be correctly re- ferred to Au. uechtritziana. 3. Aurinia rupestris (Ten.) Cullen & Dudley, comb. nov. Subsp. rupestris Alyssum rupestre Ten. Prod. Fl. Nap. 1: 37. 1811 (!). Holotype, Italy. In ssuris rupium Magellae (Mugellae a Paieieeicng Tenore (NAP, non sa aks (c-pc).— Ten. Cat. Pl. Hort. Neap. 1: 57. aie: Fl. Nap. 2 t. 60. 1820; Syll. Fl. Neap. 316. 1831. — DC. Syst. Nat. 2: 319. 1821. epee ot Fl. Ital. 6: 498. 1844. — Fiori & Paol. Fl. Anal. Ital. 1: 458. 1898; Ic. Fl. Ital. 1: 166. fig. 1451, 1899. Koniga rupestris (Ten.) Heyn. Nom. Bot. Hort. 1: 439. 1840 (!). Lobularia rupestris (Ten.) Steud. Nom. Bot. ed. 2. 68. 1841 (!). Koniga scardica Griseb. Spic. Fl. Rum. & Bith. 1: 278. 1843 (!). Holotype, Albania. In regione alpina Scardi: sparsim in pratis m. Kobelitza, alt. 1524- 2134 m., Grisebach (GOET, non vidi) ; isotypes (G, K). Ptilotrichum dea (Ten.) Boiss. Fl. Or. 1: 288. 1867 (!).— Hal. Consp. Fl. Gr. 1: 87. — Hayek, Prod. Fl. Pen. Balc. 1: 442. 1925 Ptilotrichum ee var. scardica (Griseb.) Hal. Joc. cit. (!). DISTRIBUTION AND HABITAT: An alpine confined to calcareous screes in Italy, Yugoslavia, Albania and Greece; alt. (500—) 1850-2700 m. Iraty: Aprutii, Morrone, ge Gussone (G-pc); Morrone (fr.), monte Ma- jella loca la Cimaner 4 (8). 0-2000 m., July 1905, Rigo (BM, E, GH, w); Aprutio, Mont. pene ee Majella, 1524-1829 m., 30 July 1874, Peoria c& Rigo (BM, E, GH, K, w); Aprutii, mont. “La Majella” et “Morrone”, Rigo 3908 (BM, E, GH, K); monte ee 1819, Schouw (G—pc); Prov. de Chieti, monte Majella, inter mont. Amaro, 2700 m., Fiori 1281 (BM, E, GH, K); Mt. Majella, Boissier (BM, G, GH, K). ALBANIA: Lunxherie, 1800-1900 m., Lemperg 1406 (£, GH, K, W); meee 2100 m. (haufiger), 1850-2100 m. (selten), Lemperg 786 (E, GH, K, W); Nordost-Albanien, dist. Luma, Galica Lums, Dorfler 705 (A, BM, E, K, GH, W); Koritnik, 2200 m., Dérfler 605 (BM, E, kK, GH, W The features which indicate that Alyssum rupestre and Ptilotrichum cyclocarpum (subsp. cyclocarpa) should be treated as components of Aw- rinia, rather than of Alyssum (or Ptilotrichum which is a synonym of Alyssum) are the major diagnostics of Aurinia. These characters are: spherical buds; cup-shaped calyces; long and narrow rosette leaves, with deeply grooved petioles having swollen and persistent bases conferted and rosulate on indurated caudices; very reduced cauline leaves; and sparsely foliate, nearly scapose fertile stems. Awurinia rupestris differs from the other species of the genus in having simple racemes. Subsp. cyclocarpa (Boiss.) Cullen & Dudley, comb. & stat. nov. Ptilotrichum cyclocarpum Boiss. Ann. Sci. Nat. Paris, II. 17: 159. 1842 (!). Holotype, Turkey. B7: Prov. Tunceli, Kurdistan (Mounzur Dagh), Aucher 400 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLV 279 bis (G, non vidi) ; nea (kK). — Boiss. Fl. Or. 1: 288. 1867. — Grossh. Fl, Kavkava. ed. 2. 4: . 1950. DISTRIBUTION AND HABITAT: An alpine endemic of Turkey, from the Cilician Taurus, the Anti-Taurus, and the Armenian Highlands; dry lime- stone ridges and screes; alt. 2000-2800 m. Fl. May—Aug. Turkey. A7, Prov. Giimiigane, Karahissartasch (near Elias dag), 1894, Sin- tenis 5505 (BM, E, GH); Gimiisane, Bourgeau 33 (BM, E, GH, K, W); Ardas (Torut)—Besklise, Sintenis 92 (£, w). A8, Prov. Gilmiisane Erzurum, Bay- burt-Erzurum, 1500 m., Huet (£, GH, W). BS, Prov. Kayseri, Bakir dag, near Akoluk Yayla, 2500-2700 m., Davis 19346 (BM, E, K). B6, Prov. Kayseri, Pinarbasi, 2000 m., Stainton 5167 (©, K); Prov. Maras, Berit dag, 2800 m., Balls 1087 (BM, E, K); dist. Gdksun, Binboga dag, above Yalak, 2000 m., Davis 19977 (BM, E, K); 2200-2300 m., Davis 20155 (BM, E, K); Isik dag, above Karli Yayla, 2800 m., Davis 20021B (BM, E, K). B7, Prov. Tunceli, Munzur dag, above Ovacik, 2800 m., Davis 31241 (BM, E, K). C2, Prov. Burdur, Eldirek dag, south of Dirmil, chalk, 1850-1920 m., Huber-Morath 8451 (um). C5, Prov Icel/Nigde, Kysil depe above Bulghar Maaden, 2800 m., Siehe 457 (BM, E); Bulghar Magara, 2600 m., Siehe 573 (£, K); Bulghar Maaden, Balansa 418 (Ee, K); Gysil depe-Ketsiebele, 2580 m., Kotschy 126c (£, W). E, The only consistent difference between the Italian and Balkan subsp. rupestris, and the Anatolian subsp. cyclocarpa, is that the fruits of the latter, even in the immature state, are always glabrous. The fruits of subsp. rupestris have an indumentum, at least when immature. This con- stant difference, in conjunction with completely allopatric distributions, justifies the retention of the Turkish taxon at subspecific rank. A number of specimens from Albania and Greece link the two subspecies morphologi- cally. The mature fruits of these intermediate specimens (often de- termined as Koniga scardica or Ptilotrichum rupestre var. scardicum) are glabrous, as are the mature fruits of the Turkish subsp. cyclocarpa. Their immature fruits, however, have an indumentum characteristic of that which is persistent on the fruits of the strictly European subsp. rupestris. This pattern of variation could be interpreted as an interrupted cline. JOURNAL OF THE ARNOLD ARBORETUM VoL. XLV OcTOBER 1964 NUMBER 4 A RE-EVALUATION OF THE GENUS AMBROSIA (COMPOSITAE) ! WILLARD W. PAYNE IN SPITE OF THE FOCUs in recent years on the ragweeds and their rela- tives as sources of allergenic pollen, they are poorly understood biologically and taxonomically. Systematic controversy has centered particularly about the question of the possible congeneric nature of the so-called “true rag- weeds,” Ambrosia spp., and “false ragweeds,” Franseria spp. Linnaeus (1753) established the genus Ambrosia for modern use with the description of four species, Ambrosia trifida, A. elatior, A. artemisiifolia and A. mari- tima, all of which were based upon earlier descriptions. Cavanilles (1793) established the closely allied genus, Franseria, with the description of the single species F. ambrosioides. The similarity of the two genera was recognized by Cavanilles, as is shown by the specific epithet (ambrosioides ) he chose for his species and by a statement in the description that he con- sidered the new genus to be intermediate between the ragweeds (Ambrosia) *Publication number 38 on Atmospheric Pollution by Aeroallergens, Botanical Phase, under Research Grant Number AP-00008-02 from the Division of Air Pollution, Bureau of State Services, Public Health Service. S paper was presented, in part, before the joint meeting of the American Society of Plant Taxonomists and the Systematic Section of the Botanical Society of America at the 14th annual meeting of the American Institute of Biological Sciences, held at Amherst, Massachusetts, August 26, 1963. I view to express my appreciation to Dr. Warren H. Wagner, Jr., University of Michig Dr. Bernice G. Schubert, Harvard Gate. and Dr. ‘Otto T. Solbrig, Harvard Univers wits for assistance in the prepara- tion of the manuscript of this paper; to Dr. A. G. Norman, Director of The ee of Michigan Botanical Gardens, for providing facilities ‘for culture work; Miss Annetta Carter, University of California o Dr. Elizabeth Seon’ California Academy of Sciences (cas), and Dr. E. Schuyler, the Academy of Natural Sciences of Philadelphia (pH), for as sae in locating types. -I am also r t iz) = and the Gray Herbarium (cH) of Harvard ytpabeesiened he University of Arizona (ARIz), the University of Cincinnati (crnc), the University of Illinois (LL), Iowa State University (1sc), the University of Michigan (micu), the University of North atural Sciences (pH), the University of Arkansas (vaRK), and the United States National Museum (us) 402 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV and the cockleburs (Xanthium), “genus mediam inter Ambrosiam et Xanthiam” (p. 78). Since the time of Linnaeus and Cavanilles, as we have continued to explore and catalog the floras of the New World, the close affinity of these genera has become ever more apparent, and numerous instances of confusing intermediacy have been remarked be- tween them. This is exemplified by a statement made by T. S. Brandegee, in 1889, in his description of Franseria acuminata: “This is another of the forms which foo closely connects the genus with Ambrosia” (p. 171, italics mine). Recently Shinners (1949) reviewed the genus Franseria for central Texas, and stated, “This genus cannot be distinguished from Ambrosia except by mature fruit, and the difference . . . warrants the recognition of a subgenus at most” (p. 173). The work on ragweed biology undertaken at Michigan during the past several years (Gebben, Payne and Wagner, 1963; Payne, 1962a, 1962b, 1963; Payne, Raven and Kyhos, 1964) has convinced me that Shinners is correct and has provided further evidence to support the hypothesis that Ambrosia and Franseria are, in- deed, congeneric. It is the purpose of this paper to present the evidence and to make the necessary nomenclatural changes. The species of “ambrosioid” and “franserioid” ragweeds are wind pol- linated composites with highly specialized inflorescences (Payne, 1963). The floral heads are unisexual, and both staminate and pistillate heads are normally borne on the same plant. The involucral appendages (phyllaries) of the pistillate heads are fused about the enclosed pistillate florets or floret to form a hard, indehiscent conceptacle. The tips of the appendages are more or less evident as spiny processes which project from the surface of the involucre body at maturity. The number and arrangement of these spines provides the single criterion by means of which Ambrosia and Franseria have been distinguished. In Franseria, sensu stricto, the plants produce fruiting involucres that characteristically bear numerous spines scattered more or less uniformly over the surface, whereas fruiting in- volucres of Ambrosia, sensu stricto, are characterized as bearing few spines arranged in a single whorl below the involucral beak. This dif- ference is superficial, however, when the variation of all of the species in these complexes is considered. The spines of the pistillate involucres of species such as Ambrosia artemisiifolia, A. trifida, and A. psilostachya are arranged, in many instances, not in a single whorl, but in two, three, or more, few-spined whorls compacted near the apex of the fruiting involucre body. Intermediate involucral forms are encountered among specimens of a wide variety of franserioid species; fruiting involucres representative for both genera are commonly found among members of the Franseria con- fertiflora, F. acanthicarpa, and F. canescens complexes, and species such as F. bryantii and F. linearis can be placed almost equally well in either genus. Numerous “intergeneric” synonyms are found in the literature which re- flect this morphologic interdigitation, and the ambiguity and lack of con- stancy of the character state used to distinguish Ambrosia from Franseria constitutes, in itself, as noted by Shinners (op. cit.), a compelling argument for combination of the two. 1964 | PAYNE, THE GENUS AMBROSIA 403 Furthermore, I have produced an “intergeneric hybrid” in the greenhouse between the two very distinct species Ambrosia artemisiifolia (seed parent, m = 18) and Franseria acanthicarpa (pollen parent, n = 18). Morpho- logical features of the parental species and of the hybrid F;, are illustrated in TExtT-FIc. 1. Consideration of the hybrid series thus initiated lends ad- ditional support for combination. Chromosome pairing in the hybrid was observed to be very good; many configurations consisted of 18 bivalents at diakinesis, while others showed only one or two pairs of univalents (TEXT- FIG. 1D). Pollen viability, as estimated by staining in acetocarmine, was ap- proximately 30 per cent. Seeds, formed in abundance, were 40 per cent germinable and produced a vigorous Fs. At the time of writing, the series had been carried through the F; generation (pollen viability of individual specimens, 57 to 94.5 per cent; mean viability for 19 specimens, 81.5 per cent; seed viability for two selected seed lots, 63.3 per cent and 96 per cent), and Fy progeny were growing in the greenhouse. While the occurrence of intergeneric hybridization is not, in itself, of great taxonomic significance, the high degree of chromosomal compatibility and genome homology demonstrated by this hybrid series does not support separation of the parental species in different genera. Still further support for the union of Ambrosia and Franseria may be derived from a consideration of the evolutionary relationships of the species or species groups involved. Since the ragweeds, sensu lato, com- prise a large and morphologically diverse group, numerous characteristics may be analyzed to indicate relative species advancement and relation- ships. Progressions from primitive to advanced character states which have been worked out within the group include the following: BIT 1. From shrubs to perennial herbs and annuals. LEAVES 1. From petiolate to sessile. 2. From alternate to opposite 3. From pinnately lobed to palmately lobed or unlobed. 4. From heavily pubescent and somewhat coriaceous to less pubescent and membranaceous. STAMINATE CAPITULA 1. From stalked to sessile 2. From capitulous stalks which bear more than one head to one-headed stalks. oF ae centric attachment of the stalk to the involucre to eccentric attach- ent. 4. ate well defined involucral lobes (phyllary tips), separated by deep sinuses, to involucral lobes which are poorly defined or lackin . From large heads with many florets to small heads with few florets. rei CAPITULA . From several florets to a single floret per capitulum A From many and scattered (imbricated) spines ‘to few and localized spines. Text-FIc. 1. Ambrosia artemisiifolia * ‘“Franseria” acanthicarpa. Representa- tive leaves (subscript 1), staminate involucres (subscript 2) and fruiting in- volucres (subscript 3) for: A, F. reaping i pollen parent, ~ = 18; B, « artemistifolia, sie parent, n aap C, the hybrid (superscript numerals denote showing 4 univalents; D,, configuration showing 2 univalents. The line to the isi ial is equivalent to 10 cm. for leaf silhouettes and 10 mm. for involucre draw 1964] PAYNE, THE GENUS AMBROSIA 405 3. From flat and strongly emergent (phyllary like) spines to terete spines. Secondary modification has given rise to hooked and vestigial spines. Although the details of evolutionary pathways must still be worked out in many instances, the character progressions cited above, together with simple, nonprogressional similarities, can be used to assess relative advance- ment and to indicate infrageneric groupings and relationships. In general, species of Ambrosia, sensu stricto, appear to be more advanced than, and to have been derived from, species or species groups of Franseria, sensu Stricto. Rydberg (1922) recognizes 15 subgeneric groups among the species of Ambrosia and Franseria; however, his distinctions are often somewhat slight, serving better to distinguish small parcels of related species than to point out significant groups of evolutionarily related organisms. For the present purposes, therefore, I shall recognize three major subgeneric com- plexes, two of which include both franserioid and ambrosioid members, and a fourth containing only Ambrosia bidentata. These may be charac- terized as follows: 1. A central or core group, comprised of the majority of the species and characterized by usually quite pubescent (especially on the lower surface), pinnately to tripinnately lobed leaves. The group is, in itself, highly com- plex, with evolutionary lines apparently leading from the least specialized, shrubby, franserioid species, such as Franseria dumosa and F. ertocentra, along several derivative pathways to ambrosioid species as follows: To Ambrosia trifida along a line characterized by species with striated staminate involucres and nonhooked spines; the closest relatives of A. trifida probably include F. chamissonis and F. acanthicarpa. To A. cheiranthifolia through herbaceous species with less lobed, “bicolored” leaves, such as F. linearis. To A. bryantii via a shrubby line with much divided, irregularly lobed leaves, including such species as F. camphorata and F. magdalenae, predominant in Baja California, Mexico. To the South American species, A. polystachya, through the remarkably similar F. artemisioides Willd. 2. A small, derivative group, found in the desert regions of the south- western United States and western Mexico, made up of shrubby species which are characterized by unlobed (or nearly so) leaves, a heavily glandular indument, and fruiting involucres which closely resemble those of Xanthium, the cockleburs. This subgroup includes Franseria ambro- stoides, F. ilicifolia, F. chenopodiifolia and F. deltoidea. This complex displays no obvious, direct affinities with any of the ambrosioid species, although it is connected to franserioid members of group 1, above, and may be closely related to the Xanthium strumarium (sensu lato) complex. 3. A highly specialized, derivative group of perennial herbs and annuals which are characterized by membranaceous, regularly pinnately to tetra- pinnately lobed leaves and small staminate and pistillate involucres, the former without striations and the latter usually one-flowered and with re- duced spines. This complex includes Franseria canescens and F. conferti- 406 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV flora, which appear to be related to ambrosioid species as follows: F. canescens is probably closely related to the very similar A. pumila, the latter being smaller in stature and with reduced fruit spines. Ambrosia hispida, a species of the Caribbean strand, closely resembles A. conferti- bidentata artemisiifolia trifida psilostachya cumanensis hispida cheiranthifolia pumila polystachya bryantii XT-FIG ° Schem: es representation of the polyphyletic nature of Ambrosia, Sensu stricto. The stippled body represents the franserioid assemblage; blac lobes sidaniet ambrosiod species or species groups, as indicated, which have been derived from 1964 | PAYNE, THE GENUS AMBROSIA 407 flora, having more hispid leaves, reduced fruit spines and a slightly different habit. The large and highly integrated ambrosioid complex which includes A. artemisiifolia, A. cumanensis, A. peruviana and A. psilostachya, is probably connected to A. confertiflora along a different line from that which leads to A. hispida, through intermediate species such as A. velutina and A. tenuifolia Spreng. 4. A fourth group containing the single species A. bidentata, the rela- tionships of which are obscure, can be recognized. This is one of the most highly specialized of the ragweeds, with sessile, unlobed (but basally toothed) leaves, medium-sized, one-flowered pistillate involucres with few spines, and sessile staminate involucres with extreme prolongation of the distal lobe. It hybridizes easily with A. trifida and shares with it the annual habit, relatively large fruiting involucres and the condition of aneuploidy, but displays none of the morphological peculiarities charac- teristic of A. trifida. On the latter basis, it would appear to be more closely related to group 3, above, but its progenitor cannot be distinguished at present The relationships suggested in the foregoing paragraphs are summarized diagrammatically in Text-Fic. 2, in which the franserioid complex is repre- sented as a stippled, lobed mass and the derivative, ambrosioid complexes are shown as black, peripheral lobes. This arrangement is supported by the distributional pattern obtained by designating three artificial groups con- taining the apparently most primitive species, the intermediate species, and the most advanced species. The first group is located in the desert South- west, the second in less arid regions surrounding it, and the third extends to the northern and eastern periphery of the Ambrosia-Franseria range, providing a classic picture of the diversification and outward spread from a center of origin of ever more advanced members of a genus. It is evident that, if the suggested evolutionary lines are even approximately correct, Ambrosia, in the strict sense, is a highly polyphyletic group, its divisions involving as ancestral forms different members of Franseria, a situation which is unacceptable by present generic standards. In view of the above arguments in support of combining Ambrosia and Franseria, including the poorly defined and undependable criteria for separating them, an apparently higher degree of genome and chromosome homology than is ordinarily found between distinct genera, and the polyphyletic nature of Ambrosia, sensu stricto, they are combined below, and the necessary nomenclatural changes are made in the following pages. The treatment of the species and the synonymy presented do not constitute a monograph of the genus as newly defined but are intended only to clarify the nomenclature of the “franserioid’”’ members listed. Ambrosia Linnaeus, Sp. Pl. 2: 987. 1753; Gen. Pl. ed. 5. 425. 1754, non Ambrosia sensu Hort. ex Bedevian, Ill. Polyglot. Dict. 170. 1936 (Chenopodiaceae). Type species: A. maritima L. Sp. Pl. 2: 988. 1753. Gaertneria Medicus, Phil. Bot. 1: 45. 1789, non Gaertnera Schreb. 1789 (Mal- 408 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV phigiaceae), nec Gaertneria Neck. 1790 (Gentianaceae), nec Gaertnera Retz. 1791 (Campanulaceae), nec Gaertnera Lam. 1791 (Loganiaceae). Type species: Ambrosia fruticosa Medic. Hist. Comment. Acad. Elect., Theodoro-Palatinae 3: 244. 1775, nomen illeg., Ambrosia arborescens Mill. 1768 Franseria Cavanilles, Ic. Descr. Pl. 2: 78. 1793, nomen cons., Int. Code Bot. Nomencl., Reg. Veg. 23: 318. 1961. Type species: F. ambrosioides Cav. Ibid. Note: This name is erroneously equated in the Code (loc. cit.) with Ambrosia arborescens Mill. (1768), an error which may be based on the fact that A. arborescens Mill. is synonymous with Franseria artemisioides Willd. (1805) and there is, as a result, confusion between the epithets “arte- misioides” and “ambrosioides.” Hemiambrosia Delpino, Studi Lign. Anem. Comp. Artem. 16. 1871. Type species: a heterocephala Delp. Ibid., nomen illeg., Ambrosia pumila (Nutt.) Gray ea Delpino, /bid. 17. Type species: H. paradoxum Delp. Jbid., nomen illeg., Ambrosia acanthicarpa Hook. ; Xanthidium eager Ibid. 17. Type species: X. tenuifolium Delp. Ibid., , Ambrosia confertiflora DC. 1836. Sigua Rydberg, N. Am. Fl. 33: 22. 1922. Type species: A. bryanti (Curran) Rydb., nomen files. Franseria bryantii Curran, 1888. The genus Ambrosia, in the new sense, may be characterized briefly as follows: Shrubs or subshrubs to perennial herbs spreading by means of stem and (or) root proliferations, or strictly annual herbs. Entire plant glandular, aromatic with resinous odor, glandular trichomes (colleters) biseriate, sessile or long-stalked. Leaves diverse (often with considerable intraspecific variation), pinnately lobed or dissected, palmately lobed or unlobed, petiolate or sessile. Anemophilous; staminate and pistillate florets in distinct heads normally borne on the same plant. Staminate heads stalked to subsessile or sessile, borne in usually aphyllous, spicate or racemose clusters at the tips of stems and branches, the phyllaries laterally connate and forming a broadly funnel-shaped to cupulate or saucer-shaped involucre with the phyllary tips more or less prominent as marginal lobes; receptacle paleaceous, pales various; many to several. flowered. Staminate florets modified: calyx or pappus lacking; corolla hyaline, campanulate, 5-lobed (rarely 4-lobed), only comissural veins present; stamens 5, alternating with corolla lobes, filaments sometimes connivent, anthers without “tails,” weakly connivent, sometimes separating during pollen discharge, terminal appendages deltoid to long-attenuate; pistil reduced (pistillodium), ovary lacking, style short, truncate (stigmatic lobes lack- ing) with terminal, circular “brush” of hairs (penicillate). Pollen sub- globose, echinate, the spines short, 3-pored (rarely 2-6-pored) with bladder-like air chambers developed within the exine between the pores. Pistillate heads borne in sessile or stalked clusters in axils of leaves sub- tending staminate racemes or spikes; bases of phyllaries fused to form a hard, vase shaped conceptacle; phyllary tips forming more or less spiny processes, straight or uncinate, flattened or terete, strongly emergent or vestigial, rarely lacking, variously arranged or scattered over surface of 1964] PAYNE, THE GENUS AMBROSIA 409 fruiting involucre; without pales; florets few (6-7) to one, if more than one, interior of fruiting involucre compartmentalized, each floret in a separate chamber. Pistillate florets reduced: perianth lacking; androecium lacking; mature ovary obovate, base somewhat oblique, abruptly rounded to base of short style; stigmatic lobes elongate, linear, spreading, inner stigmatic surface minutely papillose, exserted through spine-like beaks at apex of involucre, several-flowered involucres usually with distinct beak for each stigma but occasionally with a communal beak. Basic chromosome number, x = 18, the majority of species diploid but outbreeding populations of fe raploids: hexaploids, and octoploids known, in some instances within what are here considered single species; two aneuploid derivatives known, A. bidentata (n = 17), A. trifida (n = 12). Predominantly New World, mostly North American, including species of arid and semi-arid habitats and weedy species of ruderal sites. Virtually all species exhibit extreme morphological variability. In numerous instances variants which appear to be within the natural limits of variation for given species have received separate names. Measurements cited in the species descriptions which follow are usually given as maxima found in non-teratological specimens from populations growing under favorable conditions. os . Ambrosia acanthicarpa Hook. Fl. Bor. Am. 1: 309. 1834. Type: Banks of the Saskatchewan and Red River; Douglas (probably in x but not seen). Note: Although I have net seen the type, Douglas’ description leaves no doubt of its identity. Among other characters Douglas refers to the bipinnatifid leaves and the “singularly spinescent fruits,” the fruit spines being long, rigid and subulate. No other species from the northwestern United States displays this combination of characteristics. Franseria montana Nutt. Trans. Am. Phil. Soc. ser. 2. 7: 345. 1840. Type: In ree Mountains, near the Colorado of the West, Nuttall (possibly in BM t seen] or PH!; isotype, GH! Franseria ie eee Nutt. Ibid. Type: Based upon Ambrosia acanthicarpa ok. 00 Hemixanthidium paradoxum Delp. Studi ea Anem. Comp. Artem, 17, 1871. Type: Based upon Franseria montana ; Gaertnera hookeriana (Nutt.) Kuntze, Rev. Gen. Pl. 1: 339. 1891. Franseria acanthicarpa (Hook.) Cov. Contr. U.S. Nat. Herb. 4: 129. 1893. Gaertnera acanthicarpa (Hook.) Britt. Mem. Torrey Club 5: 332. 1894. Franseria californica Gandoger, Bull. Soc. Bot. Fr. 65: 54. 1918, non Ambrosia californica Rydb. 1922. Type: California, ad San Bernardino, Parish, October, 1893 (us!). Franseria palmeri Rydb. N. Am. FI. 33: 25. 1922. Type: California, San Diego, E. Palmer, 1875 (us!). Erect, annual herb, to 15 dm. high, becoming much branched. Leaves opposite below, alternate above, petiolate, petioles to 6 cm. long, laminar tissue narrowly decurrent to base. Lamina extremely variable, broadly 410 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV deltoid to narrowly lanceolate; to ca. 8 cm. long and 6 cm. broad, pinnately to tripinnately lobed, lobes from narrowly linear to subovate and abruptly acute; both surfaces green, sparsely tomentulose to strigose or hispid- scabrous. Inflorescence racemose-spicate. Staminate involucres stalked, stalks to 15 mm. long, usually shorter, 3-12 mm. broad, 3—-9-lobed, lobes shallow to well-defined with sinuses separating lobes reaching halfway or more to point of stalk attachment, upper surface of three distal lobes usually marked with heavy, black striations along midveins. Fruiting involucres in sessile clusters below staminate racemes, to 10 mm. long and 14 mm. wide, 1-flowered; spines to 30, flattened, linear to lanceolate, sharply pointed, to 8 mm. long, scattered, occasionally vestigial’or lacking. Haploid chromosome number, ” = 18 Flowering: June to December. Range: Prairie and semi-arid regions of the western United States, from the Dakotas, Nebraska and Kansas, west to New Mexico, Arizona, Cali- fornia, Washington and Oregon. Adventive in eastern Minnesota. Notes: As here circumscribed, this is one of the most heteromorphic of the ragweeds. Most segregates have been delimited on the basis of leaf and fruit spine differences which do not appear at present to be sufficiently correlated to permit recognition of distinct taxa within the complex. 2. Ambrosia acuminata (Brandeg.) Payne, comb. nov. PLaTE I Franseria acuminata Brandeg. Proc. Calif. Acad. ser. 2. 2: 171. 1889. Type: Mexico, Lower California, Purisima [Purissima]. T. S. Brandegee, Feb- ruary 11, 1889 (uc, no. 88798; Ny, photograph!; isotype, US! no. 47364). Erect shrub, probably over 3 dm. high. Leaves alternate, petiolate, petiole to 18 mm., laminar tissue narrowly decurrent to base, usually with one or more pairs of small lobes below the main laminar region. Lamina lanceolate to ovate-deltoid in outline; to 8 cm. long, pinnately lobed, somewhat irregular, rachis of the lamina broad above first sinus, terminal lobe large, attenuated and comprising major portion of blade. Staminate involucre stalked, stalks to 5 mm. long, lowermost stalks some- times bearing more than one head, involucre to 3 mm. broad, cupulate; marginal lobes 3—7, shallow. Fruiting involucre to 6 mm. long and 3 mm. broad, 2-flowered; spines few (ca. 4-8), less than 1 mm. long, uncinate, appressed, scattered, Chromosome number unknown. Flowering: February. Range: Baja California, Mexico, known only from the type specimen. 3. Ambrosia ambrosioides (Cav.) Payne, comb. nov. PLATE [ Franseria ambrosioides Cav. Ic. Descr. Pl. 2: 79. 1793. Type: Not seen, prob- mA. The description of this species is based upon material culti- vated at the Royal Bot. Gard., Madrid, and represented by Plate 200, Cavs loc. cit. This plate is very well done and, in view of the unusual morphology of this species, leaves no question as to the identity of the taxon. 1964 ] PAYNE, THE GENUS AMBROSIA 411 Xanthidium ambrosioides (Cav.) Delp. Studi Lign. Anem. Comp. Artem. 18. 1871 Gaertnera ambrosiodes (Cav.) Kuntze, Rev. Gen. Pl. 1: 339. 1891. Erect shrub, to 25 dm. high. Leaves alternate, petiolate, petiole to 8 cm. long, usually without obvious decurrent tissue from the lamina. Lamina lanceolate-attenuate (rarely to ovate) 2-18 cm. long or longer, 1-6 cm. wide, unlobed or with marginal teeth developed as weak lobes, both surfaces bright green, tomentose, glandular, sticky when fresh with a strongly resinous odor. Inflorescence racemose, usually much branched below. Staminate involucres stalked, stalks to 14 mm., the lower often bearing mofe than one head; involucre to 8 mm. broad, lobes 7-12, teeth prominent, sinuses extending 1/3-1/2 way to point of stalk attachment, midveins of lobes darkened upon drying, occasionally marked with obscure striations on upper surface. Fruiting involucre to 15 mm. long and 12 mm. broad, florets 3-5; spines numerous, usually more than 30, scattered thickly over surface of involucre, slender, uncinate, 2-4 mm. long; mature involucres closely resembling those of X anthium, the cocklebur. Haploid chromosome number, » = 18 Flowering: February to May. Range: Sonoran Desert region, common through southern Arizona, extending into San Diego County, California, and to Sonora and Baja California, Mexico; rare in the Mexican states of Sinaloa and Coahuila. Notes: The type species of Franseria, as mentioned earlier, was efro- neously synonymized with 4. arborescens Mill. in the International Code of Botanical Nomenclature, Reg. Veg. 23: 318. 1961, probably through equation with F. artemisioides Willd. which is synonymous with 4. arborescens Mill. 4. Ambrosia arborescens Mill. Gard. Dict. ed. 8. 1768, non Franseria arborescens Brandeg. 1903. Lectotype: Garden specimen from Philip Miller collections (sm; photograph micu!, BH, negative no. 5274). PLATE IT Ambrosia fruticosa Medic. Hist. Comment. Acad. Elect., Theodoro-Palatinae 3: 1775, non Ambrosia fruticosa DC. 1836, nec Franseria fruticosa Phil. Xanthium fruticosum L.f. Suppl. Plant. Syst. Veg. 418. 1781. Type: Based upon A, arborescens Mill. Note: This name change was apparently made because of the somewhat Yanthium-like fruiting involucres of the species and the seeming impossibility of including a plant with such fruits in Ambrosia delimited at that time. Ambrosia frutescens Lam. Encyc. Meth. 1: 128. 1783, im synon. Franseria artemisioides Willd., Linn. Sp. Pl. ed. 4. 4: 378. 1805, non Ambrosia artemisioides Meyen & Walp. ex Meyen, 1843. Type: Based upon Xanthium fruticosum L.f. and upon Ambrosia arborescens Mill. as described by 412 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Lamarck (loc. cit.). Note: A tracing of material of this species from the Willdenow collections (B) was made by F. W. Klatt before 1898 and is now ai; Xanthidium artemisioides (Willd.) Delp. Studi Lign. Anem. Comp. Artem. 18. 71. Gaertnera artemisiodes (Willd.) Kuntze, Rev. Gen. Pl. 1: 339. 1891. Franseria conwayi Rusby, Bull. N.Y. Bot. Gard. 8: 130. 1912. Type: Bolivia, Santa Cruz, R. S. Williams 1464, Aug. 25, 1902 (ny!). Erect shrub, to 40 dm. high. Leaves alternate, petiolate, petioles to 10 cm. long, laminar tissue decurrent to base, several small lobes usually formed below main lamina. Lamina ovate-lanceolate to broadly ovate, deltoid or nearly orbicular in outline, to 20 cm. long and 24 cm. wide, bi- or tripinnately lobed, the second or third order lobes intergrading with the laciniate-serrate marginal teeth; upper surface green, puberulous to glabrate, lower surface gray-green, woolly-tomentose. Inflorescence panic- ulate-racemose, lower staminate stalks often with several heads. Staminate involucre stalked, stalk to 10 mm., often much shorter, involucre to 6 mm. broad, 5-9-lobed, sinuses between lobes often extending more than 4 way to point of stalk attachment; veins on upper surface of lobes darkened, sometimes with black striations near tips. Fruiting involucres in often massive, sessile clusters below the staminate heads, 5—8 mm. long, 4-7 mm. broad, body thick and heavy, often nearly black, (1) 2—4-flowered, stigmas discharged through a communal beak or through poorly defined distinct beaks; spines 12-20, scattered over surface, 1-3 mm. long, bases thick, abruptly narrowed to slender, uncinate tips, somewhat appressed. Chromo- some number unknown. Flowering: Throughout year. Range: Interandean plains of Ecuador, north to north-central Colombia, south through Peru to western Bolivia. Notes: A variable species in regard to habit, leaf size, and morphology. Franseria conwayi appears to be based upon a specimen with aberrant leaves. Ambrosia arborescens is apparently a plant well known to the inhabitants of the regions in which it occurs (colloquial name “Altimisa,” sometimes erroneously transliterated to “Artemisia”). Notes of W. H. Camp (specimens E-2498 and E-5030, Ny) contain the following remarks “ “it ig good to sweep out a house which has too many fleas’ . . . it seems to be the only thing used to sweep out ovens after heating and before the bread is put inside. It is sold on the Cuenca [Ecuador] market for this purpose . . . Used as a local application in skin infections.” This widespread use may explain arrival of the species in Europe early enough to be included in Miller’s Gardeners Dictionary. 5. Ambrosia artemisioides Meyen & Walp. ex Meyen, Nov. Act. Acad. Caes. Leopold. Carol. 19 (suppl. 1, Nat. Cur., Obs. Bot.) : 268. 1843, non Franseria artemisioides Willd. 1805, nec Xanthidium artemisioides (Willd.) Delp. 1871, nec Gaertnera artemisiodes (Willd.) Kuntze, 1891. Neotype: Peru, Tacna, Maisin, in glareosis joropi, Lechler 1964] PAYNE, THE GENUS AMBROSIA 413 1568, in 1854 (Ny!; duplicate in GH). Note: The Meyen and Walpers material upon which this taxon was based (Peru: in planitie circa Tacoram, alt. 14-17,000 ped.) was presumably destroyed in Berlin during World War II. Their description is, however, diagnostic “Plant shrubby, erect, pilose, branched; all leaves narrowly linearly pinnatifid, lobes obtuse, entire to pinnately incised, oe puberu- ent, ... rachis ... barely 1/2 a line wide, roun .. 3 fruit- ing irivoluere Spiny; spines uncinate; stem... cere glutinous.” The fact that Hohenacker evidently examined their material in Berlin and equated it with the Lechler specimen which is the type for the following synonym, provides a basis for acceptance of Meyen’s name for this species and for the selection of Lechler’s specimen as the neotype. PLATE I Franseria meyeniana Sch. Bip. ex Hohenacker, Bonplandia 4: 54. 1856. Type: Lechler 1568 (as above) (w, not seen; isotypes GH, NY!). Franseria Pitas Phil. Anal. Mus. Nac. Chi le, seg. secc. Botanica. 50. 1891, non Ambrosia fruticosa Medic. 1775, nec Noni fruticosum L.f. 1781, nec Ambrosia fruticosa DC. 1836. Type: Chile, Tarapaca, Philippi, 1888 (B, not seen; photograph, cH!). Gaertnera meyeniana (Sch. Bip.) Kuntze, Rev. Gen. Pl. 1: 339. 1891. Franseria recurva Rusby, Bull. N.Y. Bot. Gard. 8: 131. 1912. Type: Peru, Arequipa, 7500 ft., Rusby 2527, Aug. 8, 1901 (Ny!; isotype us! no. 1098696) Erect shrub, becoming much branched, new growth stem tending to be roe black, glutinous, somewhat pilose, bark of older stem gray, lacking pubescence. Leaves alternate, petiolate, petioles to 4 cm. long (usually much shorter) laminar tissue decurrent to base. Lamina ovate in outline to 4 cm. long and 3 cm. wide, irregularly bipinnately lobed, lobes linear; both surfaces green, sparingly puberulent to glabrate, minutely glandular. Inflorescence racemose-spicate. Staminate involucre stalked, stalks to 8 m. long, occasionally more than one head borne on lower stalks, in- volucre to 5 mm. broad, 7—12-lobed, lobes often marked with black stria- tions along midveins, sinuses separating lobes extending ca. 1/2 way to point of stalk attachment. Fruiting involucres in small axillary clusters below staminate capitula, 4-6 mm. long, 2-3 mm. wide, 1(?)-fruited: spines to 20, scattered, slender, uncinate, to 2.5 mm. long, easily broken from body. Chromosome number unknown. Flowering: March to August. Range: Southern Peru (Arequipa province) to northern Chile (Tarapaca province). Notes: I have seen only six specimens of this species, including the types, and these display considerable leaf variation, tending to merge with what seems to be an undescribed species of the same area with larger, more broadly lobed leaves, larger staminate heads and fruiting involucres of quite different morphology, viz. F. W. Pennell 13107 (Ny), Mr. & Mrs. F. E. Hinkley 79 (Ny). The spines of the fruiting involucre of A. artemi- 414 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV sioides appear to be easily broken from the body at an abscission zone, a characteristic unknown elsewhere in the genus. - Ambrosia bryantii (Curran) Payne, Madrofio 16: 234. 1962. Franseria bryantii Curran, Proc. Calif. Acad. ser. 2, 1: 232. 1888. Type: Mexico, Baja California, Magdalena Bay, Santa Margarita Island, Walter E, Bryant, in 1888 (cas, not seen; isotype, uc). Note: Although I have not examined the type material of this species, Dr. Elizabeth McClintock (cas) and Miss Annetta Carter (uc) have kindly verified its nature and location. Both of the specimens cited probably comprise the original holotype and they have been labelled respectively “part of the type” and “part of type” by Mrs. Curran who, at the time Franseria bryantii was published, was the editor of the Proceedings, and a member of the staff of the California Academy of Sciences. It is assumed that the material was divided and in- corporated into the T. S. Brandegee collections at the University of California after Mrs. Curran’s marriage to Brandegee. It is, therefore, appropriate to consider the cas material as the holotype Acanthambrosia bryantii (Curran) Rydb. N. Am. Fl. 33: 22. 1922. Erect shrub, reaching ca. 8 dm. Leaves alternate, petiolate, petioles to 3 cm. long, laminar tissue decurrent to base, usually forming several small lobes below main laminar region. Lamina ovate in outline, to 8 cm. long and 6 cm. broad, irregularly bi- to tripinnately lobed, lobes linear; both surfaces green, lower becoming lighter, nearly glabrous to sparsely hispid- ulous. Inflorescence racemose-spicate. Staminate involucres subsessile to stalked, stalks to 2 mm. long; involucre to 7 mm. broad, 8—12-lobed, lobes shallow, not noticeably striate. Body of fruiting involucre to 10 mm. long and 5 mm. broad, 2—7-flowered, the stigmas discharged through a common beak; spines 5—9, chalk-white, 12-40 mm. long, sharp, adi near apex of involucre body. Haploid sheomncenin number, ” = Flowering: February to April. Range: Central to southern Baja California, Mexico. Notes: A species remarkable for the long, thorny spines of the fruiting involucre, and for retention of a percentage of the fruiting involucres to function as armature through several growing seasons. Segregated in the monotypic genus Acanthambrosia by Rydberg because of the discharge of the stigmatic lobes of all contained florets through the single beak, a phenomenon also displayed by other species. 7. Ambrosia camphorata (Greene) Payne, comb. nov. PLATE IIT Franseria hispida Benth. in Hinds, Bot. Voy. Sulphur, 25. 1844, non Ambrosia 1 Pursh, 1814. Type: Mexico, Lower California, Magdalena Bay, herd ie not seen; ARIZ! photograph; cH! drawing of type by Smith, July Franseria Bana oem sensu Gray ex Wats. Proc. Am, Acad. 11: 115. 1876, non Franseria bipinnatifida Nutt. 1840, nec Ambrosia bipinnatifida (Nutt.) 1964 | PAYNE, THE GENUS AMBROSIA 415 Greene, 1894. Mexico, Baja California, Guadalupe Island, E. Palmer 40, in 1875 (ny!; ! sai ikta camphorata Greene, Bull. Calif. Acad. 4: 192. 1886. Type: Mexico, r California, Guadalupe Island, E. L. Greene, April 24, 1885 (cas, a seen; isotype, NY! Franseria camphorata var. leptophylla Gray, Proc. Am. Acad. 22: 309. 1887. ype: Mexico, Lower California, near San Fernando, C. R. Orcutt 1352, May 4, 1886 (us! . Gaertnera hispida (Benth.) Kuntze, Rev. Gen. Pl. 1: 339. 1891. Note: Kuntze credits the original name to ‘“Barcl.,”’ a misconception probably resulting from the fact that the type of F. haha Benth., collected by Barclay, is ed “Franseria hispida sp. n. Barclay Franseria leptophylla (Gray) Rydb. N. Am. “FL. 33: 32, 1922. Erect shrub, becoming much branched from the base; to 6 dm. tall; glandular throughout, viscid-sticky when fresh, with strongly resinous odor. Leaves alternate, petiolate, petioles to 10 cm. long, usually shorter, laminar tissue decurrent to base, usually with several medium to small pairs of lateral lobes below main laminar region. Lamina deltoid-lanceolate to ovate or lanceolate in outline, to 10 cm. long and broad, irregularly bi- or tripinnately lobed, lobes various; upper surface gray-green to green, minutely glandular, sparingly puberulent to glabrate, lower surface gray- green to white woolly-tomentose. Inflorescence racemose-spicate. Staminate involucres stalked, stalks to 5 mm. long, involucre to 7 mm. broad 7-12- lobed, upper surface of lobes often marked with obscure striations, sinuses separating lobes extending approximately halfway to point of stalk attach- ment. Fruiting involucres variable, 1—7-flowered, 5-10 mm. long, 3-11 mm. broad; spines 3-15 (rarely none), terete, variously arranged or scattered, strongly emergent to vestigial. Haploid chromosome number, N= 36; Flowering: September to May. Range: Throughout Baja California, sporadic in western Sonora and southern San Luis Potosi, Mexico. Notes: An extremely variable species, especially in regard to fruiting involucre morphology; possibly an aggregate or with several intergraded varieties, needing further study. 8. Ambrosia canescens Gray, Proc. Am. Acad. 17: 217. 1882. Type: Mexico, Aguas Calientes, In arvis communis, Hartweg 121, in 1839 (kx; photograph micH!, Kew negative number 715, 715 A; isotypes, GH, NY! PLATE IV Ambrosia fruticosa var. canescens Benth. Pl. Hartweg. 17. 1839, nomen nudum. Franseria canescens (Gray) Rydb. N. Am. Fl. 33: 27. 1922 Erect perennial herb, perennating by means of shoots from roots. Leaves alternate, petiolate, petioles 1-5 cm. long, laminar tissue decurrent to base, often with 1 or more reduced pairs of lobes below main laminar region. Lamina broadly ovate to lanceolate in outline, to 8 cm. long and 6 cm. wide, finely bi-, tri- or tetrapinnately lobed; both surfaces gray-canescent. 416 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV aime ga racemose-spicate. Staminate involucre stalked, stalks to 3 m. long, involucre 3-6 mm. in diameter, 5—10-lobed, the sinuses shallow. aida involucres 1-fruited, 3-6 mm. long. 2-5 mm. broad; spines 0-10, strongly emergent (to 2 Tnm. long) to vestigial, usually flattened, occa- sionally uncinate, variously scattered or grouped over surface. Haploid chromosome number, oe 18, Flowering: April to October. Range: East-central Mexico, from San Luis Potosi to Durango, south to Aguascalientes. Also reported from Chihuahua (C. G. Pringle 1284, GH, NY) and Arizona (H. H. Rusby 680, Ny). Notes: A species of considerable variability, appearing to intergrade somewhat with A. confertiflora in the eastern portion of its range. 9. Ambrosia carduacea (Greene) Payne, comb. nov. PLATE II Franseria arborescens Brandeg. Zoe 5: 162. 1903, non Ambrosia arborescens Mill. 1768. Type: Mexico, Baja California, Ascension, 7. S. Brandegee, aaa 18, 1889 (uc, no. 88767, not seen; type photograph, Nv!). Nore: ecimens commonly identified with this name ies carduacea Greene, Leaflets Bot. Obs. Crit. 2: 156. 1912. Type: Mexico, r California, Aquaje de Santana, 35 miles N. of San Ignacio, alt. 3400 ft., E. W. Nelson & E. A. Goldman 7187, October 4, 1905 (us! no. 565273; photograph, NyY!). Franseria sanctae-gertrudis Rydb. N. Am. Fl. 33: 35. 1922. Type: Mexico, Baja California, rocky slopes, Santa Gertrudis, alt. 6-700 ft., C. A. Purpus 116, January-March, 1898 (us! no. 383387; photograph, ny!). Franseria acerifolia M. E. Jones, Contrib. West. Bot. 18: 78. 1935. Syntypes: exico, Lower California, Arroyo Undo Ranch, Loreto, Marcus E. Jones 27764, October 26, 1930 (NY!); Cayuca Ranch, Loreto, M. E. Jones 27765, October 23, 1930 (us! no. 1589558). Erect shrub to 30 dm. high or higher. Leaves alternate, petiolate, petioles to 3 cm. long, without obvious decurrent laminar tissue. Lamina ovate-lanceolate to lanceolate in outline, nearly unlobed to pinnately lobed, leaf segments broad, marginal teeth often forming weak secondary lobes; upper surface green, strigose to hispid-scabrous, lower surface lighter. to densely tomentose. Inflorescence racemose-paniculate. Staminate involucre stalked, stalks to 10 mm., the lower often bearing more than one head, involucre to 5 mm. broad, 5—9-lobed, lobes with darkened veins on upper surface but rarely with obvious striations, sinuses separating lobes extend- ing halfway or more to point of stalk attachment. Fruiting involucres 6—10 mm. long, 3-5 mm. wide, 2-flowered; spines 6-12, somewhat flattened, usually uncinate, less than 3 mm. long, scattered over surface of body. Chromosome number unknown. Flowering: December to March, possibly throughout year. Range: Central to southern Baja California, Mexico. Notes: A species with considerable leaf variation, the synonymous taxa being distinguished primarily on this basis. Another complex requiring further study. 1964 | PAYNE, THE GENUS AMBROSIA 417 10. Ambrosia chamissonis (Less.) Greene, Man. Bot. Reg. San Fran- cisco Bay 188. 1894. PLATE V Franseria chamissonis Less. Linnaea 6: 507. 1831. Type: California, Cel. de Chamisso (probably in LE, not seen). Note: Plants so named by Lessing were collected on a Russian expedition (‘“Expeditione Romanzoffiana”’) to California in the 1820’s. Although I have not seen the types, Lessing’s detailed description leaves no question as to the identity of the plants in- cluded. Among the plants collected by Chamisso, Lessing recognized two varieties, var. malvaefolia and var. bipinnatisecta, characterized re- casei! by unlobed and bipinnately lobed leaves (one specimen was noted o have leaves of malvaefolia below and bipinnatisecta above). Much of ihe nomenclatural confusion, indicated by the numerous synonyms which follow, has resulted from attempts to distinguish these and other leaf forms within the complex Franseria palate Nutt. Trans. Am. Phil. Soc. ser. 2. 7: 344. 1840. Type: alifornia, Nuttall, in 1835 (2?) labeled with an unpublished name (PH!). It is Some that this is an isotype and that the holotype is at BM. NOTE: A sheet labeled ‘“‘Franseria bipinnatifida Nutt.” from the Nuttall collections in pH bears three specimens. Two of these, C. G. Pringle, April 27, 1882, and E. Palmer 40, 1875 (see synonymy of A. camphorata) were collected after Nuttall’s death and could not have figured in his concept. The third is labeled, by Nuttall, with an unpublished name oe the decumbent abit. However, it may be that this specimen was at hand when Nuttall described F. bipinnatifida, as the description states (loc cit.): “perennial, herbaceous, decumbent and diffusely branched.”’ Nuttall then equated the species with F. chamissonis 8 bipinnatisecta Less., and described the habit (which is unusual in this species) and miorphological features in such a way as to leave no doubt that he was, in fact, dealing with the bipinnatisect form of A. chamissonis as recognized here, and that he had seen and examined it in the field, along the Pacific Coast of California. Franseria cuneifolia Nutt. Ibid. 345. Type: Oregon (?), Columbia estuary, Nuttall, in 1835 (?) (possibly in Bm [not seen] or PH!). Franseria chamissonis var. cuneifolia (Nutt.) Torr. & Gray, Fl. N. Am. 2: 293. 1842. nee ea Meyen & Walp. ex sang Nov. Act. Acad. oe Leopold. Ca 9 (suppl. 1, Nat. Cur., Obs. Bot.): 267. 1843. Type: This is a oui nes name for the bipinnatifid form of F. Sees possibly based upon isotypic material for var. bipinnatisectis in the herbarium at Berlin, examined by Meyen and Walpers. Gaertnera chamissonis (Less.) Kuntze, Rev. Gen. Pl. 1: 339. 1891. Ambrosia bipinnatifida (Nutt.) Greene, Man. Bot. Reg. San Francisco Bay 187. : Franseria bipinnatifida var. dubia Eastw. Proc. Cal. Acad. ser. 3. 1: 117. 1898. Type: California, Ventura County, San Nicolas Island, Mrs. Blanche Trask, April, 1897 (cas, not seen). Note: Although I have not seen this specimen, the type material has been verified by Dr. Elizabeth McClintock. Material from the same collector and locality serves to typify the following synonym. Franseria chamissonis var. viscida Eastw. Ibid. Type: See preceding note. Gaertneria bipinnatifida dubia (Eastw.) Heller, Muhlenbergia 1: 6. 1900. Gaertneria chamissonis viscida (Eastw.) Heller, ibid. 418 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Franseria bipinnatifida var. insularis Reiche, Anal. Univ. Chile 112: 131. 1903. Type: Chile, Isla de la Mocha, en la playa del NE, Karl Reiche, in 1892. I have not located this type. Nore: Distinguished from typical F. bipinnati- da Nutt., according to Reiche, principally on the basis of pubescence differences. Reiche notes that the plant is “probablemente introducida.” Franseria bipinnatifida villosa Eastw. ex Rydb. N. Am. Fl. 33: 26. 1922, im synon., Franseria villosa (Eastw.) Rydb. [bid. Type: California, Mendocino County, near Mendocino, H. E. Brown 843, June, 1898 (Ny! isotype, Us! no. 343666). Franseria chamissonis subsp. typica Wigg. & Stockw. Madrono 4: 120. 1937. Type: Based on F. chamissonis var. malvaefolia Less. Franseria chamissonis subsp. bipinnatisecta Wigg. & Stockw. Ibid. Type: Based on F. chamissonis var. bipinnatisecta Less. Erect, decumbent or trailing subshrub, becoming very large since stems and leaves trap sand and continue growth by emergent branch tips. Leaves alternate, petiolate, petioles to 8 cm. long, usually shorter, laminar tissue decurrent to base. Lamina broadly ovate to lanceolate in outline, lobing extremely variable, from unlobed to tripinnately lobed; both sur- faces gray-green, silky canescent. Inflorescence racemose. Staminate in- volucres stalked, stalks to 5 mm. long; involucre 3-8 mm. broad, 7—15- lobed, the sinuses shallow, upper surface of lobes often marked with black striations along veins. Fruiting involucres clustered in often heavy masses below staminate racemes, 6—12 mm. long and 3—9 mm. broad; spines 9-20, to 3 mm. long, terete or flattened above, strongly emergent, sharply pointed, scattered over involucre surface. Haploid chromosome number, n=. 16. Flowering: Throughout year. ange: Along Pacific coast of North America from Vancouver Island, British Columbia to Baja California, Mexico. Adventive along South American coast. Notes: The variability of this species, especially in regard to leaf morphology, has led to considerable taxonomic confusion, as is evident from the synonymy. The basic problem is that two very different leaf forms are found, one unlobed and one pinnately dissected. As Eastwood (op. cit., p. 117) remarked concerning these, “They are most puzzling to the sys- tematist who endeavors to make boundary lines, because they not only appear to run into each other but each is variable even in regard to what are supposed to be its own individual characteristics. Especially this is true as regards foliage, pubescence and size of the heads.’’ Within what must be considered “normal” populations in every locality in which the species exists in abundance, one encounters both extremes as well as all intermediate forms, I have raised self-pollination progeny of the extremes and of intermediates in the greenhouse and it appears that any form is capable of giving rise to any other form. At the same time, however, there may be regional tendencies which need further investigation. Certainly the forces acting to maintain the observed variability, within what must be recarded as a stable and specialized site preference, are very subtle indeed. 1964 | PAYNE, THE GENUS AMBROSIA 419 11. Ambrosia chenopodiifolia (Benth.) Payne, comb. nov. PLATE IV Franseria chenopodiifolia Benth. in Hinds, Bot. Voy. Sulphur 26. 1844. Type: exico, Baja California, Bay of Magdalena, Barclay (kK, not seen; pho- tograph, GH). Gaertneria chenopodifolia Abrams, Bull. N.Y. Bot. Gard. 6: 461. 1910. Franseria lancifolia Rydb. N. Am. Fl. 33: 36. 1922. Type: Mexico, Baja Cali- fornia, Cedros Island, A. W. Anthony 290, March—June, 1897, (us! no. 313846; photograph, ny!). Erect shrub, becoming much branched at base, to 35 dm. high. Leaves alternate, petiolate, petioles to 6 cm. long, laminar tissue decurrent to base. Lamina ovate-lanceolate to broadly ovate in outline, to 10 cm. long and 3- nerved, tending to become trilobed, lobes weak, sinuses shallow, rarely somewhat pinnately lobed; both surfaces glandular and pubescent, upper gray-green, tomentose, lower becoming white-tomentose. Inflorescence racemose-spicate, becoming paniculate. Staminate involucre stalked, stalks to 8 mm. long, occasionally with more than one head borne on lower stalks; involucre to 6 mm. broad, 5—10-lobed, sinuses between lobes shallow and often irregular, lobes occasionally marked with brownish to black striations on upper surface. Fruiting involucres clustered on lower inflorescence branches, to 7 mm. long and broad, 2—3-fruited, the body usually thickly clothed with woolly white tomentum from which spines and beaks emerge; spines 12—25, slender, uncinate, to 3 mm. long, scattered over surface of involucre. Haploid chromosome number, ” = 36 Flowering: January to June Range: Baja California, Mexico, from Comondu north into San Diego Co., California. Notes: This species is very similar in aspect, habit, and leaf morphology to A. deltoidea, a species with which it undoubtedly has close affinity. Ordinarily the two can be distinguished at least on the basis of the characters of the fruiting involucre. A. chenopodiifolia is characterized by woolly tomentose involucre bodies with terete, uncinate spines and A. deltoidea is characterized by non-woolly involucres with flattened, straight spines (see PLATE IV). However, intermediate specimens ae oolly involucre bodies and straight spines or with uncinate spines and non- woolly bodies are frequent. This complex needs further study. 12. Ambrosia confertiflora DC. Prodr. 5: 526. 1836. Type: Mexico, Matamoros, Berlandier 2297, October, 1830 (G, not seen; photograph from Hb. Delessert, series 33702, GH, MICH! ). PLATE V aiid Ambrosia fruticosa DC. Ibid. 525, non Ambrosia fruticosa Medic. 1775, nec anthium fruticosum L.f. 1781, nec Franseria fruticosa Phil. 1891. Type: ese. Tamaulipas, circa Reynosa, Berlandier 2112, April, 1830 (G, not seen; photograph from Hb. Delessert, series 33700, GH, MICH!; isotypes, NY!, GH!). Franseria tenuifolia Harv. & Gray in Gray, Pl. Fendl. in Mem. Am. Acad. ser. 2. 4: 80. 1849, non Ambrosia tenuifolia Spreng. 1826. Type: California, 420 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Poni Creek, between Bent’s Fort and Sante Fé, Fendler 406 (not seen). Note: Although I have been unable to locate this type, the original descrip- tion applies to A. confertiflora as defined here. Also, see discussion for F. tenuifolia var. tripinnatifida below Franseria tenuifolia var. tripinnatifida Gray, Pl. Lindheim. 2. in Boston Jour. at. Hist. 6: 227. 1850. Type: Texas, Mountain prairies of the Liano [Llano], Lindheimer 640 (not seen). Note: I have been unable to locate this type, but the description fits A. confertiflora, and a specimen labeled F. tenuifolia var. tripinnatifida by Gray (New Mexico, C. Wright 1210, 1851. Ny!) is A. confertiflora. Also, isotypes of A. fruticosa DC. (Ber- landier 2112, GH!, Ny!) have been annotated by Gray, leaving little uestion as to the identity of this taxon. Xanthidium tenuifolium Delp. Studi. Lign. Anem. Comp. Artem. 17. 1871. ; d upon Franseria tenuifolia Harv. & Gray. Gaertnera tenuifolia (Gray) Kuntze. Rev. Gen. Pl. 1: 339. 1891. Franseria confertiflora (DC.) Rydb. N. Am. Fl. 33: 28. 1922. Franseria caudata Rydb. Ibid. 29. Type: New Mexico, — County, Gray, Josephine Skehan 108, September 1, 1898 (Ny!; isotypes, NY!, Us!). Franseria hispidissima Rydb. Ibid. Type: Mexico, pe Real de Pinos, n Diquet (NY!). eo Prasat incana Rydb, Ibid. 30. Type: Arizona, Ft. Huachuca, Dr. T. E. Wilcox, Sept. 1892 (Ny!; isotype, Ny!). Franseria strigulosa Rydb. Ibid. 28. Type: California, Los Angeles County, L. T. Chamberlain, Sept. (Ny!). Franseria pringlei Rydb. Ibid. Type: Mexico, Coahuila, Valley Jimulco, C. G. Pringle 192, May 17, 1885 (Ny!; isotype, us!). Ambrosia simulans Shinners, Field Lab. 17: 173. 1949. Type: A new name assigned to F. tenuifolia Harv. & Gray because of existence of earlier homonym, Ambrosia tenuifolia Spreng. Ambrosia caudata (Rydb.) Shinners, ibid. 174. Erect, perennial herb, perennating from base of stem or by means of adventitious shoots from roots, tending to form large clones; to 18 dm high. Leaves alternate, petiolate, petioles to 15 cm. long; laminar tissue decurrent, usually with one or more pairs of small lobes below the main laminar region. Lamina ovate-lanceolate to lanceolate in outline, pinnately to tetrapinnately lobed, the lobes lanceolate to linear, attenuated or abruptly terminated, blade to 16 cm, long and 15 cm. broad, usually smaller; both surfaces green to somewhat gray-green, glabrous or nearly Se) to densely strigose, becoming velutinous. Inflorescence racemose-spicate, often much branched, Staminate involucres stalked, stalks to 8 mm. long, involucre to 10 mm. broad, 5—9-lobed (or more), the lobes obscure and poorly defined or well defined with sinuses extending nearly halfway to point of stalk attachment. Fruiting involucres in sessile, frequently massive clusters below staminate racemes, to 5 mm. long and 4 mm. broad, often much smaller, the base tending to be elongate, leaving the spines arranged on the upper 2/3 of the body, 1 (rarely 2)-flowered; spines 0-20, often fewer than 10, short, 1 mm. or less long to 2 mm. lone and slender, uncinate. Haploid chromosome numbers, ” = 36, 54. Flowering: Throughout year in various ponens of range. 1964] PAYNE, THE GENUS AMBROSIA 421 Range: Southwestern United States from Texas and Colorado to Cali- fornia, throughout central Mexico from Nuevo Leon and Tamaulipas through Guanajuato and Jalisco to Sinaloa and Sonora. Adventive in Hawaii on the islands of Oahu and Molokai, and in Puerto Rico. Notes: An extremely variable species, without doubt the most complex and difficult of the ragweeds, possibly an aggregate species as defined here, needing considerable further study before useful subgroups can be recog- nized. 13. Ambrosia cordifolia (Gray) Payne, comb. nov. PLATE VI Franseria cordifolia Gray, Syn. Fl. N. Am. 1: 445. 1884. Type: Arizona, Sierra Tucson, C. G. Pringle, April 26, 1884 (GH!; isotypes, NY!, US!, MICH Gaertnera cordifolia (Gray) Kuntze, Rev. Gen. Pl. 1: 339. 1891. Franseria malvacea Rydb. N. Am. Fl. 33: 34. 1922. Type: Mexico, Sinaloa, Culiacan, Dr. Edward Palmer 1770, October 25 to November 18, 1891 (ny!; isotype, US!). Erect shrub, becoming much branched at base, to 13 dm. high. Leaves alternate, petiolate, petioles to 4.5 cm. long, laminar tissue very narrowly decurrent, rarely with a few minute lobes below blade. Lamina ovate- cordate to broadly lanceolate in outline, to 13 cm. long and 10 cm. broad, usually smaller, margin dentate-serrate, 3-veined, weakly trilobed be- coming obscurely pinnately lobed, base cordate to cuneate; both surfaces green, upper puberulous, lower lighter, tomentose. Inflorescence racemose- spicate, becoming paniculate. Staminate involucre stalked to subsessile, stalks to 4 mm. long; involucre to 7 mm. broad, shallowly 5—10-lobed, lobes without obvious striations, Fruiting involucres clustered on branches below staminate spikes, to 9 mm. long and 6 mm. broad, 2-fruited; spines 6-30, to 3 mm. long, scattered below the usually divergent beaks, some- times swollen below, slender above, uncinate. Chromosome number un- known. Flowering: November to May. Range: Southern Arizona, south and west into San Luis Potosi, Sonora, Sinaloa, and Baja California, Mexico. 14. Ambrosia deltoidea (Torr.) Payne, comb. nov. PLaTE IV Franseria deltoidea Torr. Pl. Fremont. in Smithson. Contr. Knowl. 6: 15. 1854. Type: Southern California, on the Gila River, Fremont in 1849 (ny!). (Label includes pencil drawings of floral structures — Torrey’s?). Sheet ears two specimens in addition to the type (which is at the extreme left): vaca V. 855 (central) and C. C. ae 549 (extreme right). The latter is a paratype (duplicate of paratype, Gaertnera Beads (Torr.) Kuntze, oe ee Pl. 1: 339. 1891. Erect shrub, to 5 dm. high, occasionally higher, usually much branched from base. Leaves alternate, petiolate, petiole to 2 cm. long, occasionally longer, laminar tissue narrowly decurrent to stem. Lamina narrowly lanceolate to deltoid-lanceolate or somewhat rhombic, abruptly expanded 422 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV from cuneate (to nearly straight) base, margin serrate or doubly serrate, unlobed or occasionally with few weak lobes; both surfaces gray-green to whitish-tomentose, strongly resinous and sticky. Inflorescence spicate- racemose to paniculate. Staminate involucre stalked to subsessile, stalks to 5 mm. long; involucres to 7 mm. broad, 6—9(—13)-lobed, lobes some- times weakly defined, sinuses separating lobes usually extending less than halfway to point of stalk attachment, obscurely striated on upper surface or without striations, Fruiting involucres clustered on axis below staminate spike or on lateral branches, involucres to 10 mm. long and 12 mm. broad, 2 (rarely more)-flowered; spines to 18 (occasionally more), often the lowermost in a well defined whorl resembling the phyllaries of less specialized involucres, flattened, very sharply pointed, rarely somewhat uncinate, to 5 mm. long. Haploid chromosome number, ” = 18. Flowering: February to July. ove Southern Arizona, extending into Sonora and Baja California, Mexico Notes: A species similar to, and with intermediate forms resembling, A. chenopodiifolia (see notes for A. chenopodiifolia). 15. Ambrosia divaricata (Brandeg.) Payne, Pap. Mich. Acad. 49: Franseria divaricata Brandeg. Proc. Calif. Acad. ser. 2. 2: 171. 1889. Type: Mexico, Lower California, San Gregorio, T. S. Brandegee, Feb. 6, 1889 (uc, no. 88805, not seen; photograph, NY!; isotype, us!, no. 47803). Erect, much branched shrub, probably more than 3 dm. high. Leaves alternate, petiolate, petioles to 15 mm. long, laminar tissue decurrent to base, occasionally with one or more pairs of minute lobes below main laminar region. Lamina broadly ovate, to 4 cm. long and 3.5 cm. broad, 3-veined, shallowly 3—7-lobed, the lobes broad; green both surfaces, some- what lighter to gray-green below, strigose, becoming tomentose. Inflor- escence racemose-paniculate, the lower staminate stalks often bearing more than one head. Staminate heads stalked, stalks to 2 cm. long (or longer) ; involucres to 7 mm. broad, 7—11-lobed, the lobes well defined, upper sur- faces sometimes with weak striations. Fruiting involucres clustered on stalks below staminate racemes, to 9 mm. long and broad, 2-flowered: spines to more than 30, scattered, slender, uncinate, to 3 mm. long. Chromosome number unknown. Flowering: January to April. Range: Central Baja California, Mexico. Notes: A species of great similarity to A. magdalenae, being distinguished primarily by the differences in leaf morphology. 16. Ambrosia dumosa (Gray) Payne, comb. nov. PLATE VI Franseria dumosa Gray in Torr. & Frem. in Fremont, Rept. Expl. Exped. Rocky Mtns. 316. 1845. Type: California, sandy uplands of the Mohave River, very common in all that region of California, Fremont 374, 1844 (ny!). 1964 | PAYNE, THE GENUS AMBROSIA 423 Franseria albicaulis Torr. Pl. Frem. in Smithson. Contr. Knowl. 6: 16. 1853. Type: California, Fremont’s Expedition, Fremont, 1849 (ny!). Franseria dumosa var. albicaulis (Torr.) Gray in Torrey, Bot. Mex. Bound. Gaertnera dumosa (Gray) Kuntze, Rev. Gen. Pl. 1: 339. 1891. Erect, much branched shrub, to 9 dm. high. Leaves alternate, petiolate, petioles to 20 mm. long, usually much shorter, laminar tissue narrowly decurrent to base. Lamina broadly ovate to narrowly elliptic-lanceolate, to 40 mm. long and 30 mm. broad, usually much smaller, pinnately to bi- or tripinnately lobed, lobes linear to distally expanded; both surfaces gray-green, puberulous, becoming white-tomentose on the under surface. Inflorescence racemose-paniculate, all stalks usually bearing distal stam- inate heads and one or more proximal pistillate heads subtended by small bracts; capitulous stalks to 8 mm. long. Staminate involucre 3-7 mm. broad, 5—9-lobed, sinuses separating lobes extending about halfway to point of stalk attachment, lobes without striations. Fruiting involucres to 9 mm. long and broad, 1- or 2-fruited; spines 12-30, scattered over surface of involucre, flattened, straight. Haploid chromosome number, ” = 18, 36, 54, 63 (72?). Flowering: February to July (fruiting specimens have been collected into September ). Range: Arid regions of Utah, Arizona, Nevada and California, south into Sonora and Baja California, Mexico. Notes: The organization of the inflorescence of this species is unique for Ambrosia. Also, it is the only species which spans all ploidal levels known for the genus, from diploid to octoploid. Several variants may be recognized in the field, and the species may be of greater complexity than is indicated here. 17. Ambrosia eriocentra (Gray) Payne, comb. nov. PLaTE VI Franseria eriocentra Gray, Proc. Am. Acad. 7: 355. 1868. Type: Arizona, East slope of Providence Mountain, J. G. Cooper, May, 1861 (cH!). Gaertnera eriocentra (Gray) Kuntze, Rev. Gen. Pl. 1: 339. 1891. Erect, much branched shrub, to 18 dm. high. Leaves alternate, petiolate to subsessile, petioles to 3 mm. long. Lamina elliptical or narrowly lance- olate to deltoid-lanceolate in outline, to 90 mm. long and 30 mm. broad, usually smaller, unlobed to marginally pinnately lobed, the rachis lamina usually comprising largest portion of blade; upper surface green, puber- ulous to tomentose, main veins frequently whitened by pubescence, under surface becoming white woolly-tomentose; margin usually strongly revo- lute. Inflorescence spicate, becoming racemose. Staminate involucre subsessile to stalked, stalks to 2 mm. long, lower stalks often with two or more closely approximate heads resembling a single large head; involucres to 7 mm. broad, 5—9-lobed, the sinuses between lobes shallow to more than halfway to point of stalk attachment, without obvious striations on upper surface. Fruiting involucres to 10 mm. long and 11 mm. broad, 1(-—2?)- 424 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV flowered; spines to 20, usually emerging near equator of fruit body, flat, sharply pointed, more or less covered with long, white, silky hairs. Hap- loid chromosome number, ” = 18. Flowering: April to July. Range: Southern Arizona, scattered in arid regions of Nevada, Utah, and California. 18. Ambrosia flexuosa (Gray) Payne, comb. nov. — flexuosa Gray, Proc. Am. Acad. ser. 2. 12 (whole series 20): 298. 1885. Mexico, Lower California, H. C. & C. R. Orcutt 1103, July 8, 1884 one type photograph, ny!). Erect shrub, similar in general aspect to A. carduacea. Leaves alternate, petiolate, petioles to 1 cm. long. Lamina narrowly lanceolate to deltoid- lanceolate in outline, to 5 cm. long and 3 cm. broad, cuneate at base, un- lobed to weakly pinnately lobed by prolongation of marginal teeth, teeth stiff; both surfaces green, under surface somewhat woolly before maturity, veins prominent on under surface, yellowish. Inflorescence racemose-panic- ulate. Staminate involucres stalked, stalks to 7 mm. long; involucres to 5 mm. broad, 5—10-lobed, lobes marked on the upper surface with dark stria- tions. Fruiting involucres in clusters on short or long stalks below stami- nate racemes, to 8 mm. long and 6 mm. broad, 2-flowered; spines to 15 (or more?), terete, uncinate, broadened at base. Chromosome number un- known. Flowering: July. Range: Baja California, Mexico, known only from type collection. Notes: A species of questionable nature because of the limited material. Gray noted that the foliage was somewhat intermediate in character between A. ambrosioides and A. ilicifolia, but the fruiting involucres are very different from both, more closely resembling those of A. cordifolia. The abundance with which fruiting involucres have been matured on this specimen would argue against a hybrid nature. 19. Ambrosia grayi (Nels.) Shinners, Field Lab. 17: 174. 1949. Franseria tomentosa Gray, Pl. Fendl. in Mem. Am. Acad. ser. 2. 4: 80. 1849, non A. tomentosa Nutt. 1818, nec Gaertneria tomentosa (Nutt.) Nels. 1909. Type: Kansas (?), high banks of Walnut Creek, between Council Grove and Fort Mann, of the Arkansas River, Fendler 412 (Gu, not seen). pambaie nits (Gray) Kuntze, Rev. Gen. Pl. 1: 339. 1891. Gaer ayi Nels. Bot. Gaz. 34: 35. 1902. Nore: Based upon Franseria Bibel Gray. New name proposed by Nelson because of earlier . Gaertneria tomentosa Nels. (= A. tomentosa Nutt Franseria grayi (Nels.) Nels. New Man. Bot. Centr. Rocky Mtns. 542. 1909, Upright, perennial herb, perennating by means of adventitious shoots from the roots, forming large clonal populations. Leaves alternate, petiolate, petioles to 5 cm. long, laminar tissue decurrent to base, with several small lobes below main laminar region. Lamina ovate- deltoid to lanceolate in 1964 | PAYNE, THE GENUS AMBROSIA 425 outline, to 10 cm. long and 8 cm. wide, irregularly pinnately lobed, major basal lobes and terminal lobe much the largest, large lobes lanceolate- elliptical, serrate, both surfaces (and entire plant) silvery gray-canescent. Inflorescence racemose-spicate. Staminate involucre stalked, stalks to 5 mm., involucres to 5 mm. broad, 5—9-lobed; lobes often somewhat irreg- ular, shallow, upper surface with dark striations along midveins obscured by pubescence. Fruiting involucres in small axillary clusters below staminate spikes, to 7 mm. long and 4 mm. broad, 2-flowered; spines to 15, scattered, narrowed to slender uncinate tips. Haploid chromosome ey f= 15. Flowering: July to October. Range: Low, moist areas throughout Kansas and Nebraska, sporadically in Oklahoma, Colorado, and Texas. Notes: A species well known under the name Franseria tomentosa; one of the more handsome ragweeds by virtue of the silvery pubescence. 20. Ambrosia ilicifolia (Gray) Payne, comb. nov. PLATE VII Franseria ilicifolia Gray, Proc. Am. Acad. 11: 77. 1876. Type: California, n part of San Diego County, Tantillas Cafion (cH!). Gaertnera ilicifolia (Gray) Kuntze, Rev. Gen. Pl. 1: 339. 1891. Erect or semi-prostrate shrub, to 12 dm. high. Leaves alternate, sessile, clasping the stem. Lamina broadly ovate, to 9 cm. long and 7.5 cm. broad, usually smaller, typically unlobed but with strong, prickly, marginal teeth, the teeth often prolonged into weak lobes; both surfaces gray-green, minutely glandular-punctate, with ciliate hairs on margins and along veins; stiff and coriaceous, veins whitened and prominent on under surface, sticky and with a strongly resinous odor when fresh. Inflorescence race- mose. Staminate involucres stalked, stalks to 12 mm.; involucres to 35 mm. broad including the usually prominent attenuate lobes, lobes to 15, without striations. Fruiting involucres in axillary clusters below staminate racemes, nearly globose, to 20 mm. long and broad; spines numerous, to 50 or more, scattered, to 6 mm. long, slender, uncinate. Haploid chromo- some nimber, = ie Flowering: heen to June. Range: Desert regions of southern Arizona and California, south into Sonora and Baja California, Mexico. 21. Ambrosia linearis (Rydb.) Payne, comb. nov. PiaTE VIII Gaertneria linearis Rydb. Bull. Torrey Club re 133. 1905. Type: Colorado, Calhan, DeAlton Saunders, July, 1893 (Ny Franseria linearis (Rydb.) Rydb. N. Am. FI. a 27. 1922. Erect, perennial, shrubby herb, perennating by means of adventitious shoots from the base of the stem and from roots. Leaves alternate, sessile. Lamina ovate to lanceolate in outline, to 25 mm. long and 15 mm. broad, pinnately or bipinnately lobed, lobes and rachis more or less linear, 426 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV margin revolute; upper surface green, strigose, lower surface white woolly- tomentose. Inflorescence racemose-spicate. Staminate involucres occasion- ally subtended by bracts, subsessile with stalks to ca. 1 mm., involucres to 5 mm. broad, 5—9-lobed, sinuses separating lobes usually extending less than halfway to point of stalk attachment; lobes without obvious stria- tions. Fruiting involucres borne singly or in small clusters in axils of leaves below staminate spike, ca. 3.5 mm. long and 2.5 mm. broad, 1- flowered; spines to 9, arranged more or less in a single whorl near apex of body, terete, uncinate, to ca. 1 mm. long. Chromosome number unknown. Flowering: July Range: Colorado, Lincoln and El Paso Counties, known only from the type and one other specimen (Ownbey 1325) collected in 1937 Notes: A very distinctive species, inexplicably rarely collected. 22. Ambrosia magdalenae (Brandeg.) Payne, comb. nov. Prater VIII Franseria magdalenae Brandeg. Proc. Calif. Acad. ser. 2. 2: 170. 1889. Type: Mexico, Lower California, Magdalena Island, T. S. hae os Jan. 13, 1889 (uc, no. 88804, not seen; isotype, us! no. 47365). Franseria intricata Rydb. N. Am. Fl. 33: 33. 1922. Type: Mexico, Lower California, San Bartolome Bay, J. N. Rose 16196, March 13, 1911 (Ny!; isotype us!, no. 638267). Erect, much branched shrub, to 2 dm. high. Leaves alternate, petiolate, petioles to 4 cm. long, laminar tissue narrowly decurrent to base. Lamina ovate to ovate-lanceolate in outline, to 5.5 cm. long and 3.5 cm. broad, irregularly bipinnately lobed, often with one or more pairs of minute lobes below the main laminar region; upper surface green, sparingly tomentose, lower similiar, to white woolly-tomentose. Inflorescence racemose to race- mose-paniculate. Staminate involucres stalked, stalks to 10 mm. long, the lowermost occasionally bearing more than one head; involucre to 5 mm. broad, lobes to 12, sometimes with weak striations along midvein, sinuses between lobes usually extending less than halfway to point of stalk attach- ment. Fruiting involucres in axillary clusters or clustered on branches below staminate raceme, subglobose, to 8 mm. long and 11 mm. broad, 2-flowered; spines to more than 30, scattered, slender, uncinate, to 3 mm. long. Chromosome number unknown. Flowering: January to April. Range: Central Baja California, Mexico. Notes: A species with considerable variability, particularly in leaf mor- phology and pubescence. Closely resembling A. divaricata, a taxon dis- tinguished primarily on the basis of leaf differences. 23. Ambrosia nivea (Rob. & Fern.) Payne, comb. nov. Puate VIII Franseria nivea Rob. & Fern. Proc. Am. Acad. 30: 117. 1895. Type: Mexico, Chihuahua, Plains near Casas Grandes, C. V. Hartman 813, 10 Oct. 1891, (GH!; isotypes, NY!, US!, no. 306295). 1964 | PAYNE, THE GENUS AMBROSIA 427 Erect shrub, to 2.5 dm. high. Leaves alternate, petiolate, petioles to 4 cm. long, ramen tissue decurrent to base, often with 1—3 or more pairs of small lobes below main laminar region. Lamina ovate to triangular in outline, to 80 mm. long and 70 mm. wide or larger, irregularly bi- to tripinnately lobed, both surfaces gray-green, silky-canescent. Inflorescence racemose-spicate. Staminate involucre stalked, stalks to 5 mm. long; involucres to 4 mm. broad, lobes to 10, the distal lobes tending to be somewhat more pronounced, sinuses separating lobes shallow, upper surfaces of lobes without striations. Fruiting involucres clustered in axils of leaves subtending staminate racemes, to 7 mm. long and 6 mm. broad, 2—3-flowered; spines to 20, scattered, to 3 mm. long, terete, heavy, straight. Chromosome ‘number dnknow wn. Flowering: October. Range: Northwestern Chihuahua (known only from type collection). Notes: A distinctive, and unquestionably valid species, even though known only from the type collection. 24. Ambrosia tomentosa Nutt. Gen. Am. Pl. 2: 186. 1818., non Fran- seria tomentosa Gray, 1849, nec Gaertnera tomentosa (Gray) Kuntze, 1891. Type: In upper Louisiana on the banks of the Missouri, Nuttall (probably Bm). Norte: I have been unable to locate this type in Nuttall’s herbarium in pH or elsewhere. However, Nuttall’s short description leaves no doubt that he was dealing with the taxon de- scribed here, viz.: ‘“‘Perennial; stem low; leaves bipinnatifid, under- side white and tomentose; spikes solitary.” Franseria discolor Nutt. Trans. Am. Phil. Soc. ser. 2 7: 345. 1840. Type: In cky Mountains, near the Colorado of the West, Nuttall (pH!). Xanthidium discolor (Nutt.) Delp. Studi Lign. Anem. Comp. Artem. 17. 1871. Franseria ey Wawra in Beck, It. Princ. S. Cob. 2: 41. 1888. Type: Norda- merika, In den Prairien, Wawra J 13. I have been unable to locate this ty N Wawra has included a beautiful and detailed illustration with the description (Taf. 8, Fig. B, loc. na which leaves no question whatso- ever about the identity of this taxo Gaertnera discolor (Nutt.) Kuntze, Rev. bea Pl. 1: 339. 1891. Gaertneria tomentosa (Nutt.) Nels. Bot. Gaz. 34: 34. 1902. Franseria tomentosa (Nutt.) Nels. New Man. Bot. Centr. Rocky Mtns. 542. 1909. Erect or reclining, perennial herb, perennating by means of adventitious shoots from the roots, forming clonal populations. Leaves alternate, petiolate, petioles to 10 cm. long, laminar tissue decurrent to base, usually with several pairs of small lobes below the main laminar area often obscur- ing the distinction between petiole and lamina. Lamina ovate-lanceolate in outline, to 10 cm. long and 6 cm. broad, irregularly bi- or tripinnately lobed; upper surface dark green, minutely scabrous, lower surface white woolly-tomentose, margins revolute. Inflorescence racemose-spicate. Staminate involucres stalked, stalks to 10 mm.; involucres to 7 mm. broad, 428 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV lobes to 12, sinuses separating lobes extending halfway or less to point of stalk attachment; upper surface of lobes with black striations, mostly along the midveins. Fruiting involucres in sessile clusters in axils of leaves subtending staminate racemes, to 6 mm. long and 3.5 mm. broad, 2- flowered; spines 0-10, scattered, ca. 1 mm. long (rarely to 3 mm.), some- what appressed. Chromosome number unknown. Flowering: May to August. Range: Prairie and semi-arid regions of Wyoming, Nebraska, Colorado, New Mexico, and Arizona; sporadic in Idaho, Iowa, and South Dakota. Notes: Aptly described by Nuttall (as Franseria discolor) as “A re- markable and distinct, as well as elegant species,” a comment based, no doubt, on the striking appearance of the white under surface of the leaf, rimmed by the dark green, revolute margin. LITERATURE CITED BRANDEGEE, T. S. 1889. A aru ee of plants from Baja California, 1889. Proc. Calif. Acad. Sci. ser. 2. 2: 117-216. CAVANILLES, J. 1793. Icones et cea an Plantarum 2. Matriti GEBBEN, A. I., W. W. Payne & W. H. Wacner, Jr. 1962. Bota nical Phase. In: J. M. SHELDON and E. W. Hewson (eds.), mariage a by aeroallergens, prog. rept. 5. University of Mic ai Ann Ar LinnaEus, C. 1753. Species Plantarum, 2. Holm PAayNE, W. W. 1962a. The unique morphology of ae spines of an armed rag- weed, Ambrosia bryantii (Compositae). Madrono 16: 233-236. . 1962b. Biosystematic studies of four widespread oes te species of rag- weeds (Ambrosia: Compositae). Doctoral thesis (unpublished), University of Michigan, Ann Arbor. 1963. The morphology of the inflorescence of oe (Ambrosia — Franseria: Compositae). Am. Jour. Bot. 50: 872-8 Payne, W. W., P. H. Raven & D. W. Kyuos. 1964. ane numbers in Compositae. IV. Ambroseae. Am. Jour. Bot. 51: were Rypserc, P. A. 1922. Ambrosiaceae. Jn N. Am. FI. 33: 3-4 SHINNERS, L. H. 1949, Notes on Texas Compositae III. Field Lab. 17: 170-176. DEPARTMENT OF BOTANY, THE UNIVERSITY OF MICHIGAN, ANN Arbor, MICHIGAN EXPLANATION OF PLATES Except in Fics. 38 and 60, staminate and pistillate involucres have been illus- trated without pubescence. PLATE I Fics. 1-5, Ambrosia pict (all from Brandegee, Feb. 11, 1889, us). ages nce, X 0.5; 2, staminate head, * 3.3; 3, vad ao staminate head, ‘ ; 4, fruiting involucre, x 3 ee a leaf silhouette, x 0 ie 6-10, Ambrosia ambrosioides. 6, Inflorescence, X “nm 7, staminate head, 1964] PAYNE, THE GENUS AMBROSIA 429 x 3.3 (6 & 7, from Wiggins 6379, MICH); 8, a from staminate head, x 10 (from Peebles & Loomis 6728, "aRIZ) : 9, fru ing involucre, * 3.3 (from Peebles 15041, ariz); 10, leaf silhouette (arte teres from greenhouse specimen, Payne AJL) Fics. ie 14, Ambrosia artemisioides (all from Williams 2527, Ny). 11, In- florescence, X 0.5; 1 ; staminate head, * 3.3; 13, fruiting involucres, pa 14, leaf : silhouettes, x1 PLATE II Fics. 15-19, Ambrosia arborescens. 15, Inflorescence, * 0.5; 16, staminate head, X 3.3; 17, pale from staminate head, * 10 (15, 16 & 17 from Killip Smith 17884, NY); 18, ing involucres, X 3 from Killip & Smith rui a 19676, NY); 18b from Pennell 2636, NY); 19, leaf silhouette, * 0.33 (from Steinbeck 8817, NY). Fics. 20-24, Ambrosia ela 20, So x05; ee staminate head, nec ees pale from stam ies head, X 10 (20, 21 & 22 from Shreve 6944, ARIZ); 23, fruiting involucr 3.3 (23a from See 944; 23b from Gentry 4128, ARIz); 24, leaf Ge xX 0.33 (24a & b from Shreve 6944, ARIZ; 24c from Rempel 151, Ariz). PLATE III Fics. 25-29, Ambrosia camphorata. 25, Inflorescence, X 0.5 (from Payne 4175, MICH); 26, staminate heads, X 3.3 (26a from Payne 4175, MICH; 26b from M. E. Jones 24058, MIcH); 27, pale from staminate head, * 10 (from Payne 4175, MICH); 28, fruiting involucres, X 3.3 (28a from Payne 4000, MICH; 28b from Payne 4175, MIcH; 28c from Palmer 757, NY; 28d from usta 881, NY; 28e from Wiggins 7935, MICH; 28f from Greene, April 24, 1885, : 28g from M. E. Jones 24647, Ny; 28h from Broder 618, MICH) ; 29, leaf Ancucies X 0.33 (29a-g photoduplicated from population sample from southern San Luis Potosi, Mexico, Payne 4007, MicH; 29h photoduplicated from greenhouse specimen, Payne BCK). PLATE IV Fics. 30-34, Ambrosia canescens. 30, Inflorescence, * 0.5; 31, staminate heads, X 3.3; a2: pale from staminate head, x 10 (30, 31 & 32 from Payne 4030, MICH) : RKP fruiting involucres, X 3.3 (from a single specimen, Payne 4049, MICH): 34, leaf silhouettes, & 0.33 ta aces from population sample from Jalisco, Mexico, Payne 4121, MIc Fics. 35-39, Ambrosia ‘chenopoifolin 35, Inflorescence, X 0.5 (from M. E. Jones 3702, NY); 36, staminate head, * 3.3; 37, pale from somes head (36 & 37 from Orcutt, April 10, 1885, Ny): 38, fruiting involucres, X 3.3 (38a from Orcutt, May 7, 1886, mMicH; 38b from Palmer 731, NY); 39, leaf ainouciee x 0.33 (noted pies greenhouse population, nee BDK - 0-44, pidebiies deltoidea. 40, Inflorescence, x 0.5 (from Shreve 7506, x 3 43a from Peebles 15038, ariz; 43b from Gillespie 5436, NY); 44, leaf Gide Pe adtete cd from greenhouse population, Payne AJP). PLATE V Fic. 45, Ambrosia chamissonis, leaf silhouettes, X 0.33 (from greenhouse 430 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV grown population and progeny from Marin County, California; each leaf repre- sents a different specimen). 46, Ambrosia confertiflora, leaf silhouettes, X 0.33 (from population in eee, Coahuila, Mexico, Payne 3767, MicH; each leaf represents a dif- ferent specimen). PLATE VI Fics. 47-51, re tba adtlbbig 47, Inflorescence, < 0.5; 48, — heads, aminate x 10 (47, 48 (a) 100 CH fruit nber, May 1 50b from Turkham, April 19, 1942, Ariz); 51, leaf silhouettes, xX 0.33 (photo- duplicated from greenhouse population, Payne ‘AJH ) Fics. 52-56, Ambrosia dumosa. 52, Inflorescence, * 0.5 (from Harrison & Beldon 3577, ariz); 53, capitulous branch with mature staminate and immature pistillate heads, X 3.3; 54, pale from staminate head, 10 (53 & 54 from green- house specimen, Payne BDG-a): 55, fruiting involucres, xX 3.3; (55a from greenhouse specimen, Payne BDG-a; 55b fro m Humphrey SCS-767 7, ARIZ) ; 56, leaf silhouettes, & 0.33 (photoduplicated from greenhouse ppuiie, Payne BDG). Fics. 57-61, Ambrosia eriocentr ra. 57, Inflorescence, * 0.5; 58, staminate leaf silhouettes, x 0.33 (photoduplicated from greenhouse specimen, Payne AJR). PLATE VII Fics. 62-67, Ambrosia ese laneg 62, Inflorescence, * 0.5; 63, staminate head, x 3.3; 64, floret from staminate head, a 10; 65 ? ~~ from staminate head, x 10 (62, 63, 64 & 65 from M. E. Feb. 1924, ny); 66, fruiting involucre, 43 (from Ahles 9087, aS “67. leaf Hieenies x 0.33 (photo- duplicated from greenhouse specimen, Payne "AJO). PLATE VIII Fics. 68-72, Ambrosia ase coke 68, Inflorescence, * 0.5; 69, staminate erry X 3.33; 70, pales from staminate head, x 10 (68 69 & 70 from Ownbey 5, NY); 71, fruiting oars x 3.3 (71a from Saunders, July, 1893, ny; ne from Ownbey 1325, NY); 72, leaf sihouetts, x 1 (72a & b from Ownbey 1325, NY; 72c from Saunders, July, 1893, Fics. 73-77, Ambrosia ae io from Brandegee, Jan, 13, 1889, Ny). 73, Inflorescence, x 0.5; 74, staminate head, , pale from’ staminate head, x 10; 76, fruiting involucre, mM ot Ti, leaf silhouette, x 0 fick 78-82, Ambrosia nivea ‘on from Hartman 813, ae 78, Inflorescence, xX 0.5; 79, staminate head, * 3.3; 80, pale Sa staminate head, x 10; fruiting involucre, X 3.3; 82, leaf Ohoutie x9 p) Jour. ARNOLD Ars. VoL. XLV PLATE I PayNE, THE GENUS AMBROSIA Jour. ARNoLD Ars. VoL. XLV PiaTE II 24a 24d PAYNE, THE GENUS AMBROSIA Jour. ARNoLD Ars. VoL. XLV PrateE III PAYNE, THE GENUS AMBROSIA Jour. ARNOLD Ars. VoL. XLV PLATE IV ia, 44 fh, PAYNE, THE GENUS AMBROSIA Jour. ARNOLD Ars. VoL. XLV PLATE V 46c 46g 46h 4 46 PayNE, THE GENUS AMBROSIA Jour. ARNoLp Ars. VoL. XLV PiaTE VI PAYNE, THE GENUS AMBROSIA Jour. ARNOLD Ars. VoL. XLV PiaTE VII } \ ( / ¢ SE. 65 6 fa é 6 hb PAYNE, THE GENUS AMBROSIA Jour. ArRNoLtp Ars. VoL. XLV PiaTe VIII wie eae THE GENUS AMBROSIA 1964] BRIZICKY, GENERA OF RHAMNACEAE 439 THE GENERA OF RHAMNACEAE IN THE SOUTHEASTERN UNITED STATES 14 GEORGE K. BriIzICcKy RuHAMNACEdE A. L. de Jussieu, Gen. Pl. 376. 1789, “Rhamni,” nom. cons. (BucKTHOoRN FAmMILy.) Deciduous or evergreen trees, shrubs, or woody vines [exceptionally herbs], unarmed or armed with thorns or stipular spines. Leaves simple, alternate, subopposite, or opposite, petiolate; stipules mostly free, more rarely connate across the axil, usually minute and caducous, sometimes modified to spines [rarely absent]. Inflorescences axillary few-flowered umbel- or corymb-like cymes (sometimes reduced to a solitary flower), or axillary and/or terminal raceme- or spikelike thyrses. Flowers small or minute, regular, perigynous to epigynous, with a distinct floral tube, bi- and/or unisexual by abortion, pediceled or sessile. Floral tube (“hypan- thium”) patelliform to hemispherical or campanulate [or urceolate to cylindrical], persistent in fruit at least in part; calyx lobes usually 4 or 5, valvate in aestivation, usually deciduous either separately or with the Sometimes circumscissile upper part of the floral tube, rarely persistent. Petals 4 or 5, alternate with the calyx lobes, + concave or hooded or flat, often clawed, enfolding the stamens in aestivation, or wanting. Stamens 4 or 5, opposite the petals, adnate to them at the base and inserted at or below the margin of the disc, smaller and sterile in @ flowers; filaments usually subulate, longer or sometimes slightly shorter than the anthers: anthers dorsifixed, usually nonversatile, 2-locular at anthesis, longitudinally dehiscent; pollen usually 3-colporate, suboblate to subprolate, small to medium sized, often + triangular in polar view, + smooth to reticulate. Nectariferous disc intrastaminal, hypogynous to epigynous [rarely want- ing]. Gynoecium 2- or 3[4]-carpellate, syncarpous, rudimentary in 4 "Prepared for a generic flora of the southeastern United States, a joint project of the Gray Herbarium and the Arnold Arboretum of Harvard University made possible through the support of George R. Cooley and the National Science Foundation and under the direction of Carroll E. Wood, Jr., and Reed C. Rollins. This treatment follows the pattern established in the first paper in the series (Jour. Arnold Arb. 39: 266-346. 1958) and continued through those in volumes 40-45 (1959-1964). The area b marked by an asterisk. The author is indebted to Carroll E. Wood, Jr., for his many valuable suggestions; to J. Rzedowski, Escuela Nacional de Ciencias Bioldgicas, México, D. F., for supplying references to the distribution of Berchemia scandens in Mexico; and to Mrs. Gordon W. Dillon for her help in the preparation of the typescript. 440 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLV flowers; stigmata 2 or 3 [4] or stigma 1, usually lobed; style 1, often 2- or 3[4]-lobed to -parted ; ovary 2- or 3[4 }-locular, sometimes incompletely so at the very base [or in the middle, or rarely throughout], the septation diverse; ovules solitary [exceptionally 2] in each locule, basal [or sub- basal?], ascendent, anatropous, epitropous or (more commonly) pleurotro- pous (the raphe lateral) [very rarely apotropous?], 2- integumented, with a thick nucellus, funiculate. Fruit usually a drupe with 2 or 3 [4] dehiscent or indehiscent pyrenes, or with a (1-) 2- or 3-locular dehiscent or indehiscent stone, or rarely a commissurally winged [or unwinged] schizo- carp, with [or without | a carpophore.? Seed convex and smooth or rarely grooved on the back (abaxial side), with lateral, rarely dorsal, raphe, sometimes minutely [or conspicuously] arillate; endosperm scanty or copious, rarely ruminate, or wanting; embryo “investing,” usually large, axial, straight; cotyledons flat or plano-convex; radicle inferior. TYPE GENUS: Rhamnus L A family of nearly 55 genera and 900 species, of the Tropical and Temperate zones of both hemispheres, the northern limit of the range mostly running between 50° and 55° N. Lat. Five tribes generally are recognized. Intergeneric relationships show many reticulations, and it is likely that further studies will result in a reduction in the number of genera. The small, generally inconspicuous flowers, which always have a floral tube formed by the fusion of basal parts of the perianth and androecium, are mostly bisexual. Unisexual flowers, with consequent polygamy and/or mono- or dioecism, occur within some genera (e.g., Berchemia, Gouania, Rhamnus spp.) but only rarely, if at all, are a character of generic signifi- cance. Cross-pollination by hymenopterous, coleopterous, and dipterous insects seems to be the rule, but data are available for relatively few genera and species. Pollen morphology is relatively uniform throughout the amily ° Fruits in Rhamnaceae have wi been characterized as drupes and/or berries and/or capsules. Recently, Vikhir on the basis of the gross morphology and histology of the fruits of this fainily, Gaaiicl them as syncarpous drupes throughout. classification of geo intermediate to schizocarps seems to be especially difficult, at least in some cases. In this respect, Rusby’s suggestion (in Rusby & Jelliffe, Morphology and Histology a; Plants, pp. 110, 111. 1899) can be helpful. Since schizocarps are In concurrence with both this suggestion and Vikhireva’s con- clusions, the fruits of Rhamnaceae are classified here as drupes, except those in Gouania and its allies, which are regarded as schizocarps. 1964] BRIZICKY, GENERA OF RHAMNACEAE 441 The structure of the ovary with regard to the number of carpels, the mode of septation (‘“‘placentation”), and the fertility or sterility of the septal margins representing the margins of two adjoining connate carpels seems to be a character important for understanding the evolution of the family (see Suessenguth, pp. 31-33). As a rule, only one of the two margins of a carpel forming an ovary locule is fertile, and the sequence of the fertile and sterile margins of carpels or septa in a three-locular ovary can be different in different genera. Consequently, two general situations are possible: in the first case, one of the septa is fertile on both its margins, the second on one, and the third completely sterile (e.g., Rkamnus spp.) ; in the second case, each septum is fertile on one of its margins (e.g., Ceanothus spp.). Basal placentation has generally been regarded as one of the most dis- tinctive features of Rhamnaceae. Prichard confirmed this for at least six genera which he investigated and also remarked, “The ovules do, however, develop from the lateral [slightly incurved] margins of carpels” (p. 103). This agrees with the explanation offered by Suessenguth (loc. cit.) and seems to make it rather clear that this basal placentation is one of the variants of axile placentation, or at most a derivative of it. (See, however, Nair & Sarma, p. 54). Chromosome counts made for eleven genera (55 species) are 2n = 18, 20, 22, 24, 26, 34, 36, 37, 40, 48, 49, or 50, 72, and 96, suggesting that both aneuploidy and polyploidy have been of importance in the evolutionary de- velopment of Rhamnaceae. It is notable that 2n = 24 is the most frequent chromosome number occurring in at least some species of eight genera, and that both highly polyploid and homoploid genera occur in the family. In most of our genera the seeds are dispersed mainly by birds and mammals (e.g., Rhamnus, Berchemia, Reynosia, Sageretia, Krugioden- dron), rarely by wind (e.g., Gouania) or by water (Colubrina asiatica). In some, the fruits themselves are mechanically active (“explosive fruits”), ejecting seeds short distances (e.g., Ceanothus, Colubrina spp.). A close relationship of Rhamnaceae to Vitaceae seems to be unquestion- able. The relationship to Celastraceae presumed by Prichard and some taxonomists (cf. Suessenguth) is denied by Nair & Sarma (p. 54), who remarked, “It must be noted that floral-anatomically there is very little in common between the two families (see Berkeley, 1953). Further the embryological and palynological data do not support any relationship between the two families. . . .” Species of several native and exotic genera (e.g., Paliurus, Pomaderris, Colletia) are cultivated as ornamentals. The eastern Asiatic Hovenia dulcis Thunb., raisin tree, 2n = 24, cultivated as an ornamental and for its fleshy, edible rachis and inflorescence branches, has been recorded as escaped and spreading very slowly in Polk County, North Carolina (Free- man). Some species yield wood of local importance (e.g., the tropical African Maesopsis Eminii Engl., musizi; the East African Phyllogeiton Zeyherit (Sond.) Suesseng., pink or red ivory wood). Fruits, bark, and/or roots of some species are sources of drugs or natural dyes. 442 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV REFERENCES: Barton, H. Monographie des Célastracées et des Rhamnacées. Hist. Pl. 6: 1-92. 1875. [Rhamnaceae, 51-92. ] BEILLE, L. Recherches sur le développement floral des Disciflores. Thése. 177 pp. Bordeaux. 1902. [Floral organogeny of Rhamnus Frangula, 111-113; Ceanothus, 113, 114; Colletia, 114.] BENNEK, C. Die morphologische Beurteilung der Staub- und Blumenblatter der Rhamnaceen. Bot. Jahrb. 77: 423-457. pls. 19-27. 1958. [Each petal and the stamen opposite it form a unit, and the primordium from which they both develop is to be regarded as a phyllom- or a stamen-petal-unit. | BENTHAM, G., & J. D. Hooker. Rhamneae. Gen. Pl. 1: 371-386. 1862. BronGniart, A. Mémoire sur la famille des Rhamnées, ou histoire naturelle et médicale des genres qui composent ce groupe de plantes. 78 pp. pls. 1-6. Paris. [ca. July] 1826. [Also published in Ann. Sci. Nat. 10: 320-386. pls. 12-17. 1827, in March or later. ] Brown, C. A. Louisiana trees and shrubs. La. Forestry Commiss. Bull. No. 1. i-x + 1-262. 1945. [Rhamnaceae, 178, 179.] CANDOLLE, A. P. pe. Rhamneae. Prodr. 2: 19-42. 1825. Dotcuer, T. Ricerche embriologiche sulla famiglia delle Rhamnaceae. (In Italian; English summary.) Nuovo Gior. Bot. Ital. II. 54: 648-673. 1947. [Rhamnus Alaternus L.; Colletia cruciata Gill. & Hook., C. spinosa Lam., parthenocarpy; Paliurus aculeatus Lam.; Hovenia dulcis. | FREEMAN, O. M. New or noteworthy plants from Polk Co., North Carolina or vicinity. Castanea 21: 41-43. 1956. [“Hovenia dulcis Thunb. The raisin- tree has escaped from cultivation on the east slope of Warrior Mt. and ap- pears to be spreading rather slowly,” 42, 43.] Gemo.., K. Anatomisch-systematische Untersuchung des Blattes der Rhamneen aus den Triben: Rhamneen, Colletieen und Guanieen. Beih. Bot. Centralbl. 12: 351-424. 1902. Herzoc, T. Anatomisch-systematische Untersuchung des Blattes der Rhamneen aus den Triben: Ventilagineen, Zizypheen und Rhamneen. Beih. Bot. Centralbl. 15: 95-207. 1903. Imuorr, W. Beitrage zur Pflanzengeographie und Pflanzendkologie der Familie der Rhamnaceae. Inaug.-diss. 127 pp. Breslau. 1928. Jounston, M. C. Revision of Condalia including Microrhamnus (Rhamnaceae). Brittonia 14: 332-368. 1963. [Ziziphus, Krugiodendron, and Reynosia included in the general discussion, 333-336. Juet, H. O. Beitrage zur Morphologie und Entwickelungsgeschichte der Rham- naceen. Sv. Vet-akad. Handl. III. 7(3): 1-13. pls. 1, 2. 1929. [Ceanothus americanus; Rhamnus cathartica and R. Frangula. | LEINFELLNER, W. Besitzt die Gattung Phylica L. (Rhamnaceae) echt revolutive Rollblitter? Osterr. Bot. Zeitschr. 106: 577-603. Mrrrs, J. On the tribe Colletieae, with some observations on the structure of the seed in the family of the Rhamnaceae. Contr. Bot. 1: 232-311. pls. 33-42. 1861. [Structure of pyrenes and seeds in Colletia dumosa Miers, several spp. of Rhamnus, Ziziphus, and Alphitonia, pp. 232-250. pl. 33. See also Ann. Mag. Nat. Hist. III. 5: 76-95, 200-216, 267-273, 370-381, 482-492. 1860; 6: 5-14. 1860. ] Narr, N. C., & V. S. Sarma. Organography and floral anatomy of some members of the Rhamnaceae. Jour. Indian Bot. Soc. 40: 47-55. 1961. [Zzziphus, 1964] BRIZICKY, GENERA OF RHAMNACEAE 443 6 spp., ea subbasal; Helinus, 1 sp. and Ventilago, 1 sp., placenta- tion parieta PRICHARD, E. Cc. Morphological studies in Rhamnaceae. Jour. oo eit Sci. Soc. 71: 82-106. 1955. [Ziziphus, 1 sp.; Berchemia, 1 s (Frangula), 2 spp.; Ceanothus, 6 spp.; Colletia, 1 sp.; Gouania 1 . common origin of Rhamnaceae and Celastraceae presum REcorpD, S. American woods of the family Rhamnaceae. ae. Woods 58: 6-24. 1939. SARGENT, a S. Manual of the trees of North America (exclusive of Mexico). ed. 2. xxvi + 910 pp. map. Boston & New York. 1922. [Rhamnaceae, 718-731 SRINIVASACHAR, T. Embryological studies of some members of Rhamnaceae. Proc. Indian Acad. Sci. B. 11: 107-116. 1940. [Ziziphus, 2 spp.; Scutia, 1 sp. | SUESSENGUTH, K. Rhamnaceae. Nat. Pflanzenfam. ed. 2. 20d: 7-173. 1953. [Extensive bibliography, 7-11. ] TAKHTAJAN, A. Die Evolution der Angiospermen. viii + 344 pp. Jena. 1959. [Rhamnales, 243, =| TRELEASE, W. North American Rhamnaceae. Trans. Acad. Sci. St. Louis 5: 358-369. 1889. ———. Rhamnaceae. Jn: A. Gray, Syn. Fl. N. Am. 1(1): 401-419. 1897. Urpan, I. Rhamnaceae. Symb. Antill. 9: 218-230. 1924. [Septation of ovary in Reynosia, Auerodendron, Karwinskia, Rhamnidium, 225-229. VIKHIREVA, V. V. Seed development in the buckthorn rum (Rhamnaceae). (In Russian.) Acta Inst. Bot. Acad. Sci. URSS 7. Morphol. Anat. Pl. 2: 221-227. 1951. [Rhamnus, Ziziphus, Paliurus, Hovenia, Ceanothus, Poma- derris, Colletia; seed histology + uniform, differences only in shape and/or size of cells of outer epidermis. ] : morphological-anatomical study of the fruits of Rhamnaceae. (In Russian.) /bid. 3: 241-292. 1952. [Twenty-four genera. ] WEBERBAUER, A. Rhamnaceae. Nat. Pflanzenfam. III. 5: 393-416. 1895; 417- 7. 1896 West, E., & L. E. ARNOLD. The native trees of Florida. xx + 212 pp. Gainesville. 1946. KEY TO THE GENERA OF RHAMNACEAE General characters: woody plants; flowers small, regular, 4- or 5-merou bisexual, more rarely unisexual; floral tube always developed; petals pee concave or pee , often Re ON Fo or wanting; stamens 4 or 5, opposite the petals ; Re ctirilee cis disc intrastaminal; gynoecium syncarpous, the ovary superior to inferior, 2- or 3-locular; ovule solitary in each locule, basal, ascend- ent; fruit a syncarpous drupe or a schizocarp. A. Leaves alternate. B. Plants climbing by twining or by tendrils. C. Vines climbing by twining; leaves entire, with 9-12 pairs of straight- ish subparallel veins; ovary pate 2-locular; fruit a bluish-black ellipsoid to obovoid ae stoned drupe. .............. .5. Berchemia. C. Vines climbing by solitary ede near base of peduncle of inflor- escence; leaves coarsely and remotely crenate-serrate, with 4-7 pairs of curved. somewhat convergent lateral veins; ovary inferior, 3- 444 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV locular; fruit a 3-winged schizocarp ca. 1 cm. broad, splitting along each wing into three 2-winged mericarps............. 9. Gouania, B. Plants erect shrubs or trees . Inflorescences dichotomously branched manifestly peduncled many- flowered axillary and terminal thyrses, the branches becoming club shaped, fleshy, and reddish in fruit; nectariferous disc pubescent; drupe leathery, subglobular, light grayish or brownish, with a thin- coriaceous, indehiscent, 3-locular stone; leaves large, 3- nerved at base, the petiole 2.5-4 cm. long; deciduous ornamental tree sparingly es- caped from yen er puta daceeaueeseeesees Hovenia Thunb. ] D. Inflorescences neither dichotomous nor fleshy in fruit; nectariferous disc glabrous; petiole usually shor E. Drupe fleshy, superior, subtended at base by a flat remnant of the floral tube; inflorescences few-flowered corymb- or umbel-like fas- cicles, short peduncled to sessile. F. Branchlets bearing flowers and fruits often fascicled in axils of leaves or leaf scars, usually deciduous; stipules modified to spines; petals hooded, conspicuously clawed; nectariferous disc at, + pentagonal, encircling the ovary up to a conspicuous pe re solitary, hard, sharp pointed, indehiscent; leaves 3-nerved almost to the apex, lateral veins often — cultivated and naturalized deciduous shrub or small tree. ...... 6. Ziziphus. F. Branchlets not fascicled; stipules ee ne nectariferous disc lining walls of floral tube, free from ovary; stylopodium absent; drupe nearly black, subglobular to almost obovoid, with 3 distinct dehiscent or indehiscent pyrenes; leaves pinnately veined, finely serrulate. ..................... 1. Rhamnus. . Drupe leathery, semi-inferior, adnate to a cuplike floral tube either at base or to half its length, separating into 3 dehiscent pyrenes or pyrene-like carpels; disc + annular, encircling the semi-inferior ovary; inflorescences many- to few-flowered thyrses G. Inflorescences usually many-flowered corymb- or raceme-like ico) calyces and elongated filiform pedicels white; petals white, long clawed, much longer than the calyx lobes; leaves usually toothed, 3-nerved at least at base; drupes subglobular or obovoid, distinctly 3-lobed; deciduous low shrubs or sub- Be oes As ea ¥en 3s h oa token eee 3. Ceanothus. G. Inflorescences few-flowered contracted axillary thyrses much shorter than the subtending leaves; calyces and pedicels green; petals yellow to greenish or white, very short clawed to sessile, as long as or shorter than the calyx lobes; leaves pinnately nerved and entire or 3-nerved at base and/or serrulate; drupes subglobular, inconspicuously 3-lobed or unlobed; evergreen shrubs or small trees; southernmost Florida. .. 4. Colubrina. A. Leaves opposite ae peti osite. . Leaves usually acute or short-acuminate at apex, serrulate or serrulate- crenate, ovate; aa free, caducous; flowers petaliferous, sessile, in slender, spikelike, often panicled axillary and terminal thyrses much longer than the subtending leaves; drupe containing (2) 3 distinct, 1964] BRIZICKY, GENERA OF RHAMNACEAE 445 flattish, dehiscent pyrenes; oa spinescent shrub, eetaeecunre clamber- ing; so outhe astern Coastal Plain. ............. 2. Sageretia. . Leaves obtuse or rounded, eT notched at apex, otherwise entire ; stipules + connate across the axil, persistent; flowers apetalous, pedi- celed, in few-flowered axillary cymes much shorter than the subtending leaves or sometimes solitary; drupe subglobular, with a single hard stone; southernmost Florida. I. Leaves thin-coriaceous, not revolute at margin, ovate to broadly elliptic, obtuse to rounded at base, with 4-6 pairs of conspicuous, slightly curved lateral veins and coarsely reticulated veinlets; stipules connate only at the very base; calyx lobes crested inside. ......... No 6 OS a, Bis ce is 5.3 ny wt DG : Krugiodendron. I. Leaves stiff-coriaceous, revolute at margins, oblong or oblong-elliptic to narrowly obovate, cuneate to rounded at base, with 6-15 pairs of straightish, usually obscure lateral veins and numerous finely reticu- lated veinlets (lens!); stipules connate at least half their length; calyx lobes crestless inside. 7. Reynosia. en) Tribe RHAMNEAE 1. Rhamnus Linnaeus, Sp. Pl. 1: 193. 1753; Gen. Pl. ed. 5. 89. 1754. Deciduous [or evergreen] unarmed [or armed] shrubs or trees [rarely woody vines] with scaly or naked buds. Leaves alternate [to opposite], membranaceous [to coriaceous], pinnately veined, toothed [or entire]. Inflorescences axillary sessile or peduncled umbel-like cymes, sometimes reduced to solitary flowers |or axillary and/or terminal thyrses or rarely racemes]. Flowers small, perigynous, pediceled, bisexual and/or unisexual by abortion, the plants then usually (polygamo-) dioecious. Floral tube + campanulate [to urceolate|; calyx lobes 4 or 5, greenish or greenish yellow, deciduous separately or with the circumscissile upper part of the floral tube after anthesis. Petals 4 or 5, shorter than calyx lobes, yellow- ish or whitish, usually broad, nearly ovate or obovate |or sometimes linear to filamentous], emarginate to bilobed at apex [rarely acute], often some- what different in ¢ and @ flowers, usually short clawed, concave to flat [or hooded], inserted at upper margin of disc, or wanting. Stamens 4 or 5, as long as or somewhat longer than petals, sterile and + rudimentary [or wanting] in @ flowers; filaments slightly longer or shorter than anthers; anthers oblong-ovate to ovate in outline, + cordate at base, sometimes apiculate at apex. Nectariferous disc lining wall of floral tube, thin below, somewhat thickened above. Stigma 3-lobed or stigmata 2—3 and subcapitate-discoid; style simple or 2- or 3[4]-forked; ovary superior, 2- or 3{4]-locular, incompletely so at the very base; ovules epitropous or pleurotropous. Fruit a small subglobular to obovoid, usually black [or dark red, rarely yellow| drupe, with 2 or 3 [4] pyrenes; mesocarp fleshy, with tanniniferous idioblasts or mucilage cavities; pyrenes cartilaginous to somewhat woody, ++ convex on abaxial (dorsal), somewhat angular to almost flat on adaxial (ventral) surface, with a distinct adaxial suture, dehiscent or indehiscent. Seed narrowly or broadly obovate to subor- 446 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV bicular, plano-convex or lenticular, smooth or longitudinally grooved on abaxial side; seed coat membranaceous, sometimes very thin and closely adherent [or adnate] to endocarp; endosperm fleshy, scanty; embryo large, the cotyledons curved or flat, the radicle short. Germination epi- or hypogeal. (Including Frangula Mill.) Lectotype species: R. cathar- tica L.; typified by S. F. Gray, Nat. Arrang. Brit. Pl. 2: 621. 1824, who included only this species in sect. RAamnus. (Name from Greek, rhamnos, an ancient name of some species of the genus.) — BUCKTHORN. A genus of 150 (or fewer) species, distributed primarily in the Tem- perate and Tropical zones of both hemispheres, abundant in eastern Asia and southwestern North America, scarce in Europe and Africa, absent from Madagascar, Australia, Polynesia, and southern South America; three indigenous species in our area. The four subgenera exhibit reticulate relationships. The most distinc- tive differences are in the structure of pyrenes and seeds; in this respect subg. PSEUDOFRANGULA ® is intermediate between subgenera RHAMNUS and FraNcuta (cf. Wolf, pp. 18, 32). Subgenus ScrapopHiLa (Philippi) Benth, & Hook. (including only the Chilean Rhamnus diffusa Clos |Colle- tia maytenoides Griseb., Sciadophila maytenoides (Griseb.) Philippi] ) 1s poorly known, especially in regard to the structure of pyrenes and seeds, and perhaps represents a distinct, but closely related, genus. Subgenus RHAMNus. Winter buds scaly; leaves alternate | to opposite] ; flowers usually petaliferous and 4-merous, mostly bi- and/or unisexual; style forked; pyrenes usually dehiscent along the adaxial suture; seeds deeply grooved on the back, with an abaxial raphe on the bottom of the groove; cotyledons thin, curved around the intruded groove; germination epigeal. About 100 species, centered in eastern Asia, a few in Europe and in tropical Africa (exclusive of Madagascar), several in North America. Rhamnus lanceolata Pursh, a tall unarmed shrub, with lanceolate leaves and two-carpellate gynoecia, occurs on open wooded slopes, thickets, and glades, usually on calcareous soil, from Alabama and Tennessee west to eastern Texas and Arkansas, north to Nebraska, southern Iowa, southern Wisconsin, Indiana, southern Ohio, south-central Pennsylvania, West Vir- ginia and Virginia. In the western part of the range var. lanceolata, with the leaves short-pubescent on the lower surface, seems to predominate, while in the eastern and southeastern part var. glabrata Gleason, with the leaves glabrous or promptly glabrate, appears to prevail. The status of these variants needs further study. The European R. cathartica L., com- mon buckthorn, 2” = 24, a spinescent shrub or small tree with 3- or 4- carpellate gynoecia, has become naturalized in Canada 7nd from the Northeastern States south to Virginia. Subgenus PseuporraNcuta (Grubov) Brizicky. Winter buds scaly; leaves alternate; flowers apetalous, (4)5-merous, bi- and/or unisexual ; ’Rhamnus L. subg. Pseudofrangula (Grubov) Brizicky, stat. nov. Rhamnus L. sect. Pseudofrangula Grubov, Not. Syst. Leningrad 12: 125. 1950. TyPE SPECIES: R. alnifolia L’Hér. 1964 | BRIZICKY, GENERA OF RHAMNACEAE 447 style forked; pyrenes indehiscent; seeds not grooved, the raphe abaxial; cotyledons thin, + plane; germination epigeal. Four species, one North American and three eastern Asiatic. Rhamnus alnifolia L’Hér., a low shrub with 3-carpellate gynoecia, occurs in swamps, low woods, and meadows, often on calcareous soil, from Newfoundland to north-central Alberta, British Columbia, and Washington, south to central California and north- eastern Oregon, Idaho, Wyoming, Nebraska, Iowa to New Jersey, and West Virginia. It was formerly known from Cedar Creek, Campbell County, Tennessee, but this locality is now covered by Norris Lake. Rhamnus Ishidae Miyabe & Kudo, apparently closely related to our spe- cies, is endemic on Hokkaido, Japan; R. Arnottiana Gardner is endemic in Ceylon; and R. purpurea Edgew. is known from the Himalaya and adja- cent China. The few species and their present distribution suggest the antiquity of this group. Subgenus FrANGULA (Mill.) Reichenb. Winter buds naked; leaves alternate; flowers petaliferous, 5-merous, bisexual; style simple; pyrenes indehiscent; seeds not grooved, the raphe lateral; cotyledons plane, thickish; germination hypogeal. About 40 species, primarily American (26 species), centered in California and Mexico, occurring southward to Peru and Brazil, with one species in the West Indies; the remaining species mostly in eastern Asia, a few in Europe, North Africa, and the Azores. Rhamnus caroliniana Walt., Indian cherry, 2n = 24, a shrub or small tree, the gynoecia 3-carpellate, occurs in low woodland in valleys along streams, on rocky open wooded slopes, upland ridges, thickets, and glades, com- monly on limestone outcroppings, from Florida (Charlotte County on the west and Orange County on the east coast) to westernmost Texas, north to western Virginia, West Virginia, Ohio, Indiana, Illinois, and Nebraska. Two variants which occur throughout this range (although one may be prevalent locally) have sometimes been recognized as var. caroliniana, with the leaves glabrous or glabrescent beneath, and var. mollis Fern., with the leaves permanently pubescent (“velvety”) on the lower surface. Rhamnus caroliniana is closely related to R. betulifolia Greene, of the southwestern United States and Mexico; both “are interpreted as deriva- tives of a stock which migrated from farther south and which also probably gave rise to such a species as R. Purshiana DC., now distributed from northern California to Washington’? (McVaugh). The European R. Frangula L., alder-buckthorn, 2m = 20, 22, 26, a shrub or small tree, the gynoecia 2-carpellate, has become naturalized at least in the Northeastern States, south to New Jersey. Subgenus FRANGULA differs from our other subgenera in the mucilaginous walls of the epidermal cells and the presence of mucilage ducts in the collenchyma of veins of the leaves and in the fruits; in the uniseriate, several-celled hairs (as against unicellular) ; and in the somewhat different wood anatomy (cf. Record). On these bases the subgenus is often segre- gated (especially in Europe) as a distinct genus, Frangula Mill. Since, however, Rhamnus as a whole seems to have been inadequately studied, 448 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV especially with regard to its tropical members, generic segregation appears to be somewhat premature. The genus is poorly known in regard to its biology, floral morphology, and anatomy. The bisexual flowers of R. Frangula have been recorded as proterandrous. Chromosome counts have been made for seven species. Four species of subg. RHAMNUS have 2m = 24, and one (the tropical African R. prinoides L’Hér.), 2n = 34; two species of subg. FRANGULA have 2n = 24 and 2n = 20, 22, 26. Two different forms of karyotype not corresponding to the sibeenetic divisions have been found (cf. Dolcher). Fruit dispersal i is mostly by birds. The genus is closely allied to Sageretia Brongn. A number of species are grown as ornamentals and/or hedge plants (e.g., Rhamnus cathartica, R. Frangula, R. Purshiana), and some yield wood of local importance (e.g., R. Frangula, R. cathartica) or drugs (R. Purshiana, R. Frangula, R. cathartica). Rhamnus caroliniana and va .¢ californica Eschsch., R. Purshiana) are important honey plants. Fruits of R. alnifolia are said to be poisonous. At least R. alnifolia, R. cathartica, R. Frangula, and R. lanceolata are known to be the alternate host for Puccinia coronata Corda, crown rust of oats. REFERENCES: Under family references see GEMOLL (pp. 371-375), HeRzoc (pp. 174-199), Juet, Miers (pp. 232-250), Pricuarp (pp. 90-93), REcorD (p. 18), SARGENT (pp. 722-726), SUESSENGUTH (pp. 59-71), VIKHIREVA (1951; 1952, pp. 250- 255), WEBERBAUER (pp. 409, 410), and West & ARNOLD (p. 139). BEILMANN, A. P. Two buckthorns. Am. Bee Jour. 101: 175. 1961.* [R. caro- liniana and R. lanceolata, honey plants. | Berry, E. W. The physical conditions and age indicated by the flora of the Alum formation. U.S. Geol. Surv. Prof. Pap. 98-E: 41-59. pls. 7-10. 1916. [Includes R. apalachicolensis. | Butter, E. J., & S. G. Jones. Plant pathology. xii +979 pp. London. 1949. | Rhamnus and crown rust of oats, 411-415; references. | Coucou, V., & H. Tarpo. Sur une nouvelle substance isolée de |’écorce de bour- daine, Rhamnus Frangula L. Compt. Rend. Acad. Sci. Paris 254: 552-554. 1962. [‘‘Substance A,” in form of a ‘glucoside, presumably belonging to the group of anthraquinone derivatives. | DotcHER, T. Karyological observations on some species of Rhamnus. (In Italian; English summary.) Gior. Bot. Ital. III. 70: 147-150. sion | Two distinct forms of karyotype; the chromosome number can be the same. | FERNALD, M. L. A new variety of Rhamnus caroliniana. Rhodora 12: "19, 80. 1910. | Var. mollis. | Gopwin, H. Rhamnaceae. [Biological flora of the British Isles (8).] Jour. Ecol. 31: 66-92. 1943. [R. cathartica, 69-76; R. Frangula, 77-92.] Grusov, V. I. The monographic survey of the genus Rhamnus L. s. 1. (In Russian.) Acta Inst. Bot. Acad. Sci. URSS 1. Syst. 8: 242-423. 1949. | Frangula Mill., a segregate genus, with 52 spp. in 3 sects., 4 subsects., 19 series; Rhamnus L., sensu stricto, with 141 spp., 6 sects., 6 subsects., 36 series; descriptions of the new subdivisions of the genera are given only in 1964] BRIZICKY, GENERA OF RHAMNACEAE 449 Russian; names of several subdivisions adopted from Heppeler without mention of ce original author. | systemate generis RHAMNUS L. . (Primarily in Latin.) Not. Syst. apa eate 12: 123-133. 1950. "Latin ae of the sects., subsects., and several spp. mentioned in the previous wor HEpPELER, F. Beitrage zur Systematik der Gattung Rhamnus mit besonderer ae des Emodinvorkommens. Arch. mae Paes 152-173. 19 in Ber. Deutsch. Pharm. Ges. 38: 152-173. | Apparently a slightly ee equivalent of an inaugural cee published Jounson, C. aa & E. Hincman. Rhamnus Purshiana: its history, growth, methods of collection and bibliography. Am. Jour. Pharm. 86: 387-413. 1914.* Kozo-Potjanski, V. Ueber die Anwendung der anatomischen Methode in der Systematik der Angiospermen. Bull. Soc. Nat. Moscow Biol. II. 32: 289- 310. 1924. [Rhamnaceae, 299-303; R. cathartica, floral anatomy, 300- 302.] LacomseE, N. R., & H. W. YouNnGKEN. Studies on the anatomy of Rhamnus lanceolata Pursh and Rhamnus Frangula L. Jour. Am. Pharm. Assoc. 32 193-202. 1943.* [Anatomy of stem, leaf, and petiole, and a brief discus- sion of anthraquinone content Linpau, G. Zur Entwicklungsgeschichte einiger Samen. Ber. Deutsch. Bot. Ges. 9: 274-279. pl. 17. 1891. [R. cathartica, 274-276. ] McVaucu, R. Suggested phylogeny of Prunus serotina and other wide-ranging phylads in North America. Brittonia 7: 317-346. 1952. [R. caroliniana and R. betulifolia, 338. | Maeper, R. Pharmakochemische und physiologische Untersuchung der Cortex Frangulae. Inaug. Diss. 81 pp. St. Gallen. 1925.* Petitett, F. C. American honey plants. 297 pp. front. Hamilton, Ill. 1920. [Rhamnus, 50. | Petrurson, B. The relative prevalence of specialized forms of Puccinia coronata Corda that occur on Rhamnus cathartica L. in Canada. (Abstr.) Proc. Canad. Phytopath. Soc. 21: 15, 16. 1953.* Rospertson, C. Flowers and insects. XVII. Bot. Gaz. 22: 151-165. 1896. [| Rhamnus, 157-159. SARGENT, C. S. Rhamnus. Silva N. Am. 2: 31-40. pls. 59-63. 1891. SovEcEs, R. Embryogénie des Rhamnacées. Développement de l’embryon chez le Rhamnus Frangula L. Compt. Rend. Acad. Sci. Paris 213: 39-41. 1941. [Same type as Geum urbanum VIKHIREVA-VASILKOVA, V. V. Flower and fruit anomalies in Rhamnus dahurica Pall. (In Bucsian) Bot. Zhur. 42: 85-88. 1957. Warp, M., & J. DuNLop. On some points in the histology and physiology of the fruits of Rhamnus. Ann, Bot. 1: 1-26. pls. 1, 2. 1887. [R. infectoria L.]} Wotr, C. B. The North American species of Rhamnus. Monogr. Bot. Ser. Rancho S. Ana Bot. Gard. 1. 136 pp. front. 1938. [ Monograph. | 2. Sageretia Brongniart, Mém. Fam. Rhamn. 52. 1826. Evergreen |or deciduous| spinescent [or unarmed] shrubs [or small trees|. Leaves opposite or subopposite, small [to rather large], thin- coriaceous [rarely coriaceous], pinnately veined, with 3-5 (6) pairs of 450 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV primary veins, serrulate [or entire]. Inflorescences terminal and axillary interrupted spikelike [very rarely raceme-like] often panicled thyrses for the flowers glomerulate to solitary in leaf axils]. Flowers minute, 5-merous, bisexual, sessile |or very rarely pediceled]|, 2-bracteolate at base. Floral tube short-campanulate to patelliform [or hemispherical to urceolate (?)]; calyx lobes somewhat fleshy, keeled medially on inner surface, deciduous. Petals whitish, obovate, concave [or cucullate], minutel apiculate and often notched at apex, obscurely |or distinctly] unguiculate, much shorter than calyx lobes. Stamens slightly longer than petals; anthers ovate in outline, dorsifixed near base. Nectariferous disc fleshy, collar-like, with irregularly crenate margin, confluent with floral tube at base [or distinct to base], encircling the ovary. Stigmata 3, small, sub- capitate, sometimes + confluent adaxiaily; style short, stout, 3- sulcate: ovary superior, 3-locular; ovules pleurotropous. Fruit a small, subglobular [to obovoid], purplish-black drupe with (2) 3 pyrenes; mesocarp fleshy [or leathery], thin; pyrenes thin-cartilaginous, dorsiventrally much com- pressed, obliquely obcordate, dehiscent along adaxial suture and upper half of the abaxial median.‘ Seed rather flat, obliquely obcordate, slightly keeled submedially on ventral surface, orange-brown [or light brown], usually shining; seed coat thin, crustaceous; endosperm scanty; cotyledons suborbicular-obcordate, thin, the radicle minute. LECTOTYPE SPECIES: S. theezans (L.) Brongn.; see A. Rehder, Bibliogr. Cult. Trees Shrubs 436. 1949. (Named after Augustin Sageret, 1763-1851, French horticulturist and plant physiologist. ) A primarily tropical genus of nearly 30 species, centered in eastern Asia, extending south to Indonesia and northeastern Australia, west to southern Arabia and northeastern Africa; three species in North America, one occurring from Mexico south to Peru, Paraguay, and northern Argentina. Sageretia minutiflora (Michx.) Mohr, a trailing or straggling shrub usually six to ten feet tall, with polymorphic, relatively small leaves and very fragrant flowers, occurs on calcareous sandy soils, rocky bluffs, open dry copses, and margins of low rich woods (here sometimes ascending high *The pyrenes of Sageretia ovals ard been described as agervigaes but as early as 1907 pe stated that in at least S. Brandrethiana Aitch he pyrenes are dehiscent. Recently, Grubov (FI. USSR 14: 638. 1949) mentioned a cent pyren as a generic pine in Sageretia, and Vikhireva (p. 258) described ate ee of pyrenes in S. laetivirens (Kom.) Gontsch. I have found the pyrenes dehiscent (when dried) in S, minutiflora and S, elegans, of the New World, and in S. Henryi Drumm. & Sprague, S. i ida Merr., and S. spiciflora (A. Rich.) Chiov. ex Hutch. & Bruce, of od Old Wo ageretia mintitora (Michx.) Mohr, Contr. U. S. Natl. . 6: 609. 1901. The ‘aciearahiy of this combination has ae ten pi “Trelease (1889, = ‘ International Code of Botanical Nomenclature (ed. 1961, Art. 34, Note 2) Trelease’s mention of the combination does not constitute its valid publication. 1964 | BRIZICKY, GENERA OF RHAMNACEAE 451 trees) on the Coastal Plain from southern Florida (Lee County on the west and Brevard County on the east coast) to Mississippi and South Carolina. This species seems to be more closely related to the Mexican- South American S. elegans (HBK.) Brongn. and the wide-ranging Old World S. theezans, 2n = 24, than to the Texan-Mexican S. Wrightii S. Wats. The genus seems to be closely related to Rkamnus, combining in its fruits the dehiscent pyrenes of subg. RHAMNUS, the rather flat pyrenes and seeds of subg. PSEUDOFRANGULA, and the lateral raphe of subg. FRANGULA. It differs from Rhamnus especially in the nectariferous disc which is almost completely free from the floral tube. The fruits of some species, e.g., Sageretia theezans, are said to be edible. Leaves of the same species are reportedly used as a substitute for tea in China. REFERENCES: Under family references see GEMOLL (pp. 365-368), SUESSENGUTH (pp. 54- 58), TRELEASE (1889, p. 367; 1897, pp. 405, 406), and VIKHIREVA (1952, pp. 257, 258). Cutovenpa, E. II genere “Sageretia” Brongn. in Africa. Ann. Bot. Roma 10: 431-446, pls. 5-7. 1912. [Taxonomy and morphology. ] SpracuE, T. A. The genus Sageretia in Africa. Kew Bull. Misc. Inf. 1907: 373, 374. 1907 3. Ceanothus Linnaeus, Sp. Pl. 1: 195. 1753; Gen. Pl. ed. 5. 90. 1754. Deciduous [to evergreen] unarmed [or spinescent] shrubs [rarely small trees], with 3-lacunar nodes. Leaves alternate [or opposite], mem- branaceous [or coriaceous], serrulate or crenate-serrulate to subentire [or serrate, or spinulose|, 3-nerved at base [or pinnately veined]; stipules small, thin, deciduous [or + corky, persistent]. Inflorescences terminal and/or axillary, peduncled, raceme- or corymb-like thyrses composed of few-flowered umbel-like cymes, or the flowers solitary or in few-flowered cymes in the leaf axils at the top of leafy branchlets. Flowers small, usually 5-merous [exceptionally 6—8-merous], perigynous, bisexual, white [blue to lavender], slender pediceled, the calyx and pedicels colored like the petals. Floral tube shallow, cupular-turbinate to hemispherical; calyx lobes triangular, petaloid, usually + inflexed, deciduous. Petals hooded, clawed, longer than calyx lobes, + spreading, inserted at base of disc. Stamens exserted, equalling or exceeding petals; anthers ovate-cordate in outline. Nectariferous disc annular, tumid, often distinctive in color, en- circling the upper portion of the ovary. Stigmata 3, small, subcapitate; style 3-forked in the upper third to half; ovary semi-inferior, 3-locular, incompletely so at the very base; ovule pleurotropous. Fruit a small purplish-black drupe, + 3-lobed or -cornered, smooth or crested [ridged or horned] on back of lobes or corners at their tops, supported by and adnate to the persistent base of floral tube, containing 3 coherent pyrenes; mesocarp thin, fleshy or + leathery; pyrenes cartilaginous, obovate in 452 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV outline, convex on abaxial, angular on adaxial surface, explosively dehiscent along adaxial suture and upper portion of the abaxial median, ejecting seeds. Seed narrowly to broadly obovoid, plano-convex, dark brown to black, smooth, shining, with a lateral raphe, minutely arillate at base; seed coat crustaceous; endosperm fleshy, scarce; embryo large, the cotyle- dons broadly elliptic- to suborbicular-cordate, flat, thin, the radicle minute. TYPE SPECIES: C. americanus L.; typified by the removal of two of the three original species to other genera; see A. Brongniart, Mém. Fam. Rhamn. 62, 64. 1826. (Name from Greek, keanothos, the name of some prickly plant mentioned by Theophrastus, applied to this genus by Linnaeus.) — REDROOT. An exclusively American genus of nearly 55 species, distributed from Guatemala (one species) north to southern Canada, centered in gers (44 species). Two natural, very distinct sections, CEANoTHUS, with c 36 species, and CERASTES S. Wats., with 19 species, are recognized. Four species of sect. CEANOTHUS occur in our area. Ceanothus americanus L., New Jersey tea, 2m = 24, a low shrub or subshrub up to a meter tall, with mostly ovate leaves and axillary, mostly raceme-like thyrses on elongated leafless peduncles, occurs in woods, hill- sides, ravines, and prairies from southern Quebec and southern Ontario t to Minnesota and Nebraska, south to Texas and northern Florida. Its slender-branched variant with leaves usually 2—4 cm. long (in Gleason’s delimitation), known as var. intermedius (Pursh) Torr, & Gray,® seems to represent the species on the sterile or sandy ground of the Coastal Plain, from Massachusetts south to central Florida and west to Louisiana and Arkansas. Another variant, var. Pitcheri Torr. & Gray, with lower leaf surface densely permanently velutinous, occurs scattered throughout the range of the species, but appears to be predominant in the prairies pigs Illinois west and southwest. Dried leaves of this species were used a substitute for tea during the American Revolution. Ceanothus seepoliiiatius Nutt., a low, diffusely branched shrub 20-45 cm. tall, with ovate-elliptic leaves 0.6—1.25 cm. long and axillary, few-flowered, corymb-like thyrses on elongated leafless peduncles, occurs in a few scattered localities in pine- lands on the Coastal Plain of Georgia and Florida, south to Volusia County. It is ‘closely allied to C. americanus var. intermedius, from which it differs chiefly in size’ (Trelease, 1897, p. 410). The specific status of this plant is questionable. Ceanothus herbaceus Raf. (C. ovatus Desf.), a shrub up to 1 m. tall, with oblong-elliptic to elliptic-lanceolate leaves and terminal, short- ° This variant has been variously delimited with cee to the leaf size: e.g., leaves mostly under 1 inch long (Trelease, 1888, 1897; cMinn in Van Rensselaer & Mc- Minn) ; 2-4 cm. long (Gleason, New Britt. Brown Illus. Fl NE. U.S. 2: 514. 1952); or 2-6 cm. long (Fernald, ach Man. Bot. ed. 8. 993. 1950). Such different delimitations produce rather differ s for this variety. Until extensive popula- tion studies are made and the an i ai ie the limits of this variant, the existence of a distinct range, and, consequently, its varietal status will remain un- certain. 1964 | BRIZICKY, GENERA OF RHAMNACEAE 453 peduncled, corymb-like thyrses, occurs in upland and rocky prairies, rocky woodland, open rocky places, calcareous bluffs, and dry ground, from southern Quebec to southern Manitoba, North Dakota, and Colorado, south to Nebraska, Kansas, Oklahoma, and Texas (locally in Eddy County, in southernmost New Mexico), Arkansas, locally in central Tennessee (White County), Kentucky, West Virginia, northernmost Virginia (Arlington County), the District of Columbia, and New Jersey, and, according to McMinn, in Lake County, Florida (the specimen Nash 1041 not seen by the writer). A variant with the lower surface of the leaves permanently densely villous, has been regarded as C. herbaceus Raf. var. pubescens (S. Wats.) Shinners (C. ovatus var. pubescens S. Wats.), but Soper showed that this variant, “which has generally been considered as having a more western distribution, occurs within the range of the species even as far east as the Great Lakes and northeastern Massa- chusetts, and therefore should be considered ‘“‘as merely a pubescent form.” Ceanothus microphyllus Michx., a low, diffuse shrub or subshrub 15-60 cm. tall, with leaves 3-7 mm. long, alternate and often also fasciculate in the leaf axils, and small terminal corymb- or raceme-like panicles, occurs in dry pine- or oak-woods on the Coastal Plain from Highlands County, Florida, to southwestern Alabama and southern Georgia. The terminal inflorescences suggest a close relationship to C. herbaceus. The flowers of Ceanothus usually are 5-merous and the floral anatomy very uniform, although 6—8-merous flowers with an additional vascular bundle to each extra floral segment occur regularly in C. Jepsonit. The genus is homoploid, the chromosome number being invariably 2n = 24 in the 34 species investigated (Nobs); the karyotype is also uniform. Intersectional hybrids are extremely rare and always sterile, indicating a strong genetic barrier between the sections. On the contrary, intrasectional interspecific hybrids, both natural and artificial, are very numerous. The roots of some species (e.g., Ceanothus americanus, C. velutinus Dougl.) are known to develop mycorrhizal nodules in great abundance. The presence of nitrogen-fixing bacteria in these nodules seems to be ques- tionable (cf. Furman). The fossil records indicate that Ceanothus has been present on the Pacific slope of North America at least since the Oligocene and that the differentiation of the two sections of the genus might already have oc- curred in the Miocene (cf. Nobs, p. 78). The genus seems to be closely related to Colubrina Brongn. Numerous species and hybrids of Ceanothus are valuable ornamentals. Ceanothus americanus, C. ovatus, and C. coeruleus Lag. (C. azureus Desf.), the earliest species introduced into cultivation, provided the basis for many attractive garden hybrids. Many species are important as sources of food and shelter for wildlife; some can also be used for erosion control. REFERENCES: Under family references see GEMOLL (pp. 375-380), JUEL, PRICHARD (pp. 93- 454 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV 95), SARGENT (pp. 726-729), SUESSENGUTH (pp. 72-82), TRELEASE (1897, pp. 409-417), and VIKHIREVA (1951; 1952, pp. 262-265). Apams, L., E. Steranescu, & D. J. DuNAway. Gibberellin and thiourea break seed dormancy in California Ceanothus. U.S. Forest Serv. Pacif. SW. Forest Range Exp. Sta. Res. Note 178. 4 pp. 1961.* [See also U.S. Forest Serv. Calif. Forest Range Exp. Sta. Forest Res. Notes 143. 4 pp. 1958.*] FurMAN, T. E. The structure of the root nodules of Ceanothus sanguineus and Ceanothus velutinus, with special reference to the endophyte. Am. Jour Bot. 46: 698-703. 1959. Hoim, T. Ceanothus americanus L., and ovatus Desf.; a morphological and anatomical study. Am. Jour. Sci. IV. 22: 523-530. 1906. Howe Lt, J. T. Studies in Ceanothus —I-V. Leafl. West. Bot. 2: 159-165, 202- 208, 228-240, 259-262, 285-289. 1939-1940. [Revision of the spp. of sect. Cerastes on the Pacific Coast; includes a key to the spp. Lenz, L. W. A new garden hybrid in Ceanothus. Aliso 3: 51-53. 1954. [C. americanus 2 X C. cyanus = C. cyam Lenz, intermediate between the parents. | McMinn, H. E. The importance of field hybrids in determining species in the genus Ceanothus. Proc. Calif. Acad. Sci. IV. 25: 323-356. 1944. [See also McMInn im VAN RENSSELAER & MCMINN, Ceanothus, pp. 131-279.] Mason, H. L. Distributional history and fossil record of Ceanothus. In: VAN RENSSELAER & McCMINN, Ceanothus. pp. 281-303 Noss, M. A. Experimental studies on species relationships in Ceanothus. Carnegie Inst. Publ. 623: 1-294. 1963. Petry, E. J. Germination and growth of Ceanothus americanus as affected by heated soils. Rep. Mich. Acad. Sci. Arts Lett. 22: 135-143. pls. 13, 14. ovicx, . R., & A. S. Quick. Germination of Ceanothus seeds. Madrono 16: 23- 30. 1961. | Hot-water treatment; stratification time and temperature; tolerance to boiling water. See also C. R. Quick, ibid. 3: 135-140. 1935; 15: 79-81. 1959.] Roscoe, C. W., & N. A. Harv. A preliminary study of the oe groan of Comiatins americanus and Ceanothus velutinus. Jour Phar Assoc. Sci. Ed. 49: 108-112. 1960.* [See also A. H. sey Am. a Pharm. 98: 147-156. 1926.] SARGENT, C. S. Ceanothus. Silva N. Am. 2: 41-46. pls. 64, 65. 1891. ScuMipt, M. G. Concerning Ceanothus. Natl. Hort. Mag. 35: 59-69. 1956. [Decorative value of some spp.; cultivars mostly of garden origin, some introduced more recentl SHIMEK, B. The genus Ceanothus L. in Iowa. Proc. Iowa Acad. Sci. 28: 230-242. pl. 8. 1921. [Variation of vegetative organs in C. americanus and C. ovatus. | SHINNERS, L. H. Ceanothus herbaceus Raf. for C. ovatus: a correction of name. Field Lab. 19: 33, 34 Soper, J. H. Pubescent heidi of Ceanothus ovatus. Rhodora 43: 82, 83. 1941. Sovices, R. Embryogénie des Rhamnacées. Développement de l’embryon chez le Ceanothus azureus Desf. Compt. Rend. Acad. Sci. Paris 208: 1673-1675. 1939. [Same type as Geum urbanum. | TAYLOR, G. C. Ceanothus americanus as a hemostatic. A résumé of investiga- tions into the chemistry, pharmacology and clinical use of the drug. Am. Jour. Pharm. 99: 214— 232, 397-407. 1927.* 1964] BRIZICKY, GENERA OF RHAMNACEAE 455 TRELEASE, W. hae list of ee ce species of Ceanothus. Proc. Calif. Acad. Sci. II. 1: 106-118 VAN RENSSELAER, a " rk E. ane Ceanothus. i-xiii + 1-308. front. Santa Barbara Bot. Gard., Santa Barbara, Calif. 1942. [Includes a mono- graph of Ceanothus by McMinn. Watson, S. Revision of the genus Ceanothus, and descriptions of new plants. ; Revision of the genus Ceanothus. Proc. Am. Acad. Arts Sci. II. 2: 333- 339. 1875. [Earliest monograph of Ceanothus; subdivisions Ceanothus proper and Cerastes established. Wirth, E. H. The emia a of Ceanothus americanus. Am. Jour. Pharm. : 503-514. 1926 4. Colubrina L. C. Richard ex Brongniart, Mém. Fam. Rhamn. 61. 1826, nom. cons. Unarmed or spinescent evergreen shrubs or small [very rarely large] trees. Leaves alternate [rarely opposite], membranaceous to subcoriaceous [or coriaceous], pinnately nerved or 3-nerved from base, entire or finely [to coarsely] toothed, often with small, round glands scattered on lower surface and/or 1 or 2 glands at base of blade; stipules minute, free, caducous [or connate in the axil, persistent]. Inflorescences axillary cymes or small thyrses, sessile and umbel-like or short peduncled, few flowered, and corymb-like [or flowers rarely axillary, solitary]. Flowers small, usually 5-merous, bisexual, perigynous. Floral tube hemispherical: calyx lobes triangular-ovate, spreading, conspicuously keeled on inner surface, deciduous. Petals greenish yellow to yellow or white, cucullate, sessile to short-unguiculate, shorter than calyx lobes, inserted at exterior margin of disc. Stamens nearly as long as petals; anthers ovate in outline, subbasifixed, introrse. Nectariferous disc large, fleshy, pentagonal or in- distinctly shallowly 10-lobed, filling the floral tube, surrounding the ovary and adnate to its lower half. Stigmata 3, small, obtuse; style slender. 3-lobed to -fid; ovary semi-inferior, 3-locular:; ovules pleurotropous. Fruit a subglobular, + 3-lobed orange-red to black drupe supported at the base by the adnate floral tube; mesocarp thin and dry [or + fleshy]; stone cartilaginous or crustaceous, 3-locular, splitting septicidally into 3 pyrene- like carpels dehiscent along the whole adaxial suture and halfway down abaxially. Seeds broadly obovate in outline, convex on abaxial, flattish- angular on adaxial side, brown to black, usually lustrous, smooth or some- times pitted, sometimes with a minute aril at base; seed coat coriaceous to almost bony; endosperm fleshy, thick; embryo straight, axial, the cotyle- dons flat, + fleshy, elliptic-suborbicular, the radicle small. (Including Cormonema Reissek ex Endl., and Hybosperma Urb.) LECTOTYPE SPECIES: C. ferruginosa Brongn. = C. arborescens (Mill.) Sarg.; see N. L. Britton. N. Am. Trees 681. 1908. (Name apparently derived from bois couleuvre, snakewood or serpent tree, the French name of the type species on Martinique, translated into Latin as Arbor colubrina by Jacquin, Select. Stirp. Am. Hist. 75. 1763.) 456 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV An almost pantropical genus of nearly 30 species, centered in tropical America, with a few species in southeastern Asia, Malesia, tropical Aus- tralia and Polynesia (to Hawaii), with one species in coastal East Africa and the Mascarene Islands; three indigenous and one naturalized species in our area. The genus has recently been subdivided into subg. CoLUBRINA, with entire, pinnately veined leaves, and subg. SERRATARIA M. C. Johnst., with glandular-serrate leaves often 3-nerved at base. Each subgenus com- prises two sections. Subgenus CoLuBRINA sect. COLUBRINA, with branchlets not spinescent, glands scattered on the lower leaf surface, and pyrene-like parts of the stone dehiscing explosively, includes about five tropical American species. Colubrina arborescens (C. ferruginosa, C. Colubrina (Jacq.) Millsp.), wild-coffee, a shrub or small tree with rusty tomentose branchlets, ovate to elliptic leaves 5-15 cm. long, and stout-pediceled drupes, known from the West Indies and Mexico to Guatemala, occurs in the hammocks of southern- most peninsular Florida and on the Florida Keys. Section VELLoz1a M. C. Johnst., with branchlets spinescent, leaf blades biglandular near the base, and dehiscence of the fruits not explosive, includes seven species, all tropical American. Colubrina elliptica (Sw.) Brizicky & Stern (C. reclinata (L’Hér.) Brongn.), soldierwood or nakedwood, a shrub or small tree with puberulent branchlets, ovate-elliptic to elliptic or obovate leaves, and slender-pediceled drupes, of the West Indies and Mexico to Venezuela, occurs in southernmost peninsular Florida (Dade County) and on the Florida Keys. This species “seems to be intermediate between the two sections and to link them together” (Johnston, p. 91) Subgenus SERRATARIA M. C. Johnst. sect. SERRATARIA, with leaves more or less evenly pinnately veined |or 3-nerved at base| and fruits thick walled and tardily dehiscent, is a primarily West Indian-Mexican group of nine species, extending north to California, southern Arizona, Texas, and southernmost Florida. The West Indian Colubrina cubensis (Jacq.) Brongn. is represented in hammocks of southern peninsular Florida by var. floridana M. C. Johnst., a shrub or tree with leaves narrowly elliptic- oblong to lance-oblong, 5-10 cm. long and 1.2—3.8 cm. broad, finely tomentose on both surfaces, rounded or obtuse at apex, rounded or broad- cuneate at base, with obscurely crenulate-serrulate, somewhat revolute margins, Section BARCENA (Duges) M. C. Johnst., including two Mexican and five Old World species, is represented in the coastal hammocks of southernmost peninsular Florida (north to Palm Beach County) and on the Florida Keys by the naturalized C. asiatica (L.) Brongn., a glabrous shrub with weak diffuse or prostrate branches (rarely erect and treelike) with ovate, finely crenate-serrulate leaves 3-nerved at base. The species is widely distributed in the tropics of the Old World, in Asia, Malesia, the Pacific islands, tropical Australia, the Mascarene Islands, and coastal East Africa, and is naturalized in tropical continental America and the West Indies. Seeds of this species float and are distributed by sea currents. The genus is closely related to Ceanothus. 1964 | BRIZICKY, GENERA OF RHAMNACEAE 457 The close-grained, hard and heavy wood of some species (e.g., C. arborescens, C. elliptica) is of some local importance. REFERENCES: Under family references see GEMOLL (pp. 380-382), SARGENT (pp. 729-731), SUESSENGUTH (pp. 85-89), TRELEASE (1897, p. 418), VIKHIREVA (1952, pp. 265, 266), and West & ARNOLD (pp. 136, 137). pape G. K., & W. L. Stern. Notes from the S. J. Record Memorial Col- on. II. Trop. Woods 109: 95, 96. 1958. [C. elliptica the correct name ee of C. reclinata Jounston, M. C. Novelties in Colubrina including Cormonema and Hybosperma (Rhamnaceae). Wrightia 3: 91, 92. 1963; 93-96. 1964. SARGENT, C. S. Colubrina. Silva N. Am. 2: 47-50. pl. 66. 1891. Tribe PALIUREAE Reisseck 7 5. Berchemia Necker ex De Candolle, Prodr. 2: 22. 1825, nom. cons. prop. Glabrous, deciduous, twining woody vines [or shrubs to small trees]. Leaves alternate, membranaceous, pinnately veined, with 9-12 pairs of slightly curved, subparallel primary veins, entire, often wavy. Inflores- cences raceme-like thyrses composed of few-flowered, peduncled to sub- sessile corymb-like terminal cymes (sometimes also occurring in the axils of a few leaves just below the thyrse). Flowers small, 5-merous, perigynous, unisexual by abortion [or bisexual?], the plants apparently dioecious. Floral tube saucer- to shallowly cup-shaped; calyx lobes considerably longer than floral tube. Petals greenish white, obovate or oblanceolate, somewhat concave, -+ acute at apex, not clawed in é, often very shortly clawed in ¢ flowers, nearly as long as [or somewhat shorter than] calyx lobes, inserted at the exterior margin of the disc. Stamens longer than petals in ¢, shorter and sterile in ? flowers; anthers ovate in outline, slightly cordate, dorsifixed near base. Nectariferous disc fleshy, filling floral tube, pentagonal, broad and flat in ¢ , narrow with somewhat raised interior margin and encircling ovary base in @ flowers. Gynoecium rudi- mentary in @ flowers; stigmata 2, small, subcapitate-discoid; style nar- rowly conical, usually short-bifurcate; ovary superior, 2-locular, the 2 septa distinct, overlapping each other and extending to opposite ovary wall, accumbent; ovules probably at first epitropous, later changing to pleurotropous during development. Fruit a small, bluish-black, ellipsoid to narrowly obovoid, apiculate single-stoned drupe, supported at base by remnant of floral tube; mesocarp fleshy, often juiceless, thin; stone 2- 7 The tribe Paliureae Reisseck in Endlicher (Gen. Pl. 1095. 1840) was originally composed of two genera, Paliurus Mill. and Ventilago Gaertn. In 1862, Bentham and Hooker (Gen. Pl. 1: 372) established the tribe Ventilagineae to include Ventilago and Smythea Seem. and placed Paliurus in the new tribe Zizipheae. Inclusion of sess the type genus of the tribe mieroie in Zizipheae by Bentham and Hooker makes th latter name illegitimate. The correct name for this tribe is saietin, and the tribe is to be taxonomically emended . include all the genera of Ziziphe 458 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV locular, bony, thickish, 1- or 2-seeded, indehiscent. Seed subcylindric- ovoid: seed coat membranaceous, thin, adherent and partly adnate to endocarp; endosperm fleshy, scarce; cotyledons oblong, plane, parallel with partition, the radicle short, slightly curved. (Oenoplea (Pers.) Michx. ex Hedw. f., nom. rejic. prop.; see Taxon 12: 170. 1963.) TYPE species: B. volubilis (L. f.) DC. = B. scandens (Hill) K. Koch. (Name said to commemorate either Nikolaus Berghem, alias Berchem, 1624-1683, a Dutch landscape, plant, as animal painter, or Berthout van Berchem, a contemporary of Necker. A primarily tropical genus of 22 (or fewer) species, all except one of Asia (Japan to Afghanistan, south to Java, Celebes, and Timor). The single American species, Berchemia scandens, supple-jack or rattan vine, occurs in hammocks, in rich or low woods, or in swamps from southernmost peninsular Florida west to Texas, north and northeast to Arkansas, southern Missouri, Tennessee, and Virginia; disjunct in southern Mexico (Chiapas) and Guatemala (Baja Verapaz). It is ‘“‘one of the many vines which semi-strangle trees; the tree trunks sometimes nearly grow over this vine, thus producing an ill-shaped non-merchantable trunk”’ (Brown) The genus seems to be more or less closely related to Rhamnella Miq., Berchemiella Nakai, and Chayadaia Pitard, all small genera of the Old World. The primarily East African Phyllogeiton (Weberb.) Herzog, com- prising two species, has sometimes been included in Berchemia, but such a close relationship has not been convincingly demonstrated. A few Asiatic species are sometimes grown as ornamentals. A basic, karyophilic stain, berchemine, was recently obtained from the crushed berries of Berchentia lineata (L. REFERENCES: Under family references see Brown (pp. 178, 179), HERzoc (pp. 166-168), SUESSENGUTH (pp. 141-145), TRELEASE (1897, pp. 404, 405), and VIKHIREVA (1952, p. 259). Hatustma, S. On the genus — from Japan, Korea and Formosa. Hokuriku Jour. Bot. 7(3): 69, 1958 ae L. E. A new biological ee he (In Chinese; English sum- mary.) Acta Bot. Sinica 8: 159, 160. 1959. LetsNer, R. S. An unusually large specimen of Berchemia scandens, Jour. Elisha Mitchell Sci. Soc. 73: 448. 1957. [Stem 2.4 in. in diameter, age approximately 32 years Metcatr, F. P. Relationships between Chinese and Indian Berchemia. Peking Nat. Hist. Bull. 16: 17-28. map. 1941. [Number of eastern Asiatic spp. reduced to 9.] ®’ There were, however, two of Necker’s contemporaries with the same name: Berthout van Berchem, pére, and Berthout van Berchem, fils, both active members of the Society of Natural History in Lausanne, sarneuracs The father was the author of a few articles on agriculture (e.g., on cultivation of tatoes and on soils); the son seems to have been a more prominent zoologist pasa oe and chemist 1964] BRIZICKY, GENERA OF RHAMNACEAE 459 6. Ziziphus Miller, Gard. Dict. Abr. ed. 4. 1754. Deciduous [or evergreen] shrubs or small trees [rarely woody vines], armed with often unequal stipular spines [or + axillary solitary or paired thorns, or rarely unarmed]; a few deciduous branchlets resembling pin- nate leaves often fascicled in the leaf axils. Leaves alternate [rarely sub- opposite to opposite, or fascicled], 3-nerved to or almost to the apex [or only at base, or pinnately nerved throughout], crenate-serrulate [to crenate-serrate or entire]; stipules modified to spines [or minute, cadu- cous|. Inflorescences axillary, few-flowered, subsessile to sessile [pedun- cled|, corymb-like cymes [or terminal or axillary thyrses]. Flowers small, usually 5-merous, perigynous but appearing subepigynous, bisexual, usually yellow. Floral tube shallow, subpatelliform [to hemispherical]; calyx lobes keeled medially and swollen at apex within, deciduous. Petals hooded, unguiculate, shorter than the calyx lobes [rarely wanting]. Stamens somewhat longer than petals; anthers introrse [rarely extrorse], ovate to elliptic in outline. Nectariferous disc fleshy, + flat, + pentagonal [or + 10-lobed], surrounding ovary and + adherent to it. Stigmata 2 [3], small, subcapitate-discoid; style 2[3]-furcate, swollen toward base into a stylopodium separated from ovary by a shallow groovelike constriction; Ovary appearing to be inferior or semi-inferior (becoming superior in fruit), immersed to the stylopodium in the disc, 2[3]|-locular [or incom- pletely so], the septa meeting in the center, connate [or distinct]; ovules pleurotropous, basal (Prichard) or subbasal (Nair & Sarma). Fruit an ellipsoid, ovoid to slightly obovoid [or subglobular], dark-red to black single-stoned drupe, 1.5-3[-5] cm. long; mesocarp fleshy; stone bony, ovoid to narrowly ellipsoid, sharply pointed at apex or at both ends, smooth or furrowed, 2[3]- or 1-locular by abortion, 1-seeded, indehiscent. Seed elliptic in outline, nearly plano-convex, the raphe lateral; seed coat thin, membranaceous; endosperm fleshy, scanty; embryo straight, the cotyle- dons large, thick, plane, the radicle small. Lectotypr spectes: Rhamnus Ziziphus L. = Z. Jujuba Mill.; see Britton & Wilson, Sci. Surv. Porto Rico Virgin Is. 5(1): 535. 1923. (Latin name of the type species, first introduced by Pliny; etymology obscure.) A pantropical genus of nearly 100 species, extending locally into the Temperate Zone; centered in southeastern Asia and in tropical America, where the genus ranges from southern Brazil, Paraguay, northern Argen- tina, and Peru north to the West Indies, Mexico, and the southwestern United States. Although the species of the New World form a group rather distinct from those of the Old (Johnston), no subgeneric categories have e name has been designated by various authors as being derived from Old Phoenician, zizuf; Arabic, zezaf or zefzaf; Persian, zizafun; and Greek, ziziphos. duced into some European and reas languages from Hebrew, from which the Greek word zizyphos was also derived. 460 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV been established yet. The Old World group is represented in our area by the cultivated and naturalized Ziziphus Jujuba Mill. (Z. Ziziphus (L.) Karst.), Chinese or common jujube, 2” = 24, 26, 40, 48, 74, 96, occurring rarely in thickets and on roadsides on the Coastal Plain from Alabama to Louisiana. Ziziphus mauritiana Lam. (Z. Jujuba Lam., not Mill.), Indian jujube, 2n = 48, differing in the leaves densely tomentose beneath, is also cultivated in the southern part of our area for its edible fruits. Bees are regarded as common pollinators, and cross-pollination seems to be the rule. Although self-pollination and -fertilization sometimes occur, the fruits in such cases are small and tend to drop off prematurely. Parthenocarpy and (rarely) polyembryony have been recorded in Z. Jujuba The genus seems to be closely related to the Asiatic-South European Paliurus Mill. Some species, especially the two above, are widely cultivated as fruit trees. The wood of some is of local importance. REFERENCES: Under family references see BRowNn (p. 179), HeRzoc (pp. 127-147), Mrers, Nair & SARMA, PRICHARD (pp. 85-88), SUESSENGUTH (pp. 123-132), and VIKHIREVA (1952, pp. 269-273). ACKERMAN, W. L. Flowering, pollination, self-sterility and seed development of Chinese jujube. Proc. Am. Soc. Hort. Sci. 77: 265-269. 1961. [Z. Jujuba. | Arora, N. The embryology of Zizyphus rotundifolia Lamk. Phytomorphology 3: 88-98. 1953. Curarucl, A. Partenocarpia in Zizyphus sativa Gaertn. Nuovo Gior. Bot. Ital. II. 37: 287-312. 1930. arse M. C. The species of Ziziphus indigenous to United States and Mexico. Am. Jour. Bot. 50: 1020-1027. 1963. KAJALE, L. B. A contribution to the life history of Zyzyphus Jujuba Lamk. Proc. Natl. Inst. Sci. India 10: 387-391. 1944.* LANHAM, W. B. Jujubes in Texas. Tex. Agr. Exp. Sta. Circ. 41: 1-28. 1926.* LittLe, E. L. Miscellaneous notes on nomenclature of United States trees. Am. Midl. Nat. 33: 495-513. 1945. [Ziziphus, 511.] Mame, G. O. Uber die Dornen von Zizyphus Juss. Sv. Bot. Tidskr. 14: 190-193. 1920. Monr, C. Plant life of Alabama. Contr. U. S. Natl. _ 6: 1-921. 1901. [Z. Jujuba, escaped from cultivation in Mobile Co., Sastry, B. N. Analyses of tissues of Dodonaea viscosa en and Ziziphus Oenoplia Mill. in healthy and diseased conditions. Proc. Indian Sci. Congr. 16: 242. 1929.* SRINIVASACHAR, T. Embryological studies of some members of Rhamnaceae. Indian Acad. Sci. B. 11: 107-116. 1940.* [Z. Oenoplia (L.) Mill., Z. Jujuba Lam., Scutia myrtina Merr. | SRINIVASAN, V. K. Chromosome imal in the genus Zizyphus. Curr. Sci. Bangalore 2A 224, 225.. 1952: Tuomas, C. C. The Chinese em U. S. Dep. Agr. Bull. 1215: 1-30. 1924. 1964 | BRIZICKY, GENERA OF RHAMNACEAE 461 7. Reynosia Grisebach, Catal. Pl. Cubens. 33. 1866; emend. Urban, Symb. Antill. 9: 225. 1924 Evergreen shrubs or trees. Leaves usually opposite or subopposite [rarely partly alternate], coriaceous, entire, pinnately veined, with 6—15 pairs of inconspicuous primary veins and numerous finely reticulate vein- lets visible only with magnification; stipules connate laterally at least half their length in the axil, persistent. Inflorescences axillary sessile, sub- umbellate, cymose fascicles, sometimes reduced to solitary flowers. Flowers small, yellowish green, 5-merous, perigynous, bisexual, pediceled. Floral tube short-campanulate to hemispherical; calyx lobes ovate, acute, spread- ing, deciduous. Petals wanting [or present]. Stamens shorter than calyx lobes, inserted on upper margin of disc; anthers elliptic in outline, dorsi- fixed near base, slightly versatile. Nectariferous disc fleshy, lining floral tube. Stigmata 2; style short, stout, 2-lobed; ovary superior, 2-locular, the 2 septa meeting in the center or overlapping each other and extending nearly to the opposite ovary wall, accumbent, but distinct; ovules pleurotropous. Fruit a small, subglobular to ellipsoid, purple to black, single-stoned drupe; mesocarp fleshy, thin; stone thickish, + bony, ellipsoid-subglobular, usually 1-locular by reduction of the second locule, indehiscent. Seed ovoid to subglobular; seed coat membranaceous, thin, rugose, adherent to the endocarp; endosperm copious, hard, ruminate; embryo relatively small, the cotyledons oblong or elliptic, flat, the radicle elongate. Lectotype spectrs: R. retusa Griseb.; see N. L. Britton, N. Am. Trees 673. 1908. (Named after Alvaro Reynoso, 1830-1888, Cuban agriculturist and chemist.) A West Indian genus of about 16 species. Subgenus NEOREYNOSIA Suesseng. includes nine species with petaliferous flowers, while subg. REYNOSIA includes seven, one in our area, all with apetalous flowers. Reynosia septentrionalis Urb., darling plum, red ironwood, a shrub or small tree of Cuba and the Bahamas, occurs in hammocks on the Florida Keys and in peninsular Florida (‘‘north on east coast to central Florida,” according to Little, U. S. Dep. Agr. Handb. 41: 361. 1953). Fruits of this species are edible and are said to be pleasantly flavored. Despite its ruminate endosperm, the genus seems to be closely related to the West Indian Krugiodendron and Doerpfeldia Urb., on the one hand, and to Berchemia and its allies, on the other. The close-grained, very hard and heavy wood of a few species (e.g., R. septentrionalis, R. guama Urb., R. regia Urb. & Ekm.) seems to be of some local importance. REFERENCES: Under family references see HERzoG (pp. 154-158), SARGENT (pp. 720, 721), SUESSENGUTH (pp. 136-138), VIKHIREVA (1952, pp. 260, 261), URBAN, and West & ARNOLD (p. 138). SARGENT, C. S. Reynosia. Silva N. Am. 2: 19-22. pl. 56. 1891. 462 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV 8. Krugiodendron Urban, Symb. Antill. 3: 313. 1902. Evergreen trees or shrubs. Leaves opposite or subopposite, thin and firm to coriaceous, entire, pinnately veined, with 4—6 pairs of slightly curved primary veins and coarsely reticulate veinlets + prominent on the upper surface; petioles short, stout; stipules small, connate laterally at the very base in the axil, persistent. Inflorescences axillary, few-flowered, shortly peduncled to sessile, umbel-like cymes. Flowers 5-merous, small, greenish yellow, bisexual, pediceled. Floral tube broadly obconical, rather flat; calyx lobes tanoulae: acute, Seman) crested on inner sariane, erect or spreading, deciduous. Petals ing. Stamens inserted under margin of disc, shorter than sepals, nner in aestivation; anthers ovate to ovate-suborbicular in outline, obtuse. Nectariferous disc fleshy, broad- annular, + 5-lobed, surrounding base of ovary. Stigmata 2, small; style short, stout, 2-lobed to -fid; ovary superior, 2-locular, the 2 septa over- lapping each other, extending to the opposite ovary wall, accumbent, distinct; ovules pleurotropous. Fruit an ovoid to ovoid-subglobular, black, single-stoned drupe; mesocarp fleshy, thin; stone bony, 2-locular (one locule often + reduced), 1(2)-seeded, indehiscent. Seed ellipsoid, + compressed dorsiventrally; seed coat membranaceous, adherent to the endocarp; endosperm wanting; embryo large, the cotyledons elliptic to subobovate, plano-convex, the radicle minute, retracted between the cotyledons. Typr species: K. ferreum (Vahl) Urb. (Named in honor of Leopold Krug, 1833-1898, German botanist, explorer of the West Indian flora. ) The single species, Krugiodendron ferreum (Rhamnidium ferreum (Vahl) Sarg.), black ironwood, of the West Indies and Central America (Guatemala to southern Mexico), occurs in hammocks on the Florida Keys and in peninsular Florida from Monroe and Dade counties north to Cape Canaveral (Cape Kennedy), Brevard County. The species is closely related to and perhaps congeneric with the monotypic Cuban genus Doerpfeldia Urb., which seems to differ from Krugiodendron mainly in the alternate leaves. The close-grained, hard wood is the heaviest occurring in the United States (specific gravity 1.3—1.4). REFERENCES: Under family references see HERzocG (pp. 153, 154), SARGENT (pp. 721, 722), and SUESSENGUTH (p. 147), and West & ARNOLD (p. 138). SARGENT, C. S. Rhamnidium. Silva N. Am. 2: 27-30. pl. 58. 1891. Tribe GOUANIEAE Reisseck 9. Gouania Jacquin, Select. Stirp. Am. Hist. 263. 1763. Deciduous vines climbing by solitary tendrils located at apex of short sterile axillary shoots or laterally near the bases of inflorescences. Leaves alternate, membranaceous to subcoriaceous, pinnately veined, sometimes 1964 | BRIZICKY, GENERA OF RHAMNACEAE 463 + distinctly 3-nerved at base, with 4—7 pairs of ascendent curved primary veins, coarsely and + remotely crenate-serrate, each tooth usually with a small patelliform gland. Inflorescences axillary and terminal spike- or raceme-like thyrses composed of glomerules. Flowers small, 5-merous, epigynous, bisexual [or more commonly bi- and unisexual], short pediceled to subsessile. Floral tube broadly obconical to subcampanulate; calyx lobes persistent. Petals whitish or greenish white, ovate, concave, short clawed, inserted under the margin of the disc in its sinuses. Nectariferous disc epigynous, fleshy, 5-lobed [or pentagonal]; each lobe (opposite a sepal) extended into a distinct short and rounded or apically bilobed [or elongated and acuminate] staminode-like structure. Stigmata 3, small; style 3-lobed to -furcate; ovary inferior, 3-locular; ovule pleurotropous. Fruit a 3-locular schizocarp with 3 rounded commissural wings [or rarely unwinged], splitting septicidally through each wing into three 2-winged indehiscent mericarps separating from a 6-parted carpophore. Seed obo- vate in outline, convex abaxially, somewhat angular adaxially, or lentic- ular; seed coat bony, shining; endosperm scarce; cotyledons roundish, plane, the radicle very short. LecToTYPE SPECIES: G. tomentosa Jacq. = G. polygama (Jacq.) Urb.; see Britton & Millspaugh, Bahama Fl. 258. 1920. (Named after Antoine Gouan, 1733-1821, physician and botanist, author of the Hortus Regius Mons peliensis, 1762, and Flora Mons peliensis, 1765.) A pantropical genus of nearly 50 species, about 20 in tropical America, extending north to Mexico, the West Indies, and Florida; an almost equal number in southeastern Asia, eastern Australia, New Caledonia, and Poly- nesia; and the remainder in Africa, Madagascar, and the Mascarene Islands. Gouania lupuloides (L.) Urb., chewstick, with the range of the genus in America, occurs in hammocks on the Florida Keys and in southern peninsular Florida, north to Brevard and Manatee counties. In the West Indies, pieces of the stem are chewed to heal and harden the gums and to cleanse the teeth. The dried and powdered stems are used in making dentifrices. The genus seems to be closely allied to Reissekia Endl. of Brazil, and to the East Indian-East African Helinus E. Mey. ex Endl. It also shows somewhat more remote relationships to Ceanothus and Colubrina. REFERENCES: Under family references see GEMOLL (pp. 412-420), PRICHARD (pp. 97, 98), SUESSENGUTH (pp. 166-171), TRELEASE (1897, p. 419), and VrkHIREVA (1952, pp. 280, 281). TroLL, W. Vergleichende Morphologie der hoheren Pflanzen. Band 1 (Teil 1). i-xii + 955 pp. 1935-1937. [Gouania, tendrils modified axillary shoots, 851.] 464 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV A NEW DIOSPYROS FROM THE MISANTLA REGION IN MEXICO ArTURO GOMEZ POMPA * THIS PAPER IS A CONTRIBUTION to a critical study of the flora and vege- tation of the Misantla region in central Veracruz, México. The region is bounded on the north by the Nautla river, on the south by the Sierra of Chiconquiaco (also called Sierra of Naolinco), on the east by the Gulf of México, and on the west by the road from Misantla to Martinez de la Torre. This small region of about 1500 square kilometers ranges in altitude from sea level up to more than 2000 meters, has an annual rainfall of 1400 to 3000 millimeters, a rich superficial geology, and complex topography. These factors provide the ecological conditions for a rich flora and the varied vegetation types of the area. The region is also notable in being located near the northern limit of the warm, humid tropics in America. Many important tree genera of southern México, as, for example, Terminalia, Vochysia, Bernoullia, Dialium, and Vatairea, do not reach this area, and a few, such as Aspidosperma, Dussia; Chione, and Pseudolmedia apparently have their northern limit of distribution here. The ranges of many other genera, among them Brosimum, Achras, Carpodiptera, and Zuelania, extend farther north to San Luis Potosi and probably to southern Tamaulipas. On the other hand, some genera widespread in the eastern United States, such as Liquidambar, Fagus, Magnolia, and Juglans, are represented in the flora of the temperate, humid montane area in this region. It is of interest that several outstanding collectors, including Schiede, Deppe, Liebmann, Karwinski, and Purpus have collected in the Misantla region. The type localities of many species and a few genera, among them such well known taxa as Croton draco Schlecht., Quercus oleoides Cham. & Schlecht., Aspidosperma megalocarpon Muell., and Misanteca capitata Cham. & Schlecht. (the name of this genus derived indirectly from that of the town of Misantla) are in or near this area. One of the most interesting vegetation types in the area is a deciduous forest (bosque caducifolio) at an altitude of approximately 900 to 2000 meters above sea level, on ferrolithic acid soils rich in organic materials. The humidity is extremely high due to the frequent fogs. The original forest is rather well conserved because of the abrupt topography and the lack of roads. Among the most conspicuous tree species in the upper * While working on a series of Mexican collections I am using the facilities of the Arnold Arboretum and the Gray Herbarium of Harvard University, as a fellow of the John Simon Guggenheim Memorial Foundation. I want to express my gratitude to these institutions and their officers. 1964] GOMEZ POMPA, A NEW DIOSPYROS 465 canopy are Liquidambar styraciflua, Quercus affinis, and Meliosma alba reaching heights of more than 40 meters. Other notable tree species in this area are Befaria glauca, Dendropanax arboreus, Oreopanax lieb- mani, Podocarpus matudai, Symplocos coccinea, Ternstroemia sylvatica, Turpinia insignis, Magnolia schiedeana, Brunellia mexicana, and Wein- mannia pinnata. In the course of studying collections made in the Misantla area, I found a new persimmon from this deciduous forest. The description of the new species and a discussion of its relationship to other taxa of the genus are here presented. Diospyros riojae Gomez Pompa, sp. nov. Arbor 10-15 m. alta, foliis ellipticis vel oblongo-ellipticis, 4-12 cm. longis, 2—5.5 cm. latis, basi cuneatis breviter attenuatis, apice breviter acuminatis, acumine obtuso, lamina glabra’ supra et subtus nitida, costa supra impressa, nervis nervulisque numerosis supra et subtus prominentibus, graciliter elevatis. Flores ignoti. Fructus nitidus solitarius, subglobosus, pericarpio coriaceo-ligneo, endocarpio luteo, 4—4.5 cm. longo, 4-5 cm te calyce fructifero 5-lobato, lobis oblongis, 1.5 cm. longis, 0.5—0.7 cm. ati Tree 10-15 m. high; branchlets glabrous or the young shoots ead tomentose. Leaves elliptic or oblong-elliptic, 4-12 cm. long by ac wide, glabrous, shiny on both surfaces, when dry olive-green on the oa surface and brown-green on the lower, subcoriaceous, pellucid-punctate, cuneate at base and slightly attenuate, the apex faintly arcuate, shortly acuminate, the acumen obtusish; midrib impressed above, prominent beneath, with 6-11 pairs of primary veins prominent on both surfaces, the marginal nerves conspicuous; petioles canaliculate above, 6 mm. long. Flowers unknown. Fruit subglobose, about 4—4.5 cm. high and 4—5 cm. broad, glabrous, shiny, green when fresh, light brown when dry (probably immature when collected); the exocarp coriaceous-ligneous, the endocarp fleshy, drying orange-yellow. Calyx deeply 5-lobed, lobes oblong, acutish, glabrous, 0.5—0.7 cm. wide, 1.5 cm. long; style probably 5-parted. Ho.otyre: Mexico, State of Veracruz, between Chiconquiaco and Misantla, in Liquidambar-Quercus forest, with Magnolia, Meliosma, Juglans, and Turpinia, 1350 m. alt., Aug. 13, 1962, Arturo Gémez Pompa 789 (A). Isotypes, us and MEXU. This species is named in honor of the late Dr. Enrique Rioja LoBianco, who contributed greatly to the development of biology in México and under whom I studied ecology and biogeography at the Universidad Nacional Autonoma de México Although only this single fruiting collection is known, Diospyros riojae is distinctive in the shape and size of its leaves, in the glabrous, shining leaf blades, conspicuously reticulate on both sides; and in the large globose fruit. There is variation in the shape of the leaves, some being asymmetri- cal, others symmetrical, and in the size which may vary on one branch, but 466 JOURNAL OF THE ARNOLD ARBORETUM [| VOL. XLV their shape is predominantly elliptic to oblong-elliptic. In Standley’s Trees and Shrubs of Mexico, which includes the only relatively modern treat- ment available for the Mexican species of Diospyros, this new species will be identified as D. conzattii Standl. (D. pergamentacea Lundell) ' but it may be easily distinguished by the following obvious characters: DIOSPYROS RIOJAE LEAF SHAPE Elliptic or oblong-elliptic. LATERAL VEINS Prominent on both surfaces eaf VENULES Reticulate on both upper and lower surfaces LEAF SURFACES — Shiny throughout. FRUIT SHAPE Subglobose (4-5 cm. broad, 4-4.5 cm. high DIOSPYROS CONZATTII Lanceolate or ovate-lanceo- Not prominent on either sur- face of the lea Not reticulate on the upper surface. Dull. Depressed-globose (4 cm. broad, 2 cm. high). EXOCARP Coriaceous-ligneous (not Coriaceous (wrinkling on dry- wrinkling on drying). ing, CALYX LOBES Up to 1.5 cm. long. Up to 2.7 cm. long. It seems unlikely that Diospyros riojae and D. conzattii are at all closely related, in spite of their superficial resemblance. In a search for other possible relationships, all of the materials in the collections of the Arnold Arboretum and Gray Herbarium, including representatives of nearly half of the perhaps 400 species in the genus, were surveyed. It was soon evident that a number of species from widely scattered areas share to varying degrees some of the distinctive vegetative characters of D. riojae. In leaf characters, the combination of shiny, glabrous leaf blades, conspicuously reticulate on both surfaces is especially conspicuous in Diospyros riojae. Some 25 additional species show the same combination of characteristics, and about 27 more differ only in that the lower side of the blade was dull instead of shiny. Even though the taxonomy of many of the species of the genus is in doubt and the identifications of many are not reliable, it seems worth while to enumerate the species which share with D. riojae the combination having the leaf blades glabrous, shiny and conspicuously reticulate on both sur- faces: Mexico: Diospyros cuneata, D. anisandra. West Indies: D. crassinervis, D. tetrasperma, D. caribaea, D. leonis. Eastern Asia: D. metcalfii, D. susarticulata, D. ag D. nitida. India: D. choboénsis, D. crumenata, D. pbientirs mM, fleuryana, D. mun. Philippines: D. alata, D. curranii, D. pilosanthera, o viridifolia. ‘Both Diospyros conzattii, you from Oaxaca, and D. pergamentacea, described from Chiapas, are known only m_ fruiting specimens. The two seem to be con- specific, the differences between re being minor and mostly in size. They are there- fore combined here under the older name 1964 | GOMEZ POMPA, A NEW DIOSPYROS 467 New Caledonia: D. Polynesia: D. hillebrandié D. ferrea var. subimpressa, D. samoénsis. Australia: D. hebecarpa. Africa: D. mespiliformis. It should be noted that there is a wide variation in these species in leaf shape, venation, and size (only species with leaves from 2-15 cm. long Fic Diospyros riojae showing some details of leaf ea and calyx form ho 1/2 natural size). Drawn by Eileen K. Carr 468 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV were considered), but there is a striking resemblance between Diospyros riojae and two species from southeastern Asia (Hainan), D. metcalfii and D. susarticulata. These three species are similar, not only in the characters mentioned above, but in leaf shape, size, and petiole. On the other hand, they differ from each other in fruit and calyx characters. In a taxonomically difficult genus such as Diospyros in which all of the species are dioecious and in which many are known only from fruiting materials, the possibility of utilizing vegetative characters is an intriguing one, and the general similarity of leaf characters of species from both eastern and western hemispheres offers a number of interesting suggestions: 1. The presence of this combination of leaf characters in such a large number of species leads one to think that these may be linked characters that cannot be explained as a random distribution of individual characters. 2. Apparently these characters do not have any significant value for adaptation to the environment, for the species in which they are found occur in a wide variety of ecological situations. The large number of species with this combination in Asia, India, and the Philippines may well be related to the great number of species of Diospyros in that area, while the scarcity of these characters in extratropical regions is probably related to the paucity of species. (Only 12 species of Diospyros occur in subtropical regions and only 4 are known from truly temperate climates (White, 1956).) 3. According to Hiern in his monograph (1873), the 25 species mentioned above fall into a number of different sections. However, these need to be reevaluated, especially in view of some of the characters used by him in distinguishing some sections, as, for example, “stamens quite glabrous” versus ‘‘stamens more or less hairy.”” One can certainly join White (1956) and Wood and Channell (1960) in saying that revision at the subgeneric level is needed. The data above suggest that leaves can provide some useful characters for this purpose. This suggestion is supported by Bakhuizen van den Brink (1936-1941) who used some leaf characters (prominent venation) as an aid in distinguishing sections in the subgenus Diospyros (subgenus Eudiospyros) and by White (1956), who stated that “the leaves should contribute some useful secondary characters” for the revision of the subgeneric classification of the genus. However, so much remains to be learned of the importance of floral and fruiting characters that it is difficult to know just how to evaluate the vegetative characters. It is obvious, however, that these can be useful in the comparison of species (e.g., D. riojae vs. D. conzattit). 4. In spite of differences in calyx structure, and our lack of knowledge of the flowers, the vegetative resemblance between Diospyros riojae, D. metcalfii, and D. susarticulata, mentioned above, suggest that these three may furnish yet another example of relationships between the plants of the Liquidambar forests in Mexico with those of the eastern United States and eastern Asia (Sharp, 1951, 1959; Dressler, 1954; Li, 1957; Miranda. 1959). At present only D. virginiana (closely related to D. lotus, of 1964] GOMEZ POMPA, A NEW DIOSPYROS 469 eastern Asia), occurs in the eastern United States, but among the fossil species there is one which matches strikingly the leaves of Diospyros 10 jae Some 23 species of Diospyros have been described from Cenozoic deposits in North America (LaMotte, 1952), but many of these, described only from leaves, may not represent this genus. MacGinitie (1941) says, “The leaves of Diospyros are easily confused with those of Banisteria, Heteropteris, Chionanthus, Nyssa and the leaflets of various legumes.” To this list can be added several more Sapotaceae, Styracaceae, Moraceae, and Thymelaeaceae which could also be confused with some of the leaves of living and fossil ae Despite this, Diospyros wilcoxiana Berry (1916, pl. 101, fig. 1), described from the Lower Eocene of Tennessee is remarkably like the aa of D. riojae. This species has also been reported from the Eocene of Texas (Ball, 1931); and it is worth mentioning that Ball also reported as D. mirafloriana Berry a deeply 5-parted calyx, similar to that of D. riojae. Thus it seems evident that many of the species of Diospyros with leaf characters similar to those of D. riojae, which today are restricted to tropical or subtropical regions (as are all of the species w:th a 5-lobed calyx), have come from a stock once much more widely distributed and continuous in the northern hemisphere. It is also remarkable that vegetative parts that at first sight look very variable can remain relatively unchanged through millions of years. The relationships of Diospyros riojae with other species of the genus remain obscure, but it is to be hoped that an eventual revision of the genus, including the fossil remains, will one day clarify these. REFERENCES Batt, O. M. 1931. The paleobotany of the Eocene of Texas. Bull. Agr. Mech. Coll. Tex. IV. 2(5): 104. pl. 8 fig. 1. BAKHUIZEN VAN DEN BRINK, R. C. 1936-1941. tia Ebenacearum Malayen- sium. Bull. Jard. Bot. Buitenzorg III. 15: Berry, E. W. 1916. The Lower Eocene eae _ southeastern North America. U. S. Geol. Surv. Prof. Paper 91: 334. pl. 101, DressLer, R. L. 1954. Some floristic relationships ee Mexico and the United States. Rhodora 56: 81-96. Hiern, W. P. 1873. A monograph of the Ebenaceae. Trans. Cambridge Phil. S 0 LaMotte, R. S. 1952. Catalogue of ees plants of North America through 1950. Geol. Soc. Am. Mem. 51: —148. Li, Hur Lin. 1952. Floristic eee between eastern Asia and eastern North America. Trans. Amer. Phil. Soc. II. 42: 371-429. LUNDELL, C. L. 1942. Studies of American spermatophytes II. Contr. Univ. MacGinitTig, H. D. 1941. A middle Eocene flora from the central Sierra Nevada. Carnegie Inst. Publ. 354: 50. Miranpa, F. 1959. Possible significacién del porcentaje de eon bicontinen- tales en América Tropical. Anal. Inst. Biol. México 30: 117-150. 470 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV SHarp, A. J. 1951. The relation of the Eocene Wilcox flora to some modern floras. Evolution 5: 1-S. 1 : Some aspects of Mexican phytogeography. Mem. primer Congr. Me ex. Bot. (in press — = C. 3922, Diospyros conzattii, a new species of persimmon from Jour. Wash. Acad. 12: 399-400. ; “1928 Trees and Shrubs of Mexico. Contr. U. S. Natl. Herb. 23(4): 1126-1129. WHITE, F. cen Distribution of the African species of Diospyros. Webbia 11: Woop, C. E., Jr., and R. B. CHANNELL, 1960. The genera of the Ebenales in the southeastern United States. Jour. Arnold Arb. 41: 18-22. UNIVERSIDAD NACIONAL AUTONOMA DE MExIco, JarDin BoTANICco, Crupap UNIVERSITARIA, Mexico 21, D.F., MExiIco 1964 | BRIZICKY, NOTE ON CEANOTHUS HERBACEUS 471 A FURTHER NOTE ON CEANOTHUS HERBACEUS VERSUS C. OVATUS ! GEORGE K. BrIzICKY TORREY AND Gray (Fl. N. Am. 1: 264-268. 1838) seem to have been the first to study critically the genus Ceanothus (Rhamnaceae) in the eastern United States. The large number of species described from that area up to that time (nearly 10 by Rafinesque, three by Pursh) was re- duced by these authors to five, which have been recognized by all subse- quent authors dealing with the flora of North America. Although Torrey and Gray’s delimitation of these species was basically correct, some changes in the circumscription, or at least in the synonymy, of C. ameri- canus L. and in the nomenclature of the species generally known as C. ovatus Desf. are necessary. These two rather similar, but actually quite distinct, species are widely distributed in eastern North America, the first extending south to Florida, the second to central Tennessee and Arkansas in the southeastern United States. ““‘When compared with each other these two species resemble each other very much from seedling to mature plant. But characteristic of C. ovatus is, however, the [{short-peduncled| terminal inflorescences [in con- trast with the axillary, long-peduncled inflorescences of C. americanus | and the narrower leaves with stomata on both sides” (T. Holm, Am. Jour. Sci. 22: 530. 1906). Although the leaves of C. americanus are typically ovate, rounded or subcordate at the base, and usually more or less pubescent, while those of C. ovatus are elliptical and typically glabrous, there are specimens intermediate between the two species in the shape of the leaves and the degrees of pubescence, as well. The further feature which is characteristic of C. americanus is that ‘the flowering eee |are| herbaceous, new each year” (Fernald, Gray’s Man. Bot. ed. 8. 993. 1950), a character not recorded for C. ovatus. Apparently this last character was the reason that Ceanothus herbaceus Raf., 1808, originally described as a plant with annual stems, was included in C. americanus by Torrey and Gray as var. herbaceus (Raf.) Torr. & Gray, with C. perennis Pursh in synonymy. Most botanists in the succeed- ing century generally followed Torrey and Gray in identifying C. herbaceus Raf. with C. americanus, rather than with C. ovatus. More recently, how- ever, Shinners proposed (Field Lab. 19: 33, 34. 1951) to substitute C. herbaceus Raf., “1809,” for C. ovatus Desf., 1809, on the basis of the presumed misapplication of the latter name. Although this proposal ap- ‘Continuing a series of miscellaneous notes and papers on the flora of the south- eastern United States made possible through the interest and support ef George R. Cooley and a grant (GB-171) from the National Sane. Foundation 472 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV parently has not met with general recognition, it induced me to investigate further the taxonomic and nomenclatural status of C. herbaceus Raf. A few botanists prior to Shinners identified Rafinesque’s species, described from “near the falls of the Potowmack, between the rocks,” with that of Desfontaines, described from material cultivated in France. Thus, in 1919, Hitchcock and Standley (FI. District Columbia, p. 201) noted under C. ovatus, “Plants from our region were described by Rafinesque as C. herbaceus.” The authors of The Flora of Vermont (E. J. Dole, editor. ed. 3. p. 186. 1937) listed C. Aerbaceus in the synonymy of C. ovatus. Finally, Merrill (Index Raf. p. 162. 1949) referred C. herbaceus Raf., Med. Repos. II. 5: 360. 1808, to C. ovatus Desf., but quite inexplicably regarded C. herbaceus Raf., Jour. Bot. Desvaux 1: 227. 1809, as (. americanus, although Rafinesque’s article in the Journal de Botanique was an exact translation of that in the Medical Repository. Despite the discrepancies in the taxonomic treatment of Rafinesque’s species, no attempts were made to reinvestigate its taxonomic and nomen- clatural status until 1951, when Shinners proposed to substitute C. herbaceus Raf., 1809, for C. ovatus Desf., 1809. The conclusion he reached is correct, but the basis for the substitution is not. Having ex- amined some recent collections of Ceanothus from the type locality of C. herbaceus Raf. and having compared the original description of the latter with that of C. ovatus, Shinners concluded that 1), although the type of Rafinesque’s species was probably not preserved, his species is identifiable by the collections from the type locality; 2) C. herbaceus Raf., 1809, is conspecific with C. ovatus auct., not Desf., and with C. ovalis Bigel., 1824; 3) C. ovatus Desf. is an unidentified cultivated species (perhaps a hybrid), and “perhaps mere resemblance in name was an inducement to substitute C. ovatus for C. ovalis”; and hence, 4) C. herbaceus Raf., 1809, is to be substituted for C. ovatus Desf., 1809, the name of a different and unknown species. The circumstance that in the vicinity of Washington, D. C., Ceanothus ovatus Desf. “is confined to the Potomac shore, where it inhabits the rocks at ‘Little Falls’” (Holm, p. 523; see also L. F. Ward, Guide Fl. Wash., p. 72. 1881, and Hitchcock & Standley, loc. cit.), while C. americanus occurs in dry copses or in open fields (Holm, Joc. cit.), fully supports Shinners’ conclusion regarding the conspecificity of C. herbaceus with C. ovatus, but his statement that C. ovatus Desf. is an unknown species, different from that to which this name was applied, is incorrect. In their treatment of Ceanothus, Torrey and Gray (Fl. N. Am. 1: 264, 265. 1838) at first placed C. ovatus Desf. in the synonymy of C. ameri- canus |var.] y. herbaceus (Raf.) Torr. & Gray, considering it different from C. ovalis Bigel. After having seen an authentic specimen of C. ovatus Desf., these authors (doc. cit., p. 686) removed the latter from the synonymy of C. americanus var. herbaceus to C. ovalis Bigel. with the remark, ‘‘To this belongs C. ovatus Desf., which is the prior name, but less appropriate, as the leaves are never ovate.” This authentic specimen of C. ovatus Desf., with the label “Herb. Webbianum. Ex Herb. Desfontaines,”’ with 1964 | BRIZICKY, NOTE ON CEANOTHUS HERBACEUS 473 pencil marks on the sheet “W. T.” [William Trelease| and “‘|type],”’ is preserved in the collections of the Gray Herbarium. Since Ceanothus ovatus Desf. is clearly typified in the sense in which the name has been used since Torrey and Gray, Shinners’ second to fourth conclusions are untenable. There is, however, another reason for taking up Ceanothus herbaceus: the priority of Rafinesque’s name over that of Desfontaines. Ceanothus herbaceus Raf. was first published in 1808, not in 1809, thus antedating C. ovatus Desf., of 1809, and being the correct name for this species of northeastern and central North America. Ceanothus herbaceus Raf. Med. Repos. N. Y. II. 5: 360. 1808, “Canothus,” sphalm.; Jour. Bot. Desvaux 1: 227. 1809. C. ovatus Desf. Hist. Arbr. Arbriss. 2: 381. 1809. C. ovalis Bigel. Fl. Boston. ed. 2. 92. 1824 Topotype collections of C. herbaceus Raf. seen: In vicinis of Wash- ington, D. C., L. F. Ward, July 18, 1880 [fruiting specimen] (cH). (The same sheet also bears a flowering specimen collected June 4, 1881.) 474 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLV THE DIRECTOR’S REPORT Tue ARNOLD ARBORETUM DURING THE FISCAL YEAR ENDED June 30, 1964 FROM THE POINT OF VIEW of active participation and representation in the many areas of its interest, the year just completed was a varied and successful one for the Arnold Arboretum. Several staff members made extended and intensive field trips which yielded botanical and horticultural collections, series of photographs, and impressions of value which will con- tribute significantly to their current and future work. The number and quality of published contributions by the staff continued at a high level. The Arboretum was represented at several professional meetings, and. in turn, welcomed many distinguished visitors. At the open house scheduled for visitors on successive week-ends in May a moderate number came to Weston to the Case Estates. At Jamaica Plain, however, which is more easily accessible, the crowd was almost over- whelming in its enthusiastic pleasure with the flowering display. Staff: On January 1, Dr. Lily M. Perry retired officially from the staff of the Arnold Arboretum. Miss Perry, who joined the Arboretum staff in 1937, worked for a number of years closely with the late E. D. Merrill and was coauthor with him of a long series of papers on the flora of New Guinea. More recently, Dr. Perry has been compiling information on the medicinal plants of southeastern Asia. When this manuscript is published she plans to return to her major interest, work on the New Guinea flora, and will continue to be of assistance to the Arboretum staff on problems of the southeastern Asiatic area. Her long and devoted service has been of inestimable value to the institution and to all her colleagues. During the year also, Mrs. Susan Delano McKelvey requested her re- tirement as a member of the Committee to Visit the Arnold Arboretum and as a Research Associate on the staff. Her services, which have been continuous since 1920, are deeply appreciated. Dr. Lorin I. Nevling, Jr., was appointed Associate Curator and Super- visor of the Herbaria, jointly with the Gray Herbarium, to be in charge of the herbarium collections in Cambridge and in Jamaica Plain. Dr. Shirley A. Graham, who had recently completed her doctoral studies at the University of Michigan, joined the staffs of the Arboretum and the Gray Herbarium in September to work for a year on the Southeastern Flora Project. 1964] THE DIRECTOR’S REPORT 475 Mr. Andrey Baranoy was appointed Curatorial Assistant. His special abilities with the flora of Manchuria will aid the staff in handling the accumulated collections from eastern Asia. The appointment of Dr. Mary . Sanders as a Research Associate was renewed, as was the Mercer Fellowship awarded to Dr. Lalit Srivastava. Dr. Howard was appointed to the Scientific Advisory Committee of the Fairchild Tropical Garden and to the Committee on Pioneering Research of the National Academy of Sciences — National Research Council. Dr. Wyman was elected a Vice President of the International Dendro- logical Union and continued his services as a Director of the American Horticultural Society and as a Trustee of the Massachusetts Horticultural Society. At the International Shade Tree Conference held in Canada, the “Author’s Citation” of that organization was awarded to Dr. Wyman “‘for sustained production of excellent books in the field of arboriculture.” Horticulture: The climate of the past year was again unusual. The summer was dry with prolonged drought so that when the first rains came on October 1 there was a deficiency of ten inches in an expected normal rainfall, to that date, of 31 inches. During the winter, however, total precipitation became normal, Precipitation as snow, moreover, was the heaviest in 17 years, with a total of 63 inches. No severe storms were experienced, but an unusual one occurred on March 10 when sleet fell for eleven hours, ac- cumulating to a depth of two inches before turning to snow. The effect of these conditions on the grounds was not severe. The dry summer promoted the formation of many flower buds in the dogwoods, which produced an outstanding display in the spring. Some branch damage can be attributed to the summer dry spell, while during the winter, rabbit and mouse damage to trunks and stems of smaller shrubs was heavy under the deep snow cover. A major improvement on the grounds is now under way. The construc- tion of the first segment of a roadway to the top of Peters’ Hill, mentioned in the report of the previous year, is a costly operation, and it was expected that completion might be delayed several years. The Department of Parks and Recreation, however, recognized the need for this road and turn- In the care of the living collections a particular effort was concentrated on the area near the Forest Hills gate. Dieback of the Prunus collection necessitated the removal of many of the older plants, which now have been replaced with dogwoods to allow a period of rest for the soil. Newer accessions of cherries or propagations from the older plants are being placed elsewhere on the grounds in an effort to limit the effect of virus and nematodes. Portions of the shrub collection have also been replanted. Experience has shown that this area near the Forest Hills gate collects cold air during the winter season and that many Leguminosae have been affected by the colder environment. These have been moved to other 476 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV oan ‘e ie ye) we GF. oe - i “Winter Splendor.” The result of a February snow storm in the Arnold Arboretum. Photo. courtesy of the Boston Globe. locations, while a miscellany of shrubs is being planted experimentally to determine family or generic hardiness in that particular spot. Because of the dry fall, planting was held to a minimum, but the early, cool spring offered an extended period for moving plants from the Case Estates to Jamaica Plain. A total of 558 plants, representing 250 taxa, was planted. Included in this number were 34 taxa of Salix for improvement of the collection in the meadow area. 1964 | THE DIRECTOR’S REPORT 477 It has been possible to continue use of cocoa shells for mulch, and approximately 566 tons were used during the year, with particular success under various tree collections. A “mulching material’? being used around small-stemmed plants is sheeting of black polyethylene film, the effective- ness of which was described by Dr. Wyman in a recent issue of Arnoldia. Faulty drainage caused by an underlying layer of compacted soil necessitated special work on two nursery beds in the area of the Dana Greenhouses. After the top soil was removed and a hardpan layer broken, drainage tile was installed and covered with ten inches of sand and gravel. An enriched top soil was replaced and left over the winter, apparently completely correcting the problem. The regular activities of the greenhouse staff were carried on as usual, or increased in certain directions. During the past fiscal year the Arbo- retum received 153 shipments of whole plants of 731 taxa from 12 coun- tries, including the United States, and 159 shipments of seeds of 485 taxa from 43 different countries. The increase in these figures this year reflects the travels of many staff members who sent back plant materials or re- quested additional materials seen on their travels for study or horticultural trial. On the other hand, the Arboretum sent in response to specific re- quests, or for its own research purposes, 159 shipments of plants of 969 taxa to 7 countries and 56 shipments of 172 taxa of seeds to 13 countries. Included in the shipments of living plants were specimens of selected taxa considered worthy of more extensive growth. Under the program for co- operating nurserymen the Arboretum distributed eight taxa to 57 nurseries and botanical gardens, or a total of 549 plants. One of these, Hamamelis intermedia ‘Arnold Promise’, represented a new selection developed on the Arboretum grounds and reported in detail in a recent issue of Arnoldia. Six of the remaining seven taxa were plants recently or originally intro- duced to the United States by the Arnold Arboretum. All have been tested for hardiness in Weston and have proven to be most desirable cultivated plants not otherwise available in the United States. Also, 266 taxa were propagated for eventual use on the grounds in Jamaica Plain or in Weston: and 186 taxa were propagated for the use of the staff. Previously reported research projects of the greenhouse staff are con- tinuing. Among the projects newly initiated during the year is a study concerned with the processing of seeds of second generation hybrids of Rhododendron proven hardy at a temperature of minus 45 degrees Fahren- heit. Seeds for this project were donated by Mr. Frank Abbott, of Saxons River, Vermont. Another new project concerned study of progeny from seed of abnormal growth forms, or witches’ brooms, of pine and hemlock. Seed from cones of these abnormalities produce both normal and abnormal seedlings which are now being studied to determine whether the abnormali- ties are lost or whether they may become lethal. At the annual meeting of the American Association of Botanical Gardens and Arboretums in St. Louis, the Arnold Arboretum was reappointed for the third time national registration authority for woody plants not already represented by designated individuals or societies. During the past year 478 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV one registration list was published, for Fagus; work continues toward the completion of lists for Ulmus and Weigela within the calendar year. Dr. Dudley has continued his work on Alyssum, chiefly on floristic studies, but has initiated, also, hardiness trials of more than 40 species, in order to determine the full potential of this genus as a cultivated plant for New England. Mr. Green has organized the collections made on his trip to the Pacific. The new collections and data on various Oleaceae are being studied. A few of the recent introductions of the family formerly under plant quaran- tine restrictions have been cleared and will be distributed to other areas for trial. Case Estates: The interest in the plantings on the Case Estates in Weston continued to increase during the year. Past publicity and the maturity of some of the special collections have increased the educational value and the effective- ness of the displays. Additions were made to the trial plots of Narcissus and to the collection of Lilium. There are now more than 300 taxa of lilies, including one special gift of 52. In August, the New England section of the North American Lily Society made a special visit to the demonstration area. Additional beds have been prepared for display collections of Hemerocallis, Iris, and Allium, in the hope of developing a representation of the units of classification of each group, both for study by classes and for exhibition. The American Hemerocallis Society has offered material and financial support for work on those plants in which its members have special interest. The perennial garden and the ground covers have had the majority of old labels replaced with engraved plastic labels. Roads at 84-86 Wellesley Street were resurfaced in accordance with obligations to maintain access to this property, and one section was re- located for easier maintenance in winter snows. A commercial survey of property lines along the Ash Street section was necessary following a petition by a neighboring property owner for con- struction at variance with town regulations. The boundaries of the Case Estates in this area, where 86 cultivars of Chaenomeles obtained in connec- tion with the work of Dr. Weber have been planted, are now marked. An open house was held on May 10 Herbarium: The most conspicuous changes in the herbarium were made in the facilities for housing and studying cultivated plants in the Administration Building in Jamaica Plain. Forty-three new herbarium cases were pur- chased and installed on the second floor gallery, doubling the case capacity of the herbarium in the past three years. This entire section of the herbarium has been redecorated and new fluorescent lighting has been installed. The third floor gallery has also been thoroughly cleaned and painted, and the unmounted collections stored there have been re-examined 1964] THE DIRECTOR’S REPORT 479 for study, mounting, and insertion in the general herbarium as soon as possible, in order to make them available to the scientific community. The entire horticultural herbarium has been redistributed in the cases now available, both to increase working space and to allow for more rapid insertion of specim During the year 21,951 specimens were mounted and added to the herbarium collections in Jamaica Plain and Cambridge. The total number of sheets in the herbarium is now 785,660. During the year, 14,931 specimens were received, the greater portion by exchange. The areas represented in order of size were India, Papuasia, the West Indies, and Western Malaysia. A large collection was received from the Royal Botanic Garden in Edinburgh, comprising the collections of G. H. Cave from the mountainous areas of India. This is a particularly valuable addi- tion to our Asiatic collections. During the past year, staff members and students requested or received for identification 6,219 herbarium sheets representing 81 loans from 4 institutions, 23 in the New World, 17 in the Old World. The staff filled loan requests for 141 loans to 67 institutions (41 in the United States and 26 foreign), and sent out 14,617 specimens. The taxonomists of the staff have continued and intensified their re- search studies, partly as a result of recent field work on special groups. Dr. George K. Brizicky, who has completed work on the Celastrales and Sapindales for the generic flora of the southeastern United States, has ready for publication a study of the genera of Rhamnaceae of the same region, and is now studying the genera of Vitaceae. Dr. T. R. Dudley is continuing studies of Alyssum and its relatives, of the family Cruciferae, particularly the representatives from Turkey and adjacent areas. Dr. Shirley A. Graham has completed treatments of the Lythraceae, Elaeagna- ceae, Rhizophoraceae, and Combretaceae for the generic flora of the south- eastern United States, and is continuing work on the Polygonaceae and Araliaceae for the same project, as well as studies on the genus Cuphea. Mr. Peter Green’s work has been augmented by his new Pacific collec- tions of Oleaceae which will provide material basic for solving some of the numerous problems in that family. Dr. Howard is continuing taxonomic studies on various West Indian genera, toward completion of a flora of the region. Dr. Hu is continuing studies in the genus //ex and on other prob- lems in the Asiatic flora. Dr. Nevling is carrying on his studies in the Thymelaeaceae, a family in which many genera are poorly understood because of inadequate material. Dr. Schubert is continuing studies in the genera Dioscorea, Desmodium, and Begonia, material of which was col- lected in Mexico in the autumn. A collection of about 100 specimens of Desmodium was received during the year from the Department of Agronomy and Soil Science of the University of Hawaii, consisting of grown in breeding programs for development of forage and for other uses. Dr. Wood has continued his supervision and editing of the botanical studies and illustrations for the generic flora of the southeastern United 480 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV States for which he is also preparing treatments of Gentianaceae and Loganiaceae and a comprehensive glossary. Dr. Otto Solbrig, of the staff of the Gray Herbarium, contributed a treatment of the Tribes of Com- positae for the Southeastern Flora Project, and Dr. Wallace Ernst, formerly attached to the same project, completed studies on the Berberidaceae, Lardizabalaceae, and Menispermaceae, which were published during the ear. Dr. Claude Weber, a former graduate student, whose work on CéAae- nomeles has already been mentioned, completed the taxonomic and mor- phological portions of her work which are in the process of publication. Library: The task of maintaining a highly specialized library so that current needs are met and gaps in older literature are filled, is a demanding one. Thanks to the acuity and energy of Mrs. Schwarten, however, our library continues in its superior position. Two of the interesting rare volumes added in the past year are the Curioser Botanicus of Samuel Mueller, published in Dresden and Leipzig, in 1730, a notable addition to the pre-Linnaean col- lection, and not known to be in any other library in the United States; and a first edition of John Lindley’s The Vegetable Kingdom, which com- pletes our series of this work. Three hundred forty-five bound volumes, obtained by purchase, ex- change, or as gifts were added to the library during the year. Of these, one hundred concerning horticultural subjects are housed in Jamaica Plain. The total number of bound volumes is now 52,217. Three hundred eighty pamphlets and reprints increased the total of that collection to 19,824. The 1500 cards of the 1963-64 issues of the /ndex to American Botanical Literature published by the Torrey Botanical Club, 3000 cards of the Card Index of American Plants issued by the Gray Herbarium, and sets 18 and 19 of the /ndex Genericorum were all added to existing files in Jamaica Plain and in Cambridge. In a joint project with the library of the Gray Herbarium the microfiche reproductions of the historical botanical collections of various herbaria are purchased and housed in the library for maximum protection. During the year microfiche reproductions of the basic herbarium at the Botanical Museum and Herbarium at Copenhagen, Denmark, were made available, along with those of some of the special herbaria of that institution (among them the collections of Isert and Thonning, Vahl, Forskal and Rottboll). The microfiche reproductions have great value in taxonomic research, and continuation of this project will be most desirable. We are always grateful for the donations of books, manuscripts, and biographical materials to the library, and acknowledge with pleasure the useful gifts during the year from Mrs. R. W. Bliss and Mrs. E. Corning. Comparative Morphology: In continuation of his studies on the comparative anatomy of the leaf- bearing Cactaceae, Professor I. W. Bailey has found, in the genus Pereskia, 1964] THE DIRECTOR’S REPORT 481 that the characters of cuticle, epidermis, and stomata of leaves and young stems prior to periderm form ation offer significant evidence for the differentiation of taxa in Pereskia, as well as for distinguishing Pereskia from other genera. In a study of the occurrence of crystals in the leaves of Cactaceae his studies indicate that two morphological and biochemical categories of crystals are present. Further detailed investigation is called for which may elucidate the evolutionary changes in metabolism essential for the survival of these plants in arid environments. Dr. Lalit Srivastava completed a review article on the anatomy, chemis- try, and physiology of bark which will be published in the /nternational Review of Forestry Research. In addition, he has surveyed the distribution of lignin in the bark of numerous angiosperms and gymnosperms. This work will be extended to determine the site of lignin synthesis and the possible relationships of synthesis of lignin and other phenolic compounds. Dr. Srivastava is engaged also, in an electron microscope study of living cells in leaves and bark to determine the changes in cell protoplast during the period of frost hardiness. Dr. Howard has expanded his examination of the nodal structure and the vascular anatomy of the petiole to additional plant families. The abundance of material accumulated during field work in Africa, Mauritius, Australia, and New Guinea has added 15 families and over 400 genera to the survey. A new nodal-petiolar pattern has been found in the Crassula- ceae and a general occurrence of the type no. 4 nodal structure, accompanied by a unique vascular pattern, has been found in Zygophyllaceae. During the year additions to the slide collection included many prepared from woods collected by Dr. A. C. Smith in the Fiji Islands. Slides of the Passifloraceae have been obtained in a program of exchange with the United States National Museum. A program of exchange of documented wood samples has begun with the Forest Products Research Institute, Laguna, The Philippines. A large number of samples of woods from Borneo was received from the Forestry Department in Sandakan. Cytology and Genetics: Dr. Mary Sanders has completed a survey paper on artificial culture of embryos. She reports on her continuing work on the cytology and genetics of sorghum that a search is being made for cytological evidence to support the hypothesis, developed from genetic evidence, that colchicine-induced mutants result from chromosome changes. Chromosomes are being ex- amined in the parent line, in six mutant lines, and in reciprocal F, hybrids between parent and each mutant line. An investigation of colchicine treat- ment of four tetraploid sorghum lines, and of their corresponding diploid lines, also is in progress, in which diploid mutants have been obtained from both tetraploid and diploid seedlings. Mrs. L. Rudenberg, Research Associate of the Gray Herbarium, has developed an interest in verifying cytological counts for plants in the living collections of the Arnold Arboretum. Regrettably, much of the early work published on chromosome counts of cultivated plants in the Arbo- 482 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV retum was not documented. Mrs. Rudenberg’s present work involves plants under taxonomic study, such as Psychotria, or those soon to be studied. Chromosome numbers and cytological peculiarities are noted, with annota- tions made on herbarium vouchers and on the master locator cards in the Arboretum files. A portion of Dr. Nevling’s investigations of Daphnopsis species is also devoted to cytological studies which, up to now, have been inadequate for many representatives of this family. During the past year the first counts were published for species of Daphnopsis obtained as part of the field pro- gram in Puerto Rico last year. Education: No formal courses were offered by the Arboretum staff members during the year, but Messrs. Green, Howard, and Wood took part in presenting Biology 247, a course in plant geography. Graduate, undergraduate, and special students were assigned to staff members during the year for super- vision of special research projects. These involved such diverse investiga- tions as a review of the species of Hopea in the Philippine Islands, prepara- tion of a check list of the cultivated plants of St. Croix, and study of the origin of homostylous pollen-trimorphic plants of Psychotria guadalupensts. The informal Fall and Spring Field courses conducted by Dr. Wyman, Dr. Dudley, and Mr. Green in Jamaica Plain and Weston were well at- tended. Mr. Fordham again conducted a course in plant propagation using the facilities of the greenhouses. Dr. Howard, assisted by Dr. Wood, conducted a course in economic botany which considered the food plants offered by restaurants in the Boston area, for a selected group of dieticians. The full staff participated in taxonomic seminars held in the Harvard University Herbarium in Cambridge. Staff members attended scientific meetings associated with their special research interests. Dr. Howard and Mr. Fordham presented papers during the regional meetings of the American Society of Horticultural Sciences held in Cambridge. Mr. Fordham was a speaker at the annual meeting of the Tennessee Nurserymen’s Association, at the meetings of the Inter- national Plant Propagators Society in St. Louis, Missouri, and for the short course of the Connecticut Nurserymen at Storrs, Connecticut. Dr. Wyman was the speaker at the Awards Banquet of the American Horti- cultural Society in St. Louis, and at the annual meeting of the Garden Club Federation of Massachusetts. Dr. Howard gave the Laura L. Barnes lecture for the Morris Arboretum in Philadelphia, spoke in the lecture series of the Royal Canadian Academy in Toronto, and to the New York Horticultural Society. During his Australian trip he was invited to speak to the western section of the Australian and New Zealand Association for the Advance- ment of Science, in Perth. Messrs. Dudley, Green, and Howard were all speakers for the Boston Horticultural Club. Dr. Schubert served as chairman of a section of contributed papers and presented a paper on Desmodium at the Second Mexican Botanical Congress 1964 | THE DIRECTOR’S REPORT 483 An exhibit of the Arnold Arboretum, in cooperation with the ataig Cooma Committee of the Garden Club of America, at the International Flower Show i New York. Photo. by J. Hugelmeyer. Dr. Wood was invited to speak on modern distributional patterns in seed plants in a symposium on “The Origin and Evolution of the Biota of the Southeastern United States,” during the annual meeting of the Society for the Study of Evolution held at the University of North Carolina, Chapel Hill, North Carolina. These, and many other talks given by members of the staff, at the Arboretum or in other locations, form an important part of the educational contribution of the Arnold Arboretum. The preparation of mimeographed leaflets for distribution by the Garden Club of America represents another aspect of the educational program. Exhibits and Displays: Two large exhibits were prepared by the staff during the year. An exhibit on the propagation of woody plants by seeds was prepared at the request of the Horticultural Committee of the Garden Club of America for the International Flower Show in New York. This was similar to the exhibit last year at the Boston Flower Show, and was awarded the Director’s Trophy. The staff prepared for the Boston Spring Flower Show of the Massachusetts Horticultural Society, in Boston, ‘‘an exhibition of mulching materials.” A garden using different kinds of mulching materials formed the background for a display of wooden bowls containing samples 484 JOURNAL OF THE ARNOLD ARBORETUM VOL, XLV sti es oc. Krk ° —~r +," 2° £48" SF, ists OEY 8 ey x 3 » = Top: Administration building of the Royal Botanical Garden, Pamplemousses, Mauritius. A garden visited by Dr. Howard for its importance in the transfer of economic plants from the East Indies to the West Indies. Bottom: The exhibit of mulching materials presented by the Arnold Arbore- tum at the Spring Flower Show of the Massachusetts Horticultural Society, 1964. 1964 | THE DIRECTOR’S REPORT 485 of the materials at close range for visitors’ viewing. This exhibit was awarded a gold medal and an educational certificate. Travel and Exploration: The opportunity to observe, collect, and study plants in the field and under cultivation is both a privilege and a requirement for the botanist. During the past year staff members of the Arboretum worked close to the Arboretum and in distant countries. From their efforts have come speci- mens dried as herbarium material, preserved for anatomical or cytological studies, photographed for teaching and lecturing, as well as living material for growing in greenhouse or out-of-doors for scientific study and for horticultural application. Dr. Dudley undertook the task of adding to the representation of culti- vated plants in the herbarium. We appreciate the kindness of owners of private gardens who gave permission for the preparation of herbarium specimens from their cultivated plants in localities at Mt. Desert Island, Maine, Chatham, Cape Cod, Massachusetts, near Albany, New York, and Philadelphia, Pennsylvania. Specimens of representative holly cultivars were also made available after the meeting of the Holly Society of America at the Callaway Gardens, Georgia. Mr. Green, with the partial support of a grant from the National Science Foundation, made a trip to the Pacific to study plants of the Oleaceae and to make general collections in a few areas. His itinerary included the Hawaiian Islands, northern and western Fiji, New Caledonia, the north island of New Zealand, Norfolk Island, New South Wales, Victoria and Western Australia, and Lord Howe Islands. His collections from Norfolk and Lord Howe Islands were nearly all new to American herbaria and those of New Caledonia only slightly less so. Collections of certain families were made for specialists; and cuttings and seeds of Oleaceae and special groups for further study, or for possible horticultural introduction, were sent to the Arnold Arboretum, Foster Garden in Honolulu, the University of California Botanic Garden at Berkeley, the Los Angeles State & County Arboretum, Melbourne Botanic Garden, and the Royal Botanic Gardens at Kew and Edinburgh. Cytological collections made for Dr. Barbara Briggs of Sydney have yielded chromosome counts for 12 species in five genera of the Oleaceae not previously reported and a count for the puzzling genus Oceanopapaver. Dr. Otto Solbrig obtained clear counts from Sciaphila buds from New Caledonia, in a family which is scarcely known cytologically. Dr. Howard spent two and a half months in the autumn on a trip around the world. He attended the plenary sessions of the AETFAT (a society devoted to the study of African botany) in Genoa, and then with most helpful assistance of botanists in many areas, was able to visit gardens and herbaria in Egypt, Ethiopia, Kenya, Mauritius, many parts of Australia, New Guinea, and Fiji. The initiative for the trip was an invitation to take part in a botanical tour commemorating the 50th anniversary of the founding of the Kirstenbosch Botanical Garden and the Natural History 486 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Tor: Lorp Howe IsLanp; looking southeast from near Malabar towards Mt. Lidgbird (2547 ft. alt., left) and Mt. Gower (2833 ft. alt., right) Bottom: Hisiscus INSuLARIS Endl. on Philip’s Island (near Norfolk Island) where it is endemic, and of which only four bushes remain. Seed and cuttings were obtained. 1964 | THE DIRECTOR’S REPORT 487 Society of South Africa. After meetings in Capetown and local tours to places of botanical interest, the group of 41 visiting botanists representing 21 countries was taken on a 5,000 mile, 30-day tour of South Africa. As a botanical and horticultural introduction to the local vegetation and to the many introduced plants the tour could not have been better planned. Dr. Howard photographed many plant families and genera not usually available for use in classes, and also collected seeds, cuttings, and morphological material for further study. In April Dr. Howard, with Mr. Draper made a short trip to Puerto Rico to establish a small test garden for research purposes on land offered to the Arnold Arboretum on the south side of the Luquillo Mountain range. Dr. Hu studied //ex collections at the National Arboretum in Washing- ton, and in several nurseries in Maryland. She also examined collections of the Meserve Arboretum on Long Island, in May, and obtained materials of hybrid plants for the horticultural herbarium. Dr. Nevling represented the Arnold Arboretum at the dedication of the new herbarium facilities at Michigan State University in East Lansing. At a symposium held then, concerned with the role of “The Herbarium in the Modern University,” curators from many institutions had the opportunity for discussion of professional problems. oS tie eit gle > 3 a ¥ a : F. - es ¢ 7 tae << : = . = . Some botanists on Mt. Richmond, Victoria, Australia. Left to right: Clifford Beauglehole, Dr. Richard T. M. Pescott, Dr. James Willis, Fred Davies, Noel Learmont. Photo. by R. A. Howard, guest of this field party. Dr. Schubert spent five weeks after the Mexican Botanical Congress collecting in the states of San Luis Potosi, Jalisco, and Nayarit, Mexico, and a week studying in the herbarium of the National University of Mexico in Mexico City. The cooperation of the botanists of the Botanical Garden of the National University and of the Forestry Department of the Mexican Government greatly expedited her work. Gifts and Grants: The Arboretum has gifts of many kinds in support of its work. The help of the Friends of the Arnold Arboretum which has been most generous is gratefully acknowledged. Herbarium specimens of particular interest were received from the Department of Agriculture in Bermuda and from the agricultural station of St. Croix. Books and back numbers of journals have been useful in filling some gaps in the library and in making available useful second copies for the use of the staff. Various nurseries have made gifts of cultivated plants not represented in our living collections. The American Hemerocallis Society is supplying cultivars of daylilies for a collection in Weston, along with the funds for their maintenance. Mrs. F. W. Warburton of the Median Tris Society is contributing a collection of iris taxa for display. The Jan de Graf Company is helping to complete the representation of lily taxa in the collection in Weston. Grants from the National Science Foundation helped to support the field 1964 | THE DIRECTOR’S REPORT 489 a e a % f; ‘thre’ * II Mexican Botanical Congress, San Luis Potosi, Mexico, September, 1963. work of Mr. Peter Green in the southwest Pacific; support the investiga- tions of Professor Bailey, and contribute to the support of the Southeastern Flora Project conducted by Dr. C. E. Wood. Grants for travel to the X International Botanical Congress were received by Dr. S. Y. Hu and Mr. P. Green. Publications: Under the editorship of Dr. Bernice G. Schubert four issues of the Journal of the Arnold Arboretum (now in its 45th volume) were published during the fiscal year, containing 515 pages and 23 articles, 21 oi which were by members of the staff or former students. Dr. Wyman served as editor of Arnoldia (in its 24th volume) which appears at irregular intervals. Sixty pages were issued during the year In commemoration of the founding of the oldest botanical garden in the West Indies, on the island of St. Vincent, in 1765, the Arnold Arboretum is issuing a photocopy reproduction of the Rev. Lansdown Guilding’s cata- logue of the garden, originally published in 1825. Copies of the reproduc- tion will be made available for the celebration on St. Vincent, and to libraries as well. Another special publication is a photocopy reproduction of Dr. Donald Wyman’s How to Form an Arboretum, of which a revised version was published as a number of Arnoldia in 1960. The bibliography which follows lists 67 articles and books by members of the staff which were published during the past fiscal year. Bibliography of the Published Writings of the Staff and Students July 1, 1963—June 30, 1964 BaILey, Irvinc W. Comparative anatomy of the leaf-bearing Cactaceae, X. The xylem of Pereskia colombiana, Pereskia guamacho, Pereskia cubensis, and Pereskia portulacifolia. Jour. Arnold Arb. 44: 390-401. 1963. 490 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV . Comparative anatomy of the leaf-bearing Cactaceae, XI. The xylem of Pereskiopsis and Quiabentia. Jour. Arnold Arb. 45: 140-157. 1964. BARANOV, ANDREY. On the Manchurian species of Leonurus L. Jour. Jap. Bot. 34: 372-376. 1959. . On the economic use of the wild plants in N. E. China. Quart. Jour. Taiwan Mus. 15: 107-115. 1962. . Materials to the monograph of the isioey es ‘agtlles wide of N.E. China. Quart. Jour. Taiwan Mus. 16: 143- . Some observations on the vernal ei — the Manchurian forest in its a condition. Quart. Jour. Taiwan Mus. 16: 181-184. 1963. Brizicky, GEoRGE K. The genera of Celastrales in the southeastern United States. an Arnold Arb. 45: 206-234. 1963. The genera of Sapindales in the southeastern United States. Jour. Arnold Arb. 44: 462-501. Polyembryony in Euonymus (Celastraceae). Jour. Arnold Arb. 45: 251-259. 1964. (with STERN, WiLLIAM L. & TAMOLANG, FrRANcIScO N.) The woods and flora of the Florida Keys: Capparaceae. Contr. U.S. Natl. Herb. 34: 25-43. pl. 1-5 Duptey, THEODORE R. Alyssum L., Madwort. Jn: Flora of lowland Iraq by K. H. Rechinger, 305-307. 1964. ———. Alyssum turgidius: A new species from Iran. Great Basin Nat. 24: 1964. 7-12. —. New Combinations. /m: Iranian Plants collected - Per Wendelbo in 1959, Arb. Univ. Bergen. Mat.-Naturv. Serie 13: 1963. . Studies in Alyssum: near eastern ara and their allies, I. Jour. Arnold Arb. 45: 57-100. 1964. — ALFRED S. Tsuga canadensis and its multitude of variants. Arnoldia 23: 100-102. 1963. GRAHAM, SHIRLEY A. The Elaeagnaceae in the southeastern United States. Jour. Arnold Arb. 45: 274-278. 1964. ———. The genera of Lythraceae in the southeastern United States. Jour. Arnold Arb. 45: 235-250. GREEN, PETER S. Bible plants in Scottish gardens. Jour. Roy. Caledonian Hort. Soc. 1963: 38-48. The genus Nestegis — New Zealand. Jour. Arnold Arb. 44: 377-389. 1963. . Leucothoé axillaris and L. fontanesiana. a 28: 93-100. 1963. ; nihy fontanesiana. Arnoldia 23: 93- 1963. ae. In: HUTCHINSON, d: & DALZIEL, | M. Eds. Flora of West Tropical ‘Abies II. 2: 47-51. 1963 (with Howarp, R. A., BAKER, H. G. & YEo, P. F.) Comments on “Seed Lists.” Taxon 13: 90-94. 1964. ——— (with Howarp, RicHarp A., & WAGENKNECHT, BURDETTE L.) Inter- national Directory of Botanic Gardens. Regnum Veg. 28: 1-120. 1963. Howarp, Heman A. Labels in the Arnold Arboretum. Arnoldia 24: 9-12. 1964. Howarp, RicHarp A. Clarence Emmeren Kobuski, 1900-1963. Jour. Arnold Arb. 44: 411-416. port. : . Climbing plants. Bull. Gard. Club Am. 51: 67-71. 1963. . The Director’s Report. The Arnold Arboretum during the fiscal year ended June 30, 1963. Jour. Arnold Arb. 44: 502-521. 1963. 1964 | THE DIRECTOR’S REPORT 491 . Golden ate of the National Botanic Gardens of South Africa. Bio Science 14: 35. 1964. —. In appreciation — George Howard Hamor, 1887-1962. Brittonia 15: 204— 207. 1963. The International Association of Botanic Gardens. Taxon 12: 247-249. 1963. . Notes on Rosaceae in the Lesser Antilles. Jour. Arnold Arb. 45: 279-283. 1964. — & Dunpar, Henry F. Additions to the flora of Inagua, the Bahamas. Rhodora 66: 6- 15. 1964. , GREEN, P. S., BAKER, H. G. & YEo, P. F. Comments on “Seed Lists.” Taxon 13: 90-94. 1964. & PowELL, Dutcie A. The introduction of rubber producing species in the West Indies. Econ. Bot. 17: 337-349. 1963. , WAGENKNECHT, BURDETTE L. & GREEN, PETER S. International Direc- tory of Botanic Gardens. Regnum Veg. 28: 1-120. 1963. Hv, Suru-Yinc. Fruit characters in Holly. Am. Hort. Mag. 43: 21-32. 1964. . Sorbus for American garden and community landscapes. Gard. Jour. N.Y. Bot. Gard. 12: 164, 165, 190. 1963. economic botany of Hodgsonia. Econ. Bot. 17: 167-179. 1964. KosuskI, CLARENCE E. Studies in the Theaceae, XXXIV. Some Asiatic taxa of Ternsiroemia. Jour. Arnold Arb. 44: 421-433. 1963. Studies in the Theaceae, XXXV. Two new species of Ternstroemia from the Lesser Antilles. Jour. Arnold Arb. 44: 434, 435. 1963. NeEvLINnG, Lorin I., Jr. Climbing Hydrangeas and their relatives. Arnoldia 24: 17-39. 1964. ———. Documented chromosome numbers of plants. Madrono 17: 116. 1963. ——. Notes on Daphnopsis. Jour. Arnold Arb. 44: 402-410. 1963. —. Note on the genus Ovidia. Darwiniana 13: 72-86. 1964. . Typification in Dirca. Jour. Arnold Arb. 45: 158, 159. 1964. SANDERS, Mary E. & FRANZKE, CLirForD J. A proposed explanation for the origin of colchicine-induced diploid mutants in Sorghum. Jour. Arnold Arb. 45: 36-56. 1964. & ZreBUR, Nancy Kent. Artificial culture of embryos. Jn: ee ., Ed. Recent advances in the embryology of Angiosperms, 297— ee BERT, BERNICE G. Anton K. Schindler. Taxon 13: 7-10. port. 1964 The floristic work of Paul Carpenter Standley. 7a: WrLiiAms, L. O., Ed. Homage to Standley, 51-53. 1963 ——. In reply to Professor Lanjouw. Taxon 13: 83-85. 1964. SCHWARTEN, LazeLLa. Bibliography. 71: Howarp, Ricuarp A, Clarence Em- meren Kobuski, 1900-1963. Jour. Arnold Arb. 44: 417-420. 1963. (with Rocerson, CLarK T., Rickett, H. W., & BECKER, HERMAN.) Index to American Botanical literature. Bull. Torrey Club 90: 209-224, 271-286, 364-383, 422-436. 1963; 91: 59-77, 166-184, 239-262. 1964. Srivastava, Latir M. Secondary phloem in the Pinaceae. Univ. Calif. Publ. Bot. 36: 1-140. illus. 1963. Wyman, DonaLp. The best Forsythias. Horticulture 42: 39, 62, 63. 1964. Black polyethylene as a mulch. Arnoldia 24: 13-16. 1964. _ Centuries-old Rose species still merit garden use. Am. Nurseryman 118(5): 17, 18, 58, 59, 62-70. 1963 — ——. Few Blueberries of much value as ornamentals. Am. Nurseryman 118(3): 11, 35-41, 44, 45. 1963. = 492 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV . Fruiting of yews. Arnoldia 23: 119-122. 1963. Ground Cover Trials at the Arnold Arboretum. Am. Hort. Mag. 42: 4, ‘207- 218. 1963 . Hypericums offer summer bloom in poor soil. Am. Nurseryman 118(11): 11, 38-44. 1963. . Much confusion, big variety, in the widely grown Yews. Am. Nursery- man —— 12, 13, 88-91, 112-119. 1964. w plants pacieteced. Hamamelis intermedia ‘Arnold Promise’. Ar- noldia 2: 111-118. 1963. . Pears not outstanding as ornamentals. Am. Nurseryman 119(7): 13 58, 60. 1964. . Privet scores high for hedges, North to South. Am. Nurseryman 118(9): 12, 13, 55-57, 60, 61. 1963. . Pruning ornamental shrubs and trees. Arnoldia 23: 107-110. 1963. Prunus boasts some of the best flowering plants. Am. Nurseryman 119(9): 10, 11, 95-110. 1964 . Registration list of cultivar names of Fagus L. Arnoldia 24: 1-8. 1964. ——-, tie: blooming Deutzias eclipsed by better shrubs. Am. Nurseryman 118(1): 13, 70-74. ; ceed valued for fall color. Am. Nurseryman 118(7): 10, 11, 57-60. 1963. Spring flowering trees at the Flower Show. Mass. Hort. Soc. 93rd Ann. Spring Flower Show Program, pp. 48, 49. . Tree trunks. Arnoldia 23: 123-130. 1963. RicHarp A. Howarp, Director 1964 | THE DIRECTOR’S REPORT 493 Staff of the Arnold Arboretum 1963-1964 RICHARD ALDEN Howarp, Ph.D., Arnold Professor of Botany, Professor of Dendrology, and Director. Irvinc Wiper Baiey, S.D., Professor of Plant Anatomy, Emeritus. Kart Sax, $.D., Professor of Botany, Emeritus. ANDREY BarRANOv, Curatorial Assistant. GEORGE KONSTANTINE Brizicky, R.N.Dr., Botanist, Southeastern Flora Project MicHaAEL ANTHONY Canoso, M.S., Senior am Assistant.* HENRY DRAPER, Superintendent, Case Estat THEODORE RoBERT Duptey, Ph.D., ree “Teanediel Taxonomist. ALFRED JAMES ForpHAM, Propagato SHIRLEY ANN TouscH GRAHAM, Ph.D. Botanist, Southeastern Flora Project. PETER SHAW GREEN, B.Sc., Horticultural needs WILLIAM Ep Grime, B.A., Curatorial Assistant.* HEMAN ARTHUR Howanp, Assistant analy SHIU-YING Hu, Ph.D., Botanist MARGARET CATHERINE Leravour, Herbarium Secretary. Susan DeLANo McKeE vey, A.B., Research Associate.** Lorin Ives NEVLING, Jr., Ph.D., ‘Associate Curator and Supervisor of the Herbaria.* Lity May Perry, Ph.D., Botanist.*** Mary ELizABETH SANDERS, Ph.D., Research Associate. BERNICE Gipuz SCHUBERT, Ph.D., Associate Curator and Editor. LAZELLA SCHWARTEN, Librarian.* Latir Monan Srivastava, Ph.D., Mercer Research Fellow. STEPHANNE Barry Sutton, A.B., Business Secretary. ROBERT GEROW WILLIAMS, B.S., Superintendent. CARROLL Emory Woop, Jr., Ph.D., Associate Curator. Donatp Wyman, Ph.D., Horticulturist. * Appointed jointly with the Gray Herbarium. ** Resigned 19 *** Retired 1 January 1964. 494 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV INDEX Acanthambrosia, 408 Alyssum baicalicum, 368 —bryantii, 408, 414 — baumgartnerianum, 72, 366 Adamara 7 — bertolonii, 368 Adyseton, — biovulatum, —sect. Disodontea, 359 —blepharocarpum, 60-63, 361 Akebia, — bornmuelleri Alyssum: Near Eastern Representatives — borzaeanum, 368 and Their Allies, I. Studies i —bracteatum, 368 Alyssum, Synopsis of the Genus, 358 — brughieri, 372 Alyssum, 57-100, 358-373, 390 —bulbotrichum, 69, 363 —sect. Adyseton, 359 —caespitosum, 74, 363 — sect. on 63, 69, 74, 76, 362-365, —caliacrae, 368 372, 3 —callichroum, 368 — sect. | een 391 —calycinum, 63, 65 —sect. Aurinia, 391 —calycocarpum, 363 —sect. Eualyssum, 362 i 63, 64, 65, 66, 372 — -—subsect. Hygrochastica, 3 —-— micranthum, — sect. rg ae ais 70, 72, 76, 77, —canescens, 60, 363 79, 365, —caricum, 89-91, 371 — — ser. Ade 70, te 366 —cassium, 370 —— ser. Libera, 72, —cedrorum, 67 — sect. Meniocus, 60, 4 — cephalotes, 68, 69, 363 — sect. Odontarrhena, 77, a. — 92,367- —chondrogynum, 368 371, 372,375 — cilicicum 0 —-— subsect. Compressa, 86, 370 =oohteatann, 367, 372,319 — — subsect. Inflata, 79, 86, 89, 368 — condensatum, —w— subsect. Samarifera, 88, 91, 371 — — subsp. flexible, 85 — -—ser. Crenulata, 370 — — i Ri 5 —w— ser. Integra, 86, ——typl 5 — sect. Psilonema, 63, 361,362, 372 ~- aaa 87, 368 — sect. Bifiotrichum, 359 —contemptum, 363 —sect. Tetradenia, 367, 372, 373 —corningii, 72-74, 366 — subg. oo 361 —corsicum, 91, — affine —corymbosoides, 368 Bernese en 77, 362 — costei, 372 —akamasicum, 370 —crenulatum, 370 icon , 368 —cuneifolium, 363 — suffrutescens —cypricum, 368 ~alyssoides, 63, i i: 362, 373 — damascenum, 362 —amer icanu ress —dasycarpum, 60, 362 — anato a —davisianum, 81, 368 er nee 372 — degenianum, 88 — 393 —-—subcaespitosum, 88 mprocarpum, 393 — densistellatum, 363 —arenarium, 362 —denticulatum, 395 —argenteum, 370, 372 — desertorum, 363 —argyrophyllum, 363 — diffusum, 363 armenum, 3 — discolor, 80, 368 — artwinen — divrikii, —atlanticum, 363, 372 — djurdjurae, 372 —aurantiacum, 363 — doerfleri, 74, 363 —aureum, 361 —~ dubertretii, 92, 371 1964 | Alyssum elatum, 370 —emarginatum, 372 —ephesium, 397 —macrocalyx, 364 366 INDEX 495 Alyssum macropodum, 364 anoellendorianuny 364 —montanum, ee 64, 364, 372 leri, 87 — — subsp. stojanofi, 87, 88 “ee chleri, 8 “mrientm mihi vel tuberculatum,” — ttn: 369 aS megalocarpum, 397 — ovirense, 364 —oxycarpum, 369 — paphlagonicum, by ve 72, 366 — parviflorum, 63, — pateri, 369 — — subsp. prostratum, 84 — peltarioides, 89, 91, 92, 371 —— subsp. peltaricides, 89 praecox, 70 — propinquum, 364 496 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLV Alyssum prostratum, 84 — pseudo-mouradicum, 364 70 — rupestre, 399 —saxatile, 393, 395 — serpyllifolium, Acs $12) 319 — sibiricum, 60, , 371, 373 365 —strigosum, 66, 67-68, 69, — — subsp. cedrorum, 67, 6 — — subsp. strigosum, 68 — cedrorum, 68 — stylare, 361 — subbaicalicum, 373 — subspinosum, 370 — su!phureum, 4, 76, 366 m, ~teapeziorme, 89, 91, 371 8, 69, 365 — troodii, 369 Alyssum caren i , 420 ae 417, 418, 429 — chenopodiifolia, 419, 429 — confertiflora, 408, 419-421, 430 0 —dumosa, 422, , 430 — eriocentra, _ en 430 — flexuosa, 4 ~- soreaaieeg 408, 411, 419 — ilicifolia, 425, 430 — linearis, 425, 426, 430 — magdalenae, 426, 430 — maritima, 40 —nivea, 426, 427, 430 — pumila, 408 — simulans, 420 — tenuifolia, 419 — tomentosa, 424, 427, 428 Amelanchier, 166 , 391 Vela, 391 1964 | Antilles, Notes on Rosaceae in the Lesser, 279 poeseraee 227-234 Aronia, 166 rene: in Turkey, Synopsis of the Genus, 3 mere 390-4 eae. 391 , 392 — rupestris, 399, 400 — — subsp. cyclocarpa, 399 — subsp. rupestris, 399 aie 392-398 —-— subsp. megalocarpa, 397, 398 —w— subsp. orientalis, 394-397 — subsp. saxatilis, 393, 394 —uechtritziana, 398, 399 Bartey, I. W. Comparative Anatomy of e upon the Str Stomata, and Cuticle, 374 Barberry, 9 Barberry Family, 1 Beach heather, 355 Berberidaceae, Lardizabalaceae, and Menispermaceae in the Southeastern United States, The Genera of, Berberidaceae, 1-20 —subfam. Berberidoideae — subfam. Sea rts 13 oe ia Berber ee 457-458 Blue co ohosh, 19 Boke, 306 BrizicKy, GEorGE K. A Further Note on Ceanothus herbaceus versus C. ovatus, 71 4 Brizicky, GEORGE K. era of Celastrales in the Southeastern United BrizicKy, GEORGE The era of Cistaceae in the Southeastern United States, 346 paar peed K. The Genera of Rhamnaceae the Southeastern United pre 09 Brizicky, yembryony in Euonymus eae. 2511 nee 298 INDEX 497 Bucida, 297-298 Buckthorn, 446 Buckthorn ee 439 Cactaceae, Comparative Anatomy of the Leaf-bearing, XI. The Xylem of Pereskiopsis and Quiabentia, 140; XII. Preliminary Observations upon the Structure of the Epidermis, Stomata, and Cuticle, 374 Calycocarpum, 28-30 — lyonii, Calycophyllum grandiflorum, 128 Cathay, quince of, Cathayensis eae miniature, 328 Caulophyllum, 18-20 Ceanothus herbaceus versus C. ovatus, A po Bad om ot af aA _— erbaceus, ae 472 — herbaceus, 471 , — ovalis, 472, 4 heen 471, 473 — per nis ee 206-2 subfam. eet iy 220 States, The Genera of, 20 4 216 bg. ay amie 216 Chamaemeles Chaenomeles ae The Genus, 161, 3 Chaenomeles, 161-205, 302-345 — — ‘California’, 198, 330 — — ‘Cardinal’, 331 —— ‘Clarke’s Giant oe 198, 331 — ae — — Fire’, Seas 198, 331 —— ‘Masterpiece’, 331 —-— ‘Pink Beauty’, 331 —— ‘Rosemary’, 178, 331 —-— ‘Rosy Morn’, 198, 331 —— ‘Sunset Glow’, 331 498 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Chaenomeles cardinalis, 312 Chaenomeles speciosa ‘Contorta’, 198, 315 —cathayensis, 178, 320-323 — ‘Deep Pink’, 31 —— > japonica, 328 —— ine Bang’s 7 315 —— speciosa, 328 — — ‘Dwarf ri — superba, 330 —-- Echo’ —chinensis, 332 — uphrosine’ 315 —»€ clarkiana, 328 —_— ‘Exim Se 328 ng pie Chater, 315 — ‘Minerva’, 178, 328 — — ‘Fireball’, —eburnea, 312 — — ‘Flore ee —eugenioides, 312 — — ‘Flore Rubro dons 315 —— superba, 324 — — ‘Gaujardii’, 315 — extus-coccinea, 312 — — ‘Grandiflora’, 315 a on 305-310 — —‘Hanazono’, 31 — —alpin — — ‘Japanese Scarlet’, 315 —— a "Hilt, 307 ——‘Jimmy’s Choice’, 315 — — ‘Aurea’, 30 — —‘Kermesina Semiplena’, 315 —— ‘Dorothy Rowe’, 307 —-— ‘Knap Hill Radiance’, 315 — ‘Dwarf Poppy * Red’ 307 —w— ‘Leonard’s Velvety’, 315 — —genuina, 312 — — ‘Limoni’, — — japonica, 305 —— ‘Lutea Viridis’, 315 — —maulei, 306 —— ‘Macrocarpa’, 315 — — ‘Maulei’, — — ‘Mallarot’, 315 —_— ‘Orange Beauty’, 308 — — ‘Marmorata’, 315 —— ‘Pigmani’, 198, 30 — — ‘Moerloosei’, 316 — — ‘Plena’, 308 — — ‘Nivalis’, 198, 316 —— pygmaea, 310 — — ‘Nivea Extus Coccinea’, 316 __ — ‘Sargentii’, 178, 184, 308 — — ‘Pacific Red’, 316 — superba, — — ‘Papeleui’, 31 — — ‘Taiojishi’, 308 ——‘Phylis Moore’, 182, 316 —— ‘Tricolor’, 308 —— ‘Red Ruffle — — wilsonii, 320 — —‘Red Sprite’, 316 — — ‘Zoge’, 308 ——‘Rosea Grandiflora’, 316 —lagenaria, 311 —— ‘Rosea Plena’, 316 —cathayensis, 320 — — ‘Rosea Semple 316 — —wilsonii, 320 — — ‘Rubra’, — maulei, 306 — — ‘Rubra Grandoe 316 — speciosa, 311-320 — —‘Russell’s Red’, —— ‘Alarm’, 314 — — ‘Sanguinea Plen —— ‘Alba Cincta’, 314 — — ‘Sanguinea er ae 316 — — ‘Alba Floribunda’ —— ‘Shirataum’, 316 — — ‘Alba Grandiflora eae 314 — — ‘Simoni’, 184, 198, 316 —— ‘Alba Picta’, 314 — — ‘Snow’, 31 — — ‘Alba Rosea’, 314 —— ‘Snow Queen’, 316 —— ‘Angustifolia’, 314 — — ‘Spitfire’, 316 ——¥‘Apple Blossom’, 314 — — ‘Starlight’, 316 — — ‘Atrococcinea’, 31 — — ‘Sulphurea Perfecta’, 316 — — ‘Aurora’, — — ‘Taioh-Nishiki’, 316 — — ‘Baltzii’, 178, 314 — — ‘Tani-no-Yuki’, 316 — — ‘Blood Red’, 314 — — ‘Texas Pink’, —— ‘Bonfire’, 314 — — ‘Toyo-Nishiki’, 316 — — ‘Brilliant’, 315 ——‘Umbilicata’, 198, 316 — — ‘Candicans’, 315 — — ‘Versicolor’, — — ‘Candida’, 315 —— ‘Versicolor Lutescens’, 316 — —‘Candidissima’, 315 — — wilsonil, 320 — — ‘Cardinalis’, 315 — enpette, 324-328 — —cathayensis, 320 — Superba group, 324 1964] INDEX 499 Chaenomeles Es ees ‘Abricot’, 324 Chaenomeles & superba ‘Red Chief’, 198, Alba 326 Se adel 324 —— ‘Rowallane’, 198, 326 —— ‘Boule de Feu’, 324 —— ‘Roxana Foster’, 198, 326 —— ‘Cameo’, 198, 324 —— ‘Ruby Glow’, — — ‘Charming’, 324 —— ‘Salmon’, 326 — — ‘Cole’s Red’, 325 — — ‘Sanguinea’, 326 — — ‘Colette’, 325 — — ‘Scarlet’, 32 — — ‘Columbia’, 325 —-— ‘Shinonome’, 326 —— ‘Coquelicot’, 325 — — ‘Shirabotan’, 326 —— ‘Coral Beauty’, 325 —w— ‘Spring Fashion’, 326 — — ‘Corallina’, 198, 325 —— ‘Stanford Red’, 326 —— ‘Coral Sea’, 325 —— ‘Sunset’, 326 —w— ‘Crimson and Gold’, 325 —— ‘Superba’, 178, 326 —w— ‘Crimson Beauty’, 325 —— ‘Texas Scarlet’, 198, 326 —— ‘Della Robbia’, 325 — — ‘Tortuosa’, 326 —— ‘Double Red’, 325 —— ‘Ulidia’, 326 —w— ‘Double Vermilion’, 325 —— ‘Vermilion’, 326 — — ‘Dwarf Coral’, 325 — — ‘Vesuvius’, 326 — — ‘Early ee ‘aia 325 — — ‘Wakaba’, 326 — — ‘Ecarlate’ — — ‘Yaegaki’, 326 — — ‘Elly ae 325 — trichogyna, 312 —w— ‘Ernst Finken’, 325 — > vilmoriniana, aio — — ‘Etna’, 325 —— ‘Afterglow’, 178, — — ‘Extus Acumineus’, 325 — — ‘Mount Everest’, pe —w— ‘Fascination’, 325 See 329 — — ‘Fire Dance’, 325 —Clarkiana group, 328 —— ‘Foliis Rubris’, 325 — Miniature nee) hybrids, 328 —— ‘Frutico Alba’, 325 Cha-tzu, 306 — — ‘Fruitland’, 325 Chiku-ume, 306 — — ‘George Landis’, 325 Chinese quince, 32 —‘Glowing-Ember’, 198, 325 actrees eae 279 —— ‘Grandiflora Rosea’, 325 —icaco, 280 — — ‘Grenade’, 325 Siete, 3 3-35 —— ‘Harlequin’, 325 Cistaceae in the Southeastern United — — ‘Hever Castle’, 325 States, The Genera of, 346 — — ‘High Noon’, 325 eraser — Sf — — ‘Hi-no-Tsukasa’, 326 Cistu — — ‘Hollandia’, 326 Cly ase ate. 64, 65 — — ‘Incendie’, 326 —campestris, 63, 64, 65 —-— ‘Indian Chief’, 326 —minor, 63, 64, 66 —w— ‘Jane Taudevin’, 326 mentosa, 3 — Jet Trail’, 326 Cocculus, 30— — — ‘Juliet’, spe —car 3 — — ‘Kinjishi’, 3 Cognassier de Cathaye, 321 — — ‘Knap Hill Aan 326 Cognassier a fruit en gourde, 312 — — ‘Leichtlinii’, 326 Cognassier 7 Japon, 306, 312 — — ‘Mandarin’, 326 Colchic —— ‘Margaret Adams’, 326 sari 36 —w— ‘Mount Shasta’, 198, 326 oe 455-457 — — ‘Naranja’, 326 — subg. Colubrina sect. Colubrina, +56 — — ‘Nicoline’, 326 ase Vellozia, 45 — — ‘Nishikichidon’, 326 — subg. Serrataria sect. Barcena, 456 —— ‘Otto Froebel’, 326 —w— sect. Serrataria, 456 —— ‘Perfecta’, 184, 326 Combretaceae in he Southeastern — — ‘Pink Lady’, 226 United States, The Genera of Rhizo- —— ‘Porcelain Rose’, 326 phoraceae and, 285 500 seers: 293-301 m. Com bretoideae, 296 —tribe Terminalieae, 296 94 Combretum Comparative nage of the Leaf-bear- Cactaceae, XI. The Xylem of Poeiinnds er Quiabentia, 140; XII. Preliminary Observations upon the Structure of the Epidermis, Stomata, and Cuticle, 374 Cone Morphology in Pinus sabiniana, 260 Conocarpus, 298-300 Cotogno del oe 312 Cotoneaster olacpeneniae ik 166 t. Spartioides, 352 ee ans 220-221 — speciosa, 311 Decai 0 Diospyros fo the Misantla Region in Mexico, A New, Diospyros riojae, 465-470 Diphylleia, 16-17 Dirca, oo in, 158 a 58 Director’s eee The. The Arnold Arbo- ret we Fiscal Year Ended Ju une 30, 1964, Docynia, a0 Duptey, T. R. Studies in Alyssum: Near Eastern Shenae and Their Allies, I, 57 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Duptey, T. R. Synopsis of the Genus . Synopsis of the Genus Aurinia in Turkey, 390 warf Japanese quince, 306 Dzinashi, 30 Elaeagnaceae in the Southeastern United sect. Elaea eae 277 Fisclvn: NsT, WAL e Genera of Ber- beridaceae, sere pale , and Menis- permaceae in the Southeastern United States, Euonymus (Celastraceae), Polyembryony in, 251 in, Euonymus, 210-215, 251 — semiexsertus, 257 — vagans, 2 -- veupabasiiblen: 256 — verrucosus, 257 — yedo€énsis, 257 False heather, Flowering si hee ie. 312 Franseria, 401, 408 tiumnbnceiniins: 410 1964] INDEX 501 Franseria arborescens, 411, 416 FRANZKE, CLIFFoRD, J. and Mary E. — artemisioides, 408, 411, 412 SanvDerS. A Proposed Explanation for — bipinnat err 414, 417 the Origin of Colchicine-induced Di — dubia, 4 oid Mutants in Sorghum, 36 eae oes Frostweed, 35 — — villosa, 418 r Note on Ceanothus herbaceus — bryantii, 408 rsus C. ovatus, A, 47 — californica, 409 —camphorata, 415 —-—leptophylla, 415 Gaertnera, 407, — canescens, 415 eee rpa, 409 —carduacea, 416 —ambrosiodes, 411 = caudata —artemisiodes, 412 Spee 417 — chamissonis, 417 — — subsp. ee 418 — cordifolia, 421 ae ty , 418 — deltodea, 421 ae 417 — discolor, 427 — —cuneifolia, 417 — dumosa, 423 — — malvaefolia, 417, 418 — erlocentra, 423 — — viscida, 417 — hispida, 415 — chenopodiifolia, 419 ak hookeriana, 409 —confertiflora, 420 — meyeniana, 413 mace : — tenuifolia, 420 — cordifolia, 421 a tomentosa, 424, 427 aertneria, 5 eee oecae — bipinna ifida dubia, 417 — discolor, 427 — chamissonis ere 417 —divaricata, 422 — chenopodifolia, — dumosa, — grayi, 42 — —albicaulis, 423 — linearis, 425 —eriocentra, 423 — tomentosa, 424 — exigua, 427 ene, 60, a a 365 — flexuosa, 424 3 0, — fruticosa, 413, 419 —- ea ttt, 60, 70, 71, 72 — grayi, — lycaonic n, 7 — hispida, 414, 415 — paphlagonicum, 71, — hispidissima, 420 Genera of Berhalidasese i een — hookeriana, 409 and Menispermaceae in the Southeastern — ilicifolia, 425 United States, The, 1 —jincana, 420 Genera of Celastrales in the Southeastern —intricata, 426 United States, The, 206 — lancifolia, 419 Genera of Cistaceae in the Southeastern —leptophylla, 415 United States, The, 346 — lessingii, 417 Genera of Lythraceae in the Southeast- — linearis, 425 ern United States, The, 235 — malvacea, 421 Genera of Rhamnaceae in the Southeast- —meyeniana, 413 ern United States, The, 439 — montana, 409 Genera of hizophoraceae and Com —nivea, 426 bretaceae in - Sree Oniled — palmeri, 409 States, The, 2 — pringlei, 420 Genus “chatnomee (Rosaceae), The, bie i 13 1 sanctae-gertrudis, 416 Gomez aoe Arturo. A New Dios- ee 420 pyros from the Misantla Region in — tenuifolia, 419, 420 Mexico, 464 — — tripinnatifida, 420 Gouania, 462-463 — tomentosa, 424, 427 GRAHAM, SHIRLEY A. The Elaeagnaceae — villosa, 418 in the Shes tert United States, 274 502 GraHaM, SHIRLEY A. The Genera of Lythraceae in the Southeastern United rey A. The Genera Rhizophoraceae and Combretaceae in the Southeastern “United States, 285 s R. Cone Morphology in Gyminda, Halimium — — sect. nae 352 5 Hesperomeles, 166 Hiboke, 312 Hippocratea, 225-226 Hippocratea Family, 223 Hippocrateaceae, 223-226 Holly, 228 Horau, 300 Howarp, Ricwarp A. The Director’s Re- port. The Arnold Arboretum During the Fiscal Year Ended June 30, 1964, 474 ii Ricuarp A. Notes on Rosaceae in the aa. Antilles, 279 Hudsonia, 354- Hung mei, 312 Tlex, 228-234 — subg. Prinos, 229 Japanese quince, 161, 312 apanese quince, dwarf, 306 oe Qui Se . Hera aed Japonica, JAYAWEERA, oe M. A. The Rubiaceous Genus Mussaenda: ca Species of the Philippine i 10 Jeffersonia, Jinashi, 306 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Kai dan boke, 312 e, 306 Konig sect. Tetradenia, 367 Koniga sect. Piflotrichum, 359 — saree 399 — scardica, 399 Kragodendron, 462 Kusa-boke, Laguncularia, 300-301 Lardizabala, 20 Lardizabala Family, 2 Lardizabalaceae, and Stak the Southeastern United States, Genera of Berberidaceae, 1 Lardizabalaceae, 20-22 ispermaceae in The icania oligantha, 279 Lobularia, 373 —maritima, 373, 392 — rupestris, Loosestri fe, 243 Loosestrife Family, 235 Lythrum, 242-2 — sect. leat 243 Ma-boke, 321 Malaceae, 164 alus, 166 — japonica, 305 aaieenige Mangrove, Mangrove ai, Red, 285 Mangrove Family, ee 293 Mangrove, white, 300 Moenchia, 359 1964] Moonseed, 30, 33 Moonseed Family, 23 Mu-kua, 321 Mussaenda, e Rubiaceous Genus: The Species of the Philippine Islands, 101 Mussaenda acuminatissima, 104, 105, 126 — acutiflora, — albiflora, 105-107, 120, —anisophylla, 108, 109, 120, 121 —attenuifolia, 106, -1 — benguetensis, 106, 111, 112, 120, 121 —chlorantha, 106, — frondosa, — glabra, 12 — grandiflor = srancifotia, bn Te ae 126 —lanata, 106, 114, Sica ; — — brevipilosa, 117, 119 ip sarees 106, 114, 115, 119-121 — milleri, ~ libata, ee 124, 126 — nervosa, 106, ar 124, 125 — palawanensis, 10e 125-128 —philippica, 106, 120, 121, 128-131 — — aurorae, 132 — philippinensis, 106, 120, 121, 132-134 — pinatubensis, 134, 135 — vidalii, 114, 115, 138, 139 — villosa, Mu-T’ae, Mutants in Sorghum, A Proposed Expla- nation for the Origin of Colchicine- induced, 36 Myginda, 221 Nandina Near acters oo and Their Allies, I. Studies in Alyssum, 57 NEVLING, LorIn 1, Jr. Pepinea tian in Dirca, eo No-boke, Notes on ee in the Lesser Antilles, 279 Odontarrhena, ° 367 Colchicine-induced Diploid Mutants in Sorghum, A Proposed Ex- lanation for the, 36 Osteomeles, 166 INDEX 503 Pachistima, 207 Payne, Wittarp W. A Re-evaluation of enus Ambros (cana 401 Peplis, 241-24 — subg. Didiptis, 242 az-romeroana, 141, 379, 384 — grandifolia, 141, 375, 383, 384 — guamacho, 140, 380 — humboldtii, 141, 379, 383 — moorei, 375 —nicoyana, 140, 380 —pititache, 380, 383 — portulacifolia, 141, 380 —sacharosa, 141, 375, 383 — tampicans, 375, 384 — weberiana, 141, Bend ae and Quiabenta i marian mparative Anat of the eaf- ‘ earing Cactaceae, XL. 14 Pereskiopsis, 140, 374 eee 140, 149, 384 —blakeana, 140 Bonin ay 148, 149, 384 — diquettii, 1 — gatesii, ort — pititache, —porteri, 140, 144, 148, 384 — rotundifolia, ) Pinus cule 260 — jeffre Pinus a. Cone Morphology in, 260 Pinus sabiniana, 260-273 ani Poi ccak 2 504 Podophyllum, 13-16 Poirier du Japon, 306, Polyembryony Sau (Celas- traceae), 251 Potentilla, 281 —anglica, 281 —argentea, 281 — procumbens, 281 re s, 281 Prachtige Quitte, 312 Prelimi Observations the Structure of the Epidermis, Stomata, and Cuticle. Comparative Ana of the Leaf-bearing Cactaceae, XII, 374 Prinos, 228 Proposed Explanation for the Origin of Siren induced Diploid Mutants in Sorgh Pseudo chaenomeles, 168, 302 Psilonema, 60, Pillotcichum, — ti. 359 —cyclocarpum, 399 -—— halimifoliues, 392 Pyrocrataegus, 166 stig 6 Pyru - 166 ae. 305, 320 — — maulei, 30 — maulei, 306 Quiabentia, The Xylem of taerec an ompar ative Anatom he Sal be earing Cactaceae, XI, 16 Quiabentia, 140, 374 —chacoensis, 140, 144, 149 — pereziensis, 141, 149 —zehntneri, 141, 144, 149 Quince of Cathay, 321 Quince, flowering, 161, 312 Quince, Japanese, 161, 312 Red Mangrove Family, 285 Redroot, 452 Re-evaluation of the Genus Ambrosia (Compositae), 401 Reynosia, 461 — subg. Neoreynosia, 461 aaxiniibeg, Reynosia, 461 Rhamnaceae in the Southeastern United States, The Genera of, 439 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLV Rhamnaceae, 439-463 — tribe Gouanieae, 462 — tribe Paliureae, 457 — tribe Rhamneae, Rhamnus, posal — subg. Frangula, — subg. Pseudotrangul, 446 — subg. Rham — subg. Sciadophila, 446 Rhaphiolepis, 166 Rheumatism root, 18 Rhizophora, 286-293 — sect. Aérope, 287 — sect. Rhizophora, 287 — mangle, Rhizophoraceae and Combretaceae in the Southeastern United States, The Gen- era of, 28 Rhizophoracese, 285-293 Rockrose Family, 346 Rosaceae in the Lesser Antilles, Notes on, 279 Rosaceae, — subfam. a sidaxe: 164 eligi neers 164 Rotala, eee Genus Mussaenda, The: The Species of the Philippine Islands, 101 Rubus ferrugineus, 281 a Buuslan olive, 276 Sageretia, 449-451 S E AR and CLIFFORD Sorghum, A Proposed Explanation for the Origin of ‘Colchicine induced Di- ploid Mutants in, 36 Sorghum, 36 — subg. Eu-sorghum, 41 — subg. Sorghum, 41, 42 — um, 44 —halepense, 43, , 45 — oo escens, 42 — sudan eh 1964] Sorghum vulgare, 40, 41, 42, 43, 44, 45 —— caffrorum, — — hegari, 44 — — saccharatu — subglabrescens, = 42 — ‘Black Amber Cane’, 36 — ‘Dakota yes Sore’. 41, 43 — ‘Day Milo’ — Dual’, §3 Sent cd 3’, 36-44, 47-51 — ‘No rghum — ‘Winner Souther United mgitit The Elae- the, teats gous aa The Gene of tio ISS — balaceae, ed pe M Southeastern ‘United ite The Genera 2 tates, The Genera 6 ys goeerttin United States, The Genera 235 ythra ae cabana United yt The Genera of memo — the, green Uni eo hg ie Genera Rhizo phornre and Combreta- ceae in = Spindle-tree, Staff of the gree Arboretum, 1963- um: Near os ieee = ite Their Allies, 36 Synopsis of the Genus Alyssum, 358 INDEX 505 Synopsis of the Genus Aurinia in Tur- key, 390 peerage 296-297 sect. Terminalia, 297 Tich keng hait’ ang, 312 4 sian oe 60, 92, 371 Turkey, ‘ones of the Genus Aurinia in, 390 Twinleaf, 18 Typification in Dirca, 158 Umbrella-leaf, 16 Water-oleander, 239 a ep - Water-willow Weber, i The Genus Chaenomeles ae mene, ae 302 White man 300 White ae Family, 293 ’Xanthidium, 408 tenuifolium, 408, 420 Xanthium, 40. oS acai, 411, 413 —— of es eskiopsi ee Quiabentia, Com tive saga of the is bearing a a XI, Yodo-boke, 312 Ziziphus, 459-460